I0428 12:51:23.231676  9322 upgrade_proto.cpp:1082] Attempting to upgrade input file specified using deprecated 'solver_type' field (enum)': /mnt/bigdisk/DIGITS-AMB-2/digits/jobs/20210428-120150-57ed/solver.prototxt
I0428 12:51:23.231812  9322 upgrade_proto.cpp:1089] Successfully upgraded file specified using deprecated 'solver_type' field (enum) to 'type' field (string).
W0428 12:51:23.231817  9322 upgrade_proto.cpp:1091] Note that future Caffe releases will only support 'type' field (string) for a solver's type.
I0428 12:51:23.231875  9322 caffe.cpp:218] Using GPUs 3
I0428 12:51:23.270869  9322 caffe.cpp:223] GPU 3: GeForce RTX 2080
I0428 12:51:23.594192  9322 solver.cpp:44] Initializing solver from parameters:
test_iter: 51
test_interval: 102
base_lr: 0.01
display: 12
max_iter: 10200
lr_policy: "exp"
gamma: 0.99980193
momentum: 0.9
weight_decay: 0.0001
snapshot: 102
snapshot_prefix: "snapshot"
solver_mode: GPU
device_id: 3
net: "train_val.prototxt"
train_state {
level: 0
stage: ""
}
type: "SGD"
I0428 12:51:23.595065  9322 solver.cpp:87] Creating training net from net file: train_val.prototxt
I0428 12:51:23.595700  9322 net.cpp:294] The NetState phase (0) differed from the phase (1) specified by a rule in layer val-data
I0428 12:51:23.595714  9322 net.cpp:294] The NetState phase (0) differed from the phase (1) specified by a rule in layer accuracy
I0428 12:51:23.595860  9322 net.cpp:51] Initializing net from parameters:
state {
phase: TRAIN
level: 0
stage: ""
}
layer {
name: "train-data"
type: "Data"
top: "data"
top: "label"
include {
phase: TRAIN
}
transform_param {
mirror: true
crop_size: 227
mean_file: "/mnt/bigdisk/DIGITS-AMB-2/digits/jobs/20210419-113214-d311/mean.binaryproto"
}
data_param {
source: "/mnt/bigdisk/DIGITS-AMB-2/digits/jobs/20210419-113214-d311/train_db"
batch_size: 128
backend: LMDB
}
}
layer {
name: "conv1"
type: "Convolution"
bottom: "data"
top: "conv1"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 96
kernel_size: 11
stride: 4
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0
}
}
}
layer {
name: "relu1"
type: "ReLU"
bottom: "conv1"
top: "conv1"
}
layer {
name: "norm1"
type: "LRN"
bottom: "conv1"
top: "norm1"
lrn_param {
local_size: 5
alpha: 0.0001
beta: 0.75
}
}
layer {
name: "pool1"
type: "Pooling"
bottom: "norm1"
top: "pool1"
pooling_param {
pool: MAX
kernel_size: 3
stride: 2
}
}
layer {
name: "conv2"
type: "Convolution"
bottom: "pool1"
top: "conv2"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 256
pad: 2
kernel_size: 5
group: 2
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
name: "relu2"
type: "ReLU"
bottom: "conv2"
top: "conv2"
}
layer {
name: "norm2"
type: "LRN"
bottom: "conv2"
top: "norm2"
lrn_param {
local_size: 5
alpha: 0.0001
beta: 0.75
}
}
layer {
name: "pool2"
type: "Pooling"
bottom: "norm2"
top: "pool2"
pooling_param {
pool: MAX
kernel_size: 3
stride: 2
}
}
layer {
name: "conv3"
type: "Convolution"
bottom: "pool2"
top: "conv3"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 384
pad: 1
kernel_size: 3
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0
}
}
}
layer {
name: "relu3"
type: "ReLU"
bottom: "conv3"
top: "conv3"
}
layer {
name: "conv3.5"
type: "Convolution"
bottom: "conv3"
top: "conv3.5"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 384
pad: 1
kernel_size: 3
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0
}
}
}
layer {
name: "relu3.5"
type: "ReLU"
bottom: "conv3.5"
top: "conv3.5"
}
layer {
name: "conv4"
type: "Convolution"
bottom: "conv3.5"
top: "conv4"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 384
pad: 1
kernel_size: 3
group: 2
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
name: "relu4"
type: "ReLU"
bottom: "conv4"
top: "conv4"
}
layer {
name: "conv5"
type: "Convolution"
bottom: "conv4"
top: "conv5"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 256
pad: 1
kernel_size: 3
group: 2
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
name: "relu5"
type: "ReLU"
bottom: "conv5"
top: "conv5"
}
layer {
name: "pool5"
type: "Pooling"
bottom: "conv5"
top: "pool5"
pooling_param {
pool: MAX
kernel_size: 3
stride: 2
}
}
layer {
name: "fc6"
type: "InnerProduct"
bottom: "pool5"
top: "fc6"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
inner_product_param {
num_output: 4096
weight_filler {
type: "gaussian"
std: 0.005
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
name: "relu6"
type: "ReLU"
bottom: "fc6"
top: "fc6"
}
layer {
name: "drop6"
type: "Dropout"
bottom: "fc6"
top: "fc6"
dropout_param {
dropout_ratio: 0.5
}
}
layer {
name: "fc7"
type: "InnerProduct"
bottom: "fc6"
top: "fc7"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
inner_product_param {
num_output: 4096
weight_filler {
type: "gaussian"
std: 0.005
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
name: "relu7"
type: "ReLU"
bottom: "fc7"
top: "fc7"
}
layer {
name: "drop7"
type: "Dropout"
bottom: "fc7"
top: "fc7"
dropout_param {
dropout_ratio: 0.5
}
}
layer {
name: "fc8"
type: "InnerProduct"
bottom: "fc7"
top: "fc8"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
inner_product_param {
num_output: 196
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0
}
}
}
layer {
name: "loss"
type: "SoftmaxWithLoss"
bottom: "fc8"
bottom: "label"
top: "loss"
}
I0428 12:51:23.595957  9322 layer_factory.hpp:77] Creating layer train-data
I0428 12:51:23.657513  9322 db_lmdb.cpp:35] Opened lmdb /mnt/bigdisk/DIGITS-AMB-2/digits/jobs/20210419-113214-d311/train_db
I0428 12:51:23.684587  9322 net.cpp:84] Creating Layer train-data
I0428 12:51:23.684623  9322 net.cpp:380] train-data -> data
I0428 12:51:23.684661  9322 net.cpp:380] train-data -> label
I0428 12:51:23.684684  9322 data_transformer.cpp:25] Loading mean file from: /mnt/bigdisk/DIGITS-AMB-2/digits/jobs/20210419-113214-d311/mean.binaryproto
I0428 12:51:23.755913  9322 data_layer.cpp:45] output data size: 128,3,227,227
I0428 12:51:23.889533  9322 net.cpp:122] Setting up train-data
I0428 12:51:23.889557  9322 net.cpp:129] Top shape: 128 3 227 227 (19787136)
I0428 12:51:23.889562  9322 net.cpp:129] Top shape: 128 (128)
I0428 12:51:23.889565  9322 net.cpp:137] Memory required for data: 79149056
I0428 12:51:23.889575  9322 layer_factory.hpp:77] Creating layer conv1
I0428 12:51:23.889596  9322 net.cpp:84] Creating Layer conv1
I0428 12:51:23.889601  9322 net.cpp:406] conv1 <- data
I0428 12:51:23.889613  9322 net.cpp:380] conv1 -> conv1
I0428 12:51:24.806118  9322 net.cpp:122] Setting up conv1
I0428 12:51:24.806139  9322 net.cpp:129] Top shape: 128 96 55 55 (37171200)
I0428 12:51:24.806143  9322 net.cpp:137] Memory required for data: 227833856
I0428 12:51:24.806161  9322 layer_factory.hpp:77] Creating layer relu1
I0428 12:51:24.806170  9322 net.cpp:84] Creating Layer relu1
I0428 12:51:24.806174  9322 net.cpp:406] relu1 <- conv1
I0428 12:51:24.806200  9322 net.cpp:367] relu1 -> conv1 (in-place)
I0428 12:51:24.806521  9322 net.cpp:122] Setting up relu1
I0428 12:51:24.806531  9322 net.cpp:129] Top shape: 128 96 55 55 (37171200)
I0428 12:51:24.806533  9322 net.cpp:137] Memory required for data: 376518656
I0428 12:51:24.806537  9322 layer_factory.hpp:77] Creating layer norm1
I0428 12:51:24.806545  9322 net.cpp:84] Creating Layer norm1
I0428 12:51:24.806548  9322 net.cpp:406] norm1 <- conv1
I0428 12:51:24.806553  9322 net.cpp:380] norm1 -> norm1
I0428 12:51:24.807098  9322 net.cpp:122] Setting up norm1
I0428 12:51:24.807109  9322 net.cpp:129] Top shape: 128 96 55 55 (37171200)
I0428 12:51:24.807112  9322 net.cpp:137] Memory required for data: 525203456
I0428 12:51:24.807116  9322 layer_factory.hpp:77] Creating layer pool1
I0428 12:51:24.807122  9322 net.cpp:84] Creating Layer pool1
I0428 12:51:24.807126  9322 net.cpp:406] pool1 <- norm1
I0428 12:51:24.807130  9322 net.cpp:380] pool1 -> pool1
I0428 12:51:24.807166  9322 net.cpp:122] Setting up pool1
I0428 12:51:24.807173  9322 net.cpp:129] Top shape: 128 96 27 27 (8957952)
I0428 12:51:24.807174  9322 net.cpp:137] Memory required for data: 561035264
I0428 12:51:24.807178  9322 layer_factory.hpp:77] Creating layer conv2
I0428 12:51:24.807188  9322 net.cpp:84] Creating Layer conv2
I0428 12:51:24.807190  9322 net.cpp:406] conv2 <- pool1
I0428 12:51:24.807195  9322 net.cpp:380] conv2 -> conv2
I0428 12:51:24.814666  9322 net.cpp:122] Setting up conv2
I0428 12:51:24.814682  9322 net.cpp:129] Top shape: 128 256 27 27 (23887872)
I0428 12:51:24.814685  9322 net.cpp:137] Memory required for data: 656586752
I0428 12:51:24.814702  9322 layer_factory.hpp:77] Creating layer relu2
I0428 12:51:24.814708  9322 net.cpp:84] Creating Layer relu2
I0428 12:51:24.814711  9322 net.cpp:406] relu2 <- conv2
I0428 12:51:24.814716  9322 net.cpp:367] relu2 -> conv2 (in-place)
I0428 12:51:24.815199  9322 net.cpp:122] Setting up relu2
I0428 12:51:24.815209  9322 net.cpp:129] Top shape: 128 256 27 27 (23887872)
I0428 12:51:24.815212  9322 net.cpp:137] Memory required for data: 752138240
I0428 12:51:24.815215  9322 layer_factory.hpp:77] Creating layer norm2
I0428 12:51:24.815222  9322 net.cpp:84] Creating Layer norm2
I0428 12:51:24.815225  9322 net.cpp:406] norm2 <- conv2
I0428 12:51:24.815230  9322 net.cpp:380] norm2 -> norm2
I0428 12:51:24.815551  9322 net.cpp:122] Setting up norm2
I0428 12:51:24.815559  9322 net.cpp:129] Top shape: 128 256 27 27 (23887872)
I0428 12:51:24.815562  9322 net.cpp:137] Memory required for data: 847689728
I0428 12:51:24.815565  9322 layer_factory.hpp:77] Creating layer pool2
I0428 12:51:24.815573  9322 net.cpp:84] Creating Layer pool2
I0428 12:51:24.815577  9322 net.cpp:406] pool2 <- norm2
I0428 12:51:24.815580  9322 net.cpp:380] pool2 -> pool2
I0428 12:51:24.815604  9322 net.cpp:122] Setting up pool2
I0428 12:51:24.815608  9322 net.cpp:129] Top shape: 128 256 13 13 (5537792)
I0428 12:51:24.815611  9322 net.cpp:137] Memory required for data: 869840896
I0428 12:51:24.815614  9322 layer_factory.hpp:77] Creating layer conv3
I0428 12:51:24.815623  9322 net.cpp:84] Creating Layer conv3
I0428 12:51:24.815625  9322 net.cpp:406] conv3 <- pool2
I0428 12:51:24.815629  9322 net.cpp:380] conv3 -> conv3
I0428 12:51:24.826006  9322 net.cpp:122] Setting up conv3
I0428 12:51:24.826020  9322 net.cpp:129] Top shape: 128 384 13 13 (8306688)
I0428 12:51:24.826023  9322 net.cpp:137] Memory required for data: 903067648
I0428 12:51:24.826033  9322 layer_factory.hpp:77] Creating layer relu3
I0428 12:51:24.826040  9322 net.cpp:84] Creating Layer relu3
I0428 12:51:24.826043  9322 net.cpp:406] relu3 <- conv3
I0428 12:51:24.826047  9322 net.cpp:367] relu3 -> conv3 (in-place)
I0428 12:51:24.826562  9322 net.cpp:122] Setting up relu3
I0428 12:51:24.826572  9322 net.cpp:129] Top shape: 128 384 13 13 (8306688)
I0428 12:51:24.826576  9322 net.cpp:137] Memory required for data: 936294400
I0428 12:51:24.826579  9322 layer_factory.hpp:77] Creating layer conv3.5
I0428 12:51:24.826588  9322 net.cpp:84] Creating Layer conv3.5
I0428 12:51:24.826617  9322 net.cpp:406] conv3.5 <- conv3
I0428 12:51:24.826622  9322 net.cpp:380] conv3.5 -> conv3.5
I0428 12:51:24.844482  9322 net.cpp:122] Setting up conv3.5
I0428 12:51:24.844501  9322 net.cpp:129] Top shape: 128 384 13 13 (8306688)
I0428 12:51:24.844503  9322 net.cpp:137] Memory required for data: 969521152
I0428 12:51:24.844512  9322 layer_factory.hpp:77] Creating layer relu3.5
I0428 12:51:24.844521  9322 net.cpp:84] Creating Layer relu3.5
I0428 12:51:24.844523  9322 net.cpp:406] relu3.5 <- conv3.5
I0428 12:51:24.844529  9322 net.cpp:367] relu3.5 -> conv3.5 (in-place)
I0428 12:51:24.845644  9322 net.cpp:122] Setting up relu3.5
I0428 12:51:24.845654  9322 net.cpp:129] Top shape: 128 384 13 13 (8306688)
I0428 12:51:24.845657  9322 net.cpp:137] Memory required for data: 1002747904
I0428 12:51:24.845660  9322 layer_factory.hpp:77] Creating layer conv4
I0428 12:51:24.845670  9322 net.cpp:84] Creating Layer conv4
I0428 12:51:24.845674  9322 net.cpp:406] conv4 <- conv3.5
I0428 12:51:24.845679  9322 net.cpp:380] conv4 -> conv4
I0428 12:51:24.855854  9322 net.cpp:122] Setting up conv4
I0428 12:51:24.855870  9322 net.cpp:129] Top shape: 128 384 13 13 (8306688)
I0428 12:51:24.855873  9322 net.cpp:137] Memory required for data: 1035974656
I0428 12:51:24.855885  9322 layer_factory.hpp:77] Creating layer relu4
I0428 12:51:24.855893  9322 net.cpp:84] Creating Layer relu4
I0428 12:51:24.855896  9322 net.cpp:406] relu4 <- conv4
I0428 12:51:24.855902  9322 net.cpp:367] relu4 -> conv4 (in-place)
I0428 12:51:24.856238  9322 net.cpp:122] Setting up relu4
I0428 12:51:24.856248  9322 net.cpp:129] Top shape: 128 384 13 13 (8306688)
I0428 12:51:24.856251  9322 net.cpp:137] Memory required for data: 1069201408
I0428 12:51:24.856256  9322 layer_factory.hpp:77] Creating layer conv5
I0428 12:51:24.856267  9322 net.cpp:84] Creating Layer conv5
I0428 12:51:24.856271  9322 net.cpp:406] conv5 <- conv4
I0428 12:51:24.856278  9322 net.cpp:380] conv5 -> conv5
I0428 12:51:24.865660  9322 net.cpp:122] Setting up conv5
I0428 12:51:24.865674  9322 net.cpp:129] Top shape: 128 256 13 13 (5537792)
I0428 12:51:24.865679  9322 net.cpp:137] Memory required for data: 1091352576
I0428 12:51:24.865685  9322 layer_factory.hpp:77] Creating layer relu5
I0428 12:51:24.865696  9322 net.cpp:84] Creating Layer relu5
I0428 12:51:24.865700  9322 net.cpp:406] relu5 <- conv5
I0428 12:51:24.865705  9322 net.cpp:367] relu5 -> conv5 (in-place)
I0428 12:51:24.866256  9322 net.cpp:122] Setting up relu5
I0428 12:51:24.866266  9322 net.cpp:129] Top shape: 128 256 13 13 (5537792)
I0428 12:51:24.866268  9322 net.cpp:137] Memory required for data: 1113503744
I0428 12:51:24.866271  9322 layer_factory.hpp:77] Creating layer pool5
I0428 12:51:24.866281  9322 net.cpp:84] Creating Layer pool5
I0428 12:51:24.866284  9322 net.cpp:406] pool5 <- conv5
I0428 12:51:24.866289  9322 net.cpp:380] pool5 -> pool5
I0428 12:51:24.866325  9322 net.cpp:122] Setting up pool5
I0428 12:51:24.866330  9322 net.cpp:129] Top shape: 128 256 6 6 (1179648)
I0428 12:51:24.866333  9322 net.cpp:137] Memory required for data: 1118222336
I0428 12:51:24.866336  9322 layer_factory.hpp:77] Creating layer fc6
I0428 12:51:24.866343  9322 net.cpp:84] Creating Layer fc6
I0428 12:51:24.866346  9322 net.cpp:406] fc6 <- pool5
I0428 12:51:24.866351  9322 net.cpp:380] fc6 -> fc6
I0428 12:51:25.224593  9322 net.cpp:122] Setting up fc6
I0428 12:51:25.224617  9322 net.cpp:129] Top shape: 128 4096 (524288)
I0428 12:51:25.224620  9322 net.cpp:137] Memory required for data: 1120319488
I0428 12:51:25.224629  9322 layer_factory.hpp:77] Creating layer relu6
I0428 12:51:25.224637  9322 net.cpp:84] Creating Layer relu6
I0428 12:51:25.224642  9322 net.cpp:406] relu6 <- fc6
I0428 12:51:25.224647  9322 net.cpp:367] relu6 -> fc6 (in-place)
I0428 12:51:25.225549  9322 net.cpp:122] Setting up relu6
I0428 12:51:25.225560  9322 net.cpp:129] Top shape: 128 4096 (524288)
I0428 12:51:25.225564  9322 net.cpp:137] Memory required for data: 1122416640
I0428 12:51:25.225567  9322 layer_factory.hpp:77] Creating layer drop6
I0428 12:51:25.225592  9322 net.cpp:84] Creating Layer drop6
I0428 12:51:25.225595  9322 net.cpp:406] drop6 <- fc6
I0428 12:51:25.225600  9322 net.cpp:367] drop6 -> fc6 (in-place)
I0428 12:51:25.225630  9322 net.cpp:122] Setting up drop6
I0428 12:51:25.225636  9322 net.cpp:129] Top shape: 128 4096 (524288)
I0428 12:51:25.225639  9322 net.cpp:137] Memory required for data: 1124513792
I0428 12:51:25.225642  9322 layer_factory.hpp:77] Creating layer fc7
I0428 12:51:25.225648  9322 net.cpp:84] Creating Layer fc7
I0428 12:51:25.225651  9322 net.cpp:406] fc7 <- fc6
I0428 12:51:25.225656  9322 net.cpp:380] fc7 -> fc7
I0428 12:51:25.385246  9322 net.cpp:122] Setting up fc7
I0428 12:51:25.385263  9322 net.cpp:129] Top shape: 128 4096 (524288)
I0428 12:51:25.385267  9322 net.cpp:137] Memory required for data: 1126610944
I0428 12:51:25.385275  9322 layer_factory.hpp:77] Creating layer relu7
I0428 12:51:25.385285  9322 net.cpp:84] Creating Layer relu7
I0428 12:51:25.385288  9322 net.cpp:406] relu7 <- fc7
I0428 12:51:25.385294  9322 net.cpp:367] relu7 -> fc7 (in-place)
I0428 12:51:25.386190  9322 net.cpp:122] Setting up relu7
I0428 12:51:25.386202  9322 net.cpp:129] Top shape: 128 4096 (524288)
I0428 12:51:25.386205  9322 net.cpp:137] Memory required for data: 1128708096
I0428 12:51:25.386209  9322 layer_factory.hpp:77] Creating layer drop7
I0428 12:51:25.386214  9322 net.cpp:84] Creating Layer drop7
I0428 12:51:25.386217  9322 net.cpp:406] drop7 <- fc7
I0428 12:51:25.386224  9322 net.cpp:367] drop7 -> fc7 (in-place)
I0428 12:51:25.386245  9322 net.cpp:122] Setting up drop7
I0428 12:51:25.386250  9322 net.cpp:129] Top shape: 128 4096 (524288)
I0428 12:51:25.386252  9322 net.cpp:137] Memory required for data: 1130805248
I0428 12:51:25.386255  9322 layer_factory.hpp:77] Creating layer fc8
I0428 12:51:25.386263  9322 net.cpp:84] Creating Layer fc8
I0428 12:51:25.386266  9322 net.cpp:406] fc8 <- fc7
I0428 12:51:25.386272  9322 net.cpp:380] fc8 -> fc8
I0428 12:51:25.394099  9322 net.cpp:122] Setting up fc8
I0428 12:51:25.394110  9322 net.cpp:129] Top shape: 128 196 (25088)
I0428 12:51:25.394114  9322 net.cpp:137] Memory required for data: 1130905600
I0428 12:51:25.394127  9322 layer_factory.hpp:77] Creating layer loss
I0428 12:51:25.394134  9322 net.cpp:84] Creating Layer loss
I0428 12:51:25.394138  9322 net.cpp:406] loss <- fc8
I0428 12:51:25.394142  9322 net.cpp:406] loss <- label
I0428 12:51:25.394148  9322 net.cpp:380] loss -> loss
I0428 12:51:25.394160  9322 layer_factory.hpp:77] Creating layer loss
I0428 12:51:25.394696  9322 net.cpp:122] Setting up loss
I0428 12:51:25.394706  9322 net.cpp:129] Top shape: (1)
I0428 12:51:25.394708  9322 net.cpp:132]     with loss weight 1
I0428 12:51:25.394724  9322 net.cpp:137] Memory required for data: 1130905604
I0428 12:51:25.394728  9322 net.cpp:198] loss needs backward computation.
I0428 12:51:25.394734  9322 net.cpp:198] fc8 needs backward computation.
I0428 12:51:25.394737  9322 net.cpp:198] drop7 needs backward computation.
I0428 12:51:25.394740  9322 net.cpp:198] relu7 needs backward computation.
I0428 12:51:25.394742  9322 net.cpp:198] fc7 needs backward computation.
I0428 12:51:25.394745  9322 net.cpp:198] drop6 needs backward computation.
I0428 12:51:25.394747  9322 net.cpp:198] relu6 needs backward computation.
I0428 12:51:25.394750  9322 net.cpp:198] fc6 needs backward computation.
I0428 12:51:25.394753  9322 net.cpp:198] pool5 needs backward computation.
I0428 12:51:25.394757  9322 net.cpp:198] relu5 needs backward computation.
I0428 12:51:25.394759  9322 net.cpp:198] conv5 needs backward computation.
I0428 12:51:25.394762  9322 net.cpp:198] relu4 needs backward computation.
I0428 12:51:25.394764  9322 net.cpp:198] conv4 needs backward computation.
I0428 12:51:25.394767  9322 net.cpp:198] relu3.5 needs backward computation.
I0428 12:51:25.394771  9322 net.cpp:198] conv3.5 needs backward computation.
I0428 12:51:25.394774  9322 net.cpp:198] relu3 needs backward computation.
I0428 12:51:25.394778  9322 net.cpp:198] conv3 needs backward computation.
I0428 12:51:25.394780  9322 net.cpp:198] pool2 needs backward computation.
I0428 12:51:25.394802  9322 net.cpp:198] norm2 needs backward computation.
I0428 12:51:25.394806  9322 net.cpp:198] relu2 needs backward computation.
I0428 12:51:25.394809  9322 net.cpp:198] conv2 needs backward computation.
I0428 12:51:25.394811  9322 net.cpp:198] pool1 needs backward computation.
I0428 12:51:25.394814  9322 net.cpp:198] norm1 needs backward computation.
I0428 12:51:25.394817  9322 net.cpp:198] relu1 needs backward computation.
I0428 12:51:25.394820  9322 net.cpp:198] conv1 needs backward computation.
I0428 12:51:25.394824  9322 net.cpp:200] train-data does not need backward computation.
I0428 12:51:25.394826  9322 net.cpp:242] This network produces output loss
I0428 12:51:25.394841  9322 net.cpp:255] Network initialization done.
I0428 12:51:25.395319  9322 solver.cpp:172] Creating test net (#0) specified by net file: train_val.prototxt
I0428 12:51:25.395350  9322 net.cpp:294] The NetState phase (1) differed from the phase (0) specified by a rule in layer train-data
I0428 12:51:25.395494  9322 net.cpp:51] Initializing net from parameters:
state {
phase: TEST
}
layer {
name: "val-data"
type: "Data"
top: "data"
top: "label"
include {
phase: TEST
}
transform_param {
crop_size: 227
mean_file: "/mnt/bigdisk/DIGITS-AMB-2/digits/jobs/20210419-113214-d311/mean.binaryproto"
}
data_param {
source: "/mnt/bigdisk/DIGITS-AMB-2/digits/jobs/20210419-113214-d311/val_db"
batch_size: 32
backend: LMDB
}
}
layer {
name: "conv1"
type: "Convolution"
bottom: "data"
top: "conv1"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 96
kernel_size: 11
stride: 4
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0
}
}
}
layer {
name: "relu1"
type: "ReLU"
bottom: "conv1"
top: "conv1"
}
layer {
name: "norm1"
type: "LRN"
bottom: "conv1"
top: "norm1"
lrn_param {
local_size: 5
alpha: 0.0001
beta: 0.75
}
}
layer {
name: "pool1"
type: "Pooling"
bottom: "norm1"
top: "pool1"
pooling_param {
pool: MAX
kernel_size: 3
stride: 2
}
}
layer {
name: "conv2"
type: "Convolution"
bottom: "pool1"
top: "conv2"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 256
pad: 2
kernel_size: 5
group: 2
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
name: "relu2"
type: "ReLU"
bottom: "conv2"
top: "conv2"
}
layer {
name: "norm2"
type: "LRN"
bottom: "conv2"
top: "norm2"
lrn_param {
local_size: 5
alpha: 0.0001
beta: 0.75
}
}
layer {
name: "pool2"
type: "Pooling"
bottom: "norm2"
top: "pool2"
pooling_param {
pool: MAX
kernel_size: 3
stride: 2
}
}
layer {
name: "conv3"
type: "Convolution"
bottom: "pool2"
top: "conv3"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 384
pad: 1
kernel_size: 3
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0
}
}
}
layer {
name: "relu3"
type: "ReLU"
bottom: "conv3"
top: "conv3"
}
layer {
name: "conv3.5"
type: "Convolution"
bottom: "conv3"
top: "conv3.5"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 384
pad: 1
kernel_size: 3
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0
}
}
}
layer {
name: "relu3.5"
type: "ReLU"
bottom: "conv3.5"
top: "conv3.5"
}
layer {
name: "conv4"
type: "Convolution"
bottom: "conv3.5"
top: "conv4"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 384
pad: 1
kernel_size: 3
group: 2
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
name: "relu4"
type: "ReLU"
bottom: "conv4"
top: "conv4"
}
layer {
name: "conv5"
type: "Convolution"
bottom: "conv4"
top: "conv5"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
convolution_param {
num_output: 256
pad: 1
kernel_size: 3
group: 2
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
name: "relu5"
type: "ReLU"
bottom: "conv5"
top: "conv5"
}
layer {
name: "pool5"
type: "Pooling"
bottom: "conv5"
top: "pool5"
pooling_param {
pool: MAX
kernel_size: 3
stride: 2
}
}
layer {
name: "fc6"
type: "InnerProduct"
bottom: "pool5"
top: "fc6"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
inner_product_param {
num_output: 4096
weight_filler {
type: "gaussian"
std: 0.005
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
name: "relu6"
type: "ReLU"
bottom: "fc6"
top: "fc6"
}
layer {
name: "drop6"
type: "Dropout"
bottom: "fc6"
top: "fc6"
dropout_param {
dropout_ratio: 0.5
}
}
layer {
name: "fc7"
type: "InnerProduct"
bottom: "fc6"
top: "fc7"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
inner_product_param {
num_output: 4096
weight_filler {
type: "gaussian"
std: 0.005
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
name: "relu7"
type: "ReLU"
bottom: "fc7"
top: "fc7"
}
layer {
name: "drop7"
type: "Dropout"
bottom: "fc7"
top: "fc7"
dropout_param {
dropout_ratio: 0.5
}
}
layer {
name: "fc8"
type: "InnerProduct"
bottom: "fc7"
top: "fc8"
param {
lr_mult: 1
decay_mult: 1
}
param {
lr_mult: 2
decay_mult: 0
}
inner_product_param {
num_output: 196
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value: 0
}
}
}
layer {
name: "accuracy"
type: "Accuracy"
bottom: "fc8"
bottom: "label"
top: "accuracy"
include {
phase: TEST
}
}
layer {
name: "loss"
type: "SoftmaxWithLoss"
bottom: "fc8"
bottom: "label"
top: "loss"
}
I0428 12:51:25.395591  9322 layer_factory.hpp:77] Creating layer val-data
I0428 12:51:25.443591  9322 db_lmdb.cpp:35] Opened lmdb /mnt/bigdisk/DIGITS-AMB-2/digits/jobs/20210419-113214-d311/val_db
I0428 12:51:25.480090  9322 net.cpp:84] Creating Layer val-data
I0428 12:51:25.480126  9322 net.cpp:380] val-data -> data
I0428 12:51:25.480141  9322 net.cpp:380] val-data -> label
I0428 12:51:25.480152  9322 data_transformer.cpp:25] Loading mean file from: /mnt/bigdisk/DIGITS-AMB-2/digits/jobs/20210419-113214-d311/mean.binaryproto
I0428 12:51:25.489228  9322 data_layer.cpp:45] output data size: 32,3,227,227
I0428 12:51:25.522578  9322 net.cpp:122] Setting up val-data
I0428 12:51:25.522604  9322 net.cpp:129] Top shape: 32 3 227 227 (4946784)
I0428 12:51:25.522608  9322 net.cpp:129] Top shape: 32 (32)
I0428 12:51:25.522611  9322 net.cpp:137] Memory required for data: 19787264
I0428 12:51:25.522617  9322 layer_factory.hpp:77] Creating layer label_val-data_1_split
I0428 12:51:25.522627  9322 net.cpp:84] Creating Layer label_val-data_1_split
I0428 12:51:25.522631  9322 net.cpp:406] label_val-data_1_split <- label
I0428 12:51:25.522637  9322 net.cpp:380] label_val-data_1_split -> label_val-data_1_split_0
I0428 12:51:25.522645  9322 net.cpp:380] label_val-data_1_split -> label_val-data_1_split_1
I0428 12:51:25.522689  9322 net.cpp:122] Setting up label_val-data_1_split
I0428 12:51:25.522694  9322 net.cpp:129] Top shape: 32 (32)
I0428 12:51:25.522697  9322 net.cpp:129] Top shape: 32 (32)
I0428 12:51:25.522699  9322 net.cpp:137] Memory required for data: 19787520
I0428 12:51:25.522722  9322 layer_factory.hpp:77] Creating layer conv1
I0428 12:51:25.522734  9322 net.cpp:84] Creating Layer conv1
I0428 12:51:25.522737  9322 net.cpp:406] conv1 <- data
I0428 12:51:25.522742  9322 net.cpp:380] conv1 -> conv1
I0428 12:51:25.526986  9322 net.cpp:122] Setting up conv1
I0428 12:51:25.526998  9322 net.cpp:129] Top shape: 32 96 55 55 (9292800)
I0428 12:51:25.527001  9322 net.cpp:137] Memory required for data: 56958720
I0428 12:51:25.527011  9322 layer_factory.hpp:77] Creating layer relu1
I0428 12:51:25.527019  9322 net.cpp:84] Creating Layer relu1
I0428 12:51:25.527021  9322 net.cpp:406] relu1 <- conv1
I0428 12:51:25.527026  9322 net.cpp:367] relu1 -> conv1 (in-place)
I0428 12:51:25.527511  9322 net.cpp:122] Setting up relu1
I0428 12:51:25.527520  9322 net.cpp:129] Top shape: 32 96 55 55 (9292800)
I0428 12:51:25.527523  9322 net.cpp:137] Memory required for data: 94129920
I0428 12:51:25.527526  9322 layer_factory.hpp:77] Creating layer norm1
I0428 12:51:25.527534  9322 net.cpp:84] Creating Layer norm1
I0428 12:51:25.527537  9322 net.cpp:406] norm1 <- conv1
I0428 12:51:25.527542  9322 net.cpp:380] norm1 -> norm1
I0428 12:51:25.527866  9322 net.cpp:122] Setting up norm1
I0428 12:51:25.527876  9322 net.cpp:129] Top shape: 32 96 55 55 (9292800)
I0428 12:51:25.527878  9322 net.cpp:137] Memory required for data: 131301120
I0428 12:51:25.527882  9322 layer_factory.hpp:77] Creating layer pool1
I0428 12:51:25.527887  9322 net.cpp:84] Creating Layer pool1
I0428 12:51:25.527890  9322 net.cpp:406] pool1 <- norm1
I0428 12:51:25.527895  9322 net.cpp:380] pool1 -> pool1
I0428 12:51:25.527920  9322 net.cpp:122] Setting up pool1
I0428 12:51:25.527925  9322 net.cpp:129] Top shape: 32 96 27 27 (2239488)
I0428 12:51:25.527926  9322 net.cpp:137] Memory required for data: 140259072
I0428 12:51:25.527930  9322 layer_factory.hpp:77] Creating layer conv2
I0428 12:51:25.527937  9322 net.cpp:84] Creating Layer conv2
I0428 12:51:25.527940  9322 net.cpp:406] conv2 <- pool1
I0428 12:51:25.527945  9322 net.cpp:380] conv2 -> conv2
I0428 12:51:25.536059  9322 net.cpp:122] Setting up conv2
I0428 12:51:25.536072  9322 net.cpp:129] Top shape: 32 256 27 27 (5971968)
I0428 12:51:25.536075  9322 net.cpp:137] Memory required for data: 164146944
I0428 12:51:25.536085  9322 layer_factory.hpp:77] Creating layer relu2
I0428 12:51:25.536092  9322 net.cpp:84] Creating Layer relu2
I0428 12:51:25.536095  9322 net.cpp:406] relu2 <- conv2
I0428 12:51:25.536103  9322 net.cpp:367] relu2 -> conv2 (in-place)
I0428 12:51:25.537951  9322 net.cpp:122] Setting up relu2
I0428 12:51:25.537961  9322 net.cpp:129] Top shape: 32 256 27 27 (5971968)
I0428 12:51:25.537964  9322 net.cpp:137] Memory required for data: 188034816
I0428 12:51:25.537967  9322 layer_factory.hpp:77] Creating layer norm2
I0428 12:51:25.537978  9322 net.cpp:84] Creating Layer norm2
I0428 12:51:25.537981  9322 net.cpp:406] norm2 <- conv2
I0428 12:51:25.537987  9322 net.cpp:380] norm2 -> norm2
I0428 12:51:25.538576  9322 net.cpp:122] Setting up norm2
I0428 12:51:25.538586  9322 net.cpp:129] Top shape: 32 256 27 27 (5971968)
I0428 12:51:25.538589  9322 net.cpp:137] Memory required for data: 211922688
I0428 12:51:25.538594  9322 layer_factory.hpp:77] Creating layer pool2
I0428 12:51:25.538609  9322 net.cpp:84] Creating Layer pool2
I0428 12:51:25.538611  9322 net.cpp:406] pool2 <- norm2
I0428 12:51:25.538619  9322 net.cpp:380] pool2 -> pool2
I0428 12:51:25.538647  9322 net.cpp:122] Setting up pool2
I0428 12:51:25.538655  9322 net.cpp:129] Top shape: 32 256 13 13 (1384448)
I0428 12:51:25.538658  9322 net.cpp:137] Memory required for data: 217460480
I0428 12:51:25.538661  9322 layer_factory.hpp:77] Creating layer conv3
I0428 12:51:25.538669  9322 net.cpp:84] Creating Layer conv3
I0428 12:51:25.538672  9322 net.cpp:406] conv3 <- pool2
I0428 12:51:25.538678  9322 net.cpp:380] conv3 -> conv3
I0428 12:51:25.549216  9322 net.cpp:122] Setting up conv3
I0428 12:51:25.549234  9322 net.cpp:129] Top shape: 32 384 13 13 (2076672)
I0428 12:51:25.549237  9322 net.cpp:137] Memory required for data: 225767168
I0428 12:51:25.549268  9322 layer_factory.hpp:77] Creating layer relu3
I0428 12:51:25.549278  9322 net.cpp:84] Creating Layer relu3
I0428 12:51:25.549280  9322 net.cpp:406] relu3 <- conv3
I0428 12:51:25.549286  9322 net.cpp:367] relu3 -> conv3 (in-place)
I0428 12:51:25.549845  9322 net.cpp:122] Setting up relu3
I0428 12:51:25.549854  9322 net.cpp:129] Top shape: 32 384 13 13 (2076672)
I0428 12:51:25.549857  9322 net.cpp:137] Memory required for data: 234073856
I0428 12:51:25.549860  9322 layer_factory.hpp:77] Creating layer conv3.5
I0428 12:51:25.549872  9322 net.cpp:84] Creating Layer conv3.5
I0428 12:51:25.549876  9322 net.cpp:406] conv3.5 <- conv3
I0428 12:51:25.549883  9322 net.cpp:380] conv3.5 -> conv3.5
I0428 12:51:25.564728  9322 net.cpp:122] Setting up conv3.5
I0428 12:51:25.564747  9322 net.cpp:129] Top shape: 32 384 13 13 (2076672)
I0428 12:51:25.564750  9322 net.cpp:137] Memory required for data: 242380544
I0428 12:51:25.564759  9322 layer_factory.hpp:77] Creating layer relu3.5
I0428 12:51:25.564765  9322 net.cpp:84] Creating Layer relu3.5
I0428 12:51:25.564769  9322 net.cpp:406] relu3.5 <- conv3.5
I0428 12:51:25.564776  9322 net.cpp:367] relu3.5 -> conv3.5 (in-place)
I0428 12:51:25.565343  9322 net.cpp:122] Setting up relu3.5
I0428 12:51:25.565352  9322 net.cpp:129] Top shape: 32 384 13 13 (2076672)
I0428 12:51:25.565356  9322 net.cpp:137] Memory required for data: 250687232
I0428 12:51:25.565358  9322 layer_factory.hpp:77] Creating layer conv4
I0428 12:51:25.565369  9322 net.cpp:84] Creating Layer conv4
I0428 12:51:25.565373  9322 net.cpp:406] conv4 <- conv3.5
I0428 12:51:25.565379  9322 net.cpp:380] conv4 -> conv4
I0428 12:51:25.577142  9322 net.cpp:122] Setting up conv4
I0428 12:51:25.577162  9322 net.cpp:129] Top shape: 32 384 13 13 (2076672)
I0428 12:51:25.577165  9322 net.cpp:137] Memory required for data: 258993920
I0428 12:51:25.577178  9322 layer_factory.hpp:77] Creating layer relu4
I0428 12:51:25.577186  9322 net.cpp:84] Creating Layer relu4
I0428 12:51:25.577190  9322 net.cpp:406] relu4 <- conv4
I0428 12:51:25.577196  9322 net.cpp:367] relu4 -> conv4 (in-place)
I0428 12:51:25.577764  9322 net.cpp:122] Setting up relu4
I0428 12:51:25.577773  9322 net.cpp:129] Top shape: 32 384 13 13 (2076672)
I0428 12:51:25.577775  9322 net.cpp:137] Memory required for data: 267300608
I0428 12:51:25.577780  9322 layer_factory.hpp:77] Creating layer conv5
I0428 12:51:25.577793  9322 net.cpp:84] Creating Layer conv5
I0428 12:51:25.577797  9322 net.cpp:406] conv5 <- conv4
I0428 12:51:25.577805  9322 net.cpp:380] conv5 -> conv5
I0428 12:51:25.587442  9322 net.cpp:122] Setting up conv5
I0428 12:51:25.587463  9322 net.cpp:129] Top shape: 32 256 13 13 (1384448)
I0428 12:51:25.587466  9322 net.cpp:137] Memory required for data: 272838400
I0428 12:51:25.587476  9322 layer_factory.hpp:77] Creating layer relu5
I0428 12:51:25.587484  9322 net.cpp:84] Creating Layer relu5
I0428 12:51:25.587488  9322 net.cpp:406] relu5 <- conv5
I0428 12:51:25.587494  9322 net.cpp:367] relu5 -> conv5 (in-place)
I0428 12:51:25.587884  9322 net.cpp:122] Setting up relu5
I0428 12:51:25.587893  9322 net.cpp:129] Top shape: 32 256 13 13 (1384448)
I0428 12:51:25.587895  9322 net.cpp:137] Memory required for data: 278376192
I0428 12:51:25.587898  9322 layer_factory.hpp:77] Creating layer pool5
I0428 12:51:25.587906  9322 net.cpp:84] Creating Layer pool5
I0428 12:51:25.587909  9322 net.cpp:406] pool5 <- conv5
I0428 12:51:25.587914  9322 net.cpp:380] pool5 -> pool5
I0428 12:51:25.587954  9322 net.cpp:122] Setting up pool5
I0428 12:51:25.587958  9322 net.cpp:129] Top shape: 32 256 6 6 (294912)
I0428 12:51:25.587961  9322 net.cpp:137] Memory required for data: 279555840
I0428 12:51:25.587963  9322 layer_factory.hpp:77] Creating layer fc6
I0428 12:51:25.587971  9322 net.cpp:84] Creating Layer fc6
I0428 12:51:25.587975  9322 net.cpp:406] fc6 <- pool5
I0428 12:51:25.587978  9322 net.cpp:380] fc6 -> fc6
I0428 12:51:25.946208  9322 net.cpp:122] Setting up fc6
I0428 12:51:25.946228  9322 net.cpp:129] Top shape: 32 4096 (131072)
I0428 12:51:25.946249  9322 net.cpp:137] Memory required for data: 280080128
I0428 12:51:25.946259  9322 layer_factory.hpp:77] Creating layer relu6
I0428 12:51:25.946269  9322 net.cpp:84] Creating Layer relu6
I0428 12:51:25.946272  9322 net.cpp:406] relu6 <- fc6
I0428 12:51:25.946277  9322 net.cpp:367] relu6 -> fc6 (in-place)
I0428 12:51:25.947046  9322 net.cpp:122] Setting up relu6
I0428 12:51:25.947057  9322 net.cpp:129] Top shape: 32 4096 (131072)
I0428 12:51:25.947058  9322 net.cpp:137] Memory required for data: 280604416
I0428 12:51:25.947062  9322 layer_factory.hpp:77] Creating layer drop6
I0428 12:51:25.947068  9322 net.cpp:84] Creating Layer drop6
I0428 12:51:25.947072  9322 net.cpp:406] drop6 <- fc6
I0428 12:51:25.947078  9322 net.cpp:367] drop6 -> fc6 (in-place)
I0428 12:51:25.947103  9322 net.cpp:122] Setting up drop6
I0428 12:51:25.947108  9322 net.cpp:129] Top shape: 32 4096 (131072)
I0428 12:51:25.947111  9322 net.cpp:137] Memory required for data: 281128704
I0428 12:51:25.947113  9322 layer_factory.hpp:77] Creating layer fc7
I0428 12:51:25.947120  9322 net.cpp:84] Creating Layer fc7
I0428 12:51:25.947124  9322 net.cpp:406] fc7 <- fc6
I0428 12:51:25.947129  9322 net.cpp:380] fc7 -> fc7
I0428 12:51:26.106307  9322 net.cpp:122] Setting up fc7
I0428 12:51:26.106328  9322 net.cpp:129] Top shape: 32 4096 (131072)
I0428 12:51:26.106331  9322 net.cpp:137] Memory required for data: 281652992
I0428 12:51:26.106340  9322 layer_factory.hpp:77] Creating layer relu7
I0428 12:51:26.106348  9322 net.cpp:84] Creating Layer relu7
I0428 12:51:26.106353  9322 net.cpp:406] relu7 <- fc7
I0428 12:51:26.106360  9322 net.cpp:367] relu7 -> fc7 (in-place)
I0428 12:51:26.107129  9322 net.cpp:122] Setting up relu7
I0428 12:51:26.107139  9322 net.cpp:129] Top shape: 32 4096 (131072)
I0428 12:51:26.107142  9322 net.cpp:137] Memory required for data: 282177280
I0428 12:51:26.107146  9322 layer_factory.hpp:77] Creating layer drop7
I0428 12:51:26.107152  9322 net.cpp:84] Creating Layer drop7
I0428 12:51:26.107156  9322 net.cpp:406] drop7 <- fc7
I0428 12:51:26.107161  9322 net.cpp:367] drop7 -> fc7 (in-place)
I0428 12:51:26.107185  9322 net.cpp:122] Setting up drop7
I0428 12:51:26.107190  9322 net.cpp:129] Top shape: 32 4096 (131072)
I0428 12:51:26.107192  9322 net.cpp:137] Memory required for data: 282701568
I0428 12:51:26.107195  9322 layer_factory.hpp:77] Creating layer fc8
I0428 12:51:26.107203  9322 net.cpp:84] Creating Layer fc8
I0428 12:51:26.107206  9322 net.cpp:406] fc8 <- fc7
I0428 12:51:26.107210  9322 net.cpp:380] fc8 -> fc8
I0428 12:51:26.114934  9322 net.cpp:122] Setting up fc8
I0428 12:51:26.114943  9322 net.cpp:129] Top shape: 32 196 (6272)
I0428 12:51:26.114946  9322 net.cpp:137] Memory required for data: 282726656
I0428 12:51:26.114957  9322 layer_factory.hpp:77] Creating layer fc8_fc8_0_split
I0428 12:51:26.114964  9322 net.cpp:84] Creating Layer fc8_fc8_0_split
I0428 12:51:26.114966  9322 net.cpp:406] fc8_fc8_0_split <- fc8
I0428 12:51:26.114972  9322 net.cpp:380] fc8_fc8_0_split -> fc8_fc8_0_split_0
I0428 12:51:26.114979  9322 net.cpp:380] fc8_fc8_0_split -> fc8_fc8_0_split_1
I0428 12:51:26.115007  9322 net.cpp:122] Setting up fc8_fc8_0_split
I0428 12:51:26.115011  9322 net.cpp:129] Top shape: 32 196 (6272)
I0428 12:51:26.115015  9322 net.cpp:129] Top shape: 32 196 (6272)
I0428 12:51:26.115016  9322 net.cpp:137] Memory required for data: 282776832
I0428 12:51:26.115020  9322 layer_factory.hpp:77] Creating layer accuracy
I0428 12:51:26.115026  9322 net.cpp:84] Creating Layer accuracy
I0428 12:51:26.115029  9322 net.cpp:406] accuracy <- fc8_fc8_0_split_0
I0428 12:51:26.115033  9322 net.cpp:406] accuracy <- label_val-data_1_split_0
I0428 12:51:26.115038  9322 net.cpp:380] accuracy -> accuracy
I0428 12:51:26.115046  9322 net.cpp:122] Setting up accuracy
I0428 12:51:26.115048  9322 net.cpp:129] Top shape: (1)
I0428 12:51:26.115051  9322 net.cpp:137] Memory required for data: 282776836
I0428 12:51:26.115053  9322 layer_factory.hpp:77] Creating layer loss
I0428 12:51:26.115058  9322 net.cpp:84] Creating Layer loss
I0428 12:51:26.115080  9322 net.cpp:406] loss <- fc8_fc8_0_split_1
I0428 12:51:26.115084  9322 net.cpp:406] loss <- label_val-data_1_split_1
I0428 12:51:26.115088  9322 net.cpp:380] loss -> loss
I0428 12:51:26.115095  9322 layer_factory.hpp:77] Creating layer loss
I0428 12:51:26.115778  9322 net.cpp:122] Setting up loss
I0428 12:51:26.115788  9322 net.cpp:129] Top shape: (1)
I0428 12:51:26.115792  9322 net.cpp:132]     with loss weight 1
I0428 12:51:26.115800  9322 net.cpp:137] Memory required for data: 282776840
I0428 12:51:26.115804  9322 net.cpp:198] loss needs backward computation.
I0428 12:51:26.115808  9322 net.cpp:200] accuracy does not need backward computation.
I0428 12:51:26.115813  9322 net.cpp:198] fc8_fc8_0_split needs backward computation.
I0428 12:51:26.115815  9322 net.cpp:198] fc8 needs backward computation.
I0428 12:51:26.115818  9322 net.cpp:198] drop7 needs backward computation.
I0428 12:51:26.115820  9322 net.cpp:198] relu7 needs backward computation.
I0428 12:51:26.115823  9322 net.cpp:198] fc7 needs backward computation.
I0428 12:51:26.115825  9322 net.cpp:198] drop6 needs backward computation.
I0428 12:51:26.115828  9322 net.cpp:198] relu6 needs backward computation.
I0428 12:51:26.115830  9322 net.cpp:198] fc6 needs backward computation.
I0428 12:51:26.115833  9322 net.cpp:198] pool5 needs backward computation.
I0428 12:51:26.115836  9322 net.cpp:198] relu5 needs backward computation.
I0428 12:51:26.115839  9322 net.cpp:198] conv5 needs backward computation.
I0428 12:51:26.115842  9322 net.cpp:198] relu4 needs backward computation.
I0428 12:51:26.115844  9322 net.cpp:198] conv4 needs backward computation.
I0428 12:51:26.115847  9322 net.cpp:198] relu3.5 needs backward computation.
I0428 12:51:26.115850  9322 net.cpp:198] conv3.5 needs backward computation.
I0428 12:51:26.115854  9322 net.cpp:198] relu3 needs backward computation.
I0428 12:51:26.115855  9322 net.cpp:198] conv3 needs backward computation.
I0428 12:51:26.115859  9322 net.cpp:198] pool2 needs backward computation.
I0428 12:51:26.115861  9322 net.cpp:198] norm2 needs backward computation.
I0428 12:51:26.115864  9322 net.cpp:198] relu2 needs backward computation.
I0428 12:51:26.115867  9322 net.cpp:198] conv2 needs backward computation.
I0428 12:51:26.115870  9322 net.cpp:198] pool1 needs backward computation.
I0428 12:51:26.115873  9322 net.cpp:198] norm1 needs backward computation.
I0428 12:51:26.115876  9322 net.cpp:198] relu1 needs backward computation.
I0428 12:51:26.115878  9322 net.cpp:198] conv1 needs backward computation.
I0428 12:51:26.115882  9322 net.cpp:200] label_val-data_1_split does not need backward computation.
I0428 12:51:26.115886  9322 net.cpp:200] val-data does not need backward computation.
I0428 12:51:26.115890  9322 net.cpp:242] This network produces output accuracy
I0428 12:51:26.115892  9322 net.cpp:242] This network produces output loss
I0428 12:51:26.115909  9322 net.cpp:255] Network initialization done.
I0428 12:51:26.115981  9322 solver.cpp:56] Solver scaffolding done.
I0428 12:51:26.116367  9322 caffe.cpp:248] Starting Optimization
I0428 12:51:26.116375  9322 solver.cpp:272] Solving
I0428 12:51:26.116379  9322 solver.cpp:273] Learning Rate Policy: exp
I0428 12:51:26.118264  9322 solver.cpp:330] Iteration 0, Testing net (#0)
I0428 12:51:26.118274  9322 net.cpp:676] Ignoring source layer train-data
I0428 12:51:26.215270  9322 blocking_queue.cpp:49] Waiting for data
I0428 12:51:31.082444  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 12:51:31.134616  9322 solver.cpp:397]     Test net output #0: accuracy = 0.00245098
I0428 12:51:31.134651  9322 solver.cpp:397]     Test net output #1: loss = 5.27867 (* 1 = 5.27867 loss)
I0428 12:51:31.253636  9322 solver.cpp:218] Iteration 0 (0 iter/s, 5.13723s/12 iters), loss = 5.28336
I0428 12:51:31.255275  9322 solver.cpp:237]     Train net output #0: loss = 5.28336 (* 1 = 5.28336 loss)
I0428 12:51:31.255298  9322 sgd_solver.cpp:105] Iteration 0, lr = 0.01
I0428 12:51:35.366441  9322 solver.cpp:218] Iteration 12 (2.91888 iter/s, 4.11117s/12 iters), loss = 5.30195
I0428 12:51:35.366504  9322 solver.cpp:237]     Train net output #0: loss = 5.30195 (* 1 = 5.30195 loss)
I0428 12:51:35.366513  9322 sgd_solver.cpp:105] Iteration 12, lr = 0.00997626
I0428 12:51:43.435333  9322 solver.cpp:218] Iteration 24 (1.4872 iter/s, 8.06885s/12 iters), loss = 5.30481
I0428 12:51:43.435376  9322 solver.cpp:237]     Train net output #0: loss = 5.30481 (* 1 = 5.30481 loss)
I0428 12:51:43.435385  9322 sgd_solver.cpp:105] Iteration 24, lr = 0.00995257
I0428 12:52:11.539474  9322 solver.cpp:218] Iteration 36 (0.426982 iter/s, 28.1042s/12 iters), loss = 5.3199
I0428 12:52:11.539563  9322 solver.cpp:237]     Train net output #0: loss = 5.3199 (* 1 = 5.3199 loss)
I0428 12:52:11.539572  9322 sgd_solver.cpp:105] Iteration 36, lr = 0.00992894
I0428 12:52:17.020535  9322 solver.cpp:218] Iteration 48 (2.18939 iter/s, 5.48098s/12 iters), loss = 5.31497
I0428 12:52:17.020576  9322 solver.cpp:237]     Train net output #0: loss = 5.31497 (* 1 = 5.31497 loss)
I0428 12:52:17.020583  9322 sgd_solver.cpp:105] Iteration 48, lr = 0.00990537
I0428 12:52:22.461477  9322 solver.cpp:218] Iteration 60 (2.20551 iter/s, 5.44091s/12 iters), loss = 5.2794
I0428 12:52:22.461522  9322 solver.cpp:237]     Train net output #0: loss = 5.2794 (* 1 = 5.2794 loss)
I0428 12:52:22.461530  9322 sgd_solver.cpp:105] Iteration 60, lr = 0.00988185
I0428 12:52:28.003118  9322 solver.cpp:218] Iteration 72 (2.16544 iter/s, 5.5416s/12 iters), loss = 5.29524
I0428 12:52:28.003165  9322 solver.cpp:237]     Train net output #0: loss = 5.29524 (* 1 = 5.29524 loss)
I0428 12:52:28.003172  9322 sgd_solver.cpp:105] Iteration 72, lr = 0.00985839
I0428 12:52:33.360888  9322 solver.cpp:218] Iteration 84 (2.23975 iter/s, 5.35773s/12 iters), loss = 5.29212
I0428 12:52:33.360930  9322 solver.cpp:237]     Train net output #0: loss = 5.29212 (* 1 = 5.29212 loss)
I0428 12:52:33.360940  9322 sgd_solver.cpp:105] Iteration 84, lr = 0.00983498
I0428 12:52:41.857098  9322 solver.cpp:218] Iteration 96 (1.4124 iter/s, 8.49619s/12 iters), loss = 5.31094
I0428 12:52:41.857223  9322 solver.cpp:237]     Train net output #0: loss = 5.31094 (* 1 = 5.31094 loss)
I0428 12:52:41.857231  9322 sgd_solver.cpp:105] Iteration 96, lr = 0.00981163
I0428 12:52:43.851642  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 12:52:44.164657  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_102.caffemodel
I0428 12:52:47.969888  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_102.solverstate
I0428 12:52:51.027302  9322 solver.cpp:330] Iteration 102, Testing net (#0)
I0428 12:52:51.027320  9322 net.cpp:676] Ignoring source layer train-data
I0428 12:52:55.668686  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 12:52:55.753059  9322 solver.cpp:397]     Test net output #0: accuracy = 0.00551471
I0428 12:52:55.753093  9322 solver.cpp:397]     Test net output #1: loss = 5.28935 (* 1 = 5.28935 loss)
I0428 12:52:57.633689  9322 solver.cpp:218] Iteration 108 (0.760624 iter/s, 15.7765s/12 iters), loss = 5.30693
I0428 12:52:57.633735  9322 solver.cpp:237]     Train net output #0: loss = 5.30693 (* 1 = 5.30693 loss)
I0428 12:52:57.633744  9322 sgd_solver.cpp:105] Iteration 108, lr = 0.00978834
I0428 12:53:02.986394  9322 solver.cpp:218] Iteration 120 (2.24187 iter/s, 5.35266s/12 iters), loss = 5.27494
I0428 12:53:02.986439  9322 solver.cpp:237]     Train net output #0: loss = 5.27494 (* 1 = 5.27494 loss)
I0428 12:53:02.986447  9322 sgd_solver.cpp:105] Iteration 120, lr = 0.0097651
I0428 12:53:08.357836  9322 solver.cpp:218] Iteration 132 (2.23405 iter/s, 5.3714s/12 iters), loss = 5.29199
I0428 12:53:08.357879  9322 solver.cpp:237]     Train net output #0: loss = 5.29199 (* 1 = 5.29199 loss)
I0428 12:53:08.357887  9322 sgd_solver.cpp:105] Iteration 132, lr = 0.00974192
I0428 12:53:13.621965  9322 solver.cpp:218] Iteration 144 (2.27959 iter/s, 5.2641s/12 iters), loss = 5.30266
I0428 12:53:13.622121  9322 solver.cpp:237]     Train net output #0: loss = 5.30266 (* 1 = 5.30266 loss)
I0428 12:53:13.622130  9322 sgd_solver.cpp:105] Iteration 144, lr = 0.00971879
I0428 12:53:18.781347  9322 solver.cpp:218] Iteration 156 (2.32593 iter/s, 5.15924s/12 iters), loss = 5.27995
I0428 12:53:18.781391  9322 solver.cpp:237]     Train net output #0: loss = 5.27995 (* 1 = 5.27995 loss)
I0428 12:53:18.781399  9322 sgd_solver.cpp:105] Iteration 156, lr = 0.00969571
I0428 12:53:23.879576  9322 solver.cpp:218] Iteration 168 (2.35378 iter/s, 5.09819s/12 iters), loss = 5.30787
I0428 12:53:23.879622  9322 solver.cpp:237]     Train net output #0: loss = 5.30787 (* 1 = 5.30787 loss)
I0428 12:53:23.879631  9322 sgd_solver.cpp:105] Iteration 168, lr = 0.00967269
I0428 12:53:29.018390  9322 solver.cpp:218] Iteration 180 (2.33519 iter/s, 5.13876s/12 iters), loss = 5.30892
I0428 12:53:29.018468  9322 solver.cpp:237]     Train net output #0: loss = 5.30892 (* 1 = 5.30892 loss)
I0428 12:53:29.018484  9322 sgd_solver.cpp:105] Iteration 180, lr = 0.00964973
I0428 12:53:34.175382  9322 solver.cpp:218] Iteration 192 (2.32697 iter/s, 5.15693s/12 iters), loss = 5.27979
I0428 12:53:34.175424  9322 solver.cpp:237]     Train net output #0: loss = 5.27979 (* 1 = 5.27979 loss)
I0428 12:53:34.175432  9322 sgd_solver.cpp:105] Iteration 192, lr = 0.00962682
I0428 12:53:38.220278  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 12:53:38.941408  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_204.caffemodel
I0428 12:53:42.436168  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_204.solverstate
I0428 12:53:45.562937  9322 solver.cpp:330] Iteration 204, Testing net (#0)
I0428 12:53:45.563055  9322 net.cpp:676] Ignoring source layer train-data
I0428 12:53:50.404825  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 12:53:50.552712  9322 solver.cpp:397]     Test net output #0: accuracy = 0.00551471
I0428 12:53:50.552747  9322 solver.cpp:397]     Test net output #1: loss = 5.28618 (* 1 = 5.28618 loss)
I0428 12:53:50.670783  9322 solver.cpp:218] Iteration 204 (0.727475 iter/s, 16.4954s/12 iters), loss = 5.30575
I0428 12:53:50.670831  9322 solver.cpp:237]     Train net output #0: loss = 5.30575 (* 1 = 5.30575 loss)
I0428 12:53:50.670840  9322 sgd_solver.cpp:105] Iteration 204, lr = 0.00960396
I0428 12:53:54.979787  9322 solver.cpp:218] Iteration 216 (2.78489 iter/s, 4.30897s/12 iters), loss = 5.28341
I0428 12:53:54.979825  9322 solver.cpp:237]     Train net output #0: loss = 5.28341 (* 1 = 5.28341 loss)
I0428 12:53:54.979831  9322 sgd_solver.cpp:105] Iteration 216, lr = 0.00958116
I0428 12:54:00.000804  9322 solver.cpp:218] Iteration 228 (2.38997 iter/s, 5.02099s/12 iters), loss = 5.27117
I0428 12:54:00.000857  9322 solver.cpp:237]     Train net output #0: loss = 5.27117 (* 1 = 5.27117 loss)
I0428 12:54:00.000866  9322 sgd_solver.cpp:105] Iteration 228, lr = 0.00955841
I0428 12:54:05.136368  9322 solver.cpp:218] Iteration 240 (2.33666 iter/s, 5.13553s/12 iters), loss = 5.2705
I0428 12:54:05.136401  9322 solver.cpp:237]     Train net output #0: loss = 5.2705 (* 1 = 5.2705 loss)
I0428 12:54:05.136409  9322 sgd_solver.cpp:105] Iteration 240, lr = 0.00953572
I0428 12:54:10.302875  9322 solver.cpp:218] Iteration 252 (2.32267 iter/s, 5.16647s/12 iters), loss = 5.26757
I0428 12:54:10.302918  9322 solver.cpp:237]     Train net output #0: loss = 5.26757 (* 1 = 5.26757 loss)
I0428 12:54:10.302927  9322 sgd_solver.cpp:105] Iteration 252, lr = 0.00951308
I0428 12:54:15.468379  9322 solver.cpp:218] Iteration 264 (2.32312 iter/s, 5.16547s/12 iters), loss = 5.26087
I0428 12:54:15.468420  9322 solver.cpp:237]     Train net output #0: loss = 5.26087 (* 1 = 5.26087 loss)
I0428 12:54:15.468427  9322 sgd_solver.cpp:105] Iteration 264, lr = 0.00949049
I0428 12:54:20.621542  9322 solver.cpp:218] Iteration 276 (2.32868 iter/s, 5.15314s/12 iters), loss = 5.25937
I0428 12:54:20.621685  9322 solver.cpp:237]     Train net output #0: loss = 5.25937 (* 1 = 5.25937 loss)
I0428 12:54:20.621693  9322 sgd_solver.cpp:105] Iteration 276, lr = 0.00946796
I0428 12:54:25.758224  9322 solver.cpp:218] Iteration 288 (2.3362 iter/s, 5.13655s/12 iters), loss = 5.20204
I0428 12:54:25.758266  9322 solver.cpp:237]     Train net output #0: loss = 5.20204 (* 1 = 5.20204 loss)
I0428 12:54:25.758275  9322 sgd_solver.cpp:105] Iteration 288, lr = 0.00944548
I0428 12:54:30.904516  9322 solver.cpp:218] Iteration 300 (2.33179 iter/s, 5.14626s/12 iters), loss = 5.17693
I0428 12:54:30.904562  9322 solver.cpp:237]     Train net output #0: loss = 5.17693 (* 1 = 5.17693 loss)
I0428 12:54:30.904570  9322 sgd_solver.cpp:105] Iteration 300, lr = 0.00942305
I0428 12:54:31.935484  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 12:54:33.031719  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_306.caffemodel
I0428 12:54:36.175866  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_306.solverstate
I0428 12:54:39.177147  9322 solver.cpp:330] Iteration 306, Testing net (#0)
I0428 12:54:39.177168  9322 net.cpp:676] Ignoring source layer train-data
I0428 12:54:44.013622  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 12:54:44.198171  9322 solver.cpp:397]     Test net output #0: accuracy = 0.00796569
I0428 12:54:44.198220  9322 solver.cpp:397]     Test net output #1: loss = 5.18956 (* 1 = 5.18956 loss)
I0428 12:54:46.153208  9322 solver.cpp:218] Iteration 312 (0.786952 iter/s, 15.2487s/12 iters), loss = 5.24605
I0428 12:54:46.153251  9322 solver.cpp:237]     Train net output #0: loss = 5.24605 (* 1 = 5.24605 loss)
I0428 12:54:46.153259  9322 sgd_solver.cpp:105] Iteration 312, lr = 0.00940068
I0428 12:54:51.350975  9322 solver.cpp:218] Iteration 324 (2.3087 iter/s, 5.19773s/12 iters), loss = 5.18082
I0428 12:54:51.351094  9322 solver.cpp:237]     Train net output #0: loss = 5.18082 (* 1 = 5.18082 loss)
I0428 12:54:51.351104  9322 sgd_solver.cpp:105] Iteration 324, lr = 0.00937836
I0428 12:54:56.509254  9322 solver.cpp:218] Iteration 336 (2.32641 iter/s, 5.15817s/12 iters), loss = 5.15324
I0428 12:54:56.509292  9322 solver.cpp:237]     Train net output #0: loss = 5.15324 (* 1 = 5.15324 loss)
I0428 12:54:56.509301  9322 sgd_solver.cpp:105] Iteration 336, lr = 0.0093561
I0428 12:55:01.602720  9322 solver.cpp:218] Iteration 348 (2.35597 iter/s, 5.09343s/12 iters), loss = 5.17438
I0428 12:55:01.602766  9322 solver.cpp:237]     Train net output #0: loss = 5.17438 (* 1 = 5.17438 loss)
I0428 12:55:01.602773  9322 sgd_solver.cpp:105] Iteration 348, lr = 0.00933388
I0428 12:55:06.764917  9322 solver.cpp:218] Iteration 360 (2.32461 iter/s, 5.16216s/12 iters), loss = 5.14037
I0428 12:55:06.764958  9322 solver.cpp:237]     Train net output #0: loss = 5.14037 (* 1 = 5.14037 loss)
I0428 12:55:06.764966  9322 sgd_solver.cpp:105] Iteration 360, lr = 0.00931172
I0428 12:55:11.934681  9322 solver.cpp:218] Iteration 372 (2.3212 iter/s, 5.16973s/12 iters), loss = 5.11423
I0428 12:55:11.934725  9322 solver.cpp:237]     Train net output #0: loss = 5.11423 (* 1 = 5.11423 loss)
I0428 12:55:11.934734  9322 sgd_solver.cpp:105] Iteration 372, lr = 0.00928961
I0428 12:55:17.019492  9322 solver.cpp:218] Iteration 384 (2.35999 iter/s, 5.08477s/12 iters), loss = 5.13273
I0428 12:55:17.019536  9322 solver.cpp:237]     Train net output #0: loss = 5.13273 (* 1 = 5.13273 loss)
I0428 12:55:17.019543  9322 sgd_solver.cpp:105] Iteration 384, lr = 0.00926756
I0428 12:55:22.117060  9322 solver.cpp:218] Iteration 396 (2.35408 iter/s, 5.09754s/12 iters), loss = 5.19272
I0428 12:55:22.117156  9322 solver.cpp:237]     Train net output #0: loss = 5.19272 (* 1 = 5.19272 loss)
I0428 12:55:22.117166  9322 sgd_solver.cpp:105] Iteration 396, lr = 0.00924556
I0428 12:55:25.275832  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 12:55:26.733183  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_408.caffemodel
I0428 12:55:31.627501  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_408.solverstate
I0428 12:55:35.921540  9322 solver.cpp:330] Iteration 408, Testing net (#0)
I0428 12:55:35.921559  9322 net.cpp:676] Ignoring source layer train-data
I0428 12:55:40.417078  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 12:55:40.636448  9322 solver.cpp:397]     Test net output #0: accuracy = 0.00919118
I0428 12:55:40.636483  9322 solver.cpp:397]     Test net output #1: loss = 5.15122 (* 1 = 5.15122 loss)
I0428 12:55:40.754844  9322 solver.cpp:218] Iteration 408 (0.643854 iter/s, 18.6378s/12 iters), loss = 5.12212
I0428 12:55:40.754897  9322 solver.cpp:237]     Train net output #0: loss = 5.12212 (* 1 = 5.12212 loss)
I0428 12:55:40.754909  9322 sgd_solver.cpp:105] Iteration 408, lr = 0.00922361
I0428 12:55:45.048559  9322 solver.cpp:218] Iteration 420 (2.79481 iter/s, 4.29367s/12 iters), loss = 5.15702
I0428 12:55:45.048601  9322 solver.cpp:237]     Train net output #0: loss = 5.15702 (* 1 = 5.15702 loss)
I0428 12:55:45.048609  9322 sgd_solver.cpp:105] Iteration 420, lr = 0.00920171
I0428 12:55:50.224800  9322 solver.cpp:218] Iteration 432 (2.3183 iter/s, 5.1762s/12 iters), loss = 5.12039
I0428 12:55:50.224843  9322 solver.cpp:237]     Train net output #0: loss = 5.12039 (* 1 = 5.12039 loss)
I0428 12:55:50.224853  9322 sgd_solver.cpp:105] Iteration 432, lr = 0.00917986
I0428 12:55:55.382558  9322 solver.cpp:218] Iteration 444 (2.32661 iter/s, 5.15772s/12 iters), loss = 5.16486
I0428 12:55:55.382727  9322 solver.cpp:237]     Train net output #0: loss = 5.16486 (* 1 = 5.16486 loss)
I0428 12:55:55.382737  9322 sgd_solver.cpp:105] Iteration 444, lr = 0.00915807
I0428 12:56:00.502987  9322 solver.cpp:218] Iteration 456 (2.34362 iter/s, 5.12027s/12 iters), loss = 5.10218
I0428 12:56:00.503031  9322 solver.cpp:237]     Train net output #0: loss = 5.10218 (* 1 = 5.10218 loss)
I0428 12:56:00.503038  9322 sgd_solver.cpp:105] Iteration 456, lr = 0.00913632
I0428 12:56:05.595135  9322 solver.cpp:218] Iteration 468 (2.35659 iter/s, 5.09211s/12 iters), loss = 5.15001
I0428 12:56:05.595175  9322 solver.cpp:237]     Train net output #0: loss = 5.15001 (* 1 = 5.15001 loss)
I0428 12:56:05.595183  9322 sgd_solver.cpp:105] Iteration 468, lr = 0.00911463
I0428 12:56:10.807823  9322 solver.cpp:218] Iteration 480 (2.30209 iter/s, 5.21266s/12 iters), loss = 5.12724
I0428 12:56:10.807868  9322 solver.cpp:237]     Train net output #0: loss = 5.12724 (* 1 = 5.12724 loss)
I0428 12:56:10.807876  9322 sgd_solver.cpp:105] Iteration 480, lr = 0.00909299
I0428 12:56:15.964169  9322 solver.cpp:218] Iteration 492 (2.32724 iter/s, 5.15631s/12 iters), loss = 5.10755
I0428 12:56:15.964212  9322 solver.cpp:237]     Train net output #0: loss = 5.10755 (* 1 = 5.10755 loss)
I0428 12:56:15.964221  9322 sgd_solver.cpp:105] Iteration 492, lr = 0.0090714
I0428 12:56:21.139844  9322 solver.cpp:218] Iteration 504 (2.31855 iter/s, 5.17564s/12 iters), loss = 5.12157
I0428 12:56:21.139884  9322 solver.cpp:237]     Train net output #0: loss = 5.12157 (* 1 = 5.12157 loss)
I0428 12:56:21.139892  9322 sgd_solver.cpp:105] Iteration 504, lr = 0.00904986
I0428 12:56:21.386057  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 12:56:23.229439  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_510.caffemodel
I0428 12:56:32.116015  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_510.solverstate
I0428 12:56:36.058156  9322 solver.cpp:330] Iteration 510, Testing net (#0)
I0428 12:56:36.058176  9322 net.cpp:676] Ignoring source layer train-data
I0428 12:56:40.720386  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 12:56:40.994058  9322 solver.cpp:397]     Test net output #0: accuracy = 0.00919118
I0428 12:56:40.994107  9322 solver.cpp:397]     Test net output #1: loss = 5.10053 (* 1 = 5.10053 loss)
I0428 12:56:42.990705  9322 solver.cpp:218] Iteration 516 (0.549176 iter/s, 21.8509s/12 iters), loss = 5.05685
I0428 12:56:42.990748  9322 solver.cpp:237]     Train net output #0: loss = 5.05685 (* 1 = 5.05685 loss)
I0428 12:56:42.990756  9322 sgd_solver.cpp:105] Iteration 516, lr = 0.00902838
I0428 12:56:48.245772  9322 solver.cpp:218] Iteration 528 (2.28353 iter/s, 5.25503s/12 iters), loss = 5.10399
I0428 12:56:48.245817  9322 solver.cpp:237]     Train net output #0: loss = 5.10399 (* 1 = 5.10399 loss)
I0428 12:56:48.245826  9322 sgd_solver.cpp:105] Iteration 528, lr = 0.00900694
I0428 12:56:53.525332  9322 solver.cpp:218] Iteration 540 (2.27293 iter/s, 5.27952s/12 iters), loss = 5.1301
I0428 12:56:53.525372  9322 solver.cpp:237]     Train net output #0: loss = 5.1301 (* 1 = 5.1301 loss)
I0428 12:56:53.525380  9322 sgd_solver.cpp:105] Iteration 540, lr = 0.00898556
I0428 12:56:58.715890  9322 solver.cpp:218] Iteration 552 (2.31191 iter/s, 5.19052s/12 iters), loss = 4.9918
I0428 12:56:58.715950  9322 solver.cpp:237]     Train net output #0: loss = 4.9918 (* 1 = 4.9918 loss)
I0428 12:56:58.715961  9322 sgd_solver.cpp:105] Iteration 552, lr = 0.00896423
I0428 12:57:03.800514  9322 solver.cpp:218] Iteration 564 (2.36008 iter/s, 5.08458s/12 iters), loss = 5.09031
I0428 12:57:03.800666  9322 solver.cpp:237]     Train net output #0: loss = 5.09031 (* 1 = 5.09031 loss)
I0428 12:57:03.800675  9322 sgd_solver.cpp:105] Iteration 564, lr = 0.00894294
I0428 12:57:08.888883  9322 solver.cpp:218] Iteration 576 (2.35838 iter/s, 5.08823s/12 iters), loss = 5.08245
I0428 12:57:08.888922  9322 solver.cpp:237]     Train net output #0: loss = 5.08245 (* 1 = 5.08245 loss)
I0428 12:57:08.888929  9322 sgd_solver.cpp:105] Iteration 576, lr = 0.00892171
I0428 12:57:14.063666  9322 solver.cpp:218] Iteration 588 (2.31896 iter/s, 5.17474s/12 iters), loss = 5.04306
I0428 12:57:14.063714  9322 solver.cpp:237]     Train net output #0: loss = 5.04306 (* 1 = 5.04306 loss)
I0428 12:57:14.063722  9322 sgd_solver.cpp:105] Iteration 588, lr = 0.00890053
I0428 12:57:19.236835  9322 solver.cpp:218] Iteration 600 (2.31968 iter/s, 5.17313s/12 iters), loss = 5.08755
I0428 12:57:19.236882  9322 solver.cpp:237]     Train net output #0: loss = 5.08755 (* 1 = 5.08755 loss)
I0428 12:57:19.236891  9322 sgd_solver.cpp:105] Iteration 600, lr = 0.0088794
I0428 12:57:21.706529  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 12:57:23.937897  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_612.caffemodel
I0428 12:57:29.432356  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_612.solverstate
I0428 12:57:33.535828  9322 solver.cpp:330] Iteration 612, Testing net (#0)
I0428 12:57:33.535853  9322 net.cpp:676] Ignoring source layer train-data
I0428 12:57:38.200480  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 12:57:38.518023  9322 solver.cpp:397]     Test net output #0: accuracy = 0.0159314
I0428 12:57:38.518071  9322 solver.cpp:397]     Test net output #1: loss = 5.07052 (* 1 = 5.07052 loss)
I0428 12:57:38.632915  9322 solver.cpp:218] Iteration 612 (0.618681 iter/s, 19.3961s/12 iters), loss = 5.12692
I0428 12:57:38.632956  9322 solver.cpp:237]     Train net output #0: loss = 5.12692 (* 1 = 5.12692 loss)
I0428 12:57:38.632963  9322 sgd_solver.cpp:105] Iteration 612, lr = 0.00885831
I0428 12:57:42.924602  9322 solver.cpp:218] Iteration 624 (2.79613 iter/s, 4.29164s/12 iters), loss = 5.09193
I0428 12:57:42.924650  9322 solver.cpp:237]     Train net output #0: loss = 5.09193 (* 1 = 5.09193 loss)
I0428 12:57:42.924659  9322 sgd_solver.cpp:105] Iteration 624, lr = 0.00883728
I0428 12:57:48.087870  9322 solver.cpp:218] Iteration 636 (2.32413 iter/s, 5.16322s/12 iters), loss = 5.11235
I0428 12:57:48.087918  9322 solver.cpp:237]     Train net output #0: loss = 5.11235 (* 1 = 5.11235 loss)
I0428 12:57:48.087927  9322 sgd_solver.cpp:105] Iteration 636, lr = 0.0088163
I0428 12:57:53.355437  9322 solver.cpp:218] Iteration 648 (2.27811 iter/s, 5.26753s/12 iters), loss = 5.05131
I0428 12:57:53.355481  9322 solver.cpp:237]     Train net output #0: loss = 5.05131 (* 1 = 5.05131 loss)
I0428 12:57:53.355489  9322 sgd_solver.cpp:105] Iteration 648, lr = 0.00879537
I0428 12:57:58.735133  9322 solver.cpp:218] Iteration 660 (2.23062 iter/s, 5.37966s/12 iters), loss = 5.04228
I0428 12:57:58.735183  9322 solver.cpp:237]     Train net output #0: loss = 5.04228 (* 1 = 5.04228 loss)
I0428 12:57:58.735193  9322 sgd_solver.cpp:105] Iteration 660, lr = 0.00877449
I0428 12:58:03.922531  9322 solver.cpp:218] Iteration 672 (2.31332 iter/s, 5.18735s/12 iters), loss = 4.97548
I0428 12:58:03.922574  9322 solver.cpp:237]     Train net output #0: loss = 4.97548 (* 1 = 4.97548 loss)
I0428 12:58:03.922582  9322 sgd_solver.cpp:105] Iteration 672, lr = 0.00875366
I0428 12:58:09.032557  9322 solver.cpp:218] Iteration 684 (2.34834 iter/s, 5.10999s/12 iters), loss = 5.04095
I0428 12:58:09.032711  9322 solver.cpp:237]     Train net output #0: loss = 5.04095 (* 1 = 5.04095 loss)
I0428 12:58:09.032721  9322 sgd_solver.cpp:105] Iteration 684, lr = 0.00873287
I0428 12:58:09.772056  9322 blocking_queue.cpp:49] Waiting for data
I0428 12:58:14.160989  9322 solver.cpp:218] Iteration 696 (2.33996 iter/s, 5.12829s/12 iters), loss = 4.98607
I0428 12:58:14.161032  9322 solver.cpp:237]     Train net output #0: loss = 4.98607 (* 1 = 4.98607 loss)
I0428 12:58:14.161041  9322 sgd_solver.cpp:105] Iteration 696, lr = 0.00871214
I0428 12:58:18.949193  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 12:58:19.346781  9322 solver.cpp:218] Iteration 708 (2.31403 iter/s, 5.18575s/12 iters), loss = 4.98596
I0428 12:58:19.346829  9322 solver.cpp:237]     Train net output #0: loss = 4.98596 (* 1 = 4.98596 loss)
I0428 12:58:19.346837  9322 sgd_solver.cpp:105] Iteration 708, lr = 0.00869145
I0428 12:58:21.486407  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_714.caffemodel
I0428 12:58:27.020244  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_714.solverstate
I0428 12:58:29.437971  9322 solver.cpp:330] Iteration 714, Testing net (#0)
I0428 12:58:29.437990  9322 net.cpp:676] Ignoring source layer train-data
I0428 12:58:34.040763  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 12:58:34.403843  9322 solver.cpp:397]     Test net output #0: accuracy = 0.0110294
I0428 12:58:34.403892  9322 solver.cpp:397]     Test net output #1: loss = 5.02454 (* 1 = 5.02454 loss)
I0428 12:58:36.322803  9322 solver.cpp:218] Iteration 720 (0.706879 iter/s, 16.976s/12 iters), loss = 4.98044
I0428 12:58:36.322849  9322 solver.cpp:237]     Train net output #0: loss = 4.98044 (* 1 = 4.98044 loss)
I0428 12:58:36.322858  9322 sgd_solver.cpp:105] Iteration 720, lr = 0.00867082
I0428 12:58:41.482673  9322 solver.cpp:218] Iteration 732 (2.32566 iter/s, 5.15983s/12 iters), loss = 4.98395
I0428 12:58:41.482759  9322 solver.cpp:237]     Train net output #0: loss = 4.98395 (* 1 = 4.98395 loss)
I0428 12:58:41.482769  9322 sgd_solver.cpp:105] Iteration 732, lr = 0.00865023
I0428 12:58:46.654161  9322 solver.cpp:218] Iteration 744 (2.32045 iter/s, 5.17141s/12 iters), loss = 4.90349
I0428 12:58:46.654207  9322 solver.cpp:237]     Train net output #0: loss = 4.90349 (* 1 = 4.90349 loss)
I0428 12:58:46.654215  9322 sgd_solver.cpp:105] Iteration 744, lr = 0.0086297
I0428 12:58:51.825634  9322 solver.cpp:218] Iteration 756 (2.32044 iter/s, 5.17144s/12 iters), loss = 5.03902
I0428 12:58:51.825680  9322 solver.cpp:237]     Train net output #0: loss = 5.03902 (* 1 = 5.03902 loss)
I0428 12:58:51.825690  9322 sgd_solver.cpp:105] Iteration 756, lr = 0.00860921
I0428 12:58:56.995194  9322 solver.cpp:218] Iteration 768 (2.3213 iter/s, 5.16952s/12 iters), loss = 4.8976
I0428 12:58:56.995237  9322 solver.cpp:237]     Train net output #0: loss = 4.8976 (* 1 = 4.8976 loss)
I0428 12:58:56.995246  9322 sgd_solver.cpp:105] Iteration 768, lr = 0.00858877
I0428 12:59:02.144686  9322 solver.cpp:218] Iteration 780 (2.33034 iter/s, 5.14946s/12 iters), loss = 4.99212
I0428 12:59:02.144729  9322 solver.cpp:237]     Train net output #0: loss = 4.99212 (* 1 = 4.99212 loss)
I0428 12:59:02.144738  9322 sgd_solver.cpp:105] Iteration 780, lr = 0.00856838
I0428 12:59:07.306329  9322 solver.cpp:218] Iteration 792 (2.32486 iter/s, 5.16161s/12 iters), loss = 4.98318
I0428 12:59:07.306367  9322 solver.cpp:237]     Train net output #0: loss = 4.98318 (* 1 = 4.98318 loss)
I0428 12:59:07.306375  9322 sgd_solver.cpp:105] Iteration 792, lr = 0.00854803
I0428 12:59:12.512362  9322 solver.cpp:218] Iteration 804 (2.30503 iter/s, 5.206s/12 iters), loss = 4.99442
I0428 12:59:12.512532  9322 solver.cpp:237]     Train net output #0: loss = 4.99442 (* 1 = 4.99442 loss)
I0428 12:59:12.512543  9322 sgd_solver.cpp:105] Iteration 804, lr = 0.00852774
I0428 12:59:14.272359  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 12:59:17.154814  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_816.caffemodel
I0428 12:59:22.499146  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_816.solverstate
I0428 12:59:26.411170  9322 solver.cpp:330] Iteration 816, Testing net (#0)
I0428 12:59:26.411190  9322 net.cpp:676] Ignoring source layer train-data
I0428 12:59:30.925325  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 12:59:31.326685  9322 solver.cpp:397]     Test net output #0: accuracy = 0.0177696
I0428 12:59:31.326717  9322 solver.cpp:397]     Test net output #1: loss = 4.98664 (* 1 = 4.98664 loss)
I0428 12:59:31.444787  9322 solver.cpp:218] Iteration 816 (0.633837 iter/s, 18.9323s/12 iters), loss = 4.97587
I0428 12:59:31.444834  9322 solver.cpp:237]     Train net output #0: loss = 4.97587 (* 1 = 4.97587 loss)
I0428 12:59:31.444842  9322 sgd_solver.cpp:105] Iteration 816, lr = 0.00850749
I0428 12:59:35.765242  9322 solver.cpp:218] Iteration 828 (2.77751 iter/s, 4.32041s/12 iters), loss = 5.01078
I0428 12:59:35.765285  9322 solver.cpp:237]     Train net output #0: loss = 5.01078 (* 1 = 5.01078 loss)
I0428 12:59:35.765292  9322 sgd_solver.cpp:105] Iteration 828, lr = 0.00848729
I0428 12:59:40.920378  9322 solver.cpp:218] Iteration 840 (2.32779 iter/s, 5.1551s/12 iters), loss = 5.08605
I0428 12:59:40.920425  9322 solver.cpp:237]     Train net output #0: loss = 5.08605 (* 1 = 5.08605 loss)
I0428 12:59:40.920433  9322 sgd_solver.cpp:105] Iteration 840, lr = 0.00846714
I0428 12:59:46.082302  9322 solver.cpp:218] Iteration 852 (2.32474 iter/s, 5.16188s/12 iters), loss = 4.9805
I0428 12:59:46.082419  9322 solver.cpp:237]     Train net output #0: loss = 4.9805 (* 1 = 4.9805 loss)
I0428 12:59:46.082428  9322 sgd_solver.cpp:105] Iteration 852, lr = 0.00844704
I0428 12:59:51.158291  9322 solver.cpp:218] Iteration 864 (2.36412 iter/s, 5.07588s/12 iters), loss = 4.94718
I0428 12:59:51.158332  9322 solver.cpp:237]     Train net output #0: loss = 4.94718 (* 1 = 4.94718 loss)
I0428 12:59:51.158341  9322 sgd_solver.cpp:105] Iteration 864, lr = 0.00842698
I0428 12:59:56.358744  9322 solver.cpp:218] Iteration 876 (2.30751 iter/s, 5.20042s/12 iters), loss = 5.02497
I0428 12:59:56.358793  9322 solver.cpp:237]     Train net output #0: loss = 5.02497 (* 1 = 5.02497 loss)
I0428 12:59:56.358801  9322 sgd_solver.cpp:105] Iteration 876, lr = 0.00840698
I0428 13:00:01.655081  9322 solver.cpp:218] Iteration 888 (2.26573 iter/s, 5.2963s/12 iters), loss = 4.97446
I0428 13:00:01.655117  9322 solver.cpp:237]     Train net output #0: loss = 4.97446 (* 1 = 4.97446 loss)
I0428 13:00:01.655126  9322 sgd_solver.cpp:105] Iteration 888, lr = 0.00838702
I0428 13:00:06.812203  9322 solver.cpp:218] Iteration 900 (2.32689 iter/s, 5.15709s/12 iters), loss = 5.05464
I0428 13:00:06.812247  9322 solver.cpp:237]     Train net output #0: loss = 5.05464 (* 1 = 5.05464 loss)
I0428 13:00:06.812255  9322 sgd_solver.cpp:105] Iteration 900, lr = 0.0083671
I0428 13:00:10.820854  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:00:12.013168  9322 solver.cpp:218] Iteration 912 (2.30728 iter/s, 5.20093s/12 iters), loss = 4.94296
I0428 13:00:12.013211  9322 solver.cpp:237]     Train net output #0: loss = 4.94296 (* 1 = 4.94296 loss)
I0428 13:00:12.013218  9322 sgd_solver.cpp:105] Iteration 912, lr = 0.00834724
I0428 13:00:14.190539  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_918.caffemodel
I0428 13:00:18.767462  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_918.solverstate
I0428 13:00:25.720697  9322 solver.cpp:330] Iteration 918, Testing net (#0)
I0428 13:00:25.720715  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:00:30.343616  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:00:30.815265  9322 solver.cpp:397]     Test net output #0: accuracy = 0.0355392
I0428 13:00:30.815305  9322 solver.cpp:397]     Test net output #1: loss = 4.91981 (* 1 = 4.91981 loss)
I0428 13:00:32.713235  9322 solver.cpp:218] Iteration 924 (0.579707 iter/s, 20.7001s/12 iters), loss = 4.82869
I0428 13:00:32.713297  9322 solver.cpp:237]     Train net output #0: loss = 4.82869 (* 1 = 4.82869 loss)
I0428 13:00:32.713311  9322 sgd_solver.cpp:105] Iteration 924, lr = 0.00832742
I0428 13:00:37.821085  9322 solver.cpp:218] Iteration 936 (2.34935 iter/s, 5.1078s/12 iters), loss = 5.03175
I0428 13:00:37.821128  9322 solver.cpp:237]     Train net output #0: loss = 5.03175 (* 1 = 5.03175 loss)
I0428 13:00:37.821136  9322 sgd_solver.cpp:105] Iteration 936, lr = 0.00830765
I0428 13:00:42.928977  9322 solver.cpp:218] Iteration 948 (2.34932 iter/s, 5.10785s/12 iters), loss = 4.84396
I0428 13:00:42.929014  9322 solver.cpp:237]     Train net output #0: loss = 4.84396 (* 1 = 4.84396 loss)
I0428 13:00:42.929023  9322 sgd_solver.cpp:105] Iteration 948, lr = 0.00828793
I0428 13:00:48.100529  9322 solver.cpp:218] Iteration 960 (2.3204 iter/s, 5.17152s/12 iters), loss = 4.88261
I0428 13:00:48.100569  9322 solver.cpp:237]     Train net output #0: loss = 4.88261 (* 1 = 4.88261 loss)
I0428 13:00:48.100577  9322 sgd_solver.cpp:105] Iteration 960, lr = 0.00826825
I0428 13:00:53.254786  9322 solver.cpp:218] Iteration 972 (2.32819 iter/s, 5.15422s/12 iters), loss = 4.79714
I0428 13:00:53.254876  9322 solver.cpp:237]     Train net output #0: loss = 4.79714 (* 1 = 4.79714 loss)
I0428 13:00:53.254886  9322 sgd_solver.cpp:105] Iteration 972, lr = 0.00824862
I0428 13:00:58.408308  9322 solver.cpp:218] Iteration 984 (2.32854 iter/s, 5.15344s/12 iters), loss = 4.7192
I0428 13:00:58.408345  9322 solver.cpp:237]     Train net output #0: loss = 4.7192 (* 1 = 4.7192 loss)
I0428 13:00:58.408354  9322 sgd_solver.cpp:105] Iteration 984, lr = 0.00822903
I0428 13:01:03.556737  9322 solver.cpp:218] Iteration 996 (2.33082 iter/s, 5.14839s/12 iters), loss = 4.92814
I0428 13:01:03.556784  9322 solver.cpp:237]     Train net output #0: loss = 4.92814 (* 1 = 4.92814 loss)
I0428 13:01:03.556792  9322 sgd_solver.cpp:105] Iteration 996, lr = 0.0082095
I0428 13:01:08.749150  9322 solver.cpp:218] Iteration 1008 (2.31108 iter/s, 5.19237s/12 iters), loss = 4.83462
I0428 13:01:08.749195  9322 solver.cpp:237]     Train net output #0: loss = 4.83462 (* 1 = 4.83462 loss)
I0428 13:01:08.749203  9322 sgd_solver.cpp:105] Iteration 1008, lr = 0.00819001
I0428 13:01:09.793890  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:01:13.453945  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_1020.caffemodel
I0428 13:01:18.531841  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_1020.solverstate
I0428 13:01:25.592108  9322 solver.cpp:330] Iteration 1020, Testing net (#0)
I0428 13:01:25.592161  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:01:29.860344  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:01:30.325325  9322 solver.cpp:397]     Test net output #0: accuracy = 0.0367647
I0428 13:01:30.325371  9322 solver.cpp:397]     Test net output #1: loss = 4.81141 (* 1 = 4.81141 loss)
I0428 13:01:30.443739  9322 solver.cpp:218] Iteration 1020 (0.553132 iter/s, 21.6946s/12 iters), loss = 4.87743
I0428 13:01:30.443807  9322 solver.cpp:237]     Train net output #0: loss = 4.87743 (* 1 = 4.87743 loss)
I0428 13:01:30.443817  9322 sgd_solver.cpp:105] Iteration 1020, lr = 0.00817056
I0428 13:01:34.711472  9322 solver.cpp:218] Iteration 1032 (2.81183 iter/s, 4.26768s/12 iters), loss = 4.6725
I0428 13:01:34.711509  9322 solver.cpp:237]     Train net output #0: loss = 4.6725 (* 1 = 4.6725 loss)
I0428 13:01:34.711517  9322 sgd_solver.cpp:105] Iteration 1032, lr = 0.00815116
I0428 13:01:39.743783  9322 solver.cpp:218] Iteration 1044 (2.3846 iter/s, 5.03228s/12 iters), loss = 4.70474
I0428 13:01:39.743825  9322 solver.cpp:237]     Train net output #0: loss = 4.70474 (* 1 = 4.70474 loss)
I0428 13:01:39.743834  9322 sgd_solver.cpp:105] Iteration 1044, lr = 0.00813181
I0428 13:01:45.030158  9322 solver.cpp:218] Iteration 1056 (2.27 iter/s, 5.28633s/12 iters), loss = 4.82336
I0428 13:01:45.030205  9322 solver.cpp:237]     Train net output #0: loss = 4.82336 (* 1 = 4.82336 loss)
I0428 13:01:45.030213  9322 sgd_solver.cpp:105] Iteration 1056, lr = 0.0081125
I0428 13:01:50.244915  9322 solver.cpp:218] Iteration 1068 (2.30118 iter/s, 5.21471s/12 iters), loss = 4.82486
I0428 13:01:50.244961  9322 solver.cpp:237]     Train net output #0: loss = 4.82486 (* 1 = 4.82486 loss)
I0428 13:01:50.244971  9322 sgd_solver.cpp:105] Iteration 1068, lr = 0.00809324
I0428 13:01:55.372330  9322 solver.cpp:218] Iteration 1080 (2.34038 iter/s, 5.12737s/12 iters), loss = 4.66209
I0428 13:01:55.372373  9322 solver.cpp:237]     Train net output #0: loss = 4.66209 (* 1 = 4.66209 loss)
I0428 13:01:55.372382  9322 sgd_solver.cpp:105] Iteration 1080, lr = 0.00807403
I0428 13:02:00.556885  9322 solver.cpp:218] Iteration 1092 (2.31459 iter/s, 5.18451s/12 iters), loss = 4.6862
I0428 13:02:00.557024  9322 solver.cpp:237]     Train net output #0: loss = 4.6862 (* 1 = 4.6862 loss)
I0428 13:02:00.557032  9322 sgd_solver.cpp:105] Iteration 1092, lr = 0.00805486
I0428 13:02:05.726228  9322 solver.cpp:218] Iteration 1104 (2.32144 iter/s, 5.16922s/12 iters), loss = 4.84689
I0428 13:02:05.726270  9322 solver.cpp:237]     Train net output #0: loss = 4.84689 (* 1 = 4.84689 loss)
I0428 13:02:05.726279  9322 sgd_solver.cpp:105] Iteration 1104, lr = 0.00803573
I0428 13:02:08.974581  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:02:10.915556  9322 solver.cpp:218] Iteration 1116 (2.31245 iter/s, 5.18929s/12 iters), loss = 4.63986
I0428 13:02:10.915597  9322 solver.cpp:237]     Train net output #0: loss = 4.63986 (* 1 = 4.63986 loss)
I0428 13:02:10.915606  9322 sgd_solver.cpp:105] Iteration 1116, lr = 0.00801666
I0428 13:02:13.014083  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_1122.caffemodel
I0428 13:02:16.851347  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_1122.solverstate
I0428 13:02:19.266553  9322 solver.cpp:330] Iteration 1122, Testing net (#0)
I0428 13:02:19.266579  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:02:23.569870  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:02:24.080209  9322 solver.cpp:397]     Test net output #0: accuracy = 0.0379902
I0428 13:02:24.080250  9322 solver.cpp:397]     Test net output #1: loss = 4.7364 (* 1 = 4.7364 loss)
I0428 13:02:25.974902  9322 solver.cpp:218] Iteration 1128 (0.796847 iter/s, 15.0594s/12 iters), loss = 4.73668
I0428 13:02:25.974944  9322 solver.cpp:237]     Train net output #0: loss = 4.73668 (* 1 = 4.73668 loss)
I0428 13:02:25.974953  9322 sgd_solver.cpp:105] Iteration 1128, lr = 0.00799762
I0428 13:02:31.015342  9322 solver.cpp:218] Iteration 1140 (2.38076 iter/s, 5.04041s/12 iters), loss = 4.78587
I0428 13:02:31.015413  9322 solver.cpp:237]     Train net output #0: loss = 4.78587 (* 1 = 4.78587 loss)
I0428 13:02:31.015421  9322 sgd_solver.cpp:105] Iteration 1140, lr = 0.00797863
I0428 13:02:36.249894  9322 solver.cpp:218] Iteration 1152 (2.29248 iter/s, 5.2345s/12 iters), loss = 4.82534
I0428 13:02:36.249928  9322 solver.cpp:237]     Train net output #0: loss = 4.82534 (* 1 = 4.82534 loss)
I0428 13:02:36.249936  9322 sgd_solver.cpp:105] Iteration 1152, lr = 0.00795969
I0428 13:02:41.410374  9322 solver.cpp:218] Iteration 1164 (2.32537 iter/s, 5.16046s/12 iters), loss = 4.5531
I0428 13:02:41.410408  9322 solver.cpp:237]     Train net output #0: loss = 4.5531 (* 1 = 4.5531 loss)
I0428 13:02:41.410416  9322 sgd_solver.cpp:105] Iteration 1164, lr = 0.00794079
I0428 13:02:46.569120  9322 solver.cpp:218] Iteration 1176 (2.32616 iter/s, 5.15872s/12 iters), loss = 4.75692
I0428 13:02:46.569166  9322 solver.cpp:237]     Train net output #0: loss = 4.75692 (* 1 = 4.75692 loss)
I0428 13:02:46.569177  9322 sgd_solver.cpp:105] Iteration 1176, lr = 0.00792194
I0428 13:02:51.737408  9322 solver.cpp:218] Iteration 1188 (2.32187 iter/s, 5.16825s/12 iters), loss = 4.657
I0428 13:02:51.737447  9322 solver.cpp:237]     Train net output #0: loss = 4.657 (* 1 = 4.657 loss)
I0428 13:02:51.737454  9322 sgd_solver.cpp:105] Iteration 1188, lr = 0.00790313
I0428 13:02:56.907656  9322 solver.cpp:218] Iteration 1200 (2.32099 iter/s, 5.17022s/12 iters), loss = 4.53534
I0428 13:02:56.907701  9322 solver.cpp:237]     Train net output #0: loss = 4.53534 (* 1 = 4.53534 loss)
I0428 13:02:56.907711  9322 sgd_solver.cpp:105] Iteration 1200, lr = 0.00788437
I0428 13:03:02.081104  9322 solver.cpp:218] Iteration 1212 (2.31955 iter/s, 5.17341s/12 iters), loss = 4.5111
I0428 13:03:02.081230  9322 solver.cpp:237]     Train net output #0: loss = 4.5111 (* 1 = 4.5111 loss)
I0428 13:03:02.081240  9322 sgd_solver.cpp:105] Iteration 1212, lr = 0.00786565
I0428 13:03:02.357663  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:03:06.756212  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_1224.caffemodel
I0428 13:03:11.054354  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_1224.solverstate
I0428 13:03:14.290478  9322 solver.cpp:330] Iteration 1224, Testing net (#0)
I0428 13:03:14.290499  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:03:18.472036  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:03:19.007162  9322 solver.cpp:397]     Test net output #0: accuracy = 0.0563725
I0428 13:03:19.007205  9322 solver.cpp:397]     Test net output #1: loss = 4.59837 (* 1 = 4.59837 loss)
I0428 13:03:19.125303  9322 solver.cpp:218] Iteration 1224 (0.704054 iter/s, 17.0441s/12 iters), loss = 4.5307
I0428 13:03:19.125350  9322 solver.cpp:237]     Train net output #0: loss = 4.5307 (* 1 = 4.5307 loss)
I0428 13:03:19.125357  9322 sgd_solver.cpp:105] Iteration 1224, lr = 0.00784697
I0428 13:03:23.474939  9322 solver.cpp:218] Iteration 1236 (2.75888 iter/s, 4.3496s/12 iters), loss = 4.73665
I0428 13:03:23.474980  9322 solver.cpp:237]     Train net output #0: loss = 4.73665 (* 1 = 4.73665 loss)
I0428 13:03:23.474988  9322 sgd_solver.cpp:105] Iteration 1236, lr = 0.00782834
I0428 13:03:28.666867  9322 solver.cpp:218] Iteration 1248 (2.3113 iter/s, 5.19189s/12 iters), loss = 4.55997
I0428 13:03:28.666913  9322 solver.cpp:237]     Train net output #0: loss = 4.55997 (* 1 = 4.55997 loss)
I0428 13:03:28.666921  9322 sgd_solver.cpp:105] Iteration 1248, lr = 0.00780976
I0428 13:03:33.899209  9322 solver.cpp:218] Iteration 1260 (2.29345 iter/s, 5.2323s/12 iters), loss = 4.5791
I0428 13:03:33.899277  9322 solver.cpp:237]     Train net output #0: loss = 4.5791 (* 1 = 4.5791 loss)
I0428 13:03:33.899286  9322 sgd_solver.cpp:105] Iteration 1260, lr = 0.00779122
I0428 13:03:39.124611  9322 solver.cpp:218] Iteration 1272 (2.2965 iter/s, 5.22535s/12 iters), loss = 4.38972
I0428 13:03:39.124646  9322 solver.cpp:237]     Train net output #0: loss = 4.38972 (* 1 = 4.38972 loss)
I0428 13:03:39.124655  9322 sgd_solver.cpp:105] Iteration 1272, lr = 0.00777272
I0428 13:03:44.317039  9322 solver.cpp:218] Iteration 1284 (2.31107 iter/s, 5.19241s/12 iters), loss = 4.55155
I0428 13:03:44.317071  9322 solver.cpp:237]     Train net output #0: loss = 4.55155 (* 1 = 4.55155 loss)
I0428 13:03:44.317078  9322 sgd_solver.cpp:105] Iteration 1284, lr = 0.00775426
I0428 13:03:49.511909  9322 solver.cpp:218] Iteration 1296 (2.30998 iter/s, 5.19485s/12 iters), loss = 4.50882
I0428 13:03:49.511942  9322 solver.cpp:237]     Train net output #0: loss = 4.50882 (* 1 = 4.50882 loss)
I0428 13:03:49.511950  9322 sgd_solver.cpp:105] Iteration 1296, lr = 0.00773585
I0428 13:03:54.598999  9322 solver.cpp:218] Iteration 1308 (2.35893 iter/s, 5.08706s/12 iters), loss = 4.36816
I0428 13:03:54.599045  9322 solver.cpp:237]     Train net output #0: loss = 4.36816 (* 1 = 4.36816 loss)
I0428 13:03:54.599053  9322 sgd_solver.cpp:105] Iteration 1308, lr = 0.00771749
I0428 13:03:57.277654  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:03:59.874315  9322 solver.cpp:218] Iteration 1320 (2.27476 iter/s, 5.27528s/12 iters), loss = 4.46448
I0428 13:03:59.874364  9322 solver.cpp:237]     Train net output #0: loss = 4.46448 (* 1 = 4.46448 loss)
I0428 13:03:59.874372  9322 sgd_solver.cpp:105] Iteration 1320, lr = 0.00769916
I0428 13:04:01.957659  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_1326.caffemodel
I0428 13:04:06.204859  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_1326.solverstate
I0428 13:04:09.182201  9322 solver.cpp:330] Iteration 1326, Testing net (#0)
I0428 13:04:09.182227  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:04:13.437435  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:04:14.073688  9322 solver.cpp:397]     Test net output #0: accuracy = 0.057598
I0428 13:04:14.073734  9322 solver.cpp:397]     Test net output #1: loss = 4.56111 (* 1 = 4.56111 loss)
I0428 13:04:15.986980  9322 solver.cpp:218] Iteration 1332 (0.744755 iter/s, 16.1127s/12 iters), loss = 4.46555
I0428 13:04:15.987030  9322 solver.cpp:237]     Train net output #0: loss = 4.46555 (* 1 = 4.46555 loss)
I0428 13:04:15.987037  9322 sgd_solver.cpp:105] Iteration 1332, lr = 0.00768088
I0428 13:04:21.155884  9322 solver.cpp:218] Iteration 1344 (2.32159 iter/s, 5.16886s/12 iters), loss = 4.31314
I0428 13:04:21.155921  9322 solver.cpp:237]     Train net output #0: loss = 4.31314 (* 1 = 4.31314 loss)
I0428 13:04:21.155930  9322 sgd_solver.cpp:105] Iteration 1344, lr = 0.00766265
I0428 13:04:26.324527  9322 solver.cpp:218] Iteration 1356 (2.32171 iter/s, 5.16861s/12 iters), loss = 4.56838
I0428 13:04:26.324573  9322 solver.cpp:237]     Train net output #0: loss = 4.56838 (* 1 = 4.56838 loss)
I0428 13:04:26.324580  9322 sgd_solver.cpp:105] Iteration 1356, lr = 0.00764446
I0428 13:04:31.466989  9322 solver.cpp:218] Iteration 1368 (2.33353 iter/s, 5.14242s/12 iters), loss = 4.45433
I0428 13:04:31.467031  9322 solver.cpp:237]     Train net output #0: loss = 4.45433 (* 1 = 4.45433 loss)
I0428 13:04:31.467038  9322 sgd_solver.cpp:105] Iteration 1368, lr = 0.00762631
I0428 13:04:32.688915  9322 blocking_queue.cpp:49] Waiting for data
I0428 13:04:36.623628  9322 solver.cpp:218] Iteration 1380 (2.32711 iter/s, 5.1566s/12 iters), loss = 4.35784
I0428 13:04:36.623736  9322 solver.cpp:237]     Train net output #0: loss = 4.35784 (* 1 = 4.35784 loss)
I0428 13:04:36.623747  9322 sgd_solver.cpp:105] Iteration 1380, lr = 0.0076082
I0428 13:04:41.667742  9322 solver.cpp:218] Iteration 1392 (2.37906 iter/s, 5.04401s/12 iters), loss = 4.31197
I0428 13:04:41.667786  9322 solver.cpp:237]     Train net output #0: loss = 4.31197 (* 1 = 4.31197 loss)
I0428 13:04:41.667794  9322 sgd_solver.cpp:105] Iteration 1392, lr = 0.00759014
I0428 13:04:46.836524  9322 solver.cpp:218] Iteration 1404 (2.32165 iter/s, 5.16875s/12 iters), loss = 4.19832
I0428 13:04:46.836561  9322 solver.cpp:237]     Train net output #0: loss = 4.19832 (* 1 = 4.19832 loss)
I0428 13:04:46.836570  9322 sgd_solver.cpp:105] Iteration 1404, lr = 0.00757212
I0428 13:04:51.575681  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:04:51.943598  9322 solver.cpp:218] Iteration 1416 (2.3497 iter/s, 5.10704s/12 iters), loss = 4.31423
I0428 13:04:51.943645  9322 solver.cpp:237]     Train net output #0: loss = 4.31423 (* 1 = 4.31423 loss)
I0428 13:04:51.943652  9322 sgd_solver.cpp:105] Iteration 1416, lr = 0.00755414
I0428 13:04:56.624373  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_1428.caffemodel
I0428 13:05:01.679821  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_1428.solverstate
I0428 13:05:05.409708  9322 solver.cpp:330] Iteration 1428, Testing net (#0)
I0428 13:05:05.409726  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:05:09.516145  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:05:10.158751  9322 solver.cpp:397]     Test net output #0: accuracy = 0.0802696
I0428 13:05:10.158787  9322 solver.cpp:397]     Test net output #1: loss = 4.39858 (* 1 = 4.39858 loss)
I0428 13:05:10.281854  9322 solver.cpp:218] Iteration 1428 (0.654369 iter/s, 18.3383s/12 iters), loss = 4.31078
I0428 13:05:10.281925  9322 solver.cpp:237]     Train net output #0: loss = 4.31078 (* 1 = 4.31078 loss)
I0428 13:05:10.281934  9322 sgd_solver.cpp:105] Iteration 1428, lr = 0.0075362
I0428 13:05:14.566481  9322 solver.cpp:218] Iteration 1440 (2.80075 iter/s, 4.28456s/12 iters), loss = 4.24323
I0428 13:05:14.566525  9322 solver.cpp:237]     Train net output #0: loss = 4.24323 (* 1 = 4.24323 loss)
I0428 13:05:14.566534  9322 sgd_solver.cpp:105] Iteration 1440, lr = 0.00751831
I0428 13:05:19.714876  9322 solver.cpp:218] Iteration 1452 (2.33084 iter/s, 5.14836s/12 iters), loss = 4.0595
I0428 13:05:19.714926  9322 solver.cpp:237]     Train net output #0: loss = 4.0595 (* 1 = 4.0595 loss)
I0428 13:05:19.714933  9322 sgd_solver.cpp:105] Iteration 1452, lr = 0.00750046
I0428 13:05:24.891682  9322 solver.cpp:218] Iteration 1464 (2.31805 iter/s, 5.17676s/12 iters), loss = 4.34836
I0428 13:05:24.891726  9322 solver.cpp:237]     Train net output #0: loss = 4.34836 (* 1 = 4.34836 loss)
I0428 13:05:24.891736  9322 sgd_solver.cpp:105] Iteration 1464, lr = 0.00748265
I0428 13:05:30.055738  9322 solver.cpp:218] Iteration 1476 (2.32377 iter/s, 5.16401s/12 iters), loss = 4.21919
I0428 13:05:30.055785  9322 solver.cpp:237]     Train net output #0: loss = 4.21919 (* 1 = 4.21919 loss)
I0428 13:05:30.055794  9322 sgd_solver.cpp:105] Iteration 1476, lr = 0.00746489
I0428 13:05:35.206980  9322 solver.cpp:218] Iteration 1488 (2.32955 iter/s, 5.1512s/12 iters), loss = 4.2053
I0428 13:05:35.207022  9322 solver.cpp:237]     Train net output #0: loss = 4.2053 (* 1 = 4.2053 loss)
I0428 13:05:35.207031  9322 sgd_solver.cpp:105] Iteration 1488, lr = 0.00744716
I0428 13:05:40.296393  9322 solver.cpp:218] Iteration 1500 (2.35785 iter/s, 5.08937s/12 iters), loss = 4.18493
I0428 13:05:40.296458  9322 solver.cpp:237]     Train net output #0: loss = 4.18493 (* 1 = 4.18493 loss)
I0428 13:05:40.296466  9322 sgd_solver.cpp:105] Iteration 1500, lr = 0.00742948
I0428 13:05:45.393347  9322 solver.cpp:218] Iteration 1512 (2.35438 iter/s, 5.09689s/12 iters), loss = 4.11103
I0428 13:05:45.393393  9322 solver.cpp:237]     Train net output #0: loss = 4.11103 (* 1 = 4.11103 loss)
I0428 13:05:45.393401  9322 sgd_solver.cpp:105] Iteration 1512, lr = 0.00741184
I0428 13:05:47.237406  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:05:50.589687  9322 solver.cpp:218] Iteration 1524 (2.30934 iter/s, 5.1963s/12 iters), loss = 4.10186
I0428 13:05:50.589732  9322 solver.cpp:237]     Train net output #0: loss = 4.10186 (* 1 = 4.10186 loss)
I0428 13:05:50.589740  9322 sgd_solver.cpp:105] Iteration 1524, lr = 0.00739425
I0428 13:05:52.664899  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_1530.caffemodel
I0428 13:05:55.996012  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_1530.solverstate
I0428 13:05:58.409421  9322 solver.cpp:330] Iteration 1530, Testing net (#0)
I0428 13:05:58.409440  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:06:02.588335  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:06:03.313290  9322 solver.cpp:397]     Test net output #0: accuracy = 0.0949755
I0428 13:06:03.313338  9322 solver.cpp:397]     Test net output #1: loss = 4.21476 (* 1 = 4.21476 loss)
I0428 13:06:05.251047  9322 solver.cpp:218] Iteration 1536 (0.818477 iter/s, 14.6614s/12 iters), loss = 4.16696
I0428 13:06:05.251086  9322 solver.cpp:237]     Train net output #0: loss = 4.16696 (* 1 = 4.16696 loss)
I0428 13:06:05.251094  9322 sgd_solver.cpp:105] Iteration 1536, lr = 0.00737669
I0428 13:06:10.409813  9322 solver.cpp:218] Iteration 1548 (2.32615 iter/s, 5.15874s/12 iters), loss = 3.99802
I0428 13:06:10.409942  9322 solver.cpp:237]     Train net output #0: loss = 3.99802 (* 1 = 3.99802 loss)
I0428 13:06:10.409952  9322 sgd_solver.cpp:105] Iteration 1548, lr = 0.00735918
I0428 13:06:15.499730  9322 solver.cpp:218] Iteration 1560 (2.35766 iter/s, 5.0898s/12 iters), loss = 4.28105
I0428 13:06:15.499778  9322 solver.cpp:237]     Train net output #0: loss = 4.28105 (* 1 = 4.28105 loss)
I0428 13:06:15.499788  9322 sgd_solver.cpp:105] Iteration 1560, lr = 0.00734171
I0428 13:06:20.603569  9322 solver.cpp:218] Iteration 1572 (2.35119 iter/s, 5.1038s/12 iters), loss = 3.84795
I0428 13:06:20.603605  9322 solver.cpp:237]     Train net output #0: loss = 3.84795 (* 1 = 3.84795 loss)
I0428 13:06:20.603612  9322 sgd_solver.cpp:105] Iteration 1572, lr = 0.00732427
I0428 13:06:25.873169  9322 solver.cpp:218] Iteration 1584 (2.27723 iter/s, 5.26957s/12 iters), loss = 4.19879
I0428 13:06:25.873208  9322 solver.cpp:237]     Train net output #0: loss = 4.19879 (* 1 = 4.19879 loss)
I0428 13:06:25.873217  9322 sgd_solver.cpp:105] Iteration 1584, lr = 0.00730688
I0428 13:06:31.079792  9322 solver.cpp:218] Iteration 1596 (2.30477 iter/s, 5.20659s/12 iters), loss = 4.2704
I0428 13:06:31.079838  9322 solver.cpp:237]     Train net output #0: loss = 4.2704 (* 1 = 4.2704 loss)
I0428 13:06:31.079846  9322 sgd_solver.cpp:105] Iteration 1596, lr = 0.00728954
I0428 13:06:36.302819  9322 solver.cpp:218] Iteration 1608 (2.29754 iter/s, 5.22299s/12 iters), loss = 3.96558
I0428 13:06:36.302863  9322 solver.cpp:237]     Train net output #0: loss = 3.96558 (* 1 = 3.96558 loss)
I0428 13:06:36.302871  9322 sgd_solver.cpp:105] Iteration 1608, lr = 0.00727223
I0428 13:06:40.402833  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:06:41.539947  9322 solver.cpp:218] Iteration 1620 (2.29135 iter/s, 5.2371s/12 iters), loss = 4.12544
I0428 13:06:41.540076  9322 solver.cpp:237]     Train net output #0: loss = 4.12544 (* 1 = 4.12544 loss)
I0428 13:06:41.540086  9322 sgd_solver.cpp:105] Iteration 1620, lr = 0.00725496
I0428 13:06:46.212162  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_1632.caffemodel
I0428 13:06:52.438764  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_1632.solverstate
I0428 13:06:55.345973  9322 solver.cpp:330] Iteration 1632, Testing net (#0)
I0428 13:06:55.345993  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:06:59.332958  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:07:00.052839  9322 solver.cpp:397]     Test net output #0: accuracy = 0.0894608
I0428 13:07:00.052884  9322 solver.cpp:397]     Test net output #1: loss = 4.20825 (* 1 = 4.20825 loss)
I0428 13:07:00.171945  9322 solver.cpp:218] Iteration 1632 (0.644055 iter/s, 18.6319s/12 iters), loss = 3.93277
I0428 13:07:00.171993  9322 solver.cpp:237]     Train net output #0: loss = 3.93277 (* 1 = 3.93277 loss)
I0428 13:07:00.172000  9322 sgd_solver.cpp:105] Iteration 1632, lr = 0.00723774
I0428 13:07:04.432657  9322 solver.cpp:218] Iteration 1644 (2.81646 iter/s, 4.26067s/12 iters), loss = 3.87255
I0428 13:07:04.432699  9322 solver.cpp:237]     Train net output #0: loss = 3.87255 (* 1 = 3.87255 loss)
I0428 13:07:04.432708  9322 sgd_solver.cpp:105] Iteration 1644, lr = 0.00722056
I0428 13:07:09.579291  9322 solver.cpp:218] Iteration 1656 (2.33164 iter/s, 5.1466s/12 iters), loss = 4.00136
I0428 13:07:09.579332  9322 solver.cpp:237]     Train net output #0: loss = 4.00136 (* 1 = 4.00136 loss)
I0428 13:07:09.579340  9322 sgd_solver.cpp:105] Iteration 1656, lr = 0.00720341
I0428 13:07:14.750200  9322 solver.cpp:218] Iteration 1668 (2.32069 iter/s, 5.17088s/12 iters), loss = 4.15353
I0428 13:07:14.750311  9322 solver.cpp:237]     Train net output #0: loss = 4.15353 (* 1 = 4.15353 loss)
I0428 13:07:14.750320  9322 sgd_solver.cpp:105] Iteration 1668, lr = 0.00718631
I0428 13:07:19.939034  9322 solver.cpp:218] Iteration 1680 (2.3127 iter/s, 5.18873s/12 iters), loss = 3.65677
I0428 13:07:19.939071  9322 solver.cpp:237]     Train net output #0: loss = 3.65677 (* 1 = 3.65677 loss)
I0428 13:07:19.939079  9322 sgd_solver.cpp:105] Iteration 1680, lr = 0.00716925
I0428 13:07:25.100347  9322 solver.cpp:218] Iteration 1692 (2.32501 iter/s, 5.16127s/12 iters), loss = 3.88068
I0428 13:07:25.100390  9322 solver.cpp:237]     Train net output #0: loss = 3.88068 (* 1 = 3.88068 loss)
I0428 13:07:25.100399  9322 sgd_solver.cpp:105] Iteration 1692, lr = 0.00715223
I0428 13:07:30.258015  9322 solver.cpp:218] Iteration 1704 (2.32665 iter/s, 5.15764s/12 iters), loss = 4.00976
I0428 13:07:30.258055  9322 solver.cpp:237]     Train net output #0: loss = 4.00976 (* 1 = 4.00976 loss)
I0428 13:07:30.258062  9322 sgd_solver.cpp:105] Iteration 1704, lr = 0.00713525
I0428 13:07:35.445906  9322 solver.cpp:218] Iteration 1716 (2.31309 iter/s, 5.18786s/12 iters), loss = 3.80165
I0428 13:07:35.445950  9322 solver.cpp:237]     Train net output #0: loss = 3.80165 (* 1 = 3.80165 loss)
I0428 13:07:35.445958  9322 sgd_solver.cpp:105] Iteration 1716, lr = 0.00711831
I0428 13:07:36.527992  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:07:40.575608  9322 solver.cpp:218] Iteration 1728 (2.33933 iter/s, 5.12966s/12 iters), loss = 3.99305
I0428 13:07:40.575654  9322 solver.cpp:237]     Train net output #0: loss = 3.99305 (* 1 = 3.99305 loss)
I0428 13:07:40.575661  9322 sgd_solver.cpp:105] Iteration 1728, lr = 0.00710141
I0428 13:07:42.655740  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_1734.caffemodel
I0428 13:07:52.795771  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_1734.solverstate
I0428 13:07:59.348786  9322 solver.cpp:330] Iteration 1734, Testing net (#0)
I0428 13:07:59.348806  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:08:03.344852  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:08:04.152644  9322 solver.cpp:397]     Test net output #0: accuracy = 0.108456
I0428 13:08:04.152690  9322 solver.cpp:397]     Test net output #1: loss = 4.06659 (* 1 = 4.06659 loss)
I0428 13:08:06.102197  9322 solver.cpp:218] Iteration 1740 (0.470097 iter/s, 25.5266s/12 iters), loss = 3.81883
I0428 13:08:06.102241  9322 solver.cpp:237]     Train net output #0: loss = 3.81883 (* 1 = 3.81883 loss)
I0428 13:08:06.102248  9322 sgd_solver.cpp:105] Iteration 1740, lr = 0.00708455
I0428 13:08:11.288112  9322 solver.cpp:218] Iteration 1752 (2.31397 iter/s, 5.18589s/12 iters), loss = 3.92032
I0428 13:08:11.288147  9322 solver.cpp:237]     Train net output #0: loss = 3.92032 (* 1 = 3.92032 loss)
I0428 13:08:11.288156  9322 sgd_solver.cpp:105] Iteration 1752, lr = 0.00706773
I0428 13:08:16.452198  9322 solver.cpp:218] Iteration 1764 (2.32376 iter/s, 5.16405s/12 iters), loss = 3.88639
I0428 13:08:16.452257  9322 solver.cpp:237]     Train net output #0: loss = 3.88639 (* 1 = 3.88639 loss)
I0428 13:08:16.452268  9322 sgd_solver.cpp:105] Iteration 1764, lr = 0.00705094
I0428 13:08:21.680639  9322 solver.cpp:218] Iteration 1776 (2.29516 iter/s, 5.22839s/12 iters), loss = 3.76403
I0428 13:08:21.680681  9322 solver.cpp:237]     Train net output #0: loss = 3.76403 (* 1 = 3.76403 loss)
I0428 13:08:21.680689  9322 sgd_solver.cpp:105] Iteration 1776, lr = 0.0070342
I0428 13:08:26.760141  9322 solver.cpp:218] Iteration 1788 (2.36245 iter/s, 5.07946s/12 iters), loss = 3.5992
I0428 13:08:26.760267  9322 solver.cpp:237]     Train net output #0: loss = 3.5992 (* 1 = 3.5992 loss)
I0428 13:08:26.760277  9322 sgd_solver.cpp:105] Iteration 1788, lr = 0.0070175
I0428 13:08:31.912228  9322 solver.cpp:218] Iteration 1800 (2.32921 iter/s, 5.15197s/12 iters), loss = 3.55827
I0428 13:08:31.912272  9322 solver.cpp:237]     Train net output #0: loss = 3.55827 (* 1 = 3.55827 loss)
I0428 13:08:31.912281  9322 sgd_solver.cpp:105] Iteration 1800, lr = 0.00700084
I0428 13:08:37.079113  9322 solver.cpp:218] Iteration 1812 (2.3225 iter/s, 5.16685s/12 iters), loss = 3.76613
I0428 13:08:37.079160  9322 solver.cpp:237]     Train net output #0: loss = 3.76613 (* 1 = 3.76613 loss)
I0428 13:08:37.079169  9322 sgd_solver.cpp:105] Iteration 1812, lr = 0.00698422
I0428 13:08:40.364696  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:08:42.258714  9322 solver.cpp:218] Iteration 1824 (2.3168 iter/s, 5.17956s/12 iters), loss = 3.59061
I0428 13:08:42.258754  9322 solver.cpp:237]     Train net output #0: loss = 3.59061 (* 1 = 3.59061 loss)
I0428 13:08:42.258761  9322 sgd_solver.cpp:105] Iteration 1824, lr = 0.00696764
I0428 13:08:46.943796  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_1836.caffemodel
I0428 13:08:52.992946  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_1836.solverstate
I0428 13:09:00.482174  9322 solver.cpp:330] Iteration 1836, Testing net (#0)
I0428 13:09:00.482287  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:09:04.553520  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:09:05.422060  9322 solver.cpp:397]     Test net output #0: accuracy = 0.130515
I0428 13:09:05.422101  9322 solver.cpp:397]     Test net output #1: loss = 3.96618 (* 1 = 3.96618 loss)
I0428 13:09:05.540467  9322 solver.cpp:218] Iteration 1836 (0.515424 iter/s, 23.2818s/12 iters), loss = 3.83709
I0428 13:09:05.540519  9322 solver.cpp:237]     Train net output #0: loss = 3.83709 (* 1 = 3.83709 loss)
I0428 13:09:05.540529  9322 sgd_solver.cpp:105] Iteration 1836, lr = 0.0069511
I0428 13:09:09.858507  9322 solver.cpp:218] Iteration 1848 (2.77906 iter/s, 4.318s/12 iters), loss = 3.85046
I0428 13:09:09.858548  9322 solver.cpp:237]     Train net output #0: loss = 3.85046 (* 1 = 3.85046 loss)
I0428 13:09:09.858556  9322 sgd_solver.cpp:105] Iteration 1848, lr = 0.00693459
I0428 13:09:14.942709  9322 solver.cpp:218] Iteration 1860 (2.36027 iter/s, 5.08417s/12 iters), loss = 3.48551
I0428 13:09:14.942746  9322 solver.cpp:237]     Train net output #0: loss = 3.48551 (* 1 = 3.48551 loss)
I0428 13:09:14.942754  9322 sgd_solver.cpp:105] Iteration 1860, lr = 0.00691813
I0428 13:09:20.115428  9322 solver.cpp:218] Iteration 1872 (2.31988 iter/s, 5.17269s/12 iters), loss = 3.8487
I0428 13:09:20.115475  9322 solver.cpp:237]     Train net output #0: loss = 3.8487 (* 1 = 3.8487 loss)
I0428 13:09:20.115483  9322 sgd_solver.cpp:105] Iteration 1872, lr = 0.0069017
I0428 13:09:25.268002  9322 solver.cpp:218] Iteration 1884 (2.32895 iter/s, 5.15253s/12 iters), loss = 3.84438
I0428 13:09:25.268043  9322 solver.cpp:237]     Train net output #0: loss = 3.84438 (* 1 = 3.84438 loss)
I0428 13:09:25.268050  9322 sgd_solver.cpp:105] Iteration 1884, lr = 0.00688532
I0428 13:09:30.432737  9322 solver.cpp:218] Iteration 1896 (2.32347 iter/s, 5.1647s/12 iters), loss = 3.36248
I0428 13:09:30.432781  9322 solver.cpp:237]     Train net output #0: loss = 3.36248 (* 1 = 3.36248 loss)
I0428 13:09:30.432790  9322 sgd_solver.cpp:105] Iteration 1896, lr = 0.00686897
I0428 13:09:35.532128  9322 solver.cpp:218] Iteration 1908 (2.35324 iter/s, 5.09936s/12 iters), loss = 3.61031
I0428 13:09:35.532248  9322 solver.cpp:237]     Train net output #0: loss = 3.61031 (* 1 = 3.61031 loss)
I0428 13:09:35.532256  9322 sgd_solver.cpp:105] Iteration 1908, lr = 0.00685266
I0428 13:09:40.696696  9322 solver.cpp:218] Iteration 1920 (2.32357 iter/s, 5.16446s/12 iters), loss = 3.79536
I0428 13:09:40.696738  9322 solver.cpp:237]     Train net output #0: loss = 3.79536 (* 1 = 3.79536 loss)
I0428 13:09:40.696748  9322 sgd_solver.cpp:105] Iteration 1920, lr = 0.00683639
I0428 13:09:41.003053  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:09:45.861039  9322 solver.cpp:218] Iteration 1932 (2.32364 iter/s, 5.16431s/12 iters), loss = 3.69528
I0428 13:09:45.861079  9322 solver.cpp:237]     Train net output #0: loss = 3.69528 (* 1 = 3.69528 loss)
I0428 13:09:45.861086  9322 sgd_solver.cpp:105] Iteration 1932, lr = 0.00682016
I0428 13:09:47.926695  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_1938.caffemodel
I0428 13:09:51.306841  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_1938.solverstate
I0428 13:09:56.562791  9322 solver.cpp:330] Iteration 1938, Testing net (#0)
I0428 13:09:56.562810  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:10:00.477813  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:10:01.322427  9322 solver.cpp:397]     Test net output #0: accuracy = 0.158088
I0428 13:10:01.322466  9322 solver.cpp:397]     Test net output #1: loss = 3.79623 (* 1 = 3.79623 loss)
I0428 13:10:03.229732  9322 solver.cpp:218] Iteration 1944 (0.690897 iter/s, 17.3687s/12 iters), loss = 3.4729
I0428 13:10:03.229780  9322 solver.cpp:237]     Train net output #0: loss = 3.4729 (* 1 = 3.4729 loss)
I0428 13:10:03.229789  9322 sgd_solver.cpp:105] Iteration 1944, lr = 0.00680397
I0428 13:10:08.382176  9322 solver.cpp:218] Iteration 1956 (2.32901 iter/s, 5.15241s/12 iters), loss = 3.61863
I0428 13:10:08.382337  9322 solver.cpp:237]     Train net output #0: loss = 3.61863 (* 1 = 3.61863 loss)
I0428 13:10:08.382347  9322 sgd_solver.cpp:105] Iteration 1956, lr = 0.00678782
I0428 13:10:13.484779  9322 solver.cpp:218] Iteration 1968 (2.35182 iter/s, 5.10244s/12 iters), loss = 3.489
I0428 13:10:13.484833  9322 solver.cpp:237]     Train net output #0: loss = 3.489 (* 1 = 3.489 loss)
I0428 13:10:13.484844  9322 sgd_solver.cpp:105] Iteration 1968, lr = 0.0067717
I0428 13:10:18.626021  9322 solver.cpp:218] Iteration 1980 (2.33409 iter/s, 5.14119s/12 iters), loss = 3.13543
I0428 13:10:18.626065  9322 solver.cpp:237]     Train net output #0: loss = 3.13543 (* 1 = 3.13543 loss)
I0428 13:10:18.626071  9322 sgd_solver.cpp:105] Iteration 1980, lr = 0.00675562
I0428 13:10:23.764859  9322 solver.cpp:218] Iteration 1992 (2.33517 iter/s, 5.1388s/12 iters), loss = 3.43434
I0428 13:10:23.764901  9322 solver.cpp:237]     Train net output #0: loss = 3.43434 (* 1 = 3.43434 loss)
I0428 13:10:23.764909  9322 sgd_solver.cpp:105] Iteration 1992, lr = 0.00673958
I0428 13:10:28.847048  9322 solver.cpp:218] Iteration 2004 (2.3612 iter/s, 5.08215s/12 iters), loss = 3.57095
I0428 13:10:28.847093  9322 solver.cpp:237]     Train net output #0: loss = 3.57095 (* 1 = 3.57095 loss)
I0428 13:10:28.847102  9322 sgd_solver.cpp:105] Iteration 2004, lr = 0.00672358
I0428 13:10:34.001955  9322 solver.cpp:218] Iteration 2016 (2.3279 iter/s, 5.15487s/12 iters), loss = 3.17195
I0428 13:10:34.001993  9322 solver.cpp:237]     Train net output #0: loss = 3.17195 (* 1 = 3.17195 loss)
I0428 13:10:34.002000  9322 sgd_solver.cpp:105] Iteration 2016, lr = 0.00670762
I0428 13:10:36.641165  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:10:39.184787  9322 solver.cpp:218] Iteration 2028 (2.31535 iter/s, 5.1828s/12 iters), loss = 3.42637
I0428 13:10:39.184914  9322 solver.cpp:237]     Train net output #0: loss = 3.42637 (* 1 = 3.42637 loss)
I0428 13:10:39.184926  9322 sgd_solver.cpp:105] Iteration 2028, lr = 0.00669169
I0428 13:10:43.757236  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_2040.caffemodel
I0428 13:10:46.954196  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_2040.solverstate
I0428 13:10:49.725425  9322 solver.cpp:330] Iteration 2040, Testing net (#0)
I0428 13:10:49.725443  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:10:53.775935  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:10:54.710346  9322 solver.cpp:397]     Test net output #0: accuracy = 0.161152
I0428 13:10:54.710392  9322 solver.cpp:397]     Test net output #1: loss = 3.7317 (* 1 = 3.7317 loss)
I0428 13:10:54.828406  9322 solver.cpp:218] Iteration 2040 (0.767089 iter/s, 15.6436s/12 iters), loss = 3.67402
I0428 13:10:54.828454  9322 solver.cpp:237]     Train net output #0: loss = 3.67402 (* 1 = 3.67402 loss)
I0428 13:10:54.828461  9322 sgd_solver.cpp:105] Iteration 2040, lr = 0.00667581
I0428 13:10:59.144924  9322 solver.cpp:218] Iteration 2052 (2.78004 iter/s, 4.31648s/12 iters), loss = 3.1965
I0428 13:10:59.144958  9322 solver.cpp:237]     Train net output #0: loss = 3.1965 (* 1 = 3.1965 loss)
I0428 13:10:59.144966  9322 sgd_solver.cpp:105] Iteration 2052, lr = 0.00665996
I0428 13:11:00.821524  9322 blocking_queue.cpp:49] Waiting for data
I0428 13:11:04.330557  9322 solver.cpp:218] Iteration 2064 (2.3141 iter/s, 5.18561s/12 iters), loss = 3.52651
I0428 13:11:04.330603  9322 solver.cpp:237]     Train net output #0: loss = 3.52651 (* 1 = 3.52651 loss)
I0428 13:11:04.330615  9322 sgd_solver.cpp:105] Iteration 2064, lr = 0.00664414
I0428 13:11:09.509946  9322 solver.cpp:218] Iteration 2076 (2.31689 iter/s, 5.17936s/12 iters), loss = 3.4614
I0428 13:11:09.510077  9322 solver.cpp:237]     Train net output #0: loss = 3.4614 (* 1 = 3.4614 loss)
I0428 13:11:09.510087  9322 sgd_solver.cpp:105] Iteration 2076, lr = 0.00662837
I0428 13:11:14.639423  9322 solver.cpp:218] Iteration 2088 (2.33947 iter/s, 5.12936s/12 iters), loss = 3.2203
I0428 13:11:14.639468  9322 solver.cpp:237]     Train net output #0: loss = 3.2203 (* 1 = 3.2203 loss)
I0428 13:11:14.639477  9322 sgd_solver.cpp:105] Iteration 2088, lr = 0.00661263
I0428 13:11:19.773173  9322 solver.cpp:218] Iteration 2100 (2.33749 iter/s, 5.13371s/12 iters), loss = 3.41641
I0428 13:11:19.773216  9322 solver.cpp:237]     Train net output #0: loss = 3.41641 (* 1 = 3.41641 loss)
I0428 13:11:19.773224  9322 sgd_solver.cpp:105] Iteration 2100, lr = 0.00659693
I0428 13:11:24.923477  9322 solver.cpp:218] Iteration 2112 (2.32997 iter/s, 5.15027s/12 iters), loss = 3.13345
I0428 13:11:24.923521  9322 solver.cpp:237]     Train net output #0: loss = 3.13345 (* 1 = 3.13345 loss)
I0428 13:11:24.923528  9322 sgd_solver.cpp:105] Iteration 2112, lr = 0.00658127
I0428 13:11:29.745761  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:11:30.082315  9322 solver.cpp:218] Iteration 2124 (2.32612 iter/s, 5.1588s/12 iters), loss = 3.37605
I0428 13:11:30.082360  9322 solver.cpp:237]     Train net output #0: loss = 3.37605 (* 1 = 3.37605 loss)
I0428 13:11:30.082370  9322 sgd_solver.cpp:105] Iteration 2124, lr = 0.00656564
I0428 13:11:35.256481  9322 solver.cpp:218] Iteration 2136 (2.31923 iter/s, 5.17413s/12 iters), loss = 3.34847
I0428 13:11:35.256525  9322 solver.cpp:237]     Train net output #0: loss = 3.34847 (* 1 = 3.34847 loss)
I0428 13:11:35.256533  9322 sgd_solver.cpp:105] Iteration 2136, lr = 0.00655006
I0428 13:11:37.321851  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_2142.caffemodel
I0428 13:11:41.445919  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_2142.solverstate
I0428 13:11:44.524087  9322 solver.cpp:330] Iteration 2142, Testing net (#0)
I0428 13:11:44.524107  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:11:48.218405  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:11:49.159718  9322 solver.cpp:397]     Test net output #0: accuracy = 0.199755
I0428 13:11:49.159760  9322 solver.cpp:397]     Test net output #1: loss = 3.5745 (* 1 = 3.5745 loss)
I0428 13:11:51.111843  9322 solver.cpp:218] Iteration 2148 (0.756841 iter/s, 15.8554s/12 iters), loss = 3.33477
I0428 13:11:51.111886  9322 solver.cpp:237]     Train net output #0: loss = 3.33477 (* 1 = 3.33477 loss)
I0428 13:11:51.111892  9322 sgd_solver.cpp:105] Iteration 2148, lr = 0.00653451
I0428 13:11:56.188170  9322 solver.cpp:218] Iteration 2160 (2.36393 iter/s, 5.0763s/12 iters), loss = 3.01078
I0428 13:11:56.188205  9322 solver.cpp:237]     Train net output #0: loss = 3.01078 (* 1 = 3.01078 loss)
I0428 13:11:56.188211  9322 sgd_solver.cpp:105] Iteration 2160, lr = 0.00651899
I0428 13:12:01.386556  9322 solver.cpp:218] Iteration 2172 (2.30842 iter/s, 5.19836s/12 iters), loss = 3.40232
I0428 13:12:01.386608  9322 solver.cpp:237]     Train net output #0: loss = 3.40232 (* 1 = 3.40232 loss)
I0428 13:12:01.386616  9322 sgd_solver.cpp:105] Iteration 2172, lr = 0.00650351
I0428 13:12:06.558133  9322 solver.cpp:218] Iteration 2184 (2.3204 iter/s, 5.17153s/12 iters), loss = 2.98046
I0428 13:12:06.558172  9322 solver.cpp:237]     Train net output #0: loss = 2.98046 (* 1 = 2.98046 loss)
I0428 13:12:06.558182  9322 sgd_solver.cpp:105] Iteration 2184, lr = 0.00648807
I0428 13:12:11.685861  9322 solver.cpp:218] Iteration 2196 (2.34023 iter/s, 5.1277s/12 iters), loss = 3.44038
I0428 13:12:11.686007  9322 solver.cpp:237]     Train net output #0: loss = 3.44038 (* 1 = 3.44038 loss)
I0428 13:12:11.686015  9322 sgd_solver.cpp:105] Iteration 2196, lr = 0.00647267
I0428 13:12:16.853042  9322 solver.cpp:218] Iteration 2208 (2.32241 iter/s, 5.16705s/12 iters), loss = 2.94556
I0428 13:12:16.853083  9322 solver.cpp:237]     Train net output #0: loss = 2.94556 (* 1 = 2.94556 loss)
I0428 13:12:16.853091  9322 sgd_solver.cpp:105] Iteration 2208, lr = 0.0064573
I0428 13:12:22.070623  9322 solver.cpp:218] Iteration 2220 (2.29993 iter/s, 5.21754s/12 iters), loss = 3.2034
I0428 13:12:22.070667  9322 solver.cpp:237]     Train net output #0: loss = 3.2034 (* 1 = 3.2034 loss)
I0428 13:12:22.070675  9322 sgd_solver.cpp:105] Iteration 2220, lr = 0.00644197
I0428 13:12:23.947247  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:12:27.255089  9322 solver.cpp:218] Iteration 2232 (2.31463 iter/s, 5.18442s/12 iters), loss = 3.21752
I0428 13:12:27.255146  9322 solver.cpp:237]     Train net output #0: loss = 3.21752 (* 1 = 3.21752 loss)
I0428 13:12:27.255157  9322 sgd_solver.cpp:105] Iteration 2232, lr = 0.00642668
I0428 13:12:31.893877  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_2244.caffemodel
I0428 13:12:35.134012  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_2244.solverstate
I0428 13:12:37.873160  9322 solver.cpp:330] Iteration 2244, Testing net (#0)
I0428 13:12:37.873183  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:12:41.717022  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:12:42.690946  9322 solver.cpp:397]     Test net output #0: accuracy = 0.183824
I0428 13:12:42.691000  9322 solver.cpp:397]     Test net output #1: loss = 3.51116 (* 1 = 3.51116 loss)
I0428 13:12:42.808952  9322 solver.cpp:218] Iteration 2244 (0.771512 iter/s, 15.5539s/12 iters), loss = 2.97242
I0428 13:12:42.809001  9322 solver.cpp:237]     Train net output #0: loss = 2.97242 (* 1 = 2.97242 loss)
I0428 13:12:42.809010  9322 sgd_solver.cpp:105] Iteration 2244, lr = 0.00641142
I0428 13:12:47.231242  9322 solver.cpp:218] Iteration 2256 (2.71355 iter/s, 4.42225s/12 iters), loss = 3.08621
I0428 13:12:47.231282  9322 solver.cpp:237]     Train net output #0: loss = 3.08621 (* 1 = 3.08621 loss)
I0428 13:12:47.231290  9322 sgd_solver.cpp:105] Iteration 2256, lr = 0.0063962
I0428 13:12:52.393550  9322 solver.cpp:218] Iteration 2268 (2.32455 iter/s, 5.16228s/12 iters), loss = 3.24133
I0428 13:12:52.393592  9322 solver.cpp:237]     Train net output #0: loss = 3.24133 (* 1 = 3.24133 loss)
I0428 13:12:52.393601  9322 sgd_solver.cpp:105] Iteration 2268, lr = 0.00638101
I0428 13:12:57.662794  9322 solver.cpp:218] Iteration 2280 (2.27738 iter/s, 5.26921s/12 iters), loss = 3.08474
I0428 13:12:57.662833  9322 solver.cpp:237]     Train net output #0: loss = 3.08474 (* 1 = 3.08474 loss)
I0428 13:12:57.662842  9322 sgd_solver.cpp:105] Iteration 2280, lr = 0.00636586
I0428 13:13:02.853621  9322 solver.cpp:218] Iteration 2292 (2.31178 iter/s, 5.1908s/12 iters), loss = 3.3714
I0428 13:13:02.853663  9322 solver.cpp:237]     Train net output #0: loss = 3.3714 (* 1 = 3.3714 loss)
I0428 13:13:02.853672  9322 sgd_solver.cpp:105] Iteration 2292, lr = 0.00635075
I0428 13:13:08.020823  9322 solver.cpp:218] Iteration 2304 (2.32235 iter/s, 5.16717s/12 iters), loss = 3.06948
I0428 13:13:08.020864  9322 solver.cpp:237]     Train net output #0: loss = 3.06948 (* 1 = 3.06948 loss)
I0428 13:13:08.020871  9322 sgd_solver.cpp:105] Iteration 2304, lr = 0.00633567
I0428 13:13:13.172690  9322 solver.cpp:218] Iteration 2316 (2.32927 iter/s, 5.15183s/12 iters), loss = 2.77114
I0428 13:13:13.172782  9322 solver.cpp:237]     Train net output #0: loss = 2.77114 (* 1 = 2.77114 loss)
I0428 13:13:13.172792  9322 sgd_solver.cpp:105] Iteration 2316, lr = 0.00632063
I0428 13:13:17.341464  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:13:18.449254  9322 solver.cpp:218] Iteration 2328 (2.27424 iter/s, 5.27648s/12 iters), loss = 2.96379
I0428 13:13:18.449298  9322 solver.cpp:237]     Train net output #0: loss = 2.96379 (* 1 = 2.96379 loss)
I0428 13:13:18.449309  9322 sgd_solver.cpp:105] Iteration 2328, lr = 0.00630562
I0428 13:13:23.601074  9322 solver.cpp:218] Iteration 2340 (2.32929 iter/s, 5.15178s/12 iters), loss = 3.08294
I0428 13:13:23.601119  9322 solver.cpp:237]     Train net output #0: loss = 3.08294 (* 1 = 3.08294 loss)
I0428 13:13:23.601127  9322 sgd_solver.cpp:105] Iteration 2340, lr = 0.00629065
I0428 13:13:25.677791  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_2346.caffemodel
I0428 13:13:28.940474  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_2346.solverstate
I0428 13:13:32.016240  9322 solver.cpp:330] Iteration 2346, Testing net (#0)
I0428 13:13:32.016259  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:13:35.839905  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:13:36.916172  9322 solver.cpp:397]     Test net output #0: accuracy = 0.209559
I0428 13:13:36.916204  9322 solver.cpp:397]     Test net output #1: loss = 3.41801 (* 1 = 3.41801 loss)
I0428 13:13:38.819365  9322 solver.cpp:218] Iteration 2352 (0.788524 iter/s, 15.2183s/12 iters), loss = 3.04147
I0428 13:13:38.819399  9322 solver.cpp:237]     Train net output #0: loss = 3.04147 (* 1 = 3.04147 loss)
I0428 13:13:38.819407  9322 sgd_solver.cpp:105] Iteration 2352, lr = 0.00627571
I0428 13:13:44.099794  9322 solver.cpp:218] Iteration 2364 (2.27256 iter/s, 5.2804s/12 iters), loss = 2.97751
I0428 13:13:44.099941  9322 solver.cpp:237]     Train net output #0: loss = 2.97751 (* 1 = 2.97751 loss)
I0428 13:13:44.099951  9322 sgd_solver.cpp:105] Iteration 2364, lr = 0.00626081
I0428 13:13:49.376816  9322 solver.cpp:218] Iteration 2376 (2.27407 iter/s, 5.27688s/12 iters), loss = 3.09419
I0428 13:13:49.376863  9322 solver.cpp:237]     Train net output #0: loss = 3.09419 (* 1 = 3.09419 loss)
I0428 13:13:49.376870  9322 sgd_solver.cpp:105] Iteration 2376, lr = 0.00624595
I0428 13:13:54.673013  9322 solver.cpp:218] Iteration 2388 (2.26579 iter/s, 5.29616s/12 iters), loss = 2.55423
I0428 13:13:54.673056  9322 solver.cpp:237]     Train net output #0: loss = 2.55423 (* 1 = 2.55423 loss)
I0428 13:13:54.673064  9322 sgd_solver.cpp:105] Iteration 2388, lr = 0.00623112
I0428 13:13:59.893949  9322 solver.cpp:218] Iteration 2400 (2.29845 iter/s, 5.2209s/12 iters), loss = 3.07157
I0428 13:13:59.893990  9322 solver.cpp:237]     Train net output #0: loss = 3.07157 (* 1 = 3.07157 loss)
I0428 13:13:59.893997  9322 sgd_solver.cpp:105] Iteration 2400, lr = 0.00621633
I0428 13:14:05.250779  9322 solver.cpp:218] Iteration 2412 (2.24014 iter/s, 5.3568s/12 iters), loss = 2.57581
I0428 13:14:05.250820  9322 solver.cpp:237]     Train net output #0: loss = 2.57581 (* 1 = 2.57581 loss)
I0428 13:14:05.250828  9322 sgd_solver.cpp:105] Iteration 2412, lr = 0.00620157
I0428 13:14:10.536793  9322 solver.cpp:218] Iteration 2424 (2.27015 iter/s, 5.28598s/12 iters), loss = 2.83555
I0428 13:14:10.536834  9322 solver.cpp:237]     Train net output #0: loss = 2.83555 (* 1 = 2.83555 loss)
I0428 13:14:10.536842  9322 sgd_solver.cpp:105] Iteration 2424, lr = 0.00618684
I0428 13:14:11.637820  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:14:15.705819  9322 solver.cpp:218] Iteration 2436 (2.32153 iter/s, 5.169s/12 iters), loss = 3.17955
I0428 13:14:15.705958  9322 solver.cpp:237]     Train net output #0: loss = 3.17955 (* 1 = 3.17955 loss)
I0428 13:14:15.705969  9322 sgd_solver.cpp:105] Iteration 2436, lr = 0.00617215
I0428 13:14:20.356748  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_2448.caffemodel
I0428 13:14:25.446841  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_2448.solverstate
I0428 13:14:30.434324  9322 solver.cpp:330] Iteration 2448, Testing net (#0)
I0428 13:14:30.434341  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:14:34.281141  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:14:35.400182  9322 solver.cpp:397]     Test net output #0: accuracy = 0.22549
I0428 13:14:35.400228  9322 solver.cpp:397]     Test net output #1: loss = 3.34636 (* 1 = 3.34636 loss)
I0428 13:14:35.519296  9322 solver.cpp:218] Iteration 2448 (0.60565 iter/s, 19.8134s/12 iters), loss = 2.80396
I0428 13:14:35.519346  9322 solver.cpp:237]     Train net output #0: loss = 2.80396 (* 1 = 2.80396 loss)
I0428 13:14:35.519354  9322 sgd_solver.cpp:105] Iteration 2448, lr = 0.0061575
I0428 13:14:39.903354  9322 solver.cpp:218] Iteration 2460 (2.73722 iter/s, 4.38401s/12 iters), loss = 3.07097
I0428 13:14:39.903393  9322 solver.cpp:237]     Train net output #0: loss = 3.07097 (* 1 = 3.07097 loss)
I0428 13:14:39.903401  9322 sgd_solver.cpp:105] Iteration 2460, lr = 0.00614288
I0428 13:14:45.082665  9322 solver.cpp:218] Iteration 2472 (2.31692 iter/s, 5.17928s/12 iters), loss = 2.75937
I0428 13:14:45.082708  9322 solver.cpp:237]     Train net output #0: loss = 2.75937 (* 1 = 2.75937 loss)
I0428 13:14:45.082716  9322 sgd_solver.cpp:105] Iteration 2472, lr = 0.0061283
I0428 13:14:50.298811  9322 solver.cpp:218] Iteration 2484 (2.30057 iter/s, 5.21611s/12 iters), loss = 3.29404
I0428 13:14:50.298961  9322 solver.cpp:237]     Train net output #0: loss = 3.29404 (* 1 = 3.29404 loss)
I0428 13:14:50.298970  9322 sgd_solver.cpp:105] Iteration 2484, lr = 0.00611375
I0428 13:14:55.454041  9322 solver.cpp:218] Iteration 2496 (2.3278 iter/s, 5.15509s/12 iters), loss = 2.88244
I0428 13:14:55.454083  9322 solver.cpp:237]     Train net output #0: loss = 2.88244 (* 1 = 2.88244 loss)
I0428 13:14:55.454092  9322 sgd_solver.cpp:105] Iteration 2496, lr = 0.00609923
I0428 13:15:00.667819  9322 solver.cpp:218] Iteration 2508 (2.30161 iter/s, 5.21374s/12 iters), loss = 2.98855
I0428 13:15:00.667856  9322 solver.cpp:237]     Train net output #0: loss = 2.98855 (* 1 = 2.98855 loss)
I0428 13:15:00.667863  9322 sgd_solver.cpp:105] Iteration 2508, lr = 0.00608475
I0428 13:15:05.881361  9322 solver.cpp:218] Iteration 2520 (2.30171 iter/s, 5.21351s/12 iters), loss = 3.10868
I0428 13:15:05.881402  9322 solver.cpp:237]     Train net output #0: loss = 3.10868 (* 1 = 3.10868 loss)
I0428 13:15:05.881410  9322 sgd_solver.cpp:105] Iteration 2520, lr = 0.0060703
I0428 13:15:09.231621  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:15:11.119773  9322 solver.cpp:218] Iteration 2532 (2.29079 iter/s, 5.23837s/12 iters), loss = 2.40584
I0428 13:15:11.119822  9322 solver.cpp:237]     Train net output #0: loss = 2.40584 (* 1 = 2.40584 loss)
I0428 13:15:11.119830  9322 sgd_solver.cpp:105] Iteration 2532, lr = 0.00605589
I0428 13:15:16.330521  9322 solver.cpp:218] Iteration 2544 (2.30295 iter/s, 5.21071s/12 iters), loss = 2.89142
I0428 13:15:16.330567  9322 solver.cpp:237]     Train net output #0: loss = 2.89142 (* 1 = 2.89142 loss)
I0428 13:15:16.330575  9322 sgd_solver.cpp:105] Iteration 2544, lr = 0.00604151
I0428 13:15:18.439289  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_2550.caffemodel
I0428 13:15:23.266069  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_2550.solverstate
I0428 13:15:30.665289  9322 solver.cpp:330] Iteration 2550, Testing net (#0)
I0428 13:15:30.665308  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:15:34.567813  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:15:35.796667  9322 solver.cpp:397]     Test net output #0: accuracy = 0.231005
I0428 13:15:35.796694  9322 solver.cpp:397]     Test net output #1: loss = 3.40425 (* 1 = 3.40425 loss)
I0428 13:15:37.702325  9322 solver.cpp:218] Iteration 2556 (0.561487 iter/s, 21.3718s/12 iters), loss = 2.97581
I0428 13:15:37.702373  9322 solver.cpp:237]     Train net output #0: loss = 2.97581 (* 1 = 2.97581 loss)
I0428 13:15:37.702381  9322 sgd_solver.cpp:105] Iteration 2556, lr = 0.00602717
I0428 13:15:42.885864  9322 solver.cpp:218] Iteration 2568 (2.31504 iter/s, 5.1835s/12 iters), loss = 2.8873
I0428 13:15:42.885903  9322 solver.cpp:237]     Train net output #0: loss = 2.8873 (* 1 = 2.8873 loss)
I0428 13:15:42.885911  9322 sgd_solver.cpp:105] Iteration 2568, lr = 0.00601286
I0428 13:15:48.076927  9322 solver.cpp:218] Iteration 2580 (2.31168 iter/s, 5.19103s/12 iters), loss = 3.05593
I0428 13:15:48.076972  9322 solver.cpp:237]     Train net output #0: loss = 3.05593 (* 1 = 3.05593 loss)
I0428 13:15:48.076980  9322 sgd_solver.cpp:105] Iteration 2580, lr = 0.00599858
I0428 13:15:53.231398  9322 solver.cpp:218] Iteration 2592 (2.3281 iter/s, 5.15443s/12 iters), loss = 2.5807
I0428 13:15:53.231444  9322 solver.cpp:237]     Train net output #0: loss = 2.5807 (* 1 = 2.5807 loss)
I0428 13:15:53.231452  9322 sgd_solver.cpp:105] Iteration 2592, lr = 0.00598434
I0428 13:15:58.387301  9322 solver.cpp:218] Iteration 2604 (2.32745 iter/s, 5.15587s/12 iters), loss = 2.4114
I0428 13:15:58.387437  9322 solver.cpp:237]     Train net output #0: loss = 2.4114 (* 1 = 2.4114 loss)
I0428 13:15:58.387447  9322 sgd_solver.cpp:105] Iteration 2604, lr = 0.00597013
I0428 13:16:03.486804  9322 solver.cpp:218] Iteration 2616 (2.35323 iter/s, 5.09937s/12 iters), loss = 2.44876
I0428 13:16:03.486852  9322 solver.cpp:237]     Train net output #0: loss = 2.44876 (* 1 = 2.44876 loss)
I0428 13:16:03.486869  9322 sgd_solver.cpp:105] Iteration 2616, lr = 0.00595596
I0428 13:16:08.652035  9322 solver.cpp:218] Iteration 2628 (2.32324 iter/s, 5.16519s/12 iters), loss = 2.63204
I0428 13:16:08.652078  9322 solver.cpp:237]     Train net output #0: loss = 2.63204 (* 1 = 2.63204 loss)
I0428 13:16:08.652086  9322 sgd_solver.cpp:105] Iteration 2628, lr = 0.00594182
I0428 13:16:09.101473  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:16:13.821228  9322 solver.cpp:218] Iteration 2640 (2.32146 iter/s, 5.16915s/12 iters), loss = 2.49395
I0428 13:16:13.821276  9322 solver.cpp:237]     Train net output #0: loss = 2.49395 (* 1 = 2.49395 loss)
I0428 13:16:13.821285  9322 sgd_solver.cpp:105] Iteration 2640, lr = 0.00592771
I0428 13:16:18.396077  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_2652.caffemodel
I0428 13:16:21.554556  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_2652.solverstate
I0428 13:16:25.787577  9322 solver.cpp:330] Iteration 2652, Testing net (#0)
I0428 13:16:25.787595  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:16:29.499140  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:16:30.709017  9322 solver.cpp:397]     Test net output #0: accuracy = 0.258578
I0428 13:16:30.709061  9322 solver.cpp:397]     Test net output #1: loss = 3.21844 (* 1 = 3.21844 loss)
I0428 13:16:30.827082  9322 solver.cpp:218] Iteration 2652 (0.705639 iter/s, 17.0059s/12 iters), loss = 2.8134
I0428 13:16:30.827131  9322 solver.cpp:237]     Train net output #0: loss = 2.8134 (* 1 = 2.8134 loss)
I0428 13:16:30.827140  9322 sgd_solver.cpp:105] Iteration 2652, lr = 0.00591364
I0428 13:16:35.136008  9322 solver.cpp:218] Iteration 2664 (2.78495 iter/s, 4.30888s/12 iters), loss = 2.3599
I0428 13:16:35.136054  9322 solver.cpp:237]     Train net output #0: loss = 2.3599 (* 1 = 2.3599 loss)
I0428 13:16:35.136062  9322 sgd_solver.cpp:105] Iteration 2664, lr = 0.0058996
I0428 13:16:40.308581  9322 solver.cpp:218] Iteration 2676 (2.31995 iter/s, 5.17253s/12 iters), loss = 2.57789
I0428 13:16:40.308621  9322 solver.cpp:237]     Train net output #0: loss = 2.57789 (* 1 = 2.57789 loss)
I0428 13:16:40.308629  9322 sgd_solver.cpp:105] Iteration 2676, lr = 0.00588559
I0428 13:16:45.479041  9322 solver.cpp:218] Iteration 2688 (2.32089 iter/s, 5.17042s/12 iters), loss = 2.57722
I0428 13:16:45.479085  9322 solver.cpp:237]     Train net output #0: loss = 2.57722 (* 1 = 2.57722 loss)
I0428 13:16:45.479094  9322 sgd_solver.cpp:105] Iteration 2688, lr = 0.00587162
I0428 13:16:50.652032  9322 solver.cpp:218] Iteration 2700 (2.31976 iter/s, 5.17295s/12 iters), loss = 2.51432
I0428 13:16:50.652088  9322 solver.cpp:237]     Train net output #0: loss = 2.51432 (* 1 = 2.51432 loss)
I0428 13:16:50.652101  9322 sgd_solver.cpp:105] Iteration 2700, lr = 0.00585768
I0428 13:16:55.845414  9322 solver.cpp:218] Iteration 2712 (2.31065 iter/s, 5.19334s/12 iters), loss = 2.58732
I0428 13:16:55.845461  9322 solver.cpp:237]     Train net output #0: loss = 2.58732 (* 1 = 2.58732 loss)
I0428 13:16:55.845470  9322 sgd_solver.cpp:105] Iteration 2712, lr = 0.00584377
I0428 13:17:01.019475  9322 solver.cpp:218] Iteration 2724 (2.31928 iter/s, 5.17402s/12 iters), loss = 2.29073
I0428 13:17:01.019644  9322 solver.cpp:237]     Train net output #0: loss = 2.29073 (* 1 = 2.29073 loss)
I0428 13:17:01.019654  9322 sgd_solver.cpp:105] Iteration 2724, lr = 0.0058299
I0428 13:17:03.616801  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:17:06.137502  9322 solver.cpp:218] Iteration 2736 (2.34473 iter/s, 5.11786s/12 iters), loss = 2.3235
I0428 13:17:06.137547  9322 solver.cpp:237]     Train net output #0: loss = 2.3235 (* 1 = 2.3235 loss)
I0428 13:17:06.137555  9322 sgd_solver.cpp:105] Iteration 2736, lr = 0.00581605
I0428 13:17:11.294764  9322 solver.cpp:218] Iteration 2748 (2.32683 iter/s, 5.15723s/12 iters), loss = 2.63159
I0428 13:17:11.294809  9322 solver.cpp:237]     Train net output #0: loss = 2.63159 (* 1 = 2.63159 loss)
I0428 13:17:11.294817  9322 sgd_solver.cpp:105] Iteration 2748, lr = 0.00580225
I0428 13:17:13.380367  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_2754.caffemodel
I0428 13:17:17.484206  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_2754.solverstate
I0428 13:17:22.473210  9322 solver.cpp:330] Iteration 2754, Testing net (#0)
I0428 13:17:22.473229  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:17:25.879873  9322 blocking_queue.cpp:49] Waiting for data
I0428 13:17:26.156159  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:17:27.345001  9322 solver.cpp:397]     Test net output #0: accuracy = 0.269608
I0428 13:17:27.345038  9322 solver.cpp:397]     Test net output #1: loss = 3.17752 (* 1 = 3.17752 loss)
I0428 13:17:29.281143  9322 solver.cpp:218] Iteration 2760 (0.667171 iter/s, 17.9864s/12 iters), loss = 2.27672
I0428 13:17:29.281188  9322 solver.cpp:237]     Train net output #0: loss = 2.27672 (* 1 = 2.27672 loss)
I0428 13:17:29.281196  9322 sgd_solver.cpp:105] Iteration 2760, lr = 0.00578847
I0428 13:17:34.483307  9322 solver.cpp:218] Iteration 2772 (2.30675 iter/s, 5.20212s/12 iters), loss = 2.28818
I0428 13:17:34.483419  9322 solver.cpp:237]     Train net output #0: loss = 2.28818 (* 1 = 2.28818 loss)
I0428 13:17:34.483429  9322 sgd_solver.cpp:105] Iteration 2772, lr = 0.00577473
I0428 13:17:39.645833  9322 solver.cpp:218] Iteration 2784 (2.32449 iter/s, 5.16242s/12 iters), loss = 2.68341
I0428 13:17:39.645881  9322 solver.cpp:237]     Train net output #0: loss = 2.68341 (* 1 = 2.68341 loss)
I0428 13:17:39.645890  9322 sgd_solver.cpp:105] Iteration 2784, lr = 0.00576102
I0428 13:17:44.785245  9322 solver.cpp:218] Iteration 2796 (2.33492 iter/s, 5.13937s/12 iters), loss = 2.23877
I0428 13:17:44.785282  9322 solver.cpp:237]     Train net output #0: loss = 2.23877 (* 1 = 2.23877 loss)
I0428 13:17:44.785290  9322 sgd_solver.cpp:105] Iteration 2796, lr = 0.00574734
I0428 13:17:49.938593  9322 solver.cpp:218] Iteration 2808 (2.3286 iter/s, 5.15332s/12 iters), loss = 2.08391
I0428 13:17:49.938644  9322 solver.cpp:237]     Train net output #0: loss = 2.08391 (* 1 = 2.08391 loss)
I0428 13:17:49.938653  9322 sgd_solver.cpp:105] Iteration 2808, lr = 0.00573369
I0428 13:17:55.098390  9322 solver.cpp:218] Iteration 2820 (2.32569 iter/s, 5.15975s/12 iters), loss = 2.25334
I0428 13:17:55.098431  9322 solver.cpp:237]     Train net output #0: loss = 2.25334 (* 1 = 2.25334 loss)
I0428 13:17:55.098440  9322 sgd_solver.cpp:105] Iteration 2820, lr = 0.00572008
I0428 13:17:59.966555  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:18:00.275010  9322 solver.cpp:218] Iteration 2832 (2.31813 iter/s, 5.17659s/12 iters), loss = 2.36684
I0428 13:18:00.275056  9322 solver.cpp:237]     Train net output #0: loss = 2.36684 (* 1 = 2.36684 loss)
I0428 13:18:00.275065  9322 sgd_solver.cpp:105] Iteration 2832, lr = 0.0057065
I0428 13:18:05.444571  9322 solver.cpp:218] Iteration 2844 (2.3213 iter/s, 5.16953s/12 iters), loss = 2.3496
I0428 13:18:05.444698  9322 solver.cpp:237]     Train net output #0: loss = 2.3496 (* 1 = 2.3496 loss)
I0428 13:18:05.444707  9322 sgd_solver.cpp:105] Iteration 2844, lr = 0.00569295
I0428 13:18:09.962460  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_2856.caffemodel
I0428 13:18:17.471179  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_2856.solverstate
I0428 13:18:27.924641  9322 solver.cpp:330] Iteration 2856, Testing net (#0)
I0428 13:18:27.924659  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:18:31.543787  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:18:32.834661  9322 solver.cpp:397]     Test net output #0: accuracy = 0.28799
I0428 13:18:32.834694  9322 solver.cpp:397]     Test net output #1: loss = 3.10741 (* 1 = 3.10741 loss)
I0428 13:18:32.952605  9322 solver.cpp:218] Iteration 2856 (0.436236 iter/s, 27.508s/12 iters), loss = 2.24369
I0428 13:18:32.952646  9322 solver.cpp:237]     Train net output #0: loss = 2.24369 (* 1 = 2.24369 loss)
I0428 13:18:32.952654  9322 sgd_solver.cpp:105] Iteration 2856, lr = 0.00567944
I0428 13:18:37.282407  9322 solver.cpp:218] Iteration 2868 (2.77151 iter/s, 4.32976s/12 iters), loss = 2.12458
I0428 13:18:37.282527  9322 solver.cpp:237]     Train net output #0: loss = 2.12458 (* 1 = 2.12458 loss)
I0428 13:18:37.282536  9322 sgd_solver.cpp:105] Iteration 2868, lr = 0.00566595
I0428 13:18:42.440201  9322 solver.cpp:218] Iteration 2880 (2.32663 iter/s, 5.15768s/12 iters), loss = 2.35466
I0428 13:18:42.440248  9322 solver.cpp:237]     Train net output #0: loss = 2.35466 (* 1 = 2.35466 loss)
I0428 13:18:42.440256  9322 sgd_solver.cpp:105] Iteration 2880, lr = 0.0056525
I0428 13:18:47.596132  9322 solver.cpp:218] Iteration 2892 (2.32744 iter/s, 5.15589s/12 iters), loss = 1.95474
I0428 13:18:47.596180  9322 solver.cpp:237]     Train net output #0: loss = 1.95474 (* 1 = 1.95474 loss)
I0428 13:18:47.596189  9322 sgd_solver.cpp:105] Iteration 2892, lr = 0.00563908
I0428 13:18:52.732547  9322 solver.cpp:218] Iteration 2904 (2.33628 iter/s, 5.13637s/12 iters), loss = 2.25163
I0428 13:18:52.732592  9322 solver.cpp:237]     Train net output #0: loss = 2.25163 (* 1 = 2.25163 loss)
I0428 13:18:52.732601  9322 sgd_solver.cpp:105] Iteration 2904, lr = 0.00562569
I0428 13:18:57.899113  9322 solver.cpp:218] Iteration 2916 (2.32265 iter/s, 5.16652s/12 iters), loss = 2.25884
I0428 13:18:57.899169  9322 solver.cpp:237]     Train net output #0: loss = 2.25884 (* 1 = 2.25884 loss)
I0428 13:18:57.899178  9322 sgd_solver.cpp:105] Iteration 2916, lr = 0.00561233
I0428 13:19:03.014031  9322 solver.cpp:218] Iteration 2928 (2.3461 iter/s, 5.11487s/12 iters), loss = 2.25479
I0428 13:19:03.014076  9322 solver.cpp:237]     Train net output #0: loss = 2.25479 (* 1 = 2.25479 loss)
I0428 13:19:03.014084  9322 sgd_solver.cpp:105] Iteration 2928, lr = 0.00559901
I0428 13:19:04.918237  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:19:08.194677  9322 solver.cpp:218] Iteration 2940 (2.31633 iter/s, 5.18061s/12 iters), loss = 2.22685
I0428 13:19:08.194792  9322 solver.cpp:237]     Train net output #0: loss = 2.22685 (* 1 = 2.22685 loss)
I0428 13:19:08.194803  9322 sgd_solver.cpp:105] Iteration 2940, lr = 0.00558572
I0428 13:19:13.334189  9322 solver.cpp:218] Iteration 2952 (2.3349 iter/s, 5.1394s/12 iters), loss = 2.1512
I0428 13:19:13.334236  9322 solver.cpp:237]     Train net output #0: loss = 2.1512 (* 1 = 2.1512 loss)
I0428 13:19:13.334244  9322 sgd_solver.cpp:105] Iteration 2952, lr = 0.00557245
I0428 13:19:15.411190  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_2958.caffemodel
I0428 13:19:23.511162  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_2958.solverstate
I0428 13:19:32.686056  9322 solver.cpp:330] Iteration 2958, Testing net (#0)
I0428 13:19:32.686080  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:19:36.292403  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:19:37.649123  9322 solver.cpp:397]     Test net output #0: accuracy = 0.278186
I0428 13:19:37.649163  9322 solver.cpp:397]     Test net output #1: loss = 3.04256 (* 1 = 3.04256 loss)
I0428 13:19:39.551997  9322 solver.cpp:218] Iteration 2964 (0.457703 iter/s, 26.2179s/12 iters), loss = 1.95791
I0428 13:19:39.552125  9322 solver.cpp:237]     Train net output #0: loss = 1.95791 (* 1 = 1.95791 loss)
I0428 13:19:39.552135  9322 sgd_solver.cpp:105] Iteration 2964, lr = 0.00555922
I0428 13:19:44.707403  9322 solver.cpp:218] Iteration 2976 (2.32771 iter/s, 5.15527s/12 iters), loss = 2.53553
I0428 13:19:44.707465  9322 solver.cpp:237]     Train net output #0: loss = 2.53553 (* 1 = 2.53553 loss)
I0428 13:19:44.707479  9322 sgd_solver.cpp:105] Iteration 2976, lr = 0.00554603
I0428 13:19:49.794521  9322 solver.cpp:218] Iteration 2988 (2.35893 iter/s, 5.08706s/12 iters), loss = 2.36299
I0428 13:19:49.794561  9322 solver.cpp:237]     Train net output #0: loss = 2.36299 (* 1 = 2.36299 loss)
I0428 13:19:49.794569  9322 sgd_solver.cpp:105] Iteration 2988, lr = 0.00553286
I0428 13:19:54.949673  9322 solver.cpp:218] Iteration 3000 (2.32779 iter/s, 5.15511s/12 iters), loss = 2.22442
I0428 13:19:54.949720  9322 solver.cpp:237]     Train net output #0: loss = 2.22442 (* 1 = 2.22442 loss)
I0428 13:19:54.949729  9322 sgd_solver.cpp:105] Iteration 3000, lr = 0.00551972
I0428 13:20:00.125547  9322 solver.cpp:218] Iteration 3012 (2.31847 iter/s, 5.17583s/12 iters), loss = 2.18059
I0428 13:20:00.125597  9322 solver.cpp:237]     Train net output #0: loss = 2.18059 (* 1 = 2.18059 loss)
I0428 13:20:00.125605  9322 sgd_solver.cpp:105] Iteration 3012, lr = 0.00550662
I0428 13:20:05.275478  9322 solver.cpp:218] Iteration 3024 (2.33015 iter/s, 5.14989s/12 iters), loss = 2.00274
I0428 13:20:05.275521  9322 solver.cpp:237]     Train net output #0: loss = 2.00274 (* 1 = 2.00274 loss)
I0428 13:20:05.275529  9322 sgd_solver.cpp:105] Iteration 3024, lr = 0.00549354
I0428 13:20:09.377030  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:20:10.449934  9322 solver.cpp:218] Iteration 3036 (2.3191 iter/s, 5.17442s/12 iters), loss = 1.84014
I0428 13:20:10.450024  9322 solver.cpp:237]     Train net output #0: loss = 1.84014 (* 1 = 1.84014 loss)
I0428 13:20:10.450033  9322 sgd_solver.cpp:105] Iteration 3036, lr = 0.0054805
I0428 13:20:15.595346  9322 solver.cpp:218] Iteration 3048 (2.33221 iter/s, 5.14533s/12 iters), loss = 2.00478
I0428 13:20:15.595386  9322 solver.cpp:237]     Train net output #0: loss = 2.00478 (* 1 = 2.00478 loss)
I0428 13:20:15.595394  9322 sgd_solver.cpp:105] Iteration 3048, lr = 0.00546749
I0428 13:20:20.265782  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_3060.caffemodel
I0428 13:20:26.627377  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_3060.solverstate
I0428 13:20:32.864444  9322 solver.cpp:330] Iteration 3060, Testing net (#0)
I0428 13:20:32.864468  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:20:36.391345  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:20:37.808290  9322 solver.cpp:397]     Test net output #0: accuracy = 0.306373
I0428 13:20:37.808322  9322 solver.cpp:397]     Test net output #1: loss = 2.91093 (* 1 = 2.91093 loss)
I0428 13:20:37.926101  9322 solver.cpp:218] Iteration 3060 (0.537374 iter/s, 22.3308s/12 iters), loss = 1.88026
I0428 13:20:37.926149  9322 solver.cpp:237]     Train net output #0: loss = 1.88026 (* 1 = 1.88026 loss)
I0428 13:20:37.926158  9322 sgd_solver.cpp:105] Iteration 3060, lr = 0.00545451
I0428 13:20:42.139679  9322 solver.cpp:218] Iteration 3072 (2.84797 iter/s, 4.21353s/12 iters), loss = 2.05321
I0428 13:20:42.139772  9322 solver.cpp:237]     Train net output #0: loss = 2.05321 (* 1 = 2.05321 loss)
I0428 13:20:42.139781  9322 sgd_solver.cpp:105] Iteration 3072, lr = 0.00544156
I0428 13:20:47.266610  9322 solver.cpp:218] Iteration 3084 (2.34062 iter/s, 5.12684s/12 iters), loss = 2.02498
I0428 13:20:47.266655  9322 solver.cpp:237]     Train net output #0: loss = 2.02498 (* 1 = 2.02498 loss)
I0428 13:20:47.266664  9322 sgd_solver.cpp:105] Iteration 3084, lr = 0.00542864
I0428 13:20:52.567529  9322 solver.cpp:218] Iteration 3096 (2.26377 iter/s, 5.30088s/12 iters), loss = 2.10877
I0428 13:20:52.567574  9322 solver.cpp:237]     Train net output #0: loss = 2.10877 (* 1 = 2.10877 loss)
I0428 13:20:52.567581  9322 sgd_solver.cpp:105] Iteration 3096, lr = 0.00541575
I0428 13:20:57.768213  9322 solver.cpp:218] Iteration 3108 (2.30741 iter/s, 5.20063s/12 iters), loss = 1.96053
I0428 13:20:57.768287  9322 solver.cpp:237]     Train net output #0: loss = 1.96053 (* 1 = 1.96053 loss)
I0428 13:20:57.768303  9322 sgd_solver.cpp:105] Iteration 3108, lr = 0.00540289
I0428 13:21:02.978828  9322 solver.cpp:218] Iteration 3120 (2.30302 iter/s, 5.21055s/12 iters), loss = 1.84621
I0428 13:21:02.978874  9322 solver.cpp:237]     Train net output #0: loss = 1.84621 (* 1 = 1.84621 loss)
I0428 13:21:02.978883  9322 sgd_solver.cpp:105] Iteration 3120, lr = 0.00539006
I0428 13:21:08.190187  9322 solver.cpp:218] Iteration 3132 (2.30268 iter/s, 5.21132s/12 iters), loss = 1.88815
I0428 13:21:08.190238  9322 solver.cpp:237]     Train net output #0: loss = 1.88815 (* 1 = 1.88815 loss)
I0428 13:21:08.190246  9322 sgd_solver.cpp:105] Iteration 3132, lr = 0.00537727
I0428 13:21:09.342327  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:21:13.445174  9322 solver.cpp:218] Iteration 3144 (2.28356 iter/s, 5.25494s/12 iters), loss = 2.10419
I0428 13:21:13.445325  9322 solver.cpp:237]     Train net output #0: loss = 2.10419 (* 1 = 2.10419 loss)
I0428 13:21:13.445335  9322 sgd_solver.cpp:105] Iteration 3144, lr = 0.0053645
I0428 13:21:18.594733  9322 solver.cpp:218] Iteration 3156 (2.33036 iter/s, 5.14942s/12 iters), loss = 2.10565
I0428 13:21:18.594774  9322 solver.cpp:237]     Train net output #0: loss = 2.10565 (* 1 = 2.10565 loss)
I0428 13:21:18.594784  9322 sgd_solver.cpp:105] Iteration 3156, lr = 0.00535176
I0428 13:21:20.681895  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_3162.caffemodel
I0428 13:21:25.211537  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_3162.solverstate
I0428 13:21:28.085652  9322 solver.cpp:330] Iteration 3162, Testing net (#0)
I0428 13:21:28.085672  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:21:31.572899  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:21:33.010346  9322 solver.cpp:397]     Test net output #0: accuracy = 0.31924
I0428 13:21:33.010390  9322 solver.cpp:397]     Test net output #1: loss = 2.84213 (* 1 = 2.84213 loss)
I0428 13:21:34.927960  9322 solver.cpp:218] Iteration 3168 (0.734698 iter/s, 16.3332s/12 iters), loss = 2.08041
I0428 13:21:34.928009  9322 solver.cpp:237]     Train net output #0: loss = 2.08041 (* 1 = 2.08041 loss)
I0428 13:21:34.928017  9322 sgd_solver.cpp:105] Iteration 3168, lr = 0.00533906
I0428 13:21:40.088202  9322 solver.cpp:218] Iteration 3180 (2.32549 iter/s, 5.1602s/12 iters), loss = 1.6736
I0428 13:21:40.088250  9322 solver.cpp:237]     Train net output #0: loss = 1.6736 (* 1 = 1.6736 loss)
I0428 13:21:40.088259  9322 sgd_solver.cpp:105] Iteration 3180, lr = 0.00532638
I0428 13:21:45.229336  9322 solver.cpp:218] Iteration 3192 (2.33414 iter/s, 5.14109s/12 iters), loss = 2.05965
I0428 13:21:45.229434  9322 solver.cpp:237]     Train net output #0: loss = 2.05965 (* 1 = 2.05965 loss)
I0428 13:21:45.229444  9322 sgd_solver.cpp:105] Iteration 3192, lr = 0.00531374
I0428 13:21:50.311144  9322 solver.cpp:218] Iteration 3204 (2.36141 iter/s, 5.08172s/12 iters), loss = 1.78875
I0428 13:21:50.311190  9322 solver.cpp:237]     Train net output #0: loss = 1.78875 (* 1 = 1.78875 loss)
I0428 13:21:50.311199  9322 sgd_solver.cpp:105] Iteration 3204, lr = 0.00530112
I0428 13:21:55.332068  9322 solver.cpp:218] Iteration 3216 (2.39002 iter/s, 5.02089s/12 iters), loss = 1.97352
I0428 13:21:55.332111  9322 solver.cpp:237]     Train net output #0: loss = 1.97352 (* 1 = 1.97352 loss)
I0428 13:21:55.332119  9322 sgd_solver.cpp:105] Iteration 3216, lr = 0.00528853
I0428 13:22:00.488029  9322 solver.cpp:218] Iteration 3228 (2.32742 iter/s, 5.15593s/12 iters), loss = 1.96223
I0428 13:22:00.488061  9322 solver.cpp:237]     Train net output #0: loss = 1.96223 (* 1 = 1.96223 loss)
I0428 13:22:00.488070  9322 sgd_solver.cpp:105] Iteration 3228, lr = 0.00527598
I0428 13:22:03.836669  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:22:05.670992  9322 solver.cpp:218] Iteration 3240 (2.31529 iter/s, 5.18294s/12 iters), loss = 1.79016
I0428 13:22:05.671036  9322 solver.cpp:237]     Train net output #0: loss = 1.79016 (* 1 = 1.79016 loss)
I0428 13:22:05.671044  9322 sgd_solver.cpp:105] Iteration 3240, lr = 0.00526345
I0428 13:22:10.751121  9322 solver.cpp:218] Iteration 3252 (2.36216 iter/s, 5.08009s/12 iters), loss = 1.87786
I0428 13:22:10.751161  9322 solver.cpp:237]     Train net output #0: loss = 1.87786 (* 1 = 1.87786 loss)
I0428 13:22:10.751170  9322 sgd_solver.cpp:105] Iteration 3252, lr = 0.00525095
I0428 13:22:15.455761  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_3264.caffemodel
I0428 13:22:23.470929  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_3264.solverstate
I0428 13:22:27.488054  9322 solver.cpp:330] Iteration 3264, Testing net (#0)
I0428 13:22:27.488073  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:22:30.926667  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:22:32.399528  9322 solver.cpp:397]     Test net output #0: accuracy = 0.300245
I0428 13:22:32.399566  9322 solver.cpp:397]     Test net output #1: loss = 3.03049 (* 1 = 3.03049 loss)
I0428 13:22:32.517546  9322 solver.cpp:218] Iteration 3264 (0.551307 iter/s, 21.7665s/12 iters), loss = 1.82736
I0428 13:22:32.517593  9322 solver.cpp:237]     Train net output #0: loss = 1.82736 (* 1 = 1.82736 loss)
I0428 13:22:32.517601  9322 sgd_solver.cpp:105] Iteration 3264, lr = 0.00523849
I0428 13:22:36.919958  9322 solver.cpp:218] Iteration 3276 (2.7258 iter/s, 4.40237s/12 iters), loss = 1.54042
I0428 13:22:36.919996  9322 solver.cpp:237]     Train net output #0: loss = 1.54042 (* 1 = 1.54042 loss)
I0428 13:22:36.920006  9322 sgd_solver.cpp:105] Iteration 3276, lr = 0.00522605
I0428 13:22:42.073045  9322 solver.cpp:218] Iteration 3288 (2.32872 iter/s, 5.15306s/12 iters), loss = 1.60928
I0428 13:22:42.073084  9322 solver.cpp:237]     Train net output #0: loss = 1.60928 (* 1 = 1.60928 loss)
I0428 13:22:42.073092  9322 sgd_solver.cpp:105] Iteration 3288, lr = 0.00521364
I0428 13:22:47.222483  9322 solver.cpp:218] Iteration 3300 (2.33036 iter/s, 5.14941s/12 iters), loss = 1.61792
I0428 13:22:47.222601  9322 solver.cpp:237]     Train net output #0: loss = 1.61792 (* 1 = 1.61792 loss)
I0428 13:22:47.222611  9322 sgd_solver.cpp:105] Iteration 3300, lr = 0.00520126
I0428 13:22:52.371711  9322 solver.cpp:218] Iteration 3312 (2.33049 iter/s, 5.14912s/12 iters), loss = 1.46663
I0428 13:22:52.371757  9322 solver.cpp:237]     Train net output #0: loss = 1.46663 (* 1 = 1.46663 loss)
I0428 13:22:52.371765  9322 sgd_solver.cpp:105] Iteration 3312, lr = 0.00518892
I0428 13:22:57.481591  9322 solver.cpp:218] Iteration 3324 (2.34841 iter/s, 5.10984s/12 iters), loss = 1.64393
I0428 13:22:57.481642  9322 solver.cpp:237]     Train net output #0: loss = 1.64393 (* 1 = 1.64393 loss)
I0428 13:22:57.481650  9322 sgd_solver.cpp:105] Iteration 3324, lr = 0.0051766
I0428 13:23:02.644632  9322 solver.cpp:218] Iteration 3336 (2.32423 iter/s, 5.163s/12 iters), loss = 1.80662
I0428 13:23:02.644670  9322 solver.cpp:237]     Train net output #0: loss = 1.80662 (* 1 = 1.80662 loss)
I0428 13:23:02.644677  9322 sgd_solver.cpp:105] Iteration 3336, lr = 0.00516431
I0428 13:23:03.125978  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:23:07.757392  9322 solver.cpp:218] Iteration 3348 (2.34708 iter/s, 5.11273s/12 iters), loss = 1.7591
I0428 13:23:07.757433  9322 solver.cpp:237]     Train net output #0: loss = 1.7591 (* 1 = 1.7591 loss)
I0428 13:23:07.757442  9322 sgd_solver.cpp:105] Iteration 3348, lr = 0.00515204
I0428 13:23:12.901046  9322 solver.cpp:218] Iteration 3360 (2.33299 iter/s, 5.14362s/12 iters), loss = 1.85062
I0428 13:23:12.901096  9322 solver.cpp:237]     Train net output #0: loss = 1.85062 (* 1 = 1.85062 loss)
I0428 13:23:12.901105  9322 sgd_solver.cpp:105] Iteration 3360, lr = 0.00513981
I0428 13:23:14.988723  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_3366.caffemodel
I0428 13:23:18.190434  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_3366.solverstate
I0428 13:23:20.629948  9322 solver.cpp:330] Iteration 3366, Testing net (#0)
I0428 13:23:20.629967  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:23:23.833245  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:23:25.341600  9322 solver.cpp:397]     Test net output #0: accuracy = 0.327206
I0428 13:23:25.341646  9322 solver.cpp:397]     Test net output #1: loss = 2.83616 (* 1 = 2.83616 loss)
I0428 13:23:27.281726  9322 solver.cpp:218] Iteration 3372 (0.834453 iter/s, 14.3807s/12 iters), loss = 2.06291
I0428 13:23:27.281771  9322 solver.cpp:237]     Train net output #0: loss = 2.06291 (* 1 = 2.06291 loss)
I0428 13:23:27.281780  9322 sgd_solver.cpp:105] Iteration 3372, lr = 0.00512761
I0428 13:23:32.441329  9322 solver.cpp:218] Iteration 3384 (2.32578 iter/s, 5.15957s/12 iters), loss = 1.60096
I0428 13:23:32.441373  9322 solver.cpp:237]     Train net output #0: loss = 1.60096 (* 1 = 1.60096 loss)
I0428 13:23:32.441382  9322 sgd_solver.cpp:105] Iteration 3384, lr = 0.00511544
I0428 13:23:37.519945  9322 solver.cpp:218] Iteration 3396 (2.36287 iter/s, 5.07858s/12 iters), loss = 1.81971
I0428 13:23:37.519986  9322 solver.cpp:237]     Train net output #0: loss = 1.81971 (* 1 = 1.81971 loss)
I0428 13:23:37.519994  9322 sgd_solver.cpp:105] Iteration 3396, lr = 0.00510329
I0428 13:23:42.606614  9322 solver.cpp:218] Iteration 3408 (2.35913 iter/s, 5.08663s/12 iters), loss = 1.60627
I0428 13:23:42.606660  9322 solver.cpp:237]     Train net output #0: loss = 1.60627 (* 1 = 1.60627 loss)
I0428 13:23:42.606669  9322 sgd_solver.cpp:105] Iteration 3408, lr = 0.00509117
I0428 13:23:47.762161  9322 solver.cpp:218] Iteration 3420 (2.32761 iter/s, 5.15551s/12 iters), loss = 2.09001
I0428 13:23:47.762203  9322 solver.cpp:237]     Train net output #0: loss = 2.09001 (* 1 = 2.09001 loss)
I0428 13:23:47.762212  9322 sgd_solver.cpp:105] Iteration 3420, lr = 0.00507909
I0428 13:23:52.932376  9322 solver.cpp:218] Iteration 3432 (2.321 iter/s, 5.17018s/12 iters), loss = 1.70638
I0428 13:23:52.932464  9322 solver.cpp:237]     Train net output #0: loss = 1.70638 (* 1 = 1.70638 loss)
I0428 13:23:52.932473  9322 sgd_solver.cpp:105] Iteration 3432, lr = 0.00506703
I0428 13:23:55.643528  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:23:58.122967  9322 solver.cpp:218] Iteration 3444 (2.31191 iter/s, 5.19052s/12 iters), loss = 1.64908
I0428 13:23:58.123008  9322 solver.cpp:237]     Train net output #0: loss = 1.64908 (* 1 = 1.64908 loss)
I0428 13:23:58.123015  9322 sgd_solver.cpp:105] Iteration 3444, lr = 0.005055
I0428 13:24:03.204784  9322 solver.cpp:218] Iteration 3456 (2.36138 iter/s, 5.08178s/12 iters), loss = 1.42403
I0428 13:24:03.204829  9322 solver.cpp:237]     Train net output #0: loss = 1.42403 (* 1 = 1.42403 loss)
I0428 13:24:03.204838  9322 sgd_solver.cpp:105] Iteration 3456, lr = 0.005043
I0428 13:24:07.818461  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_3468.caffemodel
I0428 13:24:13.837653  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_3468.solverstate
I0428 13:24:16.998725  9322 solver.cpp:330] Iteration 3468, Testing net (#0)
I0428 13:24:16.998756  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:24:17.433179  9322 blocking_queue.cpp:49] Waiting for data
I0428 13:24:20.338430  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:24:21.905416  9322 solver.cpp:397]     Test net output #0: accuracy = 0.3125
I0428 13:24:21.905443  9322 solver.cpp:397]     Test net output #1: loss = 2.904 (* 1 = 2.904 loss)
I0428 13:24:22.023384  9322 solver.cpp:218] Iteration 3468 (0.637666 iter/s, 18.8186s/12 iters), loss = 1.49105
I0428 13:24:22.023430  9322 solver.cpp:237]     Train net output #0: loss = 1.49105 (* 1 = 1.49105 loss)
I0428 13:24:22.023438  9322 sgd_solver.cpp:105] Iteration 3468, lr = 0.00503102
I0428 13:24:26.345124  9322 solver.cpp:218] Iteration 3480 (2.77669 iter/s, 4.32169s/12 iters), loss = 1.56228
I0428 13:24:26.345281  9322 solver.cpp:237]     Train net output #0: loss = 1.56228 (* 1 = 1.56228 loss)
I0428 13:24:26.345291  9322 sgd_solver.cpp:105] Iteration 3480, lr = 0.00501908
I0428 13:24:31.494087  9322 solver.cpp:218] Iteration 3492 (2.33063 iter/s, 5.14881s/12 iters), loss = 1.76377
I0428 13:24:31.494128  9322 solver.cpp:237]     Train net output #0: loss = 1.76377 (* 1 = 1.76377 loss)
I0428 13:24:31.494136  9322 sgd_solver.cpp:105] Iteration 3492, lr = 0.00500716
I0428 13:24:36.636498  9322 solver.cpp:218] Iteration 3504 (2.33356 iter/s, 5.14237s/12 iters), loss = 1.51945
I0428 13:24:36.636559  9322 solver.cpp:237]     Train net output #0: loss = 1.51945 (* 1 = 1.51945 loss)
I0428 13:24:36.636574  9322 sgd_solver.cpp:105] Iteration 3504, lr = 0.00499527
I0428 13:24:41.646941  9322 solver.cpp:218] Iteration 3516 (2.39502 iter/s, 5.0104s/12 iters), loss = 1.60102
I0428 13:24:41.646977  9322 solver.cpp:237]     Train net output #0: loss = 1.60102 (* 1 = 1.60102 loss)
I0428 13:24:41.646984  9322 sgd_solver.cpp:105] Iteration 3516, lr = 0.00498341
I0428 13:24:46.830811  9322 solver.cpp:218] Iteration 3528 (2.31489 iter/s, 5.18383s/12 iters), loss = 1.805
I0428 13:24:46.830858  9322 solver.cpp:237]     Train net output #0: loss = 1.805 (* 1 = 1.805 loss)
I0428 13:24:46.830866  9322 sgd_solver.cpp:105] Iteration 3528, lr = 0.00497158
I0428 13:24:51.648222  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:24:51.925199  9322 solver.cpp:218] Iteration 3540 (2.35555 iter/s, 5.09435s/12 iters), loss = 1.66924
I0428 13:24:51.925240  9322 solver.cpp:237]     Train net output #0: loss = 1.66924 (* 1 = 1.66924 loss)
I0428 13:24:51.925247  9322 sgd_solver.cpp:105] Iteration 3540, lr = 0.00495978
I0428 13:24:57.086946  9322 solver.cpp:218] Iteration 3552 (2.32481 iter/s, 5.16171s/12 iters), loss = 1.1894
I0428 13:24:57.087028  9322 solver.cpp:237]     Train net output #0: loss = 1.1894 (* 1 = 1.1894 loss)
I0428 13:24:57.087038  9322 sgd_solver.cpp:105] Iteration 3552, lr = 0.004948
I0428 13:25:02.244029  9322 solver.cpp:218] Iteration 3564 (2.32693 iter/s, 5.15701s/12 iters), loss = 1.79771
I0428 13:25:02.244068  9322 solver.cpp:237]     Train net output #0: loss = 1.79771 (* 1 = 1.79771 loss)
I0428 13:25:02.244077  9322 sgd_solver.cpp:105] Iteration 3564, lr = 0.00493626
I0428 13:25:04.329811  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_3570.caffemodel
I0428 13:25:08.474758  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_3570.solverstate
I0428 13:25:13.821696  9322 solver.cpp:330] Iteration 3570, Testing net (#0)
I0428 13:25:13.821717  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:25:16.997999  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:25:18.495698  9322 solver.cpp:397]     Test net output #0: accuracy = 0.324755
I0428 13:25:18.495743  9322 solver.cpp:397]     Test net output #1: loss = 2.94519 (* 1 = 2.94519 loss)
I0428 13:25:20.403543  9322 solver.cpp:218] Iteration 3576 (0.66081 iter/s, 18.1595s/12 iters), loss = 1.4339
I0428 13:25:20.403592  9322 solver.cpp:237]     Train net output #0: loss = 1.4339 (* 1 = 1.4339 loss)
I0428 13:25:20.403600  9322 sgd_solver.cpp:105] Iteration 3576, lr = 0.00492454
I0428 13:25:25.563673  9322 solver.cpp:218] Iteration 3588 (2.32554 iter/s, 5.16009s/12 iters), loss = 1.68284
I0428 13:25:25.563719  9322 solver.cpp:237]     Train net output #0: loss = 1.68284 (* 1 = 1.68284 loss)
I0428 13:25:25.563726  9322 sgd_solver.cpp:105] Iteration 3588, lr = 0.00491284
I0428 13:25:30.730897  9322 solver.cpp:218] Iteration 3600 (2.32235 iter/s, 5.16718s/12 iters), loss = 1.46544
I0428 13:25:30.731060  9322 solver.cpp:237]     Train net output #0: loss = 1.46544 (* 1 = 1.46544 loss)
I0428 13:25:30.731068  9322 sgd_solver.cpp:105] Iteration 3600, lr = 0.00490118
I0428 13:25:35.818841  9322 solver.cpp:218] Iteration 3612 (2.35859 iter/s, 5.08779s/12 iters), loss = 1.53867
I0428 13:25:35.818888  9322 solver.cpp:237]     Train net output #0: loss = 1.53867 (* 1 = 1.53867 loss)
I0428 13:25:35.818897  9322 sgd_solver.cpp:105] Iteration 3612, lr = 0.00488954
I0428 13:25:40.970463  9322 solver.cpp:218] Iteration 3624 (2.32938 iter/s, 5.15159s/12 iters), loss = 1.17614
I0428 13:25:40.970499  9322 solver.cpp:237]     Train net output #0: loss = 1.17614 (* 1 = 1.17614 loss)
I0428 13:25:40.970506  9322 sgd_solver.cpp:105] Iteration 3624, lr = 0.00487793
I0428 13:25:46.078878  9322 solver.cpp:218] Iteration 3636 (2.34908 iter/s, 5.10839s/12 iters), loss = 1.88082
I0428 13:25:46.078924  9322 solver.cpp:237]     Train net output #0: loss = 1.88082 (* 1 = 1.88082 loss)
I0428 13:25:46.078933  9322 sgd_solver.cpp:105] Iteration 3636, lr = 0.00486635
I0428 13:25:48.006803  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:25:51.281167  9322 solver.cpp:218] Iteration 3648 (2.30669 iter/s, 5.20225s/12 iters), loss = 1.29099
I0428 13:25:51.281211  9322 solver.cpp:237]     Train net output #0: loss = 1.29099 (* 1 = 1.29099 loss)
I0428 13:25:51.281219  9322 sgd_solver.cpp:105] Iteration 3648, lr = 0.0048548
I0428 13:25:56.283351  9322 solver.cpp:218] Iteration 3660 (2.39897 iter/s, 5.00215s/12 iters), loss = 1.28071
I0428 13:25:56.283396  9322 solver.cpp:237]     Train net output #0: loss = 1.28071 (* 1 = 1.28071 loss)
I0428 13:25:56.283406  9322 sgd_solver.cpp:105] Iteration 3660, lr = 0.00484327
I0428 13:26:01.064152  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_3672.caffemodel
I0428 13:26:07.143378  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_3672.solverstate
I0428 13:26:13.124791  9322 solver.cpp:330] Iteration 3672, Testing net (#0)
I0428 13:26:13.124811  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:26:16.341480  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:26:18.002668  9322 solver.cpp:397]     Test net output #0: accuracy = 0.346814
I0428 13:26:18.002705  9322 solver.cpp:397]     Test net output #1: loss = 2.87944 (* 1 = 2.87944 loss)
I0428 13:26:18.121037  9322 solver.cpp:218] Iteration 3672 (0.549508 iter/s, 21.8377s/12 iters), loss = 1.17997
I0428 13:26:18.121078  9322 solver.cpp:237]     Train net output #0: loss = 1.17997 (* 1 = 1.17997 loss)
I0428 13:26:18.121086  9322 sgd_solver.cpp:105] Iteration 3672, lr = 0.00483177
I0428 13:26:22.438117  9322 solver.cpp:218] Iteration 3684 (2.77968 iter/s, 4.31705s/12 iters), loss = 1.50797
I0428 13:26:22.438156  9322 solver.cpp:237]     Train net output #0: loss = 1.50797 (* 1 = 1.50797 loss)
I0428 13:26:22.438165  9322 sgd_solver.cpp:105] Iteration 3684, lr = 0.0048203
I0428 13:26:27.432406  9322 solver.cpp:218] Iteration 3696 (2.40276 iter/s, 4.99426s/12 iters), loss = 1.59029
I0428 13:26:27.432453  9322 solver.cpp:237]     Train net output #0: loss = 1.59029 (* 1 = 1.59029 loss)
I0428 13:26:27.432461  9322 sgd_solver.cpp:105] Iteration 3696, lr = 0.00480886
I0428 13:26:32.488278  9322 solver.cpp:218] Iteration 3708 (2.3735 iter/s, 5.05583s/12 iters), loss = 1.56219
I0428 13:26:32.488369  9322 solver.cpp:237]     Train net output #0: loss = 1.56219 (* 1 = 1.56219 loss)
I0428 13:26:32.488379  9322 sgd_solver.cpp:105] Iteration 3708, lr = 0.00479744
I0428 13:26:37.666935  9322 solver.cpp:218] Iteration 3720 (2.31724 iter/s, 5.17858s/12 iters), loss = 1.36562
I0428 13:26:37.666978  9322 solver.cpp:237]     Train net output #0: loss = 1.36562 (* 1 = 1.36562 loss)
I0428 13:26:37.666986  9322 sgd_solver.cpp:105] Iteration 3720, lr = 0.00478605
I0428 13:26:42.795622  9322 solver.cpp:218] Iteration 3732 (2.33979 iter/s, 5.12865s/12 iters), loss = 1.46717
I0428 13:26:42.795660  9322 solver.cpp:237]     Train net output #0: loss = 1.46717 (* 1 = 1.46717 loss)
I0428 13:26:42.795667  9322 sgd_solver.cpp:105] Iteration 3732, lr = 0.00477469
I0428 13:26:46.869485  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:26:47.917948  9322 solver.cpp:218] Iteration 3744 (2.3427 iter/s, 5.12229s/12 iters), loss = 1.46048
I0428 13:26:47.917996  9322 solver.cpp:237]     Train net output #0: loss = 1.46048 (* 1 = 1.46048 loss)
I0428 13:26:47.918005  9322 sgd_solver.cpp:105] Iteration 3744, lr = 0.00476335
I0428 13:26:53.141439  9322 solver.cpp:218] Iteration 3756 (2.29733 iter/s, 5.22345s/12 iters), loss = 1.51844
I0428 13:26:53.141479  9322 solver.cpp:237]     Train net output #0: loss = 1.51844 (* 1 = 1.51844 loss)
I0428 13:26:53.141487  9322 sgd_solver.cpp:105] Iteration 3756, lr = 0.00475204
I0428 13:26:58.299896  9322 solver.cpp:218] Iteration 3768 (2.32629 iter/s, 5.15842s/12 iters), loss = 1.46575
I0428 13:26:58.299940  9322 solver.cpp:237]     Train net output #0: loss = 1.46575 (* 1 = 1.46575 loss)
I0428 13:26:58.299949  9322 sgd_solver.cpp:105] Iteration 3768, lr = 0.00474076
I0428 13:27:00.373337  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_3774.caffemodel
I0428 13:27:07.559419  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_3774.solverstate
I0428 13:27:13.384394  9322 solver.cpp:330] Iteration 3774, Testing net (#0)
I0428 13:27:13.384413  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:27:16.609059  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:27:18.228938  9322 solver.cpp:397]     Test net output #0: accuracy = 0.349265
I0428 13:27:18.228987  9322 solver.cpp:397]     Test net output #1: loss = 2.88486 (* 1 = 2.88486 loss)
I0428 13:27:20.181260  9322 solver.cpp:218] Iteration 3780 (0.548411 iter/s, 21.8814s/12 iters), loss = 1.64277
I0428 13:27:20.181306  9322 solver.cpp:237]     Train net output #0: loss = 1.64277 (* 1 = 1.64277 loss)
I0428 13:27:20.181315  9322 sgd_solver.cpp:105] Iteration 3780, lr = 0.00472951
I0428 13:27:25.418920  9322 solver.cpp:218] Iteration 3792 (2.29112 iter/s, 5.23762s/12 iters), loss = 1.82132
I0428 13:27:25.418965  9322 solver.cpp:237]     Train net output #0: loss = 1.82132 (* 1 = 1.82132 loss)
I0428 13:27:25.418974  9322 sgd_solver.cpp:105] Iteration 3792, lr = 0.00471828
I0428 13:27:30.439397  9322 solver.cpp:218] Iteration 3804 (2.39023 iter/s, 5.02044s/12 iters), loss = 1.47505
I0428 13:27:30.439445  9322 solver.cpp:237]     Train net output #0: loss = 1.47505 (* 1 = 1.47505 loss)
I0428 13:27:30.439453  9322 sgd_solver.cpp:105] Iteration 3804, lr = 0.00470707
I0428 13:27:35.578047  9322 solver.cpp:218] Iteration 3816 (2.33526 iter/s, 5.13861s/12 iters), loss = 1.42027
I0428 13:27:35.578088  9322 solver.cpp:237]     Train net output #0: loss = 1.42027 (* 1 = 1.42027 loss)
I0428 13:27:35.578096  9322 sgd_solver.cpp:105] Iteration 3816, lr = 0.0046959
I0428 13:27:40.667068  9322 solver.cpp:218] Iteration 3828 (2.35803 iter/s, 5.08899s/12 iters), loss = 1.42038
I0428 13:27:40.667176  9322 solver.cpp:237]     Train net output #0: loss = 1.42038 (* 1 = 1.42038 loss)
I0428 13:27:40.667188  9322 sgd_solver.cpp:105] Iteration 3828, lr = 0.00468475
I0428 13:27:45.830926  9322 solver.cpp:218] Iteration 3840 (2.32389 iter/s, 5.16376s/12 iters), loss = 1.23208
I0428 13:27:45.830968  9322 solver.cpp:237]     Train net output #0: loss = 1.23208 (* 1 = 1.23208 loss)
I0428 13:27:45.830976  9322 sgd_solver.cpp:105] Iteration 3840, lr = 0.00467363
I0428 13:27:46.990937  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:27:50.994102  9322 solver.cpp:218] Iteration 3852 (2.32416 iter/s, 5.16315s/12 iters), loss = 1.55588
I0428 13:27:50.994143  9322 solver.cpp:237]     Train net output #0: loss = 1.55588 (* 1 = 1.55588 loss)
I0428 13:27:50.994150  9322 sgd_solver.cpp:105] Iteration 3852, lr = 0.00466253
I0428 13:27:56.131852  9322 solver.cpp:218] Iteration 3864 (2.33567 iter/s, 5.13772s/12 iters), loss = 1.41364
I0428 13:27:56.131891  9322 solver.cpp:237]     Train net output #0: loss = 1.41364 (* 1 = 1.41364 loss)
I0428 13:27:56.131899  9322 sgd_solver.cpp:105] Iteration 3864, lr = 0.00465146
I0428 13:28:00.781117  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_3876.caffemodel
I0428 13:28:07.058423  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_3876.solverstate
I0428 13:28:10.852078  9322 solver.cpp:330] Iteration 3876, Testing net (#0)
I0428 13:28:10.852243  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:28:14.003098  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:28:15.760869  9322 solver.cpp:397]     Test net output #0: accuracy = 0.330882
I0428 13:28:15.760916  9322 solver.cpp:397]     Test net output #1: loss = 2.92349 (* 1 = 2.92349 loss)
I0428 13:28:15.879218  9322 solver.cpp:218] Iteration 3876 (0.607675 iter/s, 19.7474s/12 iters), loss = 1.68034
I0428 13:28:15.879277  9322 solver.cpp:237]     Train net output #0: loss = 1.68034 (* 1 = 1.68034 loss)
I0428 13:28:15.879289  9322 sgd_solver.cpp:105] Iteration 3876, lr = 0.00464042
I0428 13:28:20.245273  9322 solver.cpp:218] Iteration 3888 (2.74851 iter/s, 4.366s/12 iters), loss = 1.19445
I0428 13:28:20.245321  9322 solver.cpp:237]     Train net output #0: loss = 1.19445 (* 1 = 1.19445 loss)
I0428 13:28:20.245329  9322 sgd_solver.cpp:105] Iteration 3888, lr = 0.0046294
I0428 13:28:25.472133  9322 solver.cpp:218] Iteration 3900 (2.29585 iter/s, 5.22682s/12 iters), loss = 1.51373
I0428 13:28:25.472175  9322 solver.cpp:237]     Train net output #0: loss = 1.51373 (* 1 = 1.51373 loss)
I0428 13:28:25.472183  9322 sgd_solver.cpp:105] Iteration 3900, lr = 0.00461841
I0428 13:28:30.649278  9322 solver.cpp:218] Iteration 3912 (2.3179 iter/s, 5.1771s/12 iters), loss = 1.20776
I0428 13:28:30.649328  9322 solver.cpp:237]     Train net output #0: loss = 1.20776 (* 1 = 1.20776 loss)
I0428 13:28:30.649338  9322 sgd_solver.cpp:105] Iteration 3912, lr = 0.00460744
I0428 13:28:35.814954  9322 solver.cpp:218] Iteration 3924 (2.32305 iter/s, 5.16563s/12 iters), loss = 1.35698
I0428 13:28:35.814990  9322 solver.cpp:237]     Train net output #0: loss = 1.35698 (* 1 = 1.35698 loss)
I0428 13:28:35.814996  9322 sgd_solver.cpp:105] Iteration 3924, lr = 0.0045965
I0428 13:28:41.006973  9322 solver.cpp:218] Iteration 3936 (2.31125 iter/s, 5.192s/12 iters), loss = 1.30613
I0428 13:28:41.007064  9322 solver.cpp:237]     Train net output #0: loss = 1.30613 (* 1 = 1.30613 loss)
I0428 13:28:41.007073  9322 sgd_solver.cpp:105] Iteration 3936, lr = 0.00458559
I0428 13:28:44.466910  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:28:46.170755  9322 solver.cpp:218] Iteration 3948 (2.32391 iter/s, 5.1637s/12 iters), loss = 1.07542
I0428 13:28:46.170792  9322 solver.cpp:237]     Train net output #0: loss = 1.07542 (* 1 = 1.07542 loss)
I0428 13:28:46.170801  9322 sgd_solver.cpp:105] Iteration 3948, lr = 0.0045747
I0428 13:28:51.406553  9322 solver.cpp:218] Iteration 3960 (2.29193 iter/s, 5.23577s/12 iters), loss = 1.24735
I0428 13:28:51.406602  9322 solver.cpp:237]     Train net output #0: loss = 1.24735 (* 1 = 1.24735 loss)
I0428 13:28:51.406615  9322 sgd_solver.cpp:105] Iteration 3960, lr = 0.00456384
I0428 13:28:56.673560  9322 solver.cpp:218] Iteration 3972 (2.27835 iter/s, 5.26697s/12 iters), loss = 1.33235
I0428 13:28:56.673604  9322 solver.cpp:237]     Train net output #0: loss = 1.33235 (* 1 = 1.33235 loss)
I0428 13:28:56.673612  9322 sgd_solver.cpp:105] Iteration 3972, lr = 0.00455301
I0428 13:28:58.714923  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_3978.caffemodel
I0428 13:29:04.260468  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_3978.solverstate
I0428 13:29:07.315788  9322 solver.cpp:330] Iteration 3978, Testing net (#0)
I0428 13:29:07.315806  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:29:10.420861  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:29:12.206971  9322 solver.cpp:397]     Test net output #0: accuracy = 0.344976
I0428 13:29:12.207166  9322 solver.cpp:397]     Test net output #1: loss = 2.86202 (* 1 = 2.86202 loss)
I0428 13:29:14.105993  9322 solver.cpp:218] Iteration 3984 (0.688371 iter/s, 17.4325s/12 iters), loss = 1.08761
I0428 13:29:14.106032  9322 solver.cpp:237]     Train net output #0: loss = 1.08761 (* 1 = 1.08761 loss)
I0428 13:29:14.106040  9322 sgd_solver.cpp:105] Iteration 3984, lr = 0.0045422
I0428 13:29:19.266577  9322 solver.cpp:218] Iteration 3996 (2.32533 iter/s, 5.16055s/12 iters), loss = 1.17563
I0428 13:29:19.266623  9322 solver.cpp:237]     Train net output #0: loss = 1.17563 (* 1 = 1.17563 loss)
I0428 13:29:19.266630  9322 sgd_solver.cpp:105] Iteration 3996, lr = 0.00453141
I0428 13:29:24.425354  9322 solver.cpp:218] Iteration 4008 (2.32615 iter/s, 5.15874s/12 iters), loss = 1.10903
I0428 13:29:24.425391  9322 solver.cpp:237]     Train net output #0: loss = 1.10903 (* 1 = 1.10903 loss)
I0428 13:29:24.425398  9322 sgd_solver.cpp:105] Iteration 4008, lr = 0.00452066
I0428 13:29:29.583570  9322 solver.cpp:218] Iteration 4020 (2.3264 iter/s, 5.15819s/12 iters), loss = 0.972867
I0428 13:29:29.583618  9322 solver.cpp:237]     Train net output #0: loss = 0.972867 (* 1 = 0.972867 loss)
I0428 13:29:29.583627  9322 sgd_solver.cpp:105] Iteration 4020, lr = 0.00450992
I0428 13:29:34.772394  9322 solver.cpp:218] Iteration 4032 (2.31268 iter/s, 5.18878s/12 iters), loss = 0.990056
I0428 13:29:34.772441  9322 solver.cpp:237]     Train net output #0: loss = 0.990056 (* 1 = 0.990056 loss)
I0428 13:29:34.772450  9322 sgd_solver.cpp:105] Iteration 4032, lr = 0.00449921
I0428 13:29:39.938009  9322 solver.cpp:218] Iteration 4044 (2.32307 iter/s, 5.16558s/12 iters), loss = 1.10642
I0428 13:29:39.938051  9322 solver.cpp:237]     Train net output #0: loss = 1.10642 (* 1 = 1.10642 loss)
I0428 13:29:39.938060  9322 sgd_solver.cpp:105] Iteration 4044, lr = 0.00448853
I0428 13:29:40.451570  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:29:45.125818  9322 solver.cpp:218] Iteration 4056 (2.31313 iter/s, 5.18777s/12 iters), loss = 1.21127
I0428 13:29:45.125911  9322 solver.cpp:237]     Train net output #0: loss = 1.21127 (* 1 = 1.21127 loss)
I0428 13:29:45.125924  9322 sgd_solver.cpp:105] Iteration 4056, lr = 0.00447788
I0428 13:29:50.283007  9322 solver.cpp:218] Iteration 4068 (2.32689 iter/s, 5.15711s/12 iters), loss = 1.20359
I0428 13:29:50.283054  9322 solver.cpp:237]     Train net output #0: loss = 1.20359 (* 1 = 1.20359 loss)
I0428 13:29:50.283063  9322 sgd_solver.cpp:105] Iteration 4068, lr = 0.00446724
I0428 13:29:54.989833  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_4080.caffemodel
I0428 13:30:01.857703  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_4080.solverstate
I0428 13:30:04.271713  9322 solver.cpp:330] Iteration 4080, Testing net (#0)
I0428 13:30:04.271736  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:30:07.333632  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:30:09.165611  9322 solver.cpp:397]     Test net output #0: accuracy = 0.349265
I0428 13:30:09.165654  9322 solver.cpp:397]     Test net output #1: loss = 2.96623 (* 1 = 2.96623 loss)
I0428 13:30:09.283419  9322 solver.cpp:218] Iteration 4080 (0.631564 iter/s, 19.0004s/12 iters), loss = 1.25577
I0428 13:30:09.283460  9322 solver.cpp:237]     Train net output #0: loss = 1.25577 (* 1 = 1.25577 loss)
I0428 13:30:09.283468  9322 sgd_solver.cpp:105] Iteration 4080, lr = 0.00445664
I0428 13:30:13.640671  9322 solver.cpp:218] Iteration 4092 (2.75405 iter/s, 4.35722s/12 iters), loss = 1.11733
I0428 13:30:13.640712  9322 solver.cpp:237]     Train net output #0: loss = 1.11733 (* 1 = 1.11733 loss)
I0428 13:30:13.640719  9322 sgd_solver.cpp:105] Iteration 4092, lr = 0.00444606
I0428 13:30:18.794904  9322 solver.cpp:218] Iteration 4104 (2.3282 iter/s, 5.1542s/12 iters), loss = 1.32284
I0428 13:30:18.795024  9322 solver.cpp:237]     Train net output #0: loss = 1.32284 (* 1 = 1.32284 loss)
I0428 13:30:18.795033  9322 sgd_solver.cpp:105] Iteration 4104, lr = 0.0044355
I0428 13:30:23.969545  9322 solver.cpp:218] Iteration 4116 (2.31905 iter/s, 5.17453s/12 iters), loss = 1.27236
I0428 13:30:23.969590  9322 solver.cpp:237]     Train net output #0: loss = 1.27236 (* 1 = 1.27236 loss)
I0428 13:30:23.969599  9322 sgd_solver.cpp:105] Iteration 4116, lr = 0.00442497
I0428 13:30:29.117985  9322 solver.cpp:218] Iteration 4128 (2.33082 iter/s, 5.1484s/12 iters), loss = 1.00764
I0428 13:30:29.118031  9322 solver.cpp:237]     Train net output #0: loss = 1.00764 (* 1 = 1.00764 loss)
I0428 13:30:29.118041  9322 sgd_solver.cpp:105] Iteration 4128, lr = 0.00441447
I0428 13:30:34.280300  9322 solver.cpp:218] Iteration 4140 (2.32456 iter/s, 5.16227s/12 iters), loss = 1.02259
I0428 13:30:34.280349  9322 solver.cpp:237]     Train net output #0: loss = 1.02259 (* 1 = 1.02259 loss)
I0428 13:30:34.280357  9322 sgd_solver.cpp:105] Iteration 4140, lr = 0.00440398
I0428 13:30:36.951172  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:30:39.473608  9322 solver.cpp:218] Iteration 4152 (2.31069 iter/s, 5.19326s/12 iters), loss = 1.04892
I0428 13:30:39.473654  9322 solver.cpp:237]     Train net output #0: loss = 1.04892 (* 1 = 1.04892 loss)
I0428 13:30:39.473662  9322 sgd_solver.cpp:105] Iteration 4152, lr = 0.00439353
I0428 13:30:41.160856  9322 blocking_queue.cpp:49] Waiting for data
I0428 13:30:44.656177  9322 solver.cpp:218] Iteration 4164 (2.31547 iter/s, 5.18254s/12 iters), loss = 1.08129
I0428 13:30:44.656215  9322 solver.cpp:237]     Train net output #0: loss = 1.08129 (* 1 = 1.08129 loss)
I0428 13:30:44.656224  9322 sgd_solver.cpp:105] Iteration 4164, lr = 0.0043831
I0428 13:30:49.845486  9322 solver.cpp:218] Iteration 4176 (2.31246 iter/s, 5.18928s/12 iters), loss = 0.893166
I0428 13:30:49.845578  9322 solver.cpp:237]     Train net output #0: loss = 0.893166 (* 1 = 0.893166 loss)
I0428 13:30:49.845587  9322 sgd_solver.cpp:105] Iteration 4176, lr = 0.00437269
I0428 13:30:51.948190  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_4182.caffemodel
I0428 13:30:55.210258  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_4182.solverstate
I0428 13:30:58.243167  9322 solver.cpp:330] Iteration 4182, Testing net (#0)
I0428 13:30:58.243186  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:31:01.279904  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:31:03.160087  9322 solver.cpp:397]     Test net output #0: accuracy = 0.36826
I0428 13:31:03.160121  9322 solver.cpp:397]     Test net output #1: loss = 2.96533 (* 1 = 2.96533 loss)
I0428 13:31:05.069962  9322 solver.cpp:218] Iteration 4188 (0.788207 iter/s, 15.2244s/12 iters), loss = 0.900799
I0428 13:31:05.070037  9322 solver.cpp:237]     Train net output #0: loss = 0.900799 (* 1 = 0.900799 loss)
I0428 13:31:05.070055  9322 sgd_solver.cpp:105] Iteration 4188, lr = 0.00436231
I0428 13:31:10.209237  9322 solver.cpp:218] Iteration 4200 (2.33499 iter/s, 5.13921s/12 iters), loss = 1.42876
I0428 13:31:10.209296  9322 solver.cpp:237]     Train net output #0: loss = 1.42876 (* 1 = 1.42876 loss)
I0428 13:31:10.209306  9322 sgd_solver.cpp:105] Iteration 4200, lr = 0.00435195
I0428 13:31:15.428659  9322 solver.cpp:218] Iteration 4212 (2.29913 iter/s, 5.21937s/12 iters), loss = 0.971285
I0428 13:31:15.428702  9322 solver.cpp:237]     Train net output #0: loss = 0.971285 (* 1 = 0.971285 loss)
I0428 13:31:15.428712  9322 sgd_solver.cpp:105] Iteration 4212, lr = 0.00434162
I0428 13:31:20.577369  9322 solver.cpp:218] Iteration 4224 (2.3307 iter/s, 5.14868s/12 iters), loss = 0.93986
I0428 13:31:20.577462  9322 solver.cpp:237]     Train net output #0: loss = 0.93986 (* 1 = 0.93986 loss)
I0428 13:31:20.577472  9322 sgd_solver.cpp:105] Iteration 4224, lr = 0.00433131
I0428 13:31:25.758528  9322 solver.cpp:218] Iteration 4236 (2.31612 iter/s, 5.18107s/12 iters), loss = 1.04545
I0428 13:31:25.758574  9322 solver.cpp:237]     Train net output #0: loss = 1.04545 (* 1 = 1.04545 loss)
I0428 13:31:25.758584  9322 sgd_solver.cpp:105] Iteration 4236, lr = 0.00432103
I0428 13:31:30.693441  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:31:30.941177  9322 solver.cpp:218] Iteration 4248 (2.31544 iter/s, 5.1826s/12 iters), loss = 0.992225
I0428 13:31:30.941236  9322 solver.cpp:237]     Train net output #0: loss = 0.992225 (* 1 = 0.992225 loss)
I0428 13:31:30.941248  9322 sgd_solver.cpp:105] Iteration 4248, lr = 0.00431077
I0428 13:31:35.935370  9322 solver.cpp:218] Iteration 4260 (2.40282 iter/s, 4.99413s/12 iters), loss = 0.789257
I0428 13:31:35.935448  9322 solver.cpp:237]     Train net output #0: loss = 0.789257 (* 1 = 0.789257 loss)
I0428 13:31:35.935462  9322 sgd_solver.cpp:105] Iteration 4260, lr = 0.00430053
I0428 13:31:41.030975  9322 solver.cpp:218] Iteration 4272 (2.355 iter/s, 5.09554s/12 iters), loss = 1.06475
I0428 13:31:41.031013  9322 solver.cpp:237]     Train net output #0: loss = 1.06475 (* 1 = 1.06475 loss)
I0428 13:31:41.031028  9322 sgd_solver.cpp:105] Iteration 4272, lr = 0.00429032
I0428 13:31:45.701732  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_4284.caffemodel
I0428 13:31:52.150108  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_4284.solverstate
I0428 13:31:56.315196  9322 solver.cpp:330] Iteration 4284, Testing net (#0)
I0428 13:31:56.315215  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:31:59.304388  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:32:01.250967  9322 solver.cpp:397]     Test net output #0: accuracy = 0.374387
I0428 13:32:01.250998  9322 solver.cpp:397]     Test net output #1: loss = 2.86853 (* 1 = 2.86853 loss)
I0428 13:32:01.369006  9322 solver.cpp:218] Iteration 4284 (0.590026 iter/s, 20.3381s/12 iters), loss = 0.903842
I0428 13:32:01.369063  9322 solver.cpp:237]     Train net output #0: loss = 0.903842 (* 1 = 0.903842 loss)
I0428 13:32:01.369076  9322 sgd_solver.cpp:105] Iteration 4284, lr = 0.00428014
I0428 13:32:05.699784  9322 solver.cpp:218] Iteration 4296 (2.7709 iter/s, 4.33073s/12 iters), loss = 0.79316
I0428 13:32:05.699826  9322 solver.cpp:237]     Train net output #0: loss = 0.79316 (* 1 = 0.79316 loss)
I0428 13:32:05.699836  9322 sgd_solver.cpp:105] Iteration 4296, lr = 0.00426998
I0428 13:32:10.861809  9322 solver.cpp:218] Iteration 4308 (2.32468 iter/s, 5.16199s/12 iters), loss = 0.980447
I0428 13:32:10.861850  9322 solver.cpp:237]     Train net output #0: loss = 0.980447 (* 1 = 0.980447 loss)
I0428 13:32:10.861857  9322 sgd_solver.cpp:105] Iteration 4308, lr = 0.00425984
I0428 13:32:16.035725  9322 solver.cpp:218] Iteration 4320 (2.31934 iter/s, 5.17389s/12 iters), loss = 0.886632
I0428 13:32:16.035764  9322 solver.cpp:237]     Train net output #0: loss = 0.886632 (* 1 = 0.886632 loss)
I0428 13:32:16.035771  9322 sgd_solver.cpp:105] Iteration 4320, lr = 0.00424972
I0428 13:32:21.229254  9322 solver.cpp:218] Iteration 4332 (2.31058 iter/s, 5.19349s/12 iters), loss = 1.06127
I0428 13:32:21.229293  9322 solver.cpp:237]     Train net output #0: loss = 1.06127 (* 1 = 1.06127 loss)
I0428 13:32:21.229301  9322 sgd_solver.cpp:105] Iteration 4332, lr = 0.00423964
I0428 13:32:26.388084  9322 solver.cpp:218] Iteration 4344 (2.32613 iter/s, 5.15879s/12 iters), loss = 0.672814
I0428 13:32:26.388217  9322 solver.cpp:237]     Train net output #0: loss = 0.672814 (* 1 = 0.672814 loss)
I0428 13:32:26.388227  9322 sgd_solver.cpp:105] Iteration 4344, lr = 0.00422957
I0428 13:32:28.346256  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:32:31.500672  9322 solver.cpp:218] Iteration 4356 (2.34721 iter/s, 5.11246s/12 iters), loss = 0.63748
I0428 13:32:31.500705  9322 solver.cpp:237]     Train net output #0: loss = 0.63748 (* 1 = 0.63748 loss)
I0428 13:32:31.500713  9322 sgd_solver.cpp:105] Iteration 4356, lr = 0.00421953
I0428 13:32:36.664935  9322 solver.cpp:218] Iteration 4368 (2.32368 iter/s, 5.16423s/12 iters), loss = 0.841109
I0428 13:32:36.664980  9322 solver.cpp:237]     Train net output #0: loss = 0.841109 (* 1 = 0.841109 loss)
I0428 13:32:36.664988  9322 sgd_solver.cpp:105] Iteration 4368, lr = 0.00420951
I0428 13:32:41.819577  9322 solver.cpp:218] Iteration 4380 (2.32802 iter/s, 5.1546s/12 iters), loss = 0.707738
I0428 13:32:41.819612  9322 solver.cpp:237]     Train net output #0: loss = 0.707738 (* 1 = 0.707738 loss)
I0428 13:32:41.819620  9322 sgd_solver.cpp:105] Iteration 4380, lr = 0.00419952
I0428 13:32:43.910254  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_4386.caffemodel
I0428 13:32:49.866920  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_4386.solverstate
I0428 13:32:54.490499  9322 solver.cpp:330] Iteration 4386, Testing net (#0)
I0428 13:32:54.490517  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:32:57.424208  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:32:59.381115  9322 solver.cpp:397]     Test net output #0: accuracy = 0.374387
I0428 13:32:59.381142  9322 solver.cpp:397]     Test net output #1: loss = 2.78006 (* 1 = 2.78006 loss)
I0428 13:33:01.363662  9322 solver.cpp:218] Iteration 4392 (0.613996 iter/s, 19.5441s/12 iters), loss = 0.995873
I0428 13:33:01.363703  9322 solver.cpp:237]     Train net output #0: loss = 0.995873 (* 1 = 0.995873 loss)
I0428 13:33:01.363710  9322 sgd_solver.cpp:105] Iteration 4392, lr = 0.00418954
I0428 13:33:06.503443  9322 solver.cpp:218] Iteration 4404 (2.33475 iter/s, 5.13974s/12 iters), loss = 0.817859
I0428 13:33:06.503487  9322 solver.cpp:237]     Train net output #0: loss = 0.817859 (* 1 = 0.817859 loss)
I0428 13:33:06.503496  9322 sgd_solver.cpp:105] Iteration 4404, lr = 0.0041796
I0428 13:33:11.702241  9322 solver.cpp:218] Iteration 4416 (2.30824 iter/s, 5.19876s/12 iters), loss = 0.955238
I0428 13:33:11.702280  9322 solver.cpp:237]     Train net output #0: loss = 0.955238 (* 1 = 0.955238 loss)
I0428 13:33:11.702288  9322 sgd_solver.cpp:105] Iteration 4416, lr = 0.00416967
I0428 13:33:16.939059  9322 solver.cpp:218] Iteration 4428 (2.29148 iter/s, 5.23678s/12 iters), loss = 0.914131
I0428 13:33:16.939105  9322 solver.cpp:237]     Train net output #0: loss = 0.914131 (* 1 = 0.914131 loss)
I0428 13:33:16.939113  9322 sgd_solver.cpp:105] Iteration 4428, lr = 0.00415977
I0428 13:33:22.152101  9322 solver.cpp:218] Iteration 4440 (2.30194 iter/s, 5.213s/12 iters), loss = 0.713381
I0428 13:33:22.152148  9322 solver.cpp:237]     Train net output #0: loss = 0.713381 (* 1 = 0.713381 loss)
I0428 13:33:22.152156  9322 sgd_solver.cpp:105] Iteration 4440, lr = 0.0041499
I0428 13:33:26.355981  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:33:27.378578  9322 solver.cpp:218] Iteration 4452 (2.29602 iter/s, 5.22643s/12 iters), loss = 0.866084
I0428 13:33:27.378623  9322 solver.cpp:237]     Train net output #0: loss = 0.866084 (* 1 = 0.866084 loss)
I0428 13:33:27.378630  9322 sgd_solver.cpp:105] Iteration 4452, lr = 0.00414005
I0428 13:33:32.596743  9322 solver.cpp:218] Iteration 4464 (2.29968 iter/s, 5.21812s/12 iters), loss = 0.898319
I0428 13:33:32.596830  9322 solver.cpp:237]     Train net output #0: loss = 0.898319 (* 1 = 0.898319 loss)
I0428 13:33:32.596839  9322 sgd_solver.cpp:105] Iteration 4464, lr = 0.00413022
I0428 13:33:37.872644  9322 solver.cpp:218] Iteration 4476 (2.27453 iter/s, 5.27582s/12 iters), loss = 0.819846
I0428 13:33:37.872690  9322 solver.cpp:237]     Train net output #0: loss = 0.819846 (* 1 = 0.819846 loss)
I0428 13:33:37.872699  9322 sgd_solver.cpp:105] Iteration 4476, lr = 0.00412041
I0428 13:33:42.639274  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_4488.caffemodel
I0428 13:33:46.627051  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_4488.solverstate
I0428 13:33:49.040302  9322 solver.cpp:330] Iteration 4488, Testing net (#0)
I0428 13:33:49.040320  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:33:51.949097  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:33:53.915848  9322 solver.cpp:397]     Test net output #0: accuracy = 0.384191
I0428 13:33:53.915894  9322 solver.cpp:397]     Test net output #1: loss = 2.79604 (* 1 = 2.79604 loss)
I0428 13:33:54.033926  9322 solver.cpp:218] Iteration 4488 (0.742516 iter/s, 16.1613s/12 iters), loss = 0.733594
I0428 13:33:54.033969  9322 solver.cpp:237]     Train net output #0: loss = 0.733594 (* 1 = 0.733594 loss)
I0428 13:33:54.033977  9322 sgd_solver.cpp:105] Iteration 4488, lr = 0.00411063
I0428 13:33:58.317914  9322 solver.cpp:218] Iteration 4500 (2.80116 iter/s, 4.28394s/12 iters), loss = 1.00687
I0428 13:33:58.317950  9322 solver.cpp:237]     Train net output #0: loss = 1.00687 (* 1 = 1.00687 loss)
I0428 13:33:58.317957  9322 sgd_solver.cpp:105] Iteration 4500, lr = 0.00410087
I0428 13:34:03.503947  9322 solver.cpp:218] Iteration 4512 (2.31392 iter/s, 5.186s/12 iters), loss = 0.725534
I0428 13:34:03.504036  9322 solver.cpp:237]     Train net output #0: loss = 0.725534 (* 1 = 0.725534 loss)
I0428 13:34:03.504045  9322 sgd_solver.cpp:105] Iteration 4512, lr = 0.00409113
I0428 13:34:08.666128  9322 solver.cpp:218] Iteration 4524 (2.32464 iter/s, 5.16209s/12 iters), loss = 0.880313
I0428 13:34:08.666174  9322 solver.cpp:237]     Train net output #0: loss = 0.880313 (* 1 = 0.880313 loss)
I0428 13:34:08.666182  9322 sgd_solver.cpp:105] Iteration 4524, lr = 0.00408142
I0428 13:34:13.810308  9322 solver.cpp:218] Iteration 4536 (2.33275 iter/s, 5.14413s/12 iters), loss = 1.09398
I0428 13:34:13.810341  9322 solver.cpp:237]     Train net output #0: loss = 1.09398 (* 1 = 1.09398 loss)
I0428 13:34:13.810349  9322 sgd_solver.cpp:105] Iteration 4536, lr = 0.00407173
I0428 13:34:18.828994  9322 solver.cpp:218] Iteration 4548 (2.39108 iter/s, 5.01865s/12 iters), loss = 0.857493
I0428 13:34:18.829038  9322 solver.cpp:237]     Train net output #0: loss = 0.857493 (* 1 = 0.857493 loss)
I0428 13:34:18.829046  9322 sgd_solver.cpp:105] Iteration 4548, lr = 0.00406206
I0428 13:34:20.130934  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:34:24.007903  9322 solver.cpp:218] Iteration 4560 (2.31711 iter/s, 5.17886s/12 iters), loss = 0.694855
I0428 13:34:24.007948  9322 solver.cpp:237]     Train net output #0: loss = 0.694855 (* 1 = 0.694855 loss)
I0428 13:34:24.007957  9322 sgd_solver.cpp:105] Iteration 4560, lr = 0.00405242
I0428 13:34:29.152590  9322 solver.cpp:218] Iteration 4572 (2.33252 iter/s, 5.14464s/12 iters), loss = 0.778935
I0428 13:34:29.152631  9322 solver.cpp:237]     Train net output #0: loss = 0.778935 (* 1 = 0.778935 loss)
I0428 13:34:29.152638  9322 sgd_solver.cpp:105] Iteration 4572, lr = 0.0040428
I0428 13:34:34.302568  9322 solver.cpp:218] Iteration 4584 (2.33012 iter/s, 5.14994s/12 iters), loss = 0.740307
I0428 13:34:34.302650  9322 solver.cpp:237]     Train net output #0: loss = 0.740307 (* 1 = 0.740307 loss)
I0428 13:34:34.302659  9322 sgd_solver.cpp:105] Iteration 4584, lr = 0.0040332
I0428 13:34:36.391898  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_4590.caffemodel
I0428 13:34:41.883715  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_4590.solverstate
I0428 13:34:46.322849  9322 solver.cpp:330] Iteration 4590, Testing net (#0)
I0428 13:34:46.322871  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:34:49.072466  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:34:51.011847  9322 solver.cpp:397]     Test net output #0: accuracy = 0.382353
I0428 13:34:51.011884  9322 solver.cpp:397]     Test net output #1: loss = 2.88499 (* 1 = 2.88499 loss)
I0428 13:34:52.855911  9322 solver.cpp:218] Iteration 4596 (0.646785 iter/s, 18.5533s/12 iters), loss = 0.856352
I0428 13:34:52.855957  9322 solver.cpp:237]     Train net output #0: loss = 0.856352 (* 1 = 0.856352 loss)
I0428 13:34:52.855967  9322 sgd_solver.cpp:105] Iteration 4596, lr = 0.00402362
I0428 13:34:57.995450  9322 solver.cpp:218] Iteration 4608 (2.33486 iter/s, 5.13949s/12 iters), loss = 0.930013
I0428 13:34:57.995510  9322 solver.cpp:237]     Train net output #0: loss = 0.930013 (* 1 = 0.930013 loss)
I0428 13:34:57.995523  9322 sgd_solver.cpp:105] Iteration 4608, lr = 0.00401407
I0428 13:35:03.192977  9322 solver.cpp:218] Iteration 4620 (2.30881 iter/s, 5.19747s/12 iters), loss = 0.581371
I0428 13:35:03.193023  9322 solver.cpp:237]     Train net output #0: loss = 0.581371 (* 1 = 0.581371 loss)
I0428 13:35:03.193032  9322 sgd_solver.cpp:105] Iteration 4620, lr = 0.00400454
I0428 13:35:08.279601  9322 solver.cpp:218] Iteration 4632 (2.35915 iter/s, 5.08658s/12 iters), loss = 0.68351
I0428 13:35:08.279803  9322 solver.cpp:237]     Train net output #0: loss = 0.68351 (* 1 = 0.68351 loss)
I0428 13:35:08.279814  9322 sgd_solver.cpp:105] Iteration 4632, lr = 0.00399503
I0428 13:35:13.359264  9322 solver.cpp:218] Iteration 4644 (2.36245 iter/s, 5.07947s/12 iters), loss = 0.869684
I0428 13:35:13.359308  9322 solver.cpp:237]     Train net output #0: loss = 0.869684 (* 1 = 0.869684 loss)
I0428 13:35:13.359315  9322 sgd_solver.cpp:105] Iteration 4644, lr = 0.00398555
I0428 13:35:16.866094  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:35:18.524344  9322 solver.cpp:218] Iteration 4656 (2.32331 iter/s, 5.16504s/12 iters), loss = 0.759439
I0428 13:35:18.524386  9322 solver.cpp:237]     Train net output #0: loss = 0.759439 (* 1 = 0.759439 loss)
I0428 13:35:18.524394  9322 sgd_solver.cpp:105] Iteration 4656, lr = 0.00397608
I0428 13:35:23.602232  9322 solver.cpp:218] Iteration 4668 (2.36321 iter/s, 5.07784s/12 iters), loss = 0.880214
I0428 13:35:23.602277  9322 solver.cpp:237]     Train net output #0: loss = 0.880214 (* 1 = 0.880214 loss)
I0428 13:35:23.602285  9322 sgd_solver.cpp:105] Iteration 4668, lr = 0.00396664
I0428 13:35:28.706745  9322 solver.cpp:218] Iteration 4680 (2.35088 iter/s, 5.10447s/12 iters), loss = 0.619297
I0428 13:35:28.706789  9322 solver.cpp:237]     Train net output #0: loss = 0.619297 (* 1 = 0.619297 loss)
I0428 13:35:28.706797  9322 sgd_solver.cpp:105] Iteration 4680, lr = 0.00395723
I0428 13:35:33.357395  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_4692.caffemodel
I0428 13:35:40.173619  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_4692.solverstate
I0428 13:35:49.486456  9322 solver.cpp:330] Iteration 4692, Testing net (#0)
I0428 13:35:49.486474  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:35:52.294847  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:35:54.405445  9322 solver.cpp:397]     Test net output #0: accuracy = 0.387255
I0428 13:35:54.405483  9322 solver.cpp:397]     Test net output #1: loss = 2.87682 (* 1 = 2.87682 loss)
I0428 13:35:54.521754  9322 solver.cpp:218] Iteration 4692 (0.464845 iter/s, 25.815s/12 iters), loss = 0.496342
I0428 13:35:54.521804  9322 solver.cpp:237]     Train net output #0: loss = 0.496342 (* 1 = 0.496342 loss)
I0428 13:35:54.521813  9322 sgd_solver.cpp:105] Iteration 4692, lr = 0.00394783
I0428 13:35:58.830202  9322 solver.cpp:218] Iteration 4704 (2.78526 iter/s, 4.3084s/12 iters), loss = 0.945835
I0428 13:35:58.830245  9322 solver.cpp:237]     Train net output #0: loss = 0.945835 (* 1 = 0.945835 loss)
I0428 13:35:58.830255  9322 sgd_solver.cpp:105] Iteration 4704, lr = 0.00393846
I0428 13:36:03.977473  9322 solver.cpp:218] Iteration 4716 (2.33135 iter/s, 5.14723s/12 iters), loss = 0.716037
I0428 13:36:03.977517  9322 solver.cpp:237]     Train net output #0: loss = 0.716037 (* 1 = 0.716037 loss)
I0428 13:36:03.977526  9322 sgd_solver.cpp:105] Iteration 4716, lr = 0.00392911
I0428 13:36:09.116660  9322 solver.cpp:218] Iteration 4728 (2.33502 iter/s, 5.13914s/12 iters), loss = 0.495602
I0428 13:36:09.116706  9322 solver.cpp:237]     Train net output #0: loss = 0.495602 (* 1 = 0.495602 loss)
I0428 13:36:09.116714  9322 sgd_solver.cpp:105] Iteration 4728, lr = 0.00391978
I0428 13:36:14.225124  9322 solver.cpp:218] Iteration 4740 (2.34906 iter/s, 5.10842s/12 iters), loss = 0.785093
I0428 13:36:14.225286  9322 solver.cpp:237]     Train net output #0: loss = 0.785093 (* 1 = 0.785093 loss)
I0428 13:36:14.225296  9322 sgd_solver.cpp:105] Iteration 4740, lr = 0.00391047
I0428 13:36:19.342701  9322 solver.cpp:218] Iteration 4752 (2.34493 iter/s, 5.11742s/12 iters), loss = 0.940966
I0428 13:36:19.342742  9322 solver.cpp:237]     Train net output #0: loss = 0.940966 (* 1 = 0.940966 loss)
I0428 13:36:19.342751  9322 sgd_solver.cpp:105] Iteration 4752, lr = 0.00390119
I0428 13:36:19.885562  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:36:24.527402  9322 solver.cpp:218] Iteration 4764 (2.31452 iter/s, 5.18466s/12 iters), loss = 0.695059
I0428 13:36:24.527449  9322 solver.cpp:237]     Train net output #0: loss = 0.695059 (* 1 = 0.695059 loss)
I0428 13:36:24.527457  9322 sgd_solver.cpp:105] Iteration 4764, lr = 0.00389193
I0428 13:36:29.600441  9322 solver.cpp:218] Iteration 4776 (2.36547 iter/s, 5.07299s/12 iters), loss = 0.581018
I0428 13:36:29.600489  9322 solver.cpp:237]     Train net output #0: loss = 0.581018 (* 1 = 0.581018 loss)
I0428 13:36:29.600497  9322 sgd_solver.cpp:105] Iteration 4776, lr = 0.00388269
I0428 13:36:34.786557  9322 solver.cpp:218] Iteration 4788 (2.31389 iter/s, 5.18607s/12 iters), loss = 0.750308
I0428 13:36:34.786607  9322 solver.cpp:237]     Train net output #0: loss = 0.750308 (* 1 = 0.750308 loss)
I0428 13:36:34.786617  9322 sgd_solver.cpp:105] Iteration 4788, lr = 0.00387347
I0428 13:36:36.868367  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_4794.caffemodel
I0428 13:36:48.579751  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_4794.solverstate
I0428 13:36:59.964313  9322 solver.cpp:330] Iteration 4794, Testing net (#0)
I0428 13:36:59.964339  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:37:02.692433  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:37:04.825911  9322 solver.cpp:397]     Test net output #0: accuracy = 0.395221
I0428 13:37:04.825947  9322 solver.cpp:397]     Test net output #1: loss = 2.7072 (* 1 = 2.7072 loss)
I0428 13:37:06.722045  9322 solver.cpp:218] Iteration 4800 (0.375757 iter/s, 31.9355s/12 iters), loss = 0.690426
I0428 13:37:06.722088  9322 solver.cpp:237]     Train net output #0: loss = 0.690426 (* 1 = 0.690426 loss)
I0428 13:37:06.722097  9322 sgd_solver.cpp:105] Iteration 4800, lr = 0.00386427
I0428 13:37:11.891809  9322 solver.cpp:218] Iteration 4812 (2.32121 iter/s, 5.16972s/12 iters), loss = 0.795835
I0428 13:37:11.891850  9322 solver.cpp:237]     Train net output #0: loss = 0.795835 (* 1 = 0.795835 loss)
I0428 13:37:11.891858  9322 sgd_solver.cpp:105] Iteration 4812, lr = 0.0038551
I0428 13:37:17.022099  9322 solver.cpp:218] Iteration 4824 (2.33907 iter/s, 5.13024s/12 iters), loss = 0.793889
I0428 13:37:17.022145  9322 solver.cpp:237]     Train net output #0: loss = 0.793889 (* 1 = 0.793889 loss)
I0428 13:37:17.022154  9322 sgd_solver.cpp:105] Iteration 4824, lr = 0.00384594
I0428 13:37:22.170977  9322 solver.cpp:218] Iteration 4836 (2.33063 iter/s, 5.14883s/12 iters), loss = 0.610413
I0428 13:37:22.171087  9322 solver.cpp:237]     Train net output #0: loss = 0.610413 (* 1 = 0.610413 loss)
I0428 13:37:22.171095  9322 sgd_solver.cpp:105] Iteration 4836, lr = 0.00383681
I0428 13:37:24.259101  9322 blocking_queue.cpp:49] Waiting for data
I0428 13:37:27.329319  9322 solver.cpp:218] Iteration 4848 (2.32638 iter/s, 5.15823s/12 iters), loss = 0.722356
I0428 13:37:27.329365  9322 solver.cpp:237]     Train net output #0: loss = 0.722356 (* 1 = 0.722356 loss)
I0428 13:37:27.329372  9322 sgd_solver.cpp:105] Iteration 4848, lr = 0.0038277
I0428 13:37:30.079092  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:37:32.507242  9322 solver.cpp:218] Iteration 4860 (2.31755 iter/s, 5.17788s/12 iters), loss = 0.559582
I0428 13:37:32.507287  9322 solver.cpp:237]     Train net output #0: loss = 0.559582 (* 1 = 0.559582 loss)
I0428 13:37:32.507294  9322 sgd_solver.cpp:105] Iteration 4860, lr = 0.00381862
I0428 13:37:37.648953  9322 solver.cpp:218] Iteration 4872 (2.33387 iter/s, 5.14167s/12 iters), loss = 0.632631
I0428 13:37:37.649000  9322 solver.cpp:237]     Train net output #0: loss = 0.632631 (* 1 = 0.632631 loss)
I0428 13:37:37.649009  9322 sgd_solver.cpp:105] Iteration 4872, lr = 0.00380955
I0428 13:37:42.739588  9322 solver.cpp:218] Iteration 4884 (2.35729 iter/s, 5.09059s/12 iters), loss = 0.569548
I0428 13:37:42.739626  9322 solver.cpp:237]     Train net output #0: loss = 0.569548 (* 1 = 0.569548 loss)
I0428 13:37:42.739634  9322 sgd_solver.cpp:105] Iteration 4884, lr = 0.0038005
I0428 13:37:47.342325  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_4896.caffemodel
I0428 13:37:55.529940  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_4896.solverstate
I0428 13:38:05.326030  9322 solver.cpp:330] Iteration 4896, Testing net (#0)
I0428 13:38:05.326048  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:38:07.971093  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:38:10.037102  9322 solver.cpp:397]     Test net output #0: accuracy = 0.409314
I0428 13:38:10.037129  9322 solver.cpp:397]     Test net output #1: loss = 2.73407 (* 1 = 2.73407 loss)
I0428 13:38:10.155238  9322 solver.cpp:218] Iteration 4896 (0.437706 iter/s, 27.4157s/12 iters), loss = 0.449565
I0428 13:38:10.155285  9322 solver.cpp:237]     Train net output #0: loss = 0.449565 (* 1 = 0.449565 loss)
I0428 13:38:10.155293  9322 sgd_solver.cpp:105] Iteration 4896, lr = 0.00379148
I0428 13:38:14.476881  9322 solver.cpp:218] Iteration 4908 (2.77676 iter/s, 4.32159s/12 iters), loss = 0.827898
I0428 13:38:14.476931  9322 solver.cpp:237]     Train net output #0: loss = 0.827898 (* 1 = 0.827898 loss)
I0428 13:38:14.476940  9322 sgd_solver.cpp:105] Iteration 4908, lr = 0.00378248
I0428 13:38:19.604648  9322 solver.cpp:218] Iteration 4920 (2.34022 iter/s, 5.12772s/12 iters), loss = 0.573495
I0428 13:38:19.604688  9322 solver.cpp:237]     Train net output #0: loss = 0.573495 (* 1 = 0.573495 loss)
I0428 13:38:19.604696  9322 sgd_solver.cpp:105] Iteration 4920, lr = 0.0037735
I0428 13:38:24.752750  9322 solver.cpp:218] Iteration 4932 (2.33098 iter/s, 5.14806s/12 iters), loss = 0.598411
I0428 13:38:24.752799  9322 solver.cpp:237]     Train net output #0: loss = 0.598411 (* 1 = 0.598411 loss)
I0428 13:38:24.752807  9322 sgd_solver.cpp:105] Iteration 4932, lr = 0.00376454
I0428 13:38:29.909746  9322 solver.cpp:218] Iteration 4944 (2.32696 iter/s, 5.15695s/12 iters), loss = 0.724
I0428 13:38:29.909850  9322 solver.cpp:237]     Train net output #0: loss = 0.724 (* 1 = 0.724 loss)
I0428 13:38:29.909859  9322 sgd_solver.cpp:105] Iteration 4944, lr = 0.0037556
I0428 13:38:34.859972  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:38:35.078476  9322 solver.cpp:218] Iteration 4956 (2.3217 iter/s, 5.16863s/12 iters), loss = 0.698958
I0428 13:38:35.078526  9322 solver.cpp:237]     Train net output #0: loss = 0.698958 (* 1 = 0.698958 loss)
I0428 13:38:35.078533  9322 sgd_solver.cpp:105] Iteration 4956, lr = 0.00374669
I0428 13:38:40.159288  9322 solver.cpp:218] Iteration 4968 (2.36185 iter/s, 5.08076s/12 iters), loss = 0.568923
I0428 13:38:40.159337  9322 solver.cpp:237]     Train net output #0: loss = 0.568923 (* 1 = 0.568923 loss)
I0428 13:38:40.159346  9322 sgd_solver.cpp:105] Iteration 4968, lr = 0.00373779
I0428 13:38:45.300630  9322 solver.cpp:218] Iteration 4980 (2.33404 iter/s, 5.14129s/12 iters), loss = 0.404162
I0428 13:38:45.300674  9322 solver.cpp:237]     Train net output #0: loss = 0.404162 (* 1 = 0.404162 loss)
I0428 13:38:45.300683  9322 sgd_solver.cpp:105] Iteration 4980, lr = 0.00372892
I0428 13:38:50.446869  9322 solver.cpp:218] Iteration 4992 (2.33182 iter/s, 5.1462s/12 iters), loss = 0.489643
I0428 13:38:50.446913  9322 solver.cpp:237]     Train net output #0: loss = 0.489643 (* 1 = 0.489643 loss)
I0428 13:38:50.446923  9322 sgd_solver.cpp:105] Iteration 4992, lr = 0.00372006
I0428 13:38:52.536590  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_4998.caffemodel
I0428 13:38:58.793835  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_4998.solverstate
I0428 13:39:03.927314  9322 solver.cpp:330] Iteration 4998, Testing net (#0)
I0428 13:39:03.927412  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:39:06.599153  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:39:08.689445  9322 solver.cpp:397]     Test net output #0: accuracy = 0.406863
I0428 13:39:08.689493  9322 solver.cpp:397]     Test net output #1: loss = 2.80129 (* 1 = 2.80129 loss)
I0428 13:39:10.581344  9322 solver.cpp:218] Iteration 5004 (0.595992 iter/s, 20.1345s/12 iters), loss = 0.65488
I0428 13:39:10.581391  9322 solver.cpp:237]     Train net output #0: loss = 0.65488 (* 1 = 0.65488 loss)
I0428 13:39:10.581400  9322 sgd_solver.cpp:105] Iteration 5004, lr = 0.00371123
I0428 13:39:15.764667  9322 solver.cpp:218] Iteration 5016 (2.31514 iter/s, 5.18328s/12 iters), loss = 0.683627
I0428 13:39:15.764714  9322 solver.cpp:237]     Train net output #0: loss = 0.683627 (* 1 = 0.683627 loss)
I0428 13:39:15.764721  9322 sgd_solver.cpp:105] Iteration 5016, lr = 0.00370242
I0428 13:39:20.841715  9322 solver.cpp:218] Iteration 5028 (2.3636 iter/s, 5.07701s/12 iters), loss = 0.542713
I0428 13:39:20.841758  9322 solver.cpp:237]     Train net output #0: loss = 0.542713 (* 1 = 0.542713 loss)
I0428 13:39:20.841766  9322 sgd_solver.cpp:105] Iteration 5028, lr = 0.00369363
I0428 13:39:26.000041  9322 solver.cpp:218] Iteration 5040 (2.32635 iter/s, 5.15829s/12 iters), loss = 0.644397
I0428 13:39:26.000082  9322 solver.cpp:237]     Train net output #0: loss = 0.644397 (* 1 = 0.644397 loss)
I0428 13:39:26.000089  9322 sgd_solver.cpp:105] Iteration 5040, lr = 0.00368486
I0428 13:39:31.082840  9322 solver.cpp:218] Iteration 5052 (2.36092 iter/s, 5.08276s/12 iters), loss = 0.662099
I0428 13:39:31.082886  9322 solver.cpp:237]     Train net output #0: loss = 0.662099 (* 1 = 0.662099 loss)
I0428 13:39:31.082895  9322 sgd_solver.cpp:105] Iteration 5052, lr = 0.00367611
I0428 13:39:33.067920  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:39:36.261066  9322 solver.cpp:218] Iteration 5064 (2.31741 iter/s, 5.17818s/12 iters), loss = 0.447916
I0428 13:39:36.261186  9322 solver.cpp:237]     Train net output #0: loss = 0.447916 (* 1 = 0.447916 loss)
I0428 13:39:36.261195  9322 sgd_solver.cpp:105] Iteration 5064, lr = 0.00366738
I0428 13:39:41.403775  9322 solver.cpp:218] Iteration 5076 (2.33345 iter/s, 5.1426s/12 iters), loss = 0.518987
I0428 13:39:41.403815  9322 solver.cpp:237]     Train net output #0: loss = 0.518987 (* 1 = 0.518987 loss)
I0428 13:39:41.403823  9322 sgd_solver.cpp:105] Iteration 5076, lr = 0.00365868
I0428 13:39:46.539834  9322 solver.cpp:218] Iteration 5088 (2.33644 iter/s, 5.13603s/12 iters), loss = 0.533862
I0428 13:39:46.539872  9322 solver.cpp:237]     Train net output #0: loss = 0.533862 (* 1 = 0.533862 loss)
I0428 13:39:46.539880  9322 sgd_solver.cpp:105] Iteration 5088, lr = 0.00364999
I0428 13:39:51.253616  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_5100.caffemodel
I0428 13:39:55.164089  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_5100.solverstate
I0428 13:40:01.350721  9322 solver.cpp:330] Iteration 5100, Testing net (#0)
I0428 13:40:01.350749  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:40:03.978937  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:40:06.141505  9322 solver.cpp:397]     Test net output #0: accuracy = 0.398284
I0428 13:40:06.141539  9322 solver.cpp:397]     Test net output #1: loss = 2.87965 (* 1 = 2.87965 loss)
I0428 13:40:06.259516  9322 solver.cpp:218] Iteration 5100 (0.608528 iter/s, 19.7197s/12 iters), loss = 0.545617
I0428 13:40:06.259563  9322 solver.cpp:237]     Train net output #0: loss = 0.545617 (* 1 = 0.545617 loss)
I0428 13:40:06.259572  9322 sgd_solver.cpp:105] Iteration 5100, lr = 0.00364132
I0428 13:40:10.668342  9322 solver.cpp:218] Iteration 5112 (2.72184 iter/s, 4.40878s/12 iters), loss = 0.708112
I0428 13:40:10.668460  9322 solver.cpp:237]     Train net output #0: loss = 0.708112 (* 1 = 0.708112 loss)
I0428 13:40:10.668469  9322 sgd_solver.cpp:105] Iteration 5112, lr = 0.00363268
I0428 13:40:15.842942  9322 solver.cpp:218] Iteration 5124 (2.31907 iter/s, 5.17449s/12 iters), loss = 0.456258
I0428 13:40:15.842988  9322 solver.cpp:237]     Train net output #0: loss = 0.456258 (* 1 = 0.456258 loss)
I0428 13:40:15.842996  9322 sgd_solver.cpp:105] Iteration 5124, lr = 0.00362405
I0428 13:40:21.128940  9322 solver.cpp:218] Iteration 5136 (2.27017 iter/s, 5.28595s/12 iters), loss = 0.585848
I0428 13:40:21.128989  9322 solver.cpp:237]     Train net output #0: loss = 0.585848 (* 1 = 0.585848 loss)
I0428 13:40:21.128998  9322 sgd_solver.cpp:105] Iteration 5136, lr = 0.00361545
I0428 13:40:26.304782  9322 solver.cpp:218] Iteration 5148 (2.31848 iter/s, 5.1758s/12 iters), loss = 0.448413
I0428 13:40:26.304824  9322 solver.cpp:237]     Train net output #0: loss = 0.448413 (* 1 = 0.448413 loss)
I0428 13:40:26.304833  9322 sgd_solver.cpp:105] Iteration 5148, lr = 0.00360687
I0428 13:40:30.498948  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:40:31.487452  9322 solver.cpp:218] Iteration 5160 (2.31543 iter/s, 5.18263s/12 iters), loss = 0.482535
I0428 13:40:31.487494  9322 solver.cpp:237]     Train net output #0: loss = 0.482535 (* 1 = 0.482535 loss)
I0428 13:40:31.487502  9322 sgd_solver.cpp:105] Iteration 5160, lr = 0.0035983
I0428 13:40:36.641713  9322 solver.cpp:218] Iteration 5172 (2.32819 iter/s, 5.15422s/12 iters), loss = 0.393082
I0428 13:40:36.641757  9322 solver.cpp:237]     Train net output #0: loss = 0.393082 (* 1 = 0.393082 loss)
I0428 13:40:36.641765  9322 sgd_solver.cpp:105] Iteration 5172, lr = 0.00358976
I0428 13:40:41.814910  9322 solver.cpp:218] Iteration 5184 (2.31967 iter/s, 5.17316s/12 iters), loss = 0.603883
I0428 13:40:41.815058  9322 solver.cpp:237]     Train net output #0: loss = 0.603883 (* 1 = 0.603883 loss)
I0428 13:40:41.815068  9322 sgd_solver.cpp:105] Iteration 5184, lr = 0.00358124
I0428 13:40:46.959017  9322 solver.cpp:218] Iteration 5196 (2.33283 iter/s, 5.14396s/12 iters), loss = 0.556465
I0428 13:40:46.959065  9322 solver.cpp:237]     Train net output #0: loss = 0.556465 (* 1 = 0.556465 loss)
I0428 13:40:46.959074  9322 sgd_solver.cpp:105] Iteration 5196, lr = 0.00357273
I0428 13:40:49.090960  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_5202.caffemodel
I0428 13:40:53.442086  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_5202.solverstate
I0428 13:41:00.382630  9322 solver.cpp:330] Iteration 5202, Testing net (#0)
I0428 13:41:00.382649  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:41:02.949995  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:41:05.195042  9322 solver.cpp:397]     Test net output #0: accuracy = 0.411765
I0428 13:41:05.195072  9322 solver.cpp:397]     Test net output #1: loss = 2.80948 (* 1 = 2.80948 loss)
I0428 13:41:07.108321  9322 solver.cpp:218] Iteration 5208 (0.595554 iter/s, 20.1493s/12 iters), loss = 0.362914
I0428 13:41:07.108366  9322 solver.cpp:237]     Train net output #0: loss = 0.362914 (* 1 = 0.362914 loss)
I0428 13:41:07.108376  9322 sgd_solver.cpp:105] Iteration 5208, lr = 0.00356425
I0428 13:41:12.266746  9322 solver.cpp:218] Iteration 5220 (2.32631 iter/s, 5.15839s/12 iters), loss = 0.464071
I0428 13:41:12.266855  9322 solver.cpp:237]     Train net output #0: loss = 0.464071 (* 1 = 0.464071 loss)
I0428 13:41:12.266865  9322 sgd_solver.cpp:105] Iteration 5220, lr = 0.00355579
I0428 13:41:17.422564  9322 solver.cpp:218] Iteration 5232 (2.32752 iter/s, 5.15571s/12 iters), loss = 0.360028
I0428 13:41:17.422614  9322 solver.cpp:237]     Train net output #0: loss = 0.360028 (* 1 = 0.360028 loss)
I0428 13:41:17.422622  9322 sgd_solver.cpp:105] Iteration 5232, lr = 0.00354735
I0428 13:41:22.543686  9322 solver.cpp:218] Iteration 5244 (2.34326 iter/s, 5.12108s/12 iters), loss = 0.481461
I0428 13:41:22.543726  9322 solver.cpp:237]     Train net output #0: loss = 0.481461 (* 1 = 0.481461 loss)
I0428 13:41:22.543735  9322 sgd_solver.cpp:105] Iteration 5244, lr = 0.00353892
I0428 13:41:27.705762  9322 solver.cpp:218] Iteration 5256 (2.32466 iter/s, 5.16203s/12 iters), loss = 0.446569
I0428 13:41:27.705809  9322 solver.cpp:237]     Train net output #0: loss = 0.446569 (* 1 = 0.446569 loss)
I0428 13:41:27.705818  9322 sgd_solver.cpp:105] Iteration 5256, lr = 0.00353052
I0428 13:41:29.035311  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:41:32.883030  9322 solver.cpp:218] Iteration 5268 (2.31785 iter/s, 5.17722s/12 iters), loss = 0.681517
I0428 13:41:32.883078  9322 solver.cpp:237]     Train net output #0: loss = 0.681517 (* 1 = 0.681517 loss)
I0428 13:41:32.883086  9322 sgd_solver.cpp:105] Iteration 5268, lr = 0.00352214
I0428 13:41:37.965634  9322 solver.cpp:218] Iteration 5280 (2.36102 iter/s, 5.08256s/12 iters), loss = 0.433075
I0428 13:41:37.965677  9322 solver.cpp:237]     Train net output #0: loss = 0.433075 (* 1 = 0.433075 loss)
I0428 13:41:37.965687  9322 sgd_solver.cpp:105] Iteration 5280, lr = 0.00351378
I0428 13:41:43.154467  9322 solver.cpp:218] Iteration 5292 (2.31268 iter/s, 5.18879s/12 iters), loss = 0.708024
I0428 13:41:43.154635  9322 solver.cpp:237]     Train net output #0: loss = 0.708024 (* 1 = 0.708024 loss)
I0428 13:41:43.154644  9322 sgd_solver.cpp:105] Iteration 5292, lr = 0.00350544
I0428 13:41:47.822752  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_5304.caffemodel
I0428 13:41:50.974810  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_5304.solverstate
I0428 13:41:53.383579  9322 solver.cpp:330] Iteration 5304, Testing net (#0)
I0428 13:41:53.383599  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:41:55.911957  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:41:58.290904  9322 solver.cpp:397]     Test net output #0: accuracy = 0.407476
I0428 13:41:58.290949  9322 solver.cpp:397]     Test net output #1: loss = 2.84523 (* 1 = 2.84523 loss)
I0428 13:41:58.409288  9322 solver.cpp:218] Iteration 5304 (0.786643 iter/s, 15.2547s/12 iters), loss = 0.532219
I0428 13:41:58.409329  9322 solver.cpp:237]     Train net output #0: loss = 0.532219 (* 1 = 0.532219 loss)
I0428 13:41:58.409338  9322 sgd_solver.cpp:105] Iteration 5304, lr = 0.00349711
I0428 13:42:02.718500  9322 solver.cpp:218] Iteration 5316 (2.78476 iter/s, 4.30917s/12 iters), loss = 0.551718
I0428 13:42:02.718549  9322 solver.cpp:237]     Train net output #0: loss = 0.551718 (* 1 = 0.551718 loss)
I0428 13:42:02.718557  9322 sgd_solver.cpp:105] Iteration 5316, lr = 0.00348881
I0428 13:42:07.657678  9322 solver.cpp:218] Iteration 5328 (2.42958 iter/s, 4.93913s/12 iters), loss = 0.422411
I0428 13:42:07.657721  9322 solver.cpp:237]     Train net output #0: loss = 0.422411 (* 1 = 0.422411 loss)
I0428 13:42:07.657729  9322 sgd_solver.cpp:105] Iteration 5328, lr = 0.00348053
I0428 13:42:12.761464  9322 solver.cpp:218] Iteration 5340 (2.35121 iter/s, 5.10375s/12 iters), loss = 0.47779
I0428 13:42:12.761509  9322 solver.cpp:237]     Train net output #0: loss = 0.47779 (* 1 = 0.47779 loss)
I0428 13:42:12.761518  9322 sgd_solver.cpp:105] Iteration 5340, lr = 0.00347226
I0428 13:42:17.916493  9322 solver.cpp:218] Iteration 5352 (2.32784 iter/s, 5.15499s/12 iters), loss = 0.517922
I0428 13:42:17.916622  9322 solver.cpp:237]     Train net output #0: loss = 0.517922 (* 1 = 0.517922 loss)
I0428 13:42:17.916632  9322 sgd_solver.cpp:105] Iteration 5352, lr = 0.00346402
I0428 13:42:21.465620  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:42:23.113379  9322 solver.cpp:218] Iteration 5364 (2.30913 iter/s, 5.19676s/12 iters), loss = 0.507499
I0428 13:42:23.113420  9322 solver.cpp:237]     Train net output #0: loss = 0.507499 (* 1 = 0.507499 loss)
I0428 13:42:23.113427  9322 sgd_solver.cpp:105] Iteration 5364, lr = 0.0034558
I0428 13:42:28.261828  9322 solver.cpp:218] Iteration 5376 (2.33082 iter/s, 5.14841s/12 iters), loss = 0.36561
I0428 13:42:28.261873  9322 solver.cpp:237]     Train net output #0: loss = 0.36561 (* 1 = 0.36561 loss)
I0428 13:42:28.261881  9322 sgd_solver.cpp:105] Iteration 5376, lr = 0.00344759
I0428 13:42:33.451108  9322 solver.cpp:218] Iteration 5388 (2.31248 iter/s, 5.18924s/12 iters), loss = 0.252248
I0428 13:42:33.451160  9322 solver.cpp:237]     Train net output #0: loss = 0.252248 (* 1 = 0.252248 loss)
I0428 13:42:33.451169  9322 sgd_solver.cpp:105] Iteration 5388, lr = 0.00343941
I0428 13:42:38.595636  9322 solver.cpp:218] Iteration 5400 (2.3326 iter/s, 5.14448s/12 iters), loss = 0.504005
I0428 13:42:38.595681  9322 solver.cpp:237]     Train net output #0: loss = 0.504005 (* 1 = 0.504005 loss)
I0428 13:42:38.595690  9322 sgd_solver.cpp:105] Iteration 5400, lr = 0.00343124
I0428 13:42:40.707015  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_5406.caffemodel
I0428 13:42:45.598417  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_5406.solverstate
I0428 13:42:48.497359  9322 solver.cpp:330] Iteration 5406, Testing net (#0)
I0428 13:42:48.497486  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:42:51.017887  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:42:53.430990  9322 solver.cpp:397]     Test net output #0: accuracy = 0.416667
I0428 13:42:53.431031  9322 solver.cpp:397]     Test net output #1: loss = 2.82337 (* 1 = 2.82337 loss)
I0428 13:42:55.331490  9322 solver.cpp:218] Iteration 5412 (0.717023 iter/s, 16.7359s/12 iters), loss = 0.426335
I0428 13:42:55.331533  9322 solver.cpp:237]     Train net output #0: loss = 0.426335 (* 1 = 0.426335 loss)
I0428 13:42:55.331542  9322 sgd_solver.cpp:105] Iteration 5412, lr = 0.00342309
I0428 13:43:00.493234  9322 solver.cpp:218] Iteration 5424 (2.32481 iter/s, 5.16171s/12 iters), loss = 0.482432
I0428 13:43:00.493281  9322 solver.cpp:237]     Train net output #0: loss = 0.482432 (* 1 = 0.482432 loss)
I0428 13:43:00.493289  9322 sgd_solver.cpp:105] Iteration 5424, lr = 0.00341497
I0428 13:43:05.568408  9322 solver.cpp:218] Iteration 5436 (2.36447 iter/s, 5.07513s/12 iters), loss = 0.351627
I0428 13:43:05.568449  9322 solver.cpp:237]     Train net output #0: loss = 0.351627 (* 1 = 0.351627 loss)
I0428 13:43:05.568457  9322 sgd_solver.cpp:105] Iteration 5436, lr = 0.00340686
I0428 13:43:10.737888  9322 solver.cpp:218] Iteration 5448 (2.32133 iter/s, 5.16945s/12 iters), loss = 0.452109
I0428 13:43:10.737928  9322 solver.cpp:237]     Train net output #0: loss = 0.452109 (* 1 = 0.452109 loss)
I0428 13:43:10.737936  9322 sgd_solver.cpp:105] Iteration 5448, lr = 0.00339877
I0428 13:43:16.016571  9322 solver.cpp:218] Iteration 5460 (2.27331 iter/s, 5.27864s/12 iters), loss = 0.535566
I0428 13:43:16.016619  9322 solver.cpp:237]     Train net output #0: loss = 0.535566 (* 1 = 0.535566 loss)
I0428 13:43:16.016628  9322 sgd_solver.cpp:105] Iteration 5460, lr = 0.0033907
I0428 13:43:16.589709  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:43:21.200970  9322 solver.cpp:218] Iteration 5472 (2.31465 iter/s, 5.18436s/12 iters), loss = 0.397857
I0428 13:43:21.201066  9322 solver.cpp:237]     Train net output #0: loss = 0.397857 (* 1 = 0.397857 loss)
I0428 13:43:21.201076  9322 sgd_solver.cpp:105] Iteration 5472, lr = 0.00338265
I0428 13:43:26.337172  9322 solver.cpp:218] Iteration 5484 (2.3364 iter/s, 5.13611s/12 iters), loss = 0.405786
I0428 13:43:26.337219  9322 solver.cpp:237]     Train net output #0: loss = 0.405786 (* 1 = 0.405786 loss)
I0428 13:43:26.337227  9322 sgd_solver.cpp:105] Iteration 5484, lr = 0.00337462
I0428 13:43:31.510776  9322 solver.cpp:218] Iteration 5496 (2.31948 iter/s, 5.17356s/12 iters), loss = 0.628299
I0428 13:43:31.510821  9322 solver.cpp:237]     Train net output #0: loss = 0.628299 (* 1 = 0.628299 loss)
I0428 13:43:31.510828  9322 sgd_solver.cpp:105] Iteration 5496, lr = 0.00336661
I0428 13:43:36.179430  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_5508.caffemodel
I0428 13:43:40.901986  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_5508.solverstate
I0428 13:43:49.769095  9322 solver.cpp:330] Iteration 5508, Testing net (#0)
I0428 13:43:49.769116  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:43:52.234828  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:43:54.614851  9322 solver.cpp:397]     Test net output #0: accuracy = 0.433824
I0428 13:43:54.614886  9322 solver.cpp:397]     Test net output #1: loss = 2.7638 (* 1 = 2.7638 loss)
I0428 13:43:54.732846  9322 solver.cpp:218] Iteration 5508 (0.516749 iter/s, 23.2221s/12 iters), loss = 0.411842
I0428 13:43:54.732893  9322 solver.cpp:237]     Train net output #0: loss = 0.411842 (* 1 = 0.411842 loss)
I0428 13:43:54.732901  9322 sgd_solver.cpp:105] Iteration 5508, lr = 0.00335861
I0428 13:43:59.061851  9322 solver.cpp:218] Iteration 5520 (2.77203 iter/s, 4.32896s/12 iters), loss = 0.450535
I0428 13:43:59.061899  9322 solver.cpp:237]     Train net output #0: loss = 0.450535 (* 1 = 0.450535 loss)
I0428 13:43:59.061908  9322 sgd_solver.cpp:105] Iteration 5520, lr = 0.00335064
I0428 13:44:01.574877  9322 blocking_queue.cpp:49] Waiting for data
I0428 13:44:04.263543  9322 solver.cpp:218] Iteration 5532 (2.30696 iter/s, 5.20164s/12 iters), loss = 0.440659
I0428 13:44:04.263586  9322 solver.cpp:237]     Train net output #0: loss = 0.440659 (* 1 = 0.440659 loss)
I0428 13:44:04.263595  9322 sgd_solver.cpp:105] Iteration 5532, lr = 0.00334268
I0428 13:44:09.402009  9322 solver.cpp:218] Iteration 5544 (2.33535 iter/s, 5.13842s/12 iters), loss = 0.321257
I0428 13:44:09.402055  9322 solver.cpp:237]     Train net output #0: loss = 0.321257 (* 1 = 0.321257 loss)
I0428 13:44:09.402063  9322 sgd_solver.cpp:105] Iteration 5544, lr = 0.00333475
I0428 13:44:14.582644  9322 solver.cpp:218] Iteration 5556 (2.31633 iter/s, 5.1806s/12 iters), loss = 0.343792
I0428 13:44:14.582680  9322 solver.cpp:237]     Train net output #0: loss = 0.343792 (* 1 = 0.343792 loss)
I0428 13:44:14.582688  9322 sgd_solver.cpp:105] Iteration 5556, lr = 0.00332683
I0428 13:44:17.392453  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:44:19.781008  9322 solver.cpp:218] Iteration 5568 (2.30844 iter/s, 5.19832s/12 iters), loss = 0.477384
I0428 13:44:19.781065  9322 solver.cpp:237]     Train net output #0: loss = 0.477384 (* 1 = 0.477384 loss)
I0428 13:44:19.781076  9322 sgd_solver.cpp:105] Iteration 5568, lr = 0.00331893
I0428 13:44:24.927261  9322 solver.cpp:218] Iteration 5580 (2.33182 iter/s, 5.1462s/12 iters), loss = 0.569663
I0428 13:44:24.927345  9322 solver.cpp:237]     Train net output #0: loss = 0.569663 (* 1 = 0.569663 loss)
I0428 13:44:24.927356  9322 sgd_solver.cpp:105] Iteration 5580, lr = 0.00331105
I0428 13:44:30.083384  9322 solver.cpp:218] Iteration 5592 (2.32737 iter/s, 5.15604s/12 iters), loss = 0.300816
I0428 13:44:30.083432  9322 solver.cpp:237]     Train net output #0: loss = 0.300816 (* 1 = 0.300816 loss)
I0428 13:44:30.083441  9322 sgd_solver.cpp:105] Iteration 5592, lr = 0.00330319
I0428 13:44:35.259819  9322 solver.cpp:218] Iteration 5604 (2.31822 iter/s, 5.17638s/12 iters), loss = 0.308889
I0428 13:44:35.259881  9322 solver.cpp:237]     Train net output #0: loss = 0.308889 (* 1 = 0.308889 loss)
I0428 13:44:35.259898  9322 sgd_solver.cpp:105] Iteration 5604, lr = 0.00329535
I0428 13:44:37.374876  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_5610.caffemodel
I0428 13:44:41.099642  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_5610.solverstate
I0428 13:44:44.696396  9322 solver.cpp:330] Iteration 5610, Testing net (#0)
I0428 13:44:44.696419  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:44:47.108536  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:44:49.619519  9322 solver.cpp:397]     Test net output #0: accuracy = 0.423407
I0428 13:44:49.619565  9322 solver.cpp:397]     Test net output #1: loss = 2.74507 (* 1 = 2.74507 loss)
I0428 13:44:51.508277  9322 solver.cpp:218] Iteration 5616 (0.738531 iter/s, 16.2485s/12 iters), loss = 0.500853
I0428 13:44:51.508320  9322 solver.cpp:237]     Train net output #0: loss = 0.500853 (* 1 = 0.500853 loss)
I0428 13:44:51.508327  9322 sgd_solver.cpp:105] Iteration 5616, lr = 0.00328752
I0428 13:44:56.642908  9322 solver.cpp:218] Iteration 5628 (2.33709 iter/s, 5.1346s/12 iters), loss = 0.423907
I0428 13:44:56.643026  9322 solver.cpp:237]     Train net output #0: loss = 0.423907 (* 1 = 0.423907 loss)
I0428 13:44:56.643036  9322 sgd_solver.cpp:105] Iteration 5628, lr = 0.00327972
I0428 13:45:01.727351  9322 solver.cpp:218] Iteration 5640 (2.36019 iter/s, 5.08433s/12 iters), loss = 0.34078
I0428 13:45:01.727399  9322 solver.cpp:237]     Train net output #0: loss = 0.34078 (* 1 = 0.34078 loss)
I0428 13:45:01.727407  9322 sgd_solver.cpp:105] Iteration 5640, lr = 0.00327193
I0428 13:45:06.906392  9322 solver.cpp:218] Iteration 5652 (2.31705 iter/s, 5.17901s/12 iters), loss = 0.464631
I0428 13:45:06.906426  9322 solver.cpp:237]     Train net output #0: loss = 0.464631 (* 1 = 0.464631 loss)
I0428 13:45:06.906435  9322 sgd_solver.cpp:105] Iteration 5652, lr = 0.00326416
I0428 13:45:11.873828  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:45:12.061820  9322 solver.cpp:218] Iteration 5664 (2.32766 iter/s, 5.1554s/12 iters), loss = 0.259045
I0428 13:45:12.061858  9322 solver.cpp:237]     Train net output #0: loss = 0.259045 (* 1 = 0.259045 loss)
I0428 13:45:12.061866  9322 sgd_solver.cpp:105] Iteration 5664, lr = 0.00325641
I0428 13:45:17.225477  9322 solver.cpp:218] Iteration 5676 (2.32395 iter/s, 5.16363s/12 iters), loss = 0.23025
I0428 13:45:17.225519  9322 solver.cpp:237]     Train net output #0: loss = 0.23025 (* 1 = 0.23025 loss)
I0428 13:45:17.225528  9322 sgd_solver.cpp:105] Iteration 5676, lr = 0.00324868
I0428 13:45:22.365274  9322 solver.cpp:218] Iteration 5688 (2.33474 iter/s, 5.13976s/12 iters), loss = 0.289913
I0428 13:45:22.365314  9322 solver.cpp:237]     Train net output #0: loss = 0.289913 (* 1 = 0.289913 loss)
I0428 13:45:22.365321  9322 sgd_solver.cpp:105] Iteration 5688, lr = 0.00324097
I0428 13:45:27.563747  9322 solver.cpp:218] Iteration 5700 (2.30838 iter/s, 5.19844s/12 iters), loss = 0.281406
I0428 13:45:27.563875  9322 solver.cpp:237]     Train net output #0: loss = 0.281406 (* 1 = 0.281406 loss)
I0428 13:45:27.563885  9322 sgd_solver.cpp:105] Iteration 5700, lr = 0.00323328
I0428 13:45:32.255332  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_5712.caffemodel
I0428 13:45:35.399017  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_5712.solverstate
I0428 13:45:38.301611  9322 solver.cpp:330] Iteration 5712, Testing net (#0)
I0428 13:45:38.301630  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:45:40.661789  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:45:43.202898  9322 solver.cpp:397]     Test net output #0: accuracy = 0.420956
I0428 13:45:43.202940  9322 solver.cpp:397]     Test net output #1: loss = 2.93653 (* 1 = 2.93653 loss)
I0428 13:45:43.315644  9322 solver.cpp:218] Iteration 5712 (0.761817 iter/s, 15.7518s/12 iters), loss = 0.341334
I0428 13:45:43.315694  9322 solver.cpp:237]     Train net output #0: loss = 0.341334 (* 1 = 0.341334 loss)
I0428 13:45:43.315702  9322 sgd_solver.cpp:105] Iteration 5712, lr = 0.0032256
I0428 13:45:47.648205  9322 solver.cpp:218] Iteration 5724 (2.76976 iter/s, 4.33251s/12 iters), loss = 0.432404
I0428 13:45:47.648252  9322 solver.cpp:237]     Train net output #0: loss = 0.432404 (* 1 = 0.432404 loss)
I0428 13:45:47.648262  9322 sgd_solver.cpp:105] Iteration 5724, lr = 0.00321794
I0428 13:45:52.804554  9322 solver.cpp:218] Iteration 5736 (2.32724 iter/s, 5.15631s/12 iters), loss = 0.355152
I0428 13:45:52.804600  9322 solver.cpp:237]     Train net output #0: loss = 0.355152 (* 1 = 0.355152 loss)
I0428 13:45:52.804610  9322 sgd_solver.cpp:105] Iteration 5736, lr = 0.0032103
I0428 13:45:57.962265  9322 solver.cpp:218] Iteration 5748 (2.32663 iter/s, 5.15767s/12 iters), loss = 0.384096
I0428 13:45:57.962401  9322 solver.cpp:237]     Train net output #0: loss = 0.384096 (* 1 = 0.384096 loss)
I0428 13:45:57.962411  9322 sgd_solver.cpp:105] Iteration 5748, lr = 0.00320268
I0428 13:46:03.081518  9322 solver.cpp:218] Iteration 5760 (2.34415 iter/s, 5.11912s/12 iters), loss = 0.280546
I0428 13:46:03.081565  9322 solver.cpp:237]     Train net output #0: loss = 0.280546 (* 1 = 0.280546 loss)
I0428 13:46:03.081574  9322 sgd_solver.cpp:105] Iteration 5760, lr = 0.00319508
I0428 13:46:05.113317  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:46:08.277001  9322 solver.cpp:218] Iteration 5772 (2.30971 iter/s, 5.19545s/12 iters), loss = 0.404832
I0428 13:46:08.277036  9322 solver.cpp:237]     Train net output #0: loss = 0.404832 (* 1 = 0.404832 loss)
I0428 13:46:08.277043  9322 sgd_solver.cpp:105] Iteration 5772, lr = 0.00318749
I0428 13:46:13.362525  9322 solver.cpp:218] Iteration 5784 (2.35965 iter/s, 5.08549s/12 iters), loss = 0.271354
I0428 13:46:13.362571  9322 solver.cpp:237]     Train net output #0: loss = 0.271354 (* 1 = 0.271354 loss)
I0428 13:46:13.362581  9322 sgd_solver.cpp:105] Iteration 5784, lr = 0.00317992
I0428 13:46:18.516279  9322 solver.cpp:218] Iteration 5796 (2.32842 iter/s, 5.15372s/12 iters), loss = 0.504264
I0428 13:46:18.516316  9322 solver.cpp:237]     Train net output #0: loss = 0.504264 (* 1 = 0.504264 loss)
I0428 13:46:18.516324  9322 sgd_solver.cpp:105] Iteration 5796, lr = 0.00317237
I0428 13:46:23.697440  9322 solver.cpp:218] Iteration 5808 (2.3161 iter/s, 5.18113s/12 iters), loss = 0.428003
I0428 13:46:23.697487  9322 solver.cpp:237]     Train net output #0: loss = 0.428003 (* 1 = 0.428003 loss)
I0428 13:46:23.697495  9322 sgd_solver.cpp:105] Iteration 5808, lr = 0.00316484
I0428 13:46:25.771759  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_5814.caffemodel
I0428 13:46:31.540531  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_5814.solverstate
I0428 13:46:36.290573  9322 solver.cpp:330] Iteration 5814, Testing net (#0)
I0428 13:46:36.290593  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:46:38.605105  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:46:41.200376  9322 solver.cpp:397]     Test net output #0: accuracy = 0.439951
I0428 13:46:41.200421  9322 solver.cpp:397]     Test net output #1: loss = 2.89389 (* 1 = 2.89389 loss)
I0428 13:46:43.122231  9322 solver.cpp:218] Iteration 5820 (0.617767 iter/s, 19.4248s/12 iters), loss = 0.399235
I0428 13:46:43.122273  9322 solver.cpp:237]     Train net output #0: loss = 0.399235 (* 1 = 0.399235 loss)
I0428 13:46:43.122282  9322 sgd_solver.cpp:105] Iteration 5820, lr = 0.00315733
I0428 13:46:48.260532  9322 solver.cpp:218] Iteration 5832 (2.33542 iter/s, 5.13826s/12 iters), loss = 0.37869
I0428 13:46:48.260571  9322 solver.cpp:237]     Train net output #0: loss = 0.37869 (* 1 = 0.37869 loss)
I0428 13:46:48.260581  9322 sgd_solver.cpp:105] Iteration 5832, lr = 0.00314983
I0428 13:46:53.419387  9322 solver.cpp:218] Iteration 5844 (2.32611 iter/s, 5.15882s/12 iters), loss = 0.352195
I0428 13:46:53.419428  9322 solver.cpp:237]     Train net output #0: loss = 0.352195 (* 1 = 0.352195 loss)
I0428 13:46:53.419437  9322 sgd_solver.cpp:105] Iteration 5844, lr = 0.00314235
I0428 13:46:58.604069  9322 solver.cpp:218] Iteration 5856 (2.31453 iter/s, 5.18465s/12 iters), loss = 0.192868
I0428 13:46:58.604115  9322 solver.cpp:237]     Train net output #0: loss = 0.192868 (* 1 = 0.192868 loss)
I0428 13:46:58.604123  9322 sgd_solver.cpp:105] Iteration 5856, lr = 0.00313489
I0428 13:47:03.100549  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:47:03.997795  9322 solver.cpp:218] Iteration 5868 (2.22482 iter/s, 5.39369s/12 iters), loss = 0.254837
I0428 13:47:03.997836  9322 solver.cpp:237]     Train net output #0: loss = 0.254837 (* 1 = 0.254837 loss)
I0428 13:47:03.997844  9322 sgd_solver.cpp:105] Iteration 5868, lr = 0.00312745
I0428 13:47:09.354391  9322 solver.cpp:218] Iteration 5880 (2.24024 iter/s, 5.35656s/12 iters), loss = 0.194434
I0428 13:47:09.354430  9322 solver.cpp:237]     Train net output #0: loss = 0.194434 (* 1 = 0.194434 loss)
I0428 13:47:09.354439  9322 sgd_solver.cpp:105] Iteration 5880, lr = 0.00312002
I0428 13:47:14.690340  9322 solver.cpp:218] Iteration 5892 (2.24891 iter/s, 5.33592s/12 iters), loss = 0.265503
I0428 13:47:14.690382  9322 solver.cpp:237]     Train net output #0: loss = 0.265503 (* 1 = 0.265503 loss)
I0428 13:47:14.690392  9322 sgd_solver.cpp:105] Iteration 5892, lr = 0.00311262
I0428 13:47:20.000083  9322 solver.cpp:218] Iteration 5904 (2.26001 iter/s, 5.30971s/12 iters), loss = 0.273005
I0428 13:47:20.000134  9322 solver.cpp:237]     Train net output #0: loss = 0.273005 (* 1 = 0.273005 loss)
I0428 13:47:20.000142  9322 sgd_solver.cpp:105] Iteration 5904, lr = 0.00310523
I0428 13:47:24.626371  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_5916.caffemodel
I0428 13:47:30.731988  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_5916.solverstate
I0428 13:47:35.829885  9322 solver.cpp:330] Iteration 5916, Testing net (#0)
I0428 13:47:35.830003  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:47:38.110232  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:47:40.739456  9322 solver.cpp:397]     Test net output #0: accuracy = 0.427083
I0428 13:47:40.739500  9322 solver.cpp:397]     Test net output #1: loss = 2.82394 (* 1 = 2.82394 loss)
I0428 13:47:40.852905  9322 solver.cpp:218] Iteration 5916 (0.575461 iter/s, 20.8528s/12 iters), loss = 0.357744
I0428 13:47:40.852949  9322 solver.cpp:237]     Train net output #0: loss = 0.357744 (* 1 = 0.357744 loss)
I0428 13:47:40.852957  9322 sgd_solver.cpp:105] Iteration 5916, lr = 0.00309785
I0428 13:47:45.486312  9322 solver.cpp:218] Iteration 5928 (2.58991 iter/s, 4.63336s/12 iters), loss = 0.225419
I0428 13:47:45.486356  9322 solver.cpp:237]     Train net output #0: loss = 0.225419 (* 1 = 0.225419 loss)
I0428 13:47:45.486363  9322 sgd_solver.cpp:105] Iteration 5928, lr = 0.0030905
I0428 13:47:50.725252  9322 solver.cpp:218] Iteration 5940 (2.29056 iter/s, 5.2389s/12 iters), loss = 0.156762
I0428 13:47:50.725297  9322 solver.cpp:237]     Train net output #0: loss = 0.156762 (* 1 = 0.156762 loss)
I0428 13:47:50.725306  9322 sgd_solver.cpp:105] Iteration 5940, lr = 0.00308316
I0428 13:47:55.897087  9322 solver.cpp:218] Iteration 5952 (2.32028 iter/s, 5.1718s/12 iters), loss = 0.446171
I0428 13:47:55.897135  9322 solver.cpp:237]     Train net output #0: loss = 0.446171 (* 1 = 0.446171 loss)
I0428 13:47:55.897145  9322 sgd_solver.cpp:105] Iteration 5952, lr = 0.00307584
I0428 13:48:01.006592  9322 solver.cpp:218] Iteration 5964 (2.34858 iter/s, 5.10946s/12 iters), loss = 0.273079
I0428 13:48:01.006642  9322 solver.cpp:237]     Train net output #0: loss = 0.273079 (* 1 = 0.273079 loss)
I0428 13:48:01.006650  9322 sgd_solver.cpp:105] Iteration 5964, lr = 0.00306854
I0428 13:48:02.371578  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:48:06.236429  9322 solver.cpp:218] Iteration 5976 (2.29455 iter/s, 5.22979s/12 iters), loss = 0.198847
I0428 13:48:06.236495  9322 solver.cpp:237]     Train net output #0: loss = 0.198847 (* 1 = 0.198847 loss)
I0428 13:48:06.236503  9322 sgd_solver.cpp:105] Iteration 5976, lr = 0.00306125
I0428 13:48:11.362133  9322 solver.cpp:218] Iteration 5988 (2.34117 iter/s, 5.12564s/12 iters), loss = 0.170517
I0428 13:48:11.362170  9322 solver.cpp:237]     Train net output #0: loss = 0.170517 (* 1 = 0.170517 loss)
I0428 13:48:11.362179  9322 sgd_solver.cpp:105] Iteration 5988, lr = 0.00305398
I0428 13:48:16.541846  9322 solver.cpp:218] Iteration 6000 (2.31675 iter/s, 5.17968s/12 iters), loss = 0.368842
I0428 13:48:16.541893  9322 solver.cpp:237]     Train net output #0: loss = 0.368842 (* 1 = 0.368842 loss)
I0428 13:48:16.541903  9322 sgd_solver.cpp:105] Iteration 6000, lr = 0.00304673
I0428 13:48:21.703465  9322 solver.cpp:218] Iteration 6012 (2.32487 iter/s, 5.16157s/12 iters), loss = 0.265581
I0428 13:48:21.703511  9322 solver.cpp:237]     Train net output #0: loss = 0.265581 (* 1 = 0.265581 loss)
I0428 13:48:21.703521  9322 sgd_solver.cpp:105] Iteration 6012, lr = 0.0030395
I0428 13:48:23.786701  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_6018.caffemodel
I0428 13:48:29.416173  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_6018.solverstate
I0428 13:48:34.171628  9322 solver.cpp:330] Iteration 6018, Testing net (#0)
I0428 13:48:34.171648  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:48:36.436260  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:48:39.167546  9322 solver.cpp:397]     Test net output #0: accuracy = 0.427083
I0428 13:48:39.167589  9322 solver.cpp:397]     Test net output #1: loss = 2.83646 (* 1 = 2.83646 loss)
I0428 13:48:41.093842  9322 solver.cpp:218] Iteration 6024 (0.618863 iter/s, 19.3904s/12 iters), loss = 0.247806
I0428 13:48:41.093888  9322 solver.cpp:237]     Train net output #0: loss = 0.247806 (* 1 = 0.247806 loss)
I0428 13:48:41.093897  9322 sgd_solver.cpp:105] Iteration 6024, lr = 0.00303228
I0428 13:48:46.375020  9322 solver.cpp:218] Iteration 6036 (2.27224 iter/s, 5.28114s/12 iters), loss = 0.287392
I0428 13:48:46.375061  9322 solver.cpp:237]     Train net output #0: loss = 0.287392 (* 1 = 0.287392 loss)
I0428 13:48:46.375068  9322 sgd_solver.cpp:105] Iteration 6036, lr = 0.00302508
I0428 13:48:51.482136  9322 solver.cpp:218] Iteration 6048 (2.34968 iter/s, 5.10708s/12 iters), loss = 0.226727
I0428 13:48:51.482172  9322 solver.cpp:237]     Train net output #0: loss = 0.226727 (* 1 = 0.226727 loss)
I0428 13:48:51.482180  9322 sgd_solver.cpp:105] Iteration 6048, lr = 0.0030179
I0428 13:48:56.659734  9322 solver.cpp:218] Iteration 6060 (2.31769 iter/s, 5.17757s/12 iters), loss = 0.313022
I0428 13:48:56.659775  9322 solver.cpp:237]     Train net output #0: loss = 0.313022 (* 1 = 0.313022 loss)
I0428 13:48:56.659782  9322 sgd_solver.cpp:105] Iteration 6060, lr = 0.00301074
I0428 13:49:00.288290  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:49:01.895586  9322 solver.cpp:218] Iteration 6072 (2.29191 iter/s, 5.23582s/12 iters), loss = 0.263861
I0428 13:49:01.895627  9322 solver.cpp:237]     Train net output #0: loss = 0.263861 (* 1 = 0.263861 loss)
I0428 13:49:01.895634  9322 sgd_solver.cpp:105] Iteration 6072, lr = 0.00300359
I0428 13:49:06.997119  9322 solver.cpp:218] Iteration 6084 (2.35225 iter/s, 5.10151s/12 iters), loss = 0.182793
I0428 13:49:06.997227  9322 solver.cpp:237]     Train net output #0: loss = 0.182793 (* 1 = 0.182793 loss)
I0428 13:49:06.997236  9322 sgd_solver.cpp:105] Iteration 6084, lr = 0.00299646
I0428 13:49:12.192997  9322 solver.cpp:218] Iteration 6096 (2.30957 iter/s, 5.19578s/12 iters), loss = 0.384888
I0428 13:49:12.193039  9322 solver.cpp:237]     Train net output #0: loss = 0.384888 (* 1 = 0.384888 loss)
I0428 13:49:12.193048  9322 sgd_solver.cpp:105] Iteration 6096, lr = 0.00298934
I0428 13:49:17.323402  9322 solver.cpp:218] Iteration 6108 (2.33901 iter/s, 5.13037s/12 iters), loss = 0.178508
I0428 13:49:17.323444  9322 solver.cpp:237]     Train net output #0: loss = 0.178508 (* 1 = 0.178508 loss)
I0428 13:49:17.323453  9322 sgd_solver.cpp:105] Iteration 6108, lr = 0.00298225
I0428 13:49:22.020901  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_6120.caffemodel
I0428 13:49:27.152133  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_6120.solverstate
I0428 13:49:29.615763  9322 solver.cpp:330] Iteration 6120, Testing net (#0)
I0428 13:49:29.615783  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:49:31.810901  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:49:34.495337  9322 solver.cpp:397]     Test net output #0: accuracy = 0.438113
I0428 13:49:34.495374  9322 solver.cpp:397]     Test net output #1: loss = 2.82696 (* 1 = 2.82696 loss)
I0428 13:49:34.613477  9322 solver.cpp:218] Iteration 6120 (0.694039 iter/s, 17.2901s/12 iters), loss = 0.363382
I0428 13:49:34.613526  9322 solver.cpp:237]     Train net output #0: loss = 0.363382 (* 1 = 0.363382 loss)
I0428 13:49:34.613535  9322 sgd_solver.cpp:105] Iteration 6120, lr = 0.00297517
I0428 13:49:38.922083  9322 solver.cpp:218] Iteration 6132 (2.78515 iter/s, 4.30856s/12 iters), loss = 0.319541
I0428 13:49:38.922221  9322 solver.cpp:237]     Train net output #0: loss = 0.319541 (* 1 = 0.319541 loss)
I0428 13:49:38.922231  9322 sgd_solver.cpp:105] Iteration 6132, lr = 0.0029681
I0428 13:49:44.087065  9322 solver.cpp:218] Iteration 6144 (2.32339 iter/s, 5.16486s/12 iters), loss = 0.277136
I0428 13:49:44.087103  9322 solver.cpp:237]     Train net output #0: loss = 0.277136 (* 1 = 0.277136 loss)
I0428 13:49:44.087111  9322 sgd_solver.cpp:105] Iteration 6144, lr = 0.00296105
I0428 13:49:49.253850  9322 solver.cpp:218] Iteration 6156 (2.32254 iter/s, 5.16675s/12 iters), loss = 0.23841
I0428 13:49:49.253896  9322 solver.cpp:237]     Train net output #0: loss = 0.23841 (* 1 = 0.23841 loss)
I0428 13:49:49.253904  9322 sgd_solver.cpp:105] Iteration 6156, lr = 0.00295402
I0428 13:49:54.344069  9322 solver.cpp:218] Iteration 6168 (2.35748 iter/s, 5.09018s/12 iters), loss = 0.258233
I0428 13:49:54.344116  9322 solver.cpp:237]     Train net output #0: loss = 0.258233 (* 1 = 0.258233 loss)
I0428 13:49:54.344126  9322 sgd_solver.cpp:105] Iteration 6168, lr = 0.00294701
I0428 13:49:54.945770  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:49:59.578727  9322 solver.cpp:218] Iteration 6180 (2.29243 iter/s, 5.23462s/12 iters), loss = 0.372846
I0428 13:49:59.578765  9322 solver.cpp:237]     Train net output #0: loss = 0.372846 (* 1 = 0.372846 loss)
I0428 13:49:59.578774  9322 sgd_solver.cpp:105] Iteration 6180, lr = 0.00294001
I0428 13:50:04.718217  9322 solver.cpp:218] Iteration 6192 (2.33488 iter/s, 5.13946s/12 iters), loss = 0.316988
I0428 13:50:04.718263  9322 solver.cpp:237]     Train net output #0: loss = 0.316988 (* 1 = 0.316988 loss)
I0428 13:50:04.718271  9322 sgd_solver.cpp:105] Iteration 6192, lr = 0.00293303
I0428 13:50:09.887465  9322 solver.cpp:218] Iteration 6204 (2.32144 iter/s, 5.16921s/12 iters), loss = 0.212011
I0428 13:50:09.887540  9322 solver.cpp:237]     Train net output #0: loss = 0.212011 (* 1 = 0.212011 loss)
I0428 13:50:09.887548  9322 sgd_solver.cpp:105] Iteration 6204, lr = 0.00292607
I0428 13:50:14.972923  9322 solver.cpp:218] Iteration 6216 (2.3597 iter/s, 5.0854s/12 iters), loss = 0.23193
I0428 13:50:14.972959  9322 solver.cpp:237]     Train net output #0: loss = 0.23193 (* 1 = 0.23193 loss)
I0428 13:50:14.972967  9322 sgd_solver.cpp:105] Iteration 6216, lr = 0.00291912
I0428 13:50:17.041501  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_6222.caffemodel
I0428 13:50:26.281471  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_6222.solverstate
I0428 13:50:30.558282  9322 solver.cpp:330] Iteration 6222, Testing net (#0)
I0428 13:50:30.558305  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:50:32.659358  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:50:34.030300  9322 blocking_queue.cpp:49] Waiting for data
I0428 13:50:35.299630  9322 solver.cpp:397]     Test net output #0: accuracy = 0.448529
I0428 13:50:35.299675  9322 solver.cpp:397]     Test net output #1: loss = 2.83972 (* 1 = 2.83972 loss)
I0428 13:50:37.140280  9322 solver.cpp:218] Iteration 6228 (0.541335 iter/s, 22.1674s/12 iters), loss = 0.170504
I0428 13:50:37.140319  9322 solver.cpp:237]     Train net output #0: loss = 0.170504 (* 1 = 0.170504 loss)
I0428 13:50:37.140327  9322 sgd_solver.cpp:105] Iteration 6228, lr = 0.00291219
I0428 13:50:42.306630  9322 solver.cpp:218] Iteration 6240 (2.32274 iter/s, 5.1663s/12 iters), loss = 0.319302
I0428 13:50:42.306784  9322 solver.cpp:237]     Train net output #0: loss = 0.319302 (* 1 = 0.319302 loss)
I0428 13:50:42.306795  9322 sgd_solver.cpp:105] Iteration 6240, lr = 0.00290528
I0428 13:50:47.467787  9322 solver.cpp:218] Iteration 6252 (2.32512 iter/s, 5.16102s/12 iters), loss = 0.287204
I0428 13:50:47.467828  9322 solver.cpp:237]     Train net output #0: loss = 0.287204 (* 1 = 0.287204 loss)
I0428 13:50:47.467836  9322 sgd_solver.cpp:105] Iteration 6252, lr = 0.00289838
I0428 13:50:52.617600  9322 solver.cpp:218] Iteration 6264 (2.3302 iter/s, 5.14978s/12 iters), loss = 0.19807
I0428 13:50:52.617642  9322 solver.cpp:237]     Train net output #0: loss = 0.19807 (* 1 = 0.19807 loss)
I0428 13:50:52.617650  9322 sgd_solver.cpp:105] Iteration 6264, lr = 0.0028915
I0428 13:50:55.424316  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:50:57.796401  9322 solver.cpp:218] Iteration 6276 (2.31716 iter/s, 5.17876s/12 iters), loss = 0.140856
I0428 13:50:57.796449  9322 solver.cpp:237]     Train net output #0: loss = 0.140856 (* 1 = 0.140856 loss)
I0428 13:50:57.796458  9322 sgd_solver.cpp:105] Iteration 6276, lr = 0.00288463
I0428 13:51:02.982662  9322 solver.cpp:218] Iteration 6288 (2.31382 iter/s, 5.18622s/12 iters), loss = 0.298619
I0428 13:51:02.982703  9322 solver.cpp:237]     Train net output #0: loss = 0.298619 (* 1 = 0.298619 loss)
I0428 13:51:02.982712  9322 sgd_solver.cpp:105] Iteration 6288, lr = 0.00287779
I0428 13:51:08.257195  9322 solver.cpp:218] Iteration 6300 (2.2751 iter/s, 5.2745s/12 iters), loss = 0.275541
I0428 13:51:08.257236  9322 solver.cpp:237]     Train net output #0: loss = 0.275541 (* 1 = 0.275541 loss)
I0428 13:51:08.257244  9322 sgd_solver.cpp:105] Iteration 6300, lr = 0.00287095
I0428 13:51:13.347354  9322 solver.cpp:218] Iteration 6312 (2.35751 iter/s, 5.09012s/12 iters), loss = 0.219262
I0428 13:51:13.347488  9322 solver.cpp:237]     Train net output #0: loss = 0.219262 (* 1 = 0.219262 loss)
I0428 13:51:13.347498  9322 sgd_solver.cpp:105] Iteration 6312, lr = 0.00286414
I0428 13:51:17.857707  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_6324.caffemodel
I0428 13:51:32.370746  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_6324.solverstate
I0428 13:51:40.016474  9322 solver.cpp:330] Iteration 6324, Testing net (#0)
I0428 13:51:40.016492  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:51:41.999816  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:51:44.821501  9322 solver.cpp:397]     Test net output #0: accuracy = 0.436887
I0428 13:51:44.821640  9322 solver.cpp:397]     Test net output #1: loss = 2.92832 (* 1 = 2.92832 loss)
I0428 13:51:44.939754  9322 solver.cpp:218] Iteration 6324 (0.379838 iter/s, 31.5924s/12 iters), loss = 0.323721
I0428 13:51:44.939803  9322 solver.cpp:237]     Train net output #0: loss = 0.323721 (* 1 = 0.323721 loss)
I0428 13:51:44.939811  9322 sgd_solver.cpp:105] Iteration 6324, lr = 0.00285734
I0428 13:51:49.226312  9322 solver.cpp:218] Iteration 6336 (2.79948 iter/s, 4.28651s/12 iters), loss = 0.121032
I0428 13:51:49.226361  9322 solver.cpp:237]     Train net output #0: loss = 0.121032 (* 1 = 0.121032 loss)
I0428 13:51:49.226370  9322 sgd_solver.cpp:105] Iteration 6336, lr = 0.00285055
I0428 13:51:54.311007  9322 solver.cpp:218] Iteration 6348 (2.36005 iter/s, 5.08465s/12 iters), loss = 0.233191
I0428 13:51:54.311050  9322 solver.cpp:237]     Train net output #0: loss = 0.233191 (* 1 = 0.233191 loss)
I0428 13:51:54.311059  9322 sgd_solver.cpp:105] Iteration 6348, lr = 0.00284379
I0428 13:51:59.494999  9322 solver.cpp:218] Iteration 6360 (2.31484 iter/s, 5.18395s/12 iters), loss = 0.322167
I0428 13:51:59.495045  9322 solver.cpp:237]     Train net output #0: loss = 0.322167 (* 1 = 0.322167 loss)
I0428 13:51:59.495054  9322 sgd_solver.cpp:105] Iteration 6360, lr = 0.00283703
I0428 13:52:04.513056  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:52:04.671628  9322 solver.cpp:218] Iteration 6372 (2.31813 iter/s, 5.17659s/12 iters), loss = 0.275697
I0428 13:52:04.671669  9322 solver.cpp:237]     Train net output #0: loss = 0.275697 (* 1 = 0.275697 loss)
I0428 13:52:04.671676  9322 sgd_solver.cpp:105] Iteration 6372, lr = 0.0028303
I0428 13:52:09.840245  9322 solver.cpp:218] Iteration 6384 (2.32172 iter/s, 5.16858s/12 iters), loss = 0.250043
I0428 13:52:09.840291  9322 solver.cpp:237]     Train net output #0: loss = 0.250043 (* 1 = 0.250043 loss)
I0428 13:52:09.840298  9322 sgd_solver.cpp:105] Iteration 6384, lr = 0.00282358
I0428 13:52:14.989630  9322 solver.cpp:218] Iteration 6396 (2.33039 iter/s, 5.14934s/12 iters), loss = 0.189136
I0428 13:52:14.989784  9322 solver.cpp:237]     Train net output #0: loss = 0.189136 (* 1 = 0.189136 loss)
I0428 13:52:14.989794  9322 sgd_solver.cpp:105] Iteration 6396, lr = 0.00281687
I0428 13:52:20.142158  9322 solver.cpp:218] Iteration 6408 (2.32902 iter/s, 5.15238s/12 iters), loss = 0.113258
I0428 13:52:20.142202  9322 solver.cpp:237]     Train net output #0: loss = 0.113258 (* 1 = 0.113258 loss)
I0428 13:52:20.142213  9322 sgd_solver.cpp:105] Iteration 6408, lr = 0.00281019
I0428 13:52:25.244272  9322 solver.cpp:218] Iteration 6420 (2.35199 iter/s, 5.10207s/12 iters), loss = 0.255556
I0428 13:52:25.244330  9322 solver.cpp:237]     Train net output #0: loss = 0.255556 (* 1 = 0.255556 loss)
I0428 13:52:25.244343  9322 sgd_solver.cpp:105] Iteration 6420, lr = 0.00280351
I0428 13:52:27.313143  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_6426.caffemodel
I0428 13:52:31.414315  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_6426.solverstate
I0428 13:52:38.074759  9322 solver.cpp:330] Iteration 6426, Testing net (#0)
I0428 13:52:38.074776  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:52:40.149617  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:52:43.033350  9322 solver.cpp:397]     Test net output #0: accuracy = 0.441789
I0428 13:52:43.033381  9322 solver.cpp:397]     Test net output #1: loss = 2.92378 (* 1 = 2.92378 loss)
I0428 13:52:45.133330  9322 solver.cpp:218] Iteration 6432 (0.603346 iter/s, 19.8891s/12 iters), loss = 0.2663
I0428 13:52:45.133406  9322 solver.cpp:237]     Train net output #0: loss = 0.2663 (* 1 = 0.2663 loss)
I0428 13:52:45.133416  9322 sgd_solver.cpp:105] Iteration 6432, lr = 0.00279686
I0428 13:52:50.261670  9322 solver.cpp:218] Iteration 6444 (2.33997 iter/s, 5.12827s/12 iters), loss = 0.194802
I0428 13:52:50.261718  9322 solver.cpp:237]     Train net output #0: loss = 0.194802 (* 1 = 0.194802 loss)
I0428 13:52:50.261726  9322 sgd_solver.cpp:105] Iteration 6444, lr = 0.00279022
I0428 13:52:55.414230  9322 solver.cpp:218] Iteration 6456 (2.32896 iter/s, 5.15252s/12 iters), loss = 0.167026
I0428 13:52:55.414294  9322 solver.cpp:237]     Train net output #0: loss = 0.167026 (* 1 = 0.167026 loss)
I0428 13:52:55.414306  9322 sgd_solver.cpp:105] Iteration 6456, lr = 0.00278359
I0428 13:53:00.599877  9322 solver.cpp:218] Iteration 6468 (2.3141 iter/s, 5.18559s/12 iters), loss = 0.347059
I0428 13:53:00.599923  9322 solver.cpp:237]     Train net output #0: loss = 0.347059 (* 1 = 0.347059 loss)
I0428 13:53:00.599931  9322 sgd_solver.cpp:105] Iteration 6468, lr = 0.00277698
I0428 13:53:02.643005  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:53:05.796389  9322 solver.cpp:218] Iteration 6480 (2.30926 iter/s, 5.19647s/12 iters), loss = 0.219113
I0428 13:53:05.796437  9322 solver.cpp:237]     Train net output #0: loss = 0.219113 (* 1 = 0.219113 loss)
I0428 13:53:05.796445  9322 sgd_solver.cpp:105] Iteration 6480, lr = 0.00277039
I0428 13:53:10.936069  9322 solver.cpp:218] Iteration 6492 (2.33479 iter/s, 5.13964s/12 iters), loss = 0.214631
I0428 13:53:10.936112  9322 solver.cpp:237]     Train net output #0: loss = 0.214631 (* 1 = 0.214631 loss)
I0428 13:53:10.936120  9322 sgd_solver.cpp:105] Iteration 6492, lr = 0.00276381
I0428 13:53:16.097931  9322 solver.cpp:218] Iteration 6504 (2.32476 iter/s, 5.16182s/12 iters), loss = 0.171005
I0428 13:53:16.098021  9322 solver.cpp:237]     Train net output #0: loss = 0.171005 (* 1 = 0.171005 loss)
I0428 13:53:16.098029  9322 sgd_solver.cpp:105] Iteration 6504, lr = 0.00275725
I0428 13:53:21.266925  9322 solver.cpp:218] Iteration 6516 (2.32157 iter/s, 5.16891s/12 iters), loss = 0.255961
I0428 13:53:21.266970  9322 solver.cpp:237]     Train net output #0: loss = 0.255961 (* 1 = 0.255961 loss)
I0428 13:53:21.266978  9322 sgd_solver.cpp:105] Iteration 6516, lr = 0.00275071
I0428 13:53:25.909795  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_6528.caffemodel
I0428 13:53:29.141535  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_6528.solverstate
I0428 13:53:32.730559  9322 solver.cpp:330] Iteration 6528, Testing net (#0)
I0428 13:53:32.730589  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:53:34.666999  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:53:37.472899  9322 solver.cpp:397]     Test net output #0: accuracy = 0.441789
I0428 13:53:37.472947  9322 solver.cpp:397]     Test net output #1: loss = 2.97499 (* 1 = 2.97499 loss)
I0428 13:53:37.591863  9322 solver.cpp:218] Iteration 6528 (0.735071 iter/s, 16.3249s/12 iters), loss = 0.229608
I0428 13:53:37.593380  9322 solver.cpp:237]     Train net output #0: loss = 0.229608 (* 1 = 0.229608 loss)
I0428 13:53:37.593390  9322 sgd_solver.cpp:105] Iteration 6528, lr = 0.00274418
I0428 13:53:41.881418  9322 solver.cpp:218] Iteration 6540 (2.79848 iter/s, 4.28804s/12 iters), loss = 0.110869
I0428 13:53:41.881464  9322 solver.cpp:237]     Train net output #0: loss = 0.110869 (* 1 = 0.110869 loss)
I0428 13:53:41.881472  9322 sgd_solver.cpp:105] Iteration 6540, lr = 0.00273766
I0428 13:53:47.203289  9322 solver.cpp:218] Iteration 6552 (2.25486 iter/s, 5.32184s/12 iters), loss = 0.340897
I0428 13:53:47.203450  9322 solver.cpp:237]     Train net output #0: loss = 0.340897 (* 1 = 0.340897 loss)
I0428 13:53:47.203459  9322 sgd_solver.cpp:105] Iteration 6552, lr = 0.00273116
I0428 13:53:52.307642  9322 solver.cpp:218] Iteration 6564 (2.351 iter/s, 5.1042s/12 iters), loss = 0.162077
I0428 13:53:52.307689  9322 solver.cpp:237]     Train net output #0: loss = 0.162077 (* 1 = 0.162077 loss)
I0428 13:53:52.307698  9322 sgd_solver.cpp:105] Iteration 6564, lr = 0.00272468
I0428 13:53:56.687649  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:53:57.508754  9322 solver.cpp:218] Iteration 6576 (2.30722 iter/s, 5.20107s/12 iters), loss = 0.179436
I0428 13:53:57.508802  9322 solver.cpp:237]     Train net output #0: loss = 0.179436 (* 1 = 0.179436 loss)
I0428 13:53:57.508811  9322 sgd_solver.cpp:105] Iteration 6576, lr = 0.00271821
I0428 13:54:02.696930  9322 solver.cpp:218] Iteration 6588 (2.31297 iter/s, 5.18813s/12 iters), loss = 0.20514
I0428 13:54:02.696977  9322 solver.cpp:237]     Train net output #0: loss = 0.20514 (* 1 = 0.20514 loss)
I0428 13:54:02.696986  9322 sgd_solver.cpp:105] Iteration 6588, lr = 0.00271175
I0428 13:54:07.830910  9322 solver.cpp:218] Iteration 6600 (2.33739 iter/s, 5.13394s/12 iters), loss = 0.0839773
I0428 13:54:07.830960  9322 solver.cpp:237]     Train net output #0: loss = 0.0839773 (* 1 = 0.0839773 loss)
I0428 13:54:07.830969  9322 sgd_solver.cpp:105] Iteration 6600, lr = 0.00270532
I0428 13:54:12.920482  9322 solver.cpp:218] Iteration 6612 (2.35778 iter/s, 5.08953s/12 iters), loss = 0.142965
I0428 13:54:12.920529  9322 solver.cpp:237]     Train net output #0: loss = 0.142965 (* 1 = 0.142965 loss)
I0428 13:54:12.920538  9322 sgd_solver.cpp:105] Iteration 6612, lr = 0.00269889
I0428 13:54:18.013324  9322 solver.cpp:218] Iteration 6624 (2.35627 iter/s, 5.0928s/12 iters), loss = 0.12914
I0428 13:54:18.013420  9322 solver.cpp:237]     Train net output #0: loss = 0.12914 (* 1 = 0.12914 loss)
I0428 13:54:18.013429  9322 sgd_solver.cpp:105] Iteration 6624, lr = 0.00269248
I0428 13:54:20.114605  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_6630.caffemodel
I0428 13:54:24.736146  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_6630.solverstate
I0428 13:54:30.453406  9322 solver.cpp:330] Iteration 6630, Testing net (#0)
I0428 13:54:30.453424  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:54:32.451411  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:54:35.413741  9322 solver.cpp:397]     Test net output #0: accuracy = 0.445466
I0428 13:54:35.413776  9322 solver.cpp:397]     Test net output #1: loss = 2.9566 (* 1 = 2.9566 loss)
I0428 13:54:37.321334  9322 solver.cpp:218] Iteration 6636 (0.621504 iter/s, 19.308s/12 iters), loss = 0.183977
I0428 13:54:37.321380  9322 solver.cpp:237]     Train net output #0: loss = 0.183977 (* 1 = 0.183977 loss)
I0428 13:54:37.321388  9322 sgd_solver.cpp:105] Iteration 6636, lr = 0.00268609
I0428 13:54:42.497134  9322 solver.cpp:218] Iteration 6648 (2.3185 iter/s, 5.17576s/12 iters), loss = 0.14733
I0428 13:54:42.497180  9322 solver.cpp:237]     Train net output #0: loss = 0.14733 (* 1 = 0.14733 loss)
I0428 13:54:42.497189  9322 sgd_solver.cpp:105] Iteration 6648, lr = 0.00267971
I0428 13:54:47.587746  9322 solver.cpp:218] Iteration 6660 (2.3573 iter/s, 5.09057s/12 iters), loss = 0.199724
I0428 13:54:47.587790  9322 solver.cpp:237]     Train net output #0: loss = 0.199724 (* 1 = 0.199724 loss)
I0428 13:54:47.587798  9322 sgd_solver.cpp:105] Iteration 6660, lr = 0.00267335
I0428 13:54:52.753785  9322 solver.cpp:218] Iteration 6672 (2.32288 iter/s, 5.166s/12 iters), loss = 0.136644
I0428 13:54:52.753944  9322 solver.cpp:237]     Train net output #0: loss = 0.136644 (* 1 = 0.136644 loss)
I0428 13:54:52.753953  9322 sgd_solver.cpp:105] Iteration 6672, lr = 0.00266701
I0428 13:54:54.154359  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:54:57.943969  9322 solver.cpp:218] Iteration 6684 (2.31212 iter/s, 5.19003s/12 iters), loss = 0.261657
I0428 13:54:57.944005  9322 solver.cpp:237]     Train net output #0: loss = 0.261657 (* 1 = 0.261657 loss)
I0428 13:54:57.944012  9322 sgd_solver.cpp:105] Iteration 6684, lr = 0.00266067
I0428 13:55:02.951383  9322 solver.cpp:218] Iteration 6696 (2.39646 iter/s, 5.00738s/12 iters), loss = 0.0826467
I0428 13:55:02.951426  9322 solver.cpp:237]     Train net output #0: loss = 0.0826467 (* 1 = 0.0826467 loss)
I0428 13:55:02.951434  9322 sgd_solver.cpp:105] Iteration 6696, lr = 0.00265436
I0428 13:55:08.098146  9322 solver.cpp:218] Iteration 6708 (2.33158 iter/s, 5.14672s/12 iters), loss = 0.19169
I0428 13:55:08.098196  9322 solver.cpp:237]     Train net output #0: loss = 0.19169 (* 1 = 0.19169 loss)
I0428 13:55:08.098206  9322 sgd_solver.cpp:105] Iteration 6708, lr = 0.00264805
I0428 13:55:13.205942  9322 solver.cpp:218] Iteration 6720 (2.34937 iter/s, 5.10775s/12 iters), loss = 0.178922
I0428 13:55:13.205988  9322 solver.cpp:237]     Train net output #0: loss = 0.178922 (* 1 = 0.178922 loss)
I0428 13:55:13.205997  9322 sgd_solver.cpp:105] Iteration 6720, lr = 0.00264177
I0428 13:55:17.798713  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_6732.caffemodel
I0428 13:55:21.015009  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_6732.solverstate
I0428 13:55:23.425624  9322 solver.cpp:330] Iteration 6732, Testing net (#0)
I0428 13:55:23.425730  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:55:25.360698  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:55:28.235953  9322 solver.cpp:397]     Test net output #0: accuracy = 0.452819
I0428 13:55:28.235980  9322 solver.cpp:397]     Test net output #1: loss = 2.92434 (* 1 = 2.92434 loss)
I0428 13:55:28.354008  9322 solver.cpp:218] Iteration 6732 (0.79218 iter/s, 15.1481s/12 iters), loss = 0.197208
I0428 13:55:28.354058  9322 solver.cpp:237]     Train net output #0: loss = 0.197208 (* 1 = 0.197208 loss)
I0428 13:55:28.354068  9322 sgd_solver.cpp:105] Iteration 6732, lr = 0.0026355
I0428 13:55:32.688565  9322 solver.cpp:218] Iteration 6744 (2.76848 iter/s, 4.33451s/12 iters), loss = 0.101426
I0428 13:55:32.688606  9322 solver.cpp:237]     Train net output #0: loss = 0.101426 (* 1 = 0.101426 loss)
I0428 13:55:32.688614  9322 sgd_solver.cpp:105] Iteration 6744, lr = 0.00262924
I0428 13:55:37.787955  9322 solver.cpp:218] Iteration 6756 (2.35324 iter/s, 5.09935s/12 iters), loss = 0.0550761
I0428 13:55:37.787998  9322 solver.cpp:237]     Train net output #0: loss = 0.0550761 (* 1 = 0.0550761 loss)
I0428 13:55:37.788007  9322 sgd_solver.cpp:105] Iteration 6756, lr = 0.002623
I0428 13:55:42.975162  9322 solver.cpp:218] Iteration 6768 (2.3134 iter/s, 5.18717s/12 iters), loss = 0.150108
I0428 13:55:42.975203  9322 solver.cpp:237]     Train net output #0: loss = 0.150108 (* 1 = 0.150108 loss)
I0428 13:55:42.975211  9322 sgd_solver.cpp:105] Iteration 6768, lr = 0.00261677
I0428 13:55:46.576831  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:55:48.146500  9322 solver.cpp:218] Iteration 6780 (2.3205 iter/s, 5.1713s/12 iters), loss = 0.199799
I0428 13:55:48.146545  9322 solver.cpp:237]     Train net output #0: loss = 0.199799 (* 1 = 0.199799 loss)
I0428 13:55:48.146554  9322 sgd_solver.cpp:105] Iteration 6780, lr = 0.00261056
I0428 13:55:53.310922  9322 solver.cpp:218] Iteration 6792 (2.32361 iter/s, 5.16439s/12 iters), loss = 0.311571
I0428 13:55:53.310963  9322 solver.cpp:237]     Train net output #0: loss = 0.311571 (* 1 = 0.311571 loss)
I0428 13:55:53.310972  9322 sgd_solver.cpp:105] Iteration 6792, lr = 0.00260436
I0428 13:55:58.470901  9322 solver.cpp:218] Iteration 6804 (2.32561 iter/s, 5.15994s/12 iters), loss = 0.214453
I0428 13:55:58.471050  9322 solver.cpp:237]     Train net output #0: loss = 0.214453 (* 1 = 0.214453 loss)
I0428 13:55:58.471060  9322 sgd_solver.cpp:105] Iteration 6804, lr = 0.00259817
I0428 13:56:03.655813  9322 solver.cpp:218] Iteration 6816 (2.31447 iter/s, 5.18477s/12 iters), loss = 0.158945
I0428 13:56:03.655864  9322 solver.cpp:237]     Train net output #0: loss = 0.158945 (* 1 = 0.158945 loss)
I0428 13:56:03.655874  9322 sgd_solver.cpp:105] Iteration 6816, lr = 0.00259201
I0428 13:56:08.859926  9322 solver.cpp:218] Iteration 6828 (2.30589 iter/s, 5.20407s/12 iters), loss = 0.248914
I0428 13:56:08.859967  9322 solver.cpp:237]     Train net output #0: loss = 0.248914 (* 1 = 0.248914 loss)
I0428 13:56:08.859975  9322 sgd_solver.cpp:105] Iteration 6828, lr = 0.00258585
I0428 13:56:10.996541  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_6834.caffemodel
I0428 13:56:14.416520  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_6834.solverstate
I0428 13:56:17.734091  9322 solver.cpp:330] Iteration 6834, Testing net (#0)
I0428 13:56:17.734110  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:56:19.624591  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:56:22.650807  9322 solver.cpp:397]     Test net output #0: accuracy = 0.452819
I0428 13:56:22.650852  9322 solver.cpp:397]     Test net output #1: loss = 2.91516 (* 1 = 2.91516 loss)
I0428 13:56:24.550956  9322 solver.cpp:218] Iteration 6840 (0.764767 iter/s, 15.691s/12 iters), loss = 0.141324
I0428 13:56:24.551005  9322 solver.cpp:237]     Train net output #0: loss = 0.141324 (* 1 = 0.141324 loss)
I0428 13:56:24.551013  9322 sgd_solver.cpp:105] Iteration 6840, lr = 0.00257971
I0428 13:56:29.679240  9322 solver.cpp:218] Iteration 6852 (2.33998 iter/s, 5.12824s/12 iters), loss = 0.153148
I0428 13:56:29.679358  9322 solver.cpp:237]     Train net output #0: loss = 0.153148 (* 1 = 0.153148 loss)
I0428 13:56:29.679368  9322 sgd_solver.cpp:105] Iteration 6852, lr = 0.00257359
I0428 13:56:34.868122  9322 solver.cpp:218] Iteration 6864 (2.31269 iter/s, 5.18877s/12 iters), loss = 0.146885
I0428 13:56:34.868165  9322 solver.cpp:237]     Train net output #0: loss = 0.146885 (* 1 = 0.146885 loss)
I0428 13:56:34.868173  9322 sgd_solver.cpp:105] Iteration 6864, lr = 0.00256748
I0428 13:56:40.053946  9322 solver.cpp:218] Iteration 6876 (2.31402 iter/s, 5.18579s/12 iters), loss = 0.0987742
I0428 13:56:40.053992  9322 solver.cpp:237]     Train net output #0: loss = 0.0987742 (* 1 = 0.0987742 loss)
I0428 13:56:40.053999  9322 sgd_solver.cpp:105] Iteration 6876, lr = 0.00256138
I0428 13:56:40.688419  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:56:45.260442  9322 solver.cpp:218] Iteration 6888 (2.30483 iter/s, 5.20646s/12 iters), loss = 0.163252
I0428 13:56:45.260483  9322 solver.cpp:237]     Train net output #0: loss = 0.163252 (* 1 = 0.163252 loss)
I0428 13:56:45.260489  9322 sgd_solver.cpp:105] Iteration 6888, lr = 0.0025553
I0428 13:56:50.389402  9322 solver.cpp:218] Iteration 6900 (2.33967 iter/s, 5.12892s/12 iters), loss = 0.121817
I0428 13:56:50.389448  9322 solver.cpp:237]     Train net output #0: loss = 0.121817 (* 1 = 0.121817 loss)
I0428 13:56:50.389456  9322 sgd_solver.cpp:105] Iteration 6900, lr = 0.00254923
I0428 13:56:55.634444  9322 solver.cpp:218] Iteration 6912 (2.28789 iter/s, 5.24501s/12 iters), loss = 0.217205
I0428 13:56:55.634482  9322 solver.cpp:237]     Train net output #0: loss = 0.217205 (* 1 = 0.217205 loss)
I0428 13:56:55.634491  9322 sgd_solver.cpp:105] Iteration 6912, lr = 0.00254318
I0428 13:57:00.807350  9322 solver.cpp:218] Iteration 6924 (2.31979 iter/s, 5.17288s/12 iters), loss = 0.11358
I0428 13:57:00.807488  9322 solver.cpp:237]     Train net output #0: loss = 0.11358 (* 1 = 0.11358 loss)
I0428 13:57:00.807497  9322 sgd_solver.cpp:105] Iteration 6924, lr = 0.00253714
I0428 13:57:05.455305  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_6936.caffemodel
I0428 13:57:10.339172  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_6936.solverstate
I0428 13:57:13.372318  9322 solver.cpp:330] Iteration 6936, Testing net (#0)
I0428 13:57:13.372335  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:57:13.996415  9322 blocking_queue.cpp:49] Waiting for data
I0428 13:57:15.189604  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:57:18.271981  9322 solver.cpp:397]     Test net output #0: accuracy = 0.462623
I0428 13:57:18.272017  9322 solver.cpp:397]     Test net output #1: loss = 2.97309 (* 1 = 2.97309 loss)
I0428 13:57:18.386373  9322 solver.cpp:218] Iteration 6936 (0.682634 iter/s, 17.579s/12 iters), loss = 0.197823
I0428 13:57:18.386422  9322 solver.cpp:237]     Train net output #0: loss = 0.197823 (* 1 = 0.197823 loss)
I0428 13:57:18.386431  9322 sgd_solver.cpp:105] Iteration 6936, lr = 0.00253112
I0428 13:57:22.702210  9322 solver.cpp:218] Iteration 6948 (2.78048 iter/s, 4.3158s/12 iters), loss = 0.210167
I0428 13:57:22.702245  9322 solver.cpp:237]     Train net output #0: loss = 0.210167 (* 1 = 0.210167 loss)
I0428 13:57:22.702252  9322 sgd_solver.cpp:105] Iteration 6948, lr = 0.00252511
I0428 13:57:27.782933  9322 solver.cpp:218] Iteration 6960 (2.36188 iter/s, 5.08069s/12 iters), loss = 0.194889
I0428 13:57:27.782976  9322 solver.cpp:237]     Train net output #0: loss = 0.194889 (* 1 = 0.194889 loss)
I0428 13:57:27.782985  9322 sgd_solver.cpp:105] Iteration 6960, lr = 0.00251911
I0428 13:57:32.943533  9322 solver.cpp:218] Iteration 6972 (2.32533 iter/s, 5.16056s/12 iters), loss = 0.0761368
I0428 13:57:32.943631  9322 solver.cpp:237]     Train net output #0: loss = 0.0761367 (* 1 = 0.0761367 loss)
I0428 13:57:32.943640  9322 sgd_solver.cpp:105] Iteration 6972, lr = 0.00251313
I0428 13:57:35.793881  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:57:38.144913  9322 solver.cpp:218] Iteration 6984 (2.30712 iter/s, 5.20129s/12 iters), loss = 0.166722
I0428 13:57:38.144965  9322 solver.cpp:237]     Train net output #0: loss = 0.166722 (* 1 = 0.166722 loss)
I0428 13:57:38.144974  9322 sgd_solver.cpp:105] Iteration 6984, lr = 0.00250717
I0428 13:57:43.297448  9322 solver.cpp:218] Iteration 6996 (2.32897 iter/s, 5.15249s/12 iters), loss = 0.0852366
I0428 13:57:43.297490  9322 solver.cpp:237]     Train net output #0: loss = 0.0852366 (* 1 = 0.0852366 loss)
I0428 13:57:43.297498  9322 sgd_solver.cpp:105] Iteration 6996, lr = 0.00250121
I0428 13:57:48.453485  9322 solver.cpp:218] Iteration 7008 (2.32739 iter/s, 5.156s/12 iters), loss = 0.150192
I0428 13:57:48.453532  9322 solver.cpp:237]     Train net output #0: loss = 0.150192 (* 1 = 0.150192 loss)
I0428 13:57:48.453541  9322 sgd_solver.cpp:105] Iteration 7008, lr = 0.00249528
I0428 13:57:53.680327  9322 solver.cpp:218] Iteration 7020 (2.29586 iter/s, 5.2268s/12 iters), loss = 0.0722777
I0428 13:57:53.680375  9322 solver.cpp:237]     Train net output #0: loss = 0.0722777 (* 1 = 0.0722777 loss)
I0428 13:57:53.680385  9322 sgd_solver.cpp:105] Iteration 7020, lr = 0.00248935
I0428 13:57:58.838236  9322 solver.cpp:218] Iteration 7032 (2.32654 iter/s, 5.15787s/12 iters), loss = 0.148977
I0428 13:57:58.838276  9322 solver.cpp:237]     Train net output #0: loss = 0.148977 (* 1 = 0.148977 loss)
I0428 13:57:58.838285  9322 sgd_solver.cpp:105] Iteration 7032, lr = 0.00248344
I0428 13:58:00.925468  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_7038.caffemodel
I0428 13:58:05.032622  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_7038.solverstate
I0428 13:58:08.249483  9322 solver.cpp:330] Iteration 7038, Testing net (#0)
I0428 13:58:08.249500  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:58:10.061956  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:58:13.201273  9322 solver.cpp:397]     Test net output #0: accuracy = 0.46875
I0428 13:58:13.201318  9322 solver.cpp:397]     Test net output #1: loss = 2.90548 (* 1 = 2.90548 loss)
I0428 13:58:15.097128  9322 solver.cpp:218] Iteration 7044 (0.738056 iter/s, 16.2589s/12 iters), loss = 0.12146
I0428 13:58:15.097165  9322 solver.cpp:237]     Train net output #0: loss = 0.12146 (* 1 = 0.12146 loss)
I0428 13:58:15.097172  9322 sgd_solver.cpp:105] Iteration 7044, lr = 0.00247755
I0428 13:58:20.196398  9322 solver.cpp:218] Iteration 7056 (2.35329 iter/s, 5.09924s/12 iters), loss = 0.0527456
I0428 13:58:20.196441  9322 solver.cpp:237]     Train net output #0: loss = 0.0527456 (* 1 = 0.0527456 loss)
I0428 13:58:20.196450  9322 sgd_solver.cpp:105] Iteration 7056, lr = 0.00247166
I0428 13:58:25.381742  9322 solver.cpp:218] Iteration 7068 (2.31423 iter/s, 5.1853s/12 iters), loss = 0.203919
I0428 13:58:25.381803  9322 solver.cpp:237]     Train net output #0: loss = 0.203919 (* 1 = 0.203919 loss)
I0428 13:58:25.381814  9322 sgd_solver.cpp:105] Iteration 7068, lr = 0.0024658
I0428 13:58:30.418948  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:58:30.548691  9322 solver.cpp:218] Iteration 7080 (2.32248 iter/s, 5.16689s/12 iters), loss = 0.0894088
I0428 13:58:30.548763  9322 solver.cpp:237]     Train net output #0: loss = 0.0894087 (* 1 = 0.0894087 loss)
I0428 13:58:30.548779  9322 sgd_solver.cpp:105] Iteration 7080, lr = 0.00245994
I0428 13:58:35.915803  9322 solver.cpp:218] Iteration 7092 (2.23586 iter/s, 5.36705s/12 iters), loss = 0.197052
I0428 13:58:35.915891  9322 solver.cpp:237]     Train net output #0: loss = 0.197052 (* 1 = 0.197052 loss)
I0428 13:58:35.915901  9322 sgd_solver.cpp:105] Iteration 7092, lr = 0.0024541
I0428 13:58:41.233494  9322 solver.cpp:218] Iteration 7104 (2.25665 iter/s, 5.31761s/12 iters), loss = 0.155334
I0428 13:58:41.233541  9322 solver.cpp:237]     Train net output #0: loss = 0.155334 (* 1 = 0.155334 loss)
I0428 13:58:41.233548  9322 sgd_solver.cpp:105] Iteration 7104, lr = 0.00244827
I0428 13:58:46.389513  9322 solver.cpp:218] Iteration 7116 (2.3274 iter/s, 5.15598s/12 iters), loss = 0.202284
I0428 13:58:46.389559  9322 solver.cpp:237]     Train net output #0: loss = 0.202284 (* 1 = 0.202284 loss)
I0428 13:58:46.389567  9322 sgd_solver.cpp:105] Iteration 7116, lr = 0.00244246
I0428 13:58:51.563632  9322 solver.cpp:218] Iteration 7128 (2.31925 iter/s, 5.17408s/12 iters), loss = 0.125104
I0428 13:58:51.563678  9322 solver.cpp:237]     Train net output #0: loss = 0.125104 (* 1 = 0.125104 loss)
I0428 13:58:51.563685  9322 sgd_solver.cpp:105] Iteration 7128, lr = 0.00243666
I0428 13:58:56.237419  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_7140.caffemodel
I0428 13:59:03.462241  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_7140.solverstate
I0428 13:59:07.805303  9322 solver.cpp:330] Iteration 7140, Testing net (#0)
I0428 13:59:07.805366  9322 net.cpp:676] Ignoring source layer train-data
I0428 13:59:09.673084  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:59:12.785246  9322 solver.cpp:397]     Test net output #0: accuracy = 0.457108
I0428 13:59:12.785295  9322 solver.cpp:397]     Test net output #1: loss = 2.87722 (* 1 = 2.87722 loss)
I0428 13:59:12.903915  9322 solver.cpp:218] Iteration 7140 (0.562316 iter/s, 21.3403s/12 iters), loss = 0.14891
I0428 13:59:12.903959  9322 solver.cpp:237]     Train net output #0: loss = 0.148909 (* 1 = 0.148909 loss)
I0428 13:59:12.903967  9322 sgd_solver.cpp:105] Iteration 7140, lr = 0.00243088
I0428 13:59:17.336328  9322 solver.cpp:218] Iteration 7152 (2.70735 iter/s, 4.43237s/12 iters), loss = 0.183345
I0428 13:59:17.336365  9322 solver.cpp:237]     Train net output #0: loss = 0.183345 (* 1 = 0.183345 loss)
I0428 13:59:17.336374  9322 sgd_solver.cpp:105] Iteration 7152, lr = 0.00242511
I0428 13:59:22.496594  9322 solver.cpp:218] Iteration 7164 (2.32547 iter/s, 5.16024s/12 iters), loss = 0.193313
I0428 13:59:22.496635  9322 solver.cpp:237]     Train net output #0: loss = 0.193313 (* 1 = 0.193313 loss)
I0428 13:59:22.496644  9322 sgd_solver.cpp:105] Iteration 7164, lr = 0.00241935
I0428 13:59:27.759799  9322 solver.cpp:218] Iteration 7176 (2.28 iter/s, 5.26317s/12 iters), loss = 0.153075
I0428 13:59:27.759842  9322 solver.cpp:237]     Train net output #0: loss = 0.153075 (* 1 = 0.153075 loss)
I0428 13:59:27.759851  9322 sgd_solver.cpp:105] Iteration 7176, lr = 0.0024136
I0428 13:59:29.953032  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 13:59:32.947508  9322 solver.cpp:218] Iteration 7188 (2.31318 iter/s, 5.18767s/12 iters), loss = 0.181952
I0428 13:59:32.947551  9322 solver.cpp:237]     Train net output #0: loss = 0.181952 (* 1 = 0.181952 loss)
I0428 13:59:32.947559  9322 sgd_solver.cpp:105] Iteration 7188, lr = 0.00240787
I0428 13:59:38.113317  9322 solver.cpp:218] Iteration 7200 (2.32298 iter/s, 5.16577s/12 iters), loss = 0.0795393
I0428 13:59:38.113473  9322 solver.cpp:237]     Train net output #0: loss = 0.0795393 (* 1 = 0.0795393 loss)
I0428 13:59:38.113482  9322 sgd_solver.cpp:105] Iteration 7200, lr = 0.00240216
I0428 13:59:43.272966  9322 solver.cpp:218] Iteration 7212 (2.32581 iter/s, 5.1595s/12 iters), loss = 0.208727
I0428 13:59:43.273015  9322 solver.cpp:237]     Train net output #0: loss = 0.208727 (* 1 = 0.208727 loss)
I0428 13:59:43.273022  9322 sgd_solver.cpp:105] Iteration 7212, lr = 0.00239645
I0428 13:59:48.406049  9322 solver.cpp:218] Iteration 7224 (2.3378 iter/s, 5.13304s/12 iters), loss = 0.165779
I0428 13:59:48.406095  9322 solver.cpp:237]     Train net output #0: loss = 0.165779 (* 1 = 0.165779 loss)
I0428 13:59:48.406103  9322 sgd_solver.cpp:105] Iteration 7224, lr = 0.00239076
I0428 13:59:53.625864  9322 solver.cpp:218] Iteration 7236 (2.29895 iter/s, 5.21978s/12 iters), loss = 0.14739
I0428 13:59:53.625910  9322 solver.cpp:237]     Train net output #0: loss = 0.14739 (* 1 = 0.14739 loss)
I0428 13:59:53.625917  9322 sgd_solver.cpp:105] Iteration 7236, lr = 0.00238509
I0428 13:59:55.749228  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_7242.caffemodel
I0428 14:00:01.394013  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_7242.solverstate
I0428 14:00:08.049036  9322 solver.cpp:330] Iteration 7242, Testing net (#0)
I0428 14:00:08.049057  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:00:09.772053  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:00:12.998311  9322 solver.cpp:397]     Test net output #0: accuracy = 0.45527
I0428 14:00:12.998350  9322 solver.cpp:397]     Test net output #1: loss = 2.77934 (* 1 = 2.77934 loss)
I0428 14:00:14.818419  9322 solver.cpp:218] Iteration 7248 (0.566235 iter/s, 21.1926s/12 iters), loss = 0.0698457
I0428 14:00:14.818461  9322 solver.cpp:237]     Train net output #0: loss = 0.0698456 (* 1 = 0.0698456 loss)
I0428 14:00:14.818470  9322 sgd_solver.cpp:105] Iteration 7248, lr = 0.00237942
I0428 14:00:19.978482  9322 solver.cpp:218] Iteration 7260 (2.32557 iter/s, 5.16003s/12 iters), loss = 0.143033
I0428 14:00:19.978528  9322 solver.cpp:237]     Train net output #0: loss = 0.143033 (* 1 = 0.143033 loss)
I0428 14:00:19.978538  9322 sgd_solver.cpp:105] Iteration 7260, lr = 0.00237378
I0428 14:00:25.154899  9322 solver.cpp:218] Iteration 7272 (2.31822 iter/s, 5.17638s/12 iters), loss = 0.102586
I0428 14:00:25.154947  9322 solver.cpp:237]     Train net output #0: loss = 0.102586 (* 1 = 0.102586 loss)
I0428 14:00:25.154956  9322 sgd_solver.cpp:105] Iteration 7272, lr = 0.00236814
I0428 14:00:29.775725  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:00:30.613451  9322 solver.cpp:218] Iteration 7284 (2.1984 iter/s, 5.45851s/12 iters), loss = 0.127781
I0428 14:00:30.613490  9322 solver.cpp:237]     Train net output #0: loss = 0.127781 (* 1 = 0.127781 loss)
I0428 14:00:30.613498  9322 sgd_solver.cpp:105] Iteration 7284, lr = 0.00236252
I0428 14:00:35.775584  9322 solver.cpp:218] Iteration 7296 (2.32464 iter/s, 5.1621s/12 iters), loss = 0.0845246
I0428 14:00:35.775629  9322 solver.cpp:237]     Train net output #0: loss = 0.0845245 (* 1 = 0.0845245 loss)
I0428 14:00:35.775638  9322 sgd_solver.cpp:105] Iteration 7296, lr = 0.00235691
I0428 14:00:40.855079  9322 solver.cpp:218] Iteration 7308 (2.36246 iter/s, 5.07946s/12 iters), loss = 0.172329
I0428 14:00:40.855221  9322 solver.cpp:237]     Train net output #0: loss = 0.172329 (* 1 = 0.172329 loss)
I0428 14:00:40.855232  9322 sgd_solver.cpp:105] Iteration 7308, lr = 0.00235131
I0428 14:00:46.000392  9322 solver.cpp:218] Iteration 7320 (2.33229 iter/s, 5.14517s/12 iters), loss = 0.145962
I0428 14:00:46.000466  9322 solver.cpp:237]     Train net output #0: loss = 0.145962 (* 1 = 0.145962 loss)
I0428 14:00:46.000481  9322 sgd_solver.cpp:105] Iteration 7320, lr = 0.00234573
I0428 14:00:51.202263  9322 solver.cpp:218] Iteration 7332 (2.30689 iter/s, 5.2018s/12 iters), loss = 0.171607
I0428 14:00:51.202328  9322 solver.cpp:237]     Train net output #0: loss = 0.171607 (* 1 = 0.171607 loss)
I0428 14:00:51.202342  9322 sgd_solver.cpp:105] Iteration 7332, lr = 0.00234016
I0428 14:00:55.913293  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_7344.caffemodel
I0428 14:00:59.256213  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_7344.solverstate
I0428 14:01:04.153786  9322 solver.cpp:330] Iteration 7344, Testing net (#0)
I0428 14:01:04.153805  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:01:05.808290  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:01:09.076004  9322 solver.cpp:397]     Test net output #0: accuracy = 0.463848
I0428 14:01:09.076045  9322 solver.cpp:397]     Test net output #1: loss = 2.8109 (* 1 = 2.8109 loss)
I0428 14:01:09.194315  9322 solver.cpp:218] Iteration 7344 (0.666961 iter/s, 17.9921s/12 iters), loss = 0.199535
I0428 14:01:09.194365  9322 solver.cpp:237]     Train net output #0: loss = 0.199535 (* 1 = 0.199535 loss)
I0428 14:01:09.194372  9322 sgd_solver.cpp:105] Iteration 7344, lr = 0.0023346
I0428 14:01:13.523548  9322 solver.cpp:218] Iteration 7356 (2.77188 iter/s, 4.32919s/12 iters), loss = 0.189579
I0428 14:01:13.523634  9322 solver.cpp:237]     Train net output #0: loss = 0.189579 (* 1 = 0.189579 loss)
I0428 14:01:13.523643  9322 sgd_solver.cpp:105] Iteration 7356, lr = 0.00232906
I0428 14:01:18.825587  9322 solver.cpp:218] Iteration 7368 (2.26331 iter/s, 5.30196s/12 iters), loss = 0.119019
I0428 14:01:18.825625  9322 solver.cpp:237]     Train net output #0: loss = 0.119019 (* 1 = 0.119019 loss)
I0428 14:01:18.825634  9322 sgd_solver.cpp:105] Iteration 7368, lr = 0.00232353
I0428 14:01:24.146809  9322 solver.cpp:218] Iteration 7380 (2.25514 iter/s, 5.32118s/12 iters), loss = 0.118135
I0428 14:01:24.146872  9322 solver.cpp:237]     Train net output #0: loss = 0.118135 (* 1 = 0.118135 loss)
I0428 14:01:24.146886  9322 sgd_solver.cpp:105] Iteration 7380, lr = 0.00231802
I0428 14:01:25.598640  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:01:29.454713  9322 solver.cpp:218] Iteration 7392 (2.2608 iter/s, 5.30785s/12 iters), loss = 0.173503
I0428 14:01:29.454759  9322 solver.cpp:237]     Train net output #0: loss = 0.173503 (* 1 = 0.173503 loss)
I0428 14:01:29.454767  9322 sgd_solver.cpp:105] Iteration 7392, lr = 0.00231251
I0428 14:01:34.667419  9322 solver.cpp:218] Iteration 7404 (2.30209 iter/s, 5.21267s/12 iters), loss = 0.151055
I0428 14:01:34.667461  9322 solver.cpp:237]     Train net output #0: loss = 0.151055 (* 1 = 0.151055 loss)
I0428 14:01:34.667469  9322 sgd_solver.cpp:105] Iteration 7404, lr = 0.00230702
I0428 14:01:39.836952  9322 solver.cpp:218] Iteration 7416 (2.32131 iter/s, 5.1695s/12 iters), loss = 0.133581
I0428 14:01:39.836992  9322 solver.cpp:237]     Train net output #0: loss = 0.133581 (* 1 = 0.133581 loss)
I0428 14:01:39.837002  9322 sgd_solver.cpp:105] Iteration 7416, lr = 0.00230154
I0428 14:01:45.088642  9322 solver.cpp:218] Iteration 7428 (2.28499 iter/s, 5.25166s/12 iters), loss = 0.0542024
I0428 14:01:45.088778  9322 solver.cpp:237]     Train net output #0: loss = 0.0542023 (* 1 = 0.0542023 loss)
I0428 14:01:45.088786  9322 sgd_solver.cpp:105] Iteration 7428, lr = 0.00229608
I0428 14:01:50.252292  9322 solver.cpp:218] Iteration 7440 (2.324 iter/s, 5.16352s/12 iters), loss = 0.0838256
I0428 14:01:50.252355  9322 solver.cpp:237]     Train net output #0: loss = 0.0838255 (* 1 = 0.0838255 loss)
I0428 14:01:50.252367  9322 sgd_solver.cpp:105] Iteration 7440, lr = 0.00229063
I0428 14:01:52.350608  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_7446.caffemodel
I0428 14:01:55.502518  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_7446.solverstate
I0428 14:01:59.716289  9322 solver.cpp:330] Iteration 7446, Testing net (#0)
I0428 14:01:59.716308  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:02:01.325553  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:02:04.648710  9322 solver.cpp:397]     Test net output #0: accuracy = 0.470588
I0428 14:02:04.648739  9322 solver.cpp:397]     Test net output #1: loss = 2.86818 (* 1 = 2.86818 loss)
I0428 14:02:06.607890  9322 solver.cpp:218] Iteration 7452 (0.733694 iter/s, 16.3556s/12 iters), loss = 0.0867394
I0428 14:02:06.607935  9322 solver.cpp:237]     Train net output #0: loss = 0.0867393 (* 1 = 0.0867393 loss)
I0428 14:02:06.607944  9322 sgd_solver.cpp:105] Iteration 7452, lr = 0.00228519
I0428 14:02:11.725014  9322 solver.cpp:218] Iteration 7464 (2.34508 iter/s, 5.11709s/12 iters), loss = 0.0681191
I0428 14:02:11.725059  9322 solver.cpp:237]     Train net output #0: loss = 0.068119 (* 1 = 0.068119 loss)
I0428 14:02:11.725066  9322 sgd_solver.cpp:105] Iteration 7464, lr = 0.00227976
I0428 14:02:16.833906  9322 solver.cpp:218] Iteration 7476 (2.34886 iter/s, 5.10885s/12 iters), loss = 0.089547
I0428 14:02:16.834035  9322 solver.cpp:237]     Train net output #0: loss = 0.0895469 (* 1 = 0.0895469 loss)
I0428 14:02:16.834045  9322 sgd_solver.cpp:105] Iteration 7476, lr = 0.00227435
I0428 14:02:20.468011  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:02:22.016157  9322 solver.cpp:218] Iteration 7488 (2.31565 iter/s, 5.18213s/12 iters), loss = 0.161174
I0428 14:02:22.016197  9322 solver.cpp:237]     Train net output #0: loss = 0.161174 (* 1 = 0.161174 loss)
I0428 14:02:22.016206  9322 sgd_solver.cpp:105] Iteration 7488, lr = 0.00226895
I0428 14:02:27.105474  9322 solver.cpp:218] Iteration 7500 (2.3579 iter/s, 5.08928s/12 iters), loss = 0.0630992
I0428 14:02:27.105523  9322 solver.cpp:237]     Train net output #0: loss = 0.0630991 (* 1 = 0.0630991 loss)
I0428 14:02:27.105532  9322 sgd_solver.cpp:105] Iteration 7500, lr = 0.00226357
I0428 14:02:32.195863  9322 solver.cpp:218] Iteration 7512 (2.3574 iter/s, 5.09035s/12 iters), loss = 0.114678
I0428 14:02:32.195904  9322 solver.cpp:237]     Train net output #0: loss = 0.114678 (* 1 = 0.114678 loss)
I0428 14:02:32.195912  9322 sgd_solver.cpp:105] Iteration 7512, lr = 0.00225819
I0428 14:02:37.339498  9322 solver.cpp:218] Iteration 7524 (2.333 iter/s, 5.1436s/12 iters), loss = 0.0430462
I0428 14:02:37.339538  9322 solver.cpp:237]     Train net output #0: loss = 0.0430461 (* 1 = 0.0430461 loss)
I0428 14:02:37.339547  9322 sgd_solver.cpp:105] Iteration 7524, lr = 0.00225283
I0428 14:02:42.551782  9322 solver.cpp:218] Iteration 7536 (2.30227 iter/s, 5.21226s/12 iters), loss = 0.10205
I0428 14:02:42.551822  9322 solver.cpp:237]     Train net output #0: loss = 0.10205 (* 1 = 0.10205 loss)
I0428 14:02:42.551831  9322 sgd_solver.cpp:105] Iteration 7536, lr = 0.00224748
I0428 14:02:47.229359  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_7548.caffemodel
I0428 14:02:51.155417  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_7548.solverstate
I0428 14:02:55.800340  9322 solver.cpp:330] Iteration 7548, Testing net (#0)
I0428 14:02:55.800360  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:02:57.380239  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:03:00.722539  9322 solver.cpp:397]     Test net output #0: accuracy = 0.478554
I0428 14:03:00.722584  9322 solver.cpp:397]     Test net output #1: loss = 2.85304 (* 1 = 2.85304 loss)
I0428 14:03:00.839257  9322 solver.cpp:218] Iteration 7548 (0.656186 iter/s, 18.2875s/12 iters), loss = 0.150036
I0428 14:03:00.839298  9322 solver.cpp:237]     Train net output #0: loss = 0.150036 (* 1 = 0.150036 loss)
I0428 14:03:00.839306  9322 sgd_solver.cpp:105] Iteration 7548, lr = 0.00224215
I0428 14:03:05.206621  9322 solver.cpp:218] Iteration 7560 (2.74767 iter/s, 4.36733s/12 iters), loss = 0.0608968
I0428 14:03:05.206663  9322 solver.cpp:237]     Train net output #0: loss = 0.0608967 (* 1 = 0.0608967 loss)
I0428 14:03:05.206671  9322 sgd_solver.cpp:105] Iteration 7560, lr = 0.00223682
I0428 14:03:10.419045  9322 solver.cpp:218] Iteration 7572 (2.30221 iter/s, 5.21239s/12 iters), loss = 0.11804
I0428 14:03:10.419091  9322 solver.cpp:237]     Train net output #0: loss = 0.11804 (* 1 = 0.11804 loss)
I0428 14:03:10.419100  9322 sgd_solver.cpp:105] Iteration 7572, lr = 0.00223151
I0428 14:03:15.598604  9322 solver.cpp:218] Iteration 7584 (2.31682 iter/s, 5.17952s/12 iters), loss = 0.112005
I0428 14:03:15.598655  9322 solver.cpp:237]     Train net output #0: loss = 0.112004 (* 1 = 0.112004 loss)
I0428 14:03:15.598664  9322 sgd_solver.cpp:105] Iteration 7584, lr = 0.00222621
I0428 14:03:16.261333  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:03:20.796994  9322 solver.cpp:218] Iteration 7596 (2.30843 iter/s, 5.19834s/12 iters), loss = 0.0614119
I0428 14:03:20.797084  9322 solver.cpp:237]     Train net output #0: loss = 0.0614118 (* 1 = 0.0614118 loss)
I0428 14:03:20.797094  9322 sgd_solver.cpp:105] Iteration 7596, lr = 0.00222093
I0428 14:03:25.947151  9322 solver.cpp:218] Iteration 7608 (2.33006 iter/s, 5.15007s/12 iters), loss = 0.211045
I0428 14:03:25.947199  9322 solver.cpp:237]     Train net output #0: loss = 0.211045 (* 1 = 0.211045 loss)
I0428 14:03:25.947207  9322 sgd_solver.cpp:105] Iteration 7608, lr = 0.00221565
I0428 14:03:31.145462  9322 solver.cpp:218] Iteration 7620 (2.30846 iter/s, 5.19827s/12 iters), loss = 0.0889704
I0428 14:03:31.145509  9322 solver.cpp:237]     Train net output #0: loss = 0.0889703 (* 1 = 0.0889703 loss)
I0428 14:03:31.145516  9322 sgd_solver.cpp:105] Iteration 7620, lr = 0.00221039
I0428 14:03:33.658241  9322 blocking_queue.cpp:49] Waiting for data
I0428 14:03:36.315693  9322 solver.cpp:218] Iteration 7632 (2.321 iter/s, 5.17019s/12 iters), loss = 0.0799749
I0428 14:03:36.315738  9322 solver.cpp:237]     Train net output #0: loss = 0.0799749 (* 1 = 0.0799749 loss)
I0428 14:03:36.315747  9322 sgd_solver.cpp:105] Iteration 7632, lr = 0.00220515
I0428 14:03:41.466179  9322 solver.cpp:218] Iteration 7644 (2.3299 iter/s, 5.15045s/12 iters), loss = 0.0692678
I0428 14:03:41.466224  9322 solver.cpp:237]     Train net output #0: loss = 0.0692677 (* 1 = 0.0692677 loss)
I0428 14:03:41.466233  9322 sgd_solver.cpp:105] Iteration 7644, lr = 0.00219991
I0428 14:03:43.542737  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_7650.caffemodel
I0428 14:03:48.945719  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_7650.solverstate
I0428 14:03:55.075455  9322 solver.cpp:330] Iteration 7650, Testing net (#0)
I0428 14:03:55.075580  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:03:56.595163  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:03:59.984925  9322 solver.cpp:397]     Test net output #0: accuracy = 0.482843
I0428 14:03:59.984972  9322 solver.cpp:397]     Test net output #1: loss = 2.87326 (* 1 = 2.87326 loss)
I0428 14:04:01.942972  9322 solver.cpp:218] Iteration 7656 (0.586028 iter/s, 20.4768s/12 iters), loss = 0.11047
I0428 14:04:01.943014  9322 solver.cpp:237]     Train net output #0: loss = 0.11047 (* 1 = 0.11047 loss)
I0428 14:04:01.943022  9322 sgd_solver.cpp:105] Iteration 7656, lr = 0.00219469
I0428 14:04:07.181026  9322 solver.cpp:218] Iteration 7668 (2.29094 iter/s, 5.23802s/12 iters), loss = 0.143749
I0428 14:04:07.181071  9322 solver.cpp:237]     Train net output #0: loss = 0.143749 (* 1 = 0.143749 loss)
I0428 14:04:07.181078  9322 sgd_solver.cpp:105] Iteration 7668, lr = 0.00218948
I0428 14:04:12.370761  9322 solver.cpp:218] Iteration 7680 (2.31228 iter/s, 5.18969s/12 iters), loss = 0.236746
I0428 14:04:12.370811  9322 solver.cpp:237]     Train net output #0: loss = 0.236746 (* 1 = 0.236746 loss)
I0428 14:04:12.370820  9322 sgd_solver.cpp:105] Iteration 7680, lr = 0.00218428
I0428 14:04:15.258194  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:04:17.567381  9322 solver.cpp:218] Iteration 7692 (2.30921 iter/s, 5.19658s/12 iters), loss = 0.118224
I0428 14:04:17.567420  9322 solver.cpp:237]     Train net output #0: loss = 0.118224 (* 1 = 0.118224 loss)
I0428 14:04:17.567427  9322 sgd_solver.cpp:105] Iteration 7692, lr = 0.00217909
I0428 14:04:22.717854  9322 solver.cpp:218] Iteration 7704 (2.3299 iter/s, 5.15043s/12 iters), loss = 0.0745826
I0428 14:04:22.717902  9322 solver.cpp:237]     Train net output #0: loss = 0.0745825 (* 1 = 0.0745825 loss)
I0428 14:04:22.717911  9322 sgd_solver.cpp:105] Iteration 7704, lr = 0.00217392
I0428 14:04:27.918293  9322 solver.cpp:218] Iteration 7716 (2.30751 iter/s, 5.2004s/12 iters), loss = 0.0510119
I0428 14:04:27.918427  9322 solver.cpp:237]     Train net output #0: loss = 0.0510118 (* 1 = 0.0510118 loss)
I0428 14:04:27.918437  9322 sgd_solver.cpp:105] Iteration 7716, lr = 0.00216876
I0428 14:04:33.089610  9322 solver.cpp:218] Iteration 7728 (2.32055 iter/s, 5.17119s/12 iters), loss = 0.108333
I0428 14:04:33.089656  9322 solver.cpp:237]     Train net output #0: loss = 0.108333 (* 1 = 0.108333 loss)
I0428 14:04:33.089665  9322 sgd_solver.cpp:105] Iteration 7728, lr = 0.00216361
I0428 14:04:38.252995  9322 solver.cpp:218] Iteration 7740 (2.32408 iter/s, 5.16334s/12 iters), loss = 0.14849
I0428 14:04:38.253036  9322 solver.cpp:237]     Train net output #0: loss = 0.14849 (* 1 = 0.14849 loss)
I0428 14:04:38.253044  9322 sgd_solver.cpp:105] Iteration 7740, lr = 0.00215847
I0428 14:04:42.917301  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_7752.caffemodel
I0428 14:04:49.279508  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_7752.solverstate
I0428 14:04:57.728091  9322 solver.cpp:330] Iteration 7752, Testing net (#0)
I0428 14:04:57.728108  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:04:59.243139  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:05:02.540773  9322 solver.cpp:397]     Test net output #0: accuracy = 0.47549
I0428 14:05:02.540820  9322 solver.cpp:397]     Test net output #1: loss = 2.83231 (* 1 = 2.83231 loss)
I0428 14:05:02.659184  9322 solver.cpp:218] Iteration 7752 (0.491677 iter/s, 24.4062s/12 iters), loss = 0.0991216
I0428 14:05:02.659235  9322 solver.cpp:237]     Train net output #0: loss = 0.0991215 (* 1 = 0.0991215 loss)
I0428 14:05:02.659245  9322 sgd_solver.cpp:105] Iteration 7752, lr = 0.00215335
I0428 14:05:06.963300  9322 solver.cpp:218] Iteration 7764 (2.78806 iter/s, 4.30407s/12 iters), loss = 0.154827
I0428 14:05:06.963344  9322 solver.cpp:237]     Train net output #0: loss = 0.154827 (* 1 = 0.154827 loss)
I0428 14:05:06.963352  9322 sgd_solver.cpp:105] Iteration 7764, lr = 0.00214823
I0428 14:05:12.246764  9322 solver.cpp:218] Iteration 7776 (2.27125 iter/s, 5.28343s/12 iters), loss = 0.0717368
I0428 14:05:12.246809  9322 solver.cpp:237]     Train net output #0: loss = 0.0717367 (* 1 = 0.0717367 loss)
I0428 14:05:12.246817  9322 sgd_solver.cpp:105] Iteration 7776, lr = 0.00214313
I0428 14:05:17.273455  9322 solver.cpp:218] Iteration 7788 (2.38728 iter/s, 5.02665s/12 iters), loss = 0.141021
I0428 14:05:17.273502  9322 solver.cpp:237]     Train net output #0: loss = 0.141021 (* 1 = 0.141021 loss)
I0428 14:05:17.273510  9322 sgd_solver.cpp:105] Iteration 7788, lr = 0.00213805
I0428 14:05:17.281107  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:05:22.440603  9322 solver.cpp:218] Iteration 7800 (2.32238 iter/s, 5.1671s/12 iters), loss = 0.059966
I0428 14:05:22.440652  9322 solver.cpp:237]     Train net output #0: loss = 0.0599659 (* 1 = 0.0599659 loss)
I0428 14:05:22.440661  9322 sgd_solver.cpp:105] Iteration 7800, lr = 0.00213297
I0428 14:05:27.593878  9322 solver.cpp:218] Iteration 7812 (2.32864 iter/s, 5.15323s/12 iters), loss = 0.084275
I0428 14:05:27.593926  9322 solver.cpp:237]     Train net output #0: loss = 0.0842749 (* 1 = 0.0842749 loss)
I0428 14:05:27.593935  9322 sgd_solver.cpp:105] Iteration 7812, lr = 0.00212791
I0428 14:05:32.784094  9322 solver.cpp:218] Iteration 7824 (2.31206 iter/s, 5.19017s/12 iters), loss = 0.0586939
I0428 14:05:32.784265  9322 solver.cpp:237]     Train net output #0: loss = 0.0586938 (* 1 = 0.0586938 loss)
I0428 14:05:32.784274  9322 sgd_solver.cpp:105] Iteration 7824, lr = 0.00212285
I0428 14:05:37.951828  9322 solver.cpp:218] Iteration 7836 (2.32217 iter/s, 5.16758s/12 iters), loss = 0.0663468
I0428 14:05:37.951874  9322 solver.cpp:237]     Train net output #0: loss = 0.0663467 (* 1 = 0.0663467 loss)
I0428 14:05:37.951884  9322 sgd_solver.cpp:105] Iteration 7836, lr = 0.00211781
I0428 14:05:43.120904  9322 solver.cpp:218] Iteration 7848 (2.32151 iter/s, 5.16905s/12 iters), loss = 0.0485368
I0428 14:05:43.120945  9322 solver.cpp:237]     Train net output #0: loss = 0.0485367 (* 1 = 0.0485367 loss)
I0428 14:05:43.120952  9322 sgd_solver.cpp:105] Iteration 7848, lr = 0.00211279
I0428 14:05:45.181931  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_7854.caffemodel
I0428 14:05:54.789685  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_7854.solverstate
I0428 14:05:58.396430  9322 solver.cpp:330] Iteration 7854, Testing net (#0)
I0428 14:05:58.396446  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:05:59.821702  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:06:03.205183  9322 solver.cpp:397]     Test net output #0: accuracy = 0.481618
I0428 14:06:03.205339  9322 solver.cpp:397]     Test net output #1: loss = 2.8472 (* 1 = 2.8472 loss)
I0428 14:06:05.133913  9322 solver.cpp:218] Iteration 7860 (0.545131 iter/s, 22.0131s/12 iters), loss = 0.172792
I0428 14:06:05.133960  9322 solver.cpp:237]     Train net output #0: loss = 0.172792 (* 1 = 0.172792 loss)
I0428 14:06:05.133968  9322 sgd_solver.cpp:105] Iteration 7860, lr = 0.00210777
I0428 14:06:10.316926  9322 solver.cpp:218] Iteration 7872 (2.31527 iter/s, 5.18297s/12 iters), loss = 0.108896
I0428 14:06:10.316967  9322 solver.cpp:237]     Train net output #0: loss = 0.108896 (* 1 = 0.108896 loss)
I0428 14:06:10.316974  9322 sgd_solver.cpp:105] Iteration 7872, lr = 0.00210277
I0428 14:06:15.425619  9322 solver.cpp:218] Iteration 7884 (2.34895 iter/s, 5.10866s/12 iters), loss = 0.167788
I0428 14:06:15.425668  9322 solver.cpp:237]     Train net output #0: loss = 0.167788 (* 1 = 0.167788 loss)
I0428 14:06:15.425675  9322 sgd_solver.cpp:105] Iteration 7884, lr = 0.00209777
I0428 14:06:17.667593  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:06:20.626922  9322 solver.cpp:218] Iteration 7896 (2.30713 iter/s, 5.20127s/12 iters), loss = 0.0882076
I0428 14:06:20.626965  9322 solver.cpp:237]     Train net output #0: loss = 0.0882075 (* 1 = 0.0882075 loss)
I0428 14:06:20.626973  9322 sgd_solver.cpp:105] Iteration 7896, lr = 0.00209279
I0428 14:06:25.705014  9322 solver.cpp:218] Iteration 7908 (2.36311 iter/s, 5.07806s/12 iters), loss = 0.178969
I0428 14:06:25.705054  9322 solver.cpp:237]     Train net output #0: loss = 0.178968 (* 1 = 0.178968 loss)
I0428 14:06:25.705062  9322 sgd_solver.cpp:105] Iteration 7908, lr = 0.00208782
I0428 14:06:30.852874  9322 solver.cpp:218] Iteration 7920 (2.33108 iter/s, 5.14784s/12 iters), loss = 0.107415
I0428 14:06:30.852917  9322 solver.cpp:237]     Train net output #0: loss = 0.107415 (* 1 = 0.107415 loss)
I0428 14:06:30.852926  9322 sgd_solver.cpp:105] Iteration 7920, lr = 0.00208287
I0428 14:06:36.026621  9322 solver.cpp:218] Iteration 7932 (2.31941 iter/s, 5.17372s/12 iters), loss = 0.110887
I0428 14:06:36.026772  9322 solver.cpp:237]     Train net output #0: loss = 0.110886 (* 1 = 0.110886 loss)
I0428 14:06:36.026780  9322 sgd_solver.cpp:105] Iteration 7932, lr = 0.00207792
I0428 14:06:41.187870  9322 solver.cpp:218] Iteration 7944 (2.32508 iter/s, 5.16112s/12 iters), loss = 0.0474763
I0428 14:06:41.187914  9322 solver.cpp:237]     Train net output #0: loss = 0.0474762 (* 1 = 0.0474762 loss)
I0428 14:06:41.187922  9322 sgd_solver.cpp:105] Iteration 7944, lr = 0.00207299
I0428 14:06:45.864898  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_7956.caffemodel
I0428 14:06:54.271652  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_7956.solverstate
I0428 14:06:58.647006  9322 solver.cpp:330] Iteration 7956, Testing net (#0)
I0428 14:06:58.647023  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:07:00.051340  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:07:03.466908  9322 solver.cpp:397]     Test net output #0: accuracy = 0.471201
I0428 14:07:03.466939  9322 solver.cpp:397]     Test net output #1: loss = 2.82972 (* 1 = 2.82972 loss)
I0428 14:07:03.585517  9322 solver.cpp:218] Iteration 7956 (0.535768 iter/s, 22.3977s/12 iters), loss = 0.0801424
I0428 14:07:03.585561  9322 solver.cpp:237]     Train net output #0: loss = 0.0801423 (* 1 = 0.0801423 loss)
I0428 14:07:03.585569  9322 sgd_solver.cpp:105] Iteration 7956, lr = 0.00206807
I0428 14:07:08.081063  9322 solver.cpp:218] Iteration 7968 (2.66933 iter/s, 4.49552s/12 iters), loss = 0.0670097
I0428 14:07:08.081167  9322 solver.cpp:237]     Train net output #0: loss = 0.0670096 (* 1 = 0.0670096 loss)
I0428 14:07:08.081176  9322 sgd_solver.cpp:105] Iteration 7968, lr = 0.00206316
I0428 14:07:13.230049  9322 solver.cpp:218] Iteration 7980 (2.33059 iter/s, 5.1489s/12 iters), loss = 0.247135
I0428 14:07:13.230093  9322 solver.cpp:237]     Train net output #0: loss = 0.247135 (* 1 = 0.247135 loss)
I0428 14:07:13.230100  9322 sgd_solver.cpp:105] Iteration 7980, lr = 0.00205826
I0428 14:07:17.654541  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:07:18.404321  9322 solver.cpp:218] Iteration 7992 (2.31918 iter/s, 5.17425s/12 iters), loss = 0.115685
I0428 14:07:18.404366  9322 solver.cpp:237]     Train net output #0: loss = 0.115685 (* 1 = 0.115685 loss)
I0428 14:07:18.404374  9322 sgd_solver.cpp:105] Iteration 7992, lr = 0.00205337
I0428 14:07:23.609881  9322 solver.cpp:218] Iteration 8004 (2.30524 iter/s, 5.20554s/12 iters), loss = 0.0727106
I0428 14:07:23.609923  9322 solver.cpp:237]     Train net output #0: loss = 0.0727105 (* 1 = 0.0727105 loss)
I0428 14:07:23.609932  9322 sgd_solver.cpp:105] Iteration 8004, lr = 0.0020485
I0428 14:07:28.828681  9322 solver.cpp:218] Iteration 8016 (2.29939 iter/s, 5.21878s/12 iters), loss = 0.170909
I0428 14:07:28.828725  9322 solver.cpp:237]     Train net output #0: loss = 0.170908 (* 1 = 0.170908 loss)
I0428 14:07:28.828733  9322 sgd_solver.cpp:105] Iteration 8016, lr = 0.00204363
I0428 14:07:34.030819  9322 solver.cpp:218] Iteration 8028 (2.30676 iter/s, 5.20211s/12 iters), loss = 0.156434
I0428 14:07:34.030861  9322 solver.cpp:237]     Train net output #0: loss = 0.156434 (* 1 = 0.156434 loss)
I0428 14:07:34.030869  9322 sgd_solver.cpp:105] Iteration 8028, lr = 0.00203878
I0428 14:07:39.250167  9322 solver.cpp:218] Iteration 8040 (2.29915 iter/s, 5.21933s/12 iters), loss = 0.0911811
I0428 14:07:39.250320  9322 solver.cpp:237]     Train net output #0: loss = 0.091181 (* 1 = 0.091181 loss)
I0428 14:07:39.250330  9322 sgd_solver.cpp:105] Iteration 8040, lr = 0.00203394
I0428 14:07:44.480355  9322 solver.cpp:218] Iteration 8052 (2.29443 iter/s, 5.23006s/12 iters), loss = 0.167207
I0428 14:07:44.480398  9322 solver.cpp:237]     Train net output #0: loss = 0.167207 (* 1 = 0.167207 loss)
I0428 14:07:44.480407  9322 sgd_solver.cpp:105] Iteration 8052, lr = 0.00202911
I0428 14:07:46.580262  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_8058.caffemodel
I0428 14:07:56.520583  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_8058.solverstate
I0428 14:07:59.911307  9322 solver.cpp:330] Iteration 8058, Testing net (#0)
I0428 14:07:59.911325  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:08:01.290683  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:08:04.779958  9322 solver.cpp:397]     Test net output #0: accuracy = 0.474265
I0428 14:08:04.779995  9322 solver.cpp:397]     Test net output #1: loss = 2.82217 (* 1 = 2.82217 loss)
I0428 14:08:06.661782  9322 solver.cpp:218] Iteration 8064 (0.540991 iter/s, 22.1815s/12 iters), loss = 0.140515
I0428 14:08:06.661829  9322 solver.cpp:237]     Train net output #0: loss = 0.140515 (* 1 = 0.140515 loss)
I0428 14:08:06.661837  9322 sgd_solver.cpp:105] Iteration 8064, lr = 0.00202429
I0428 14:08:11.863606  9322 solver.cpp:218] Iteration 8076 (2.3069 iter/s, 5.20179s/12 iters), loss = 0.0826277
I0428 14:08:11.863700  9322 solver.cpp:237]     Train net output #0: loss = 0.0826276 (* 1 = 0.0826276 loss)
I0428 14:08:11.863710  9322 sgd_solver.cpp:105] Iteration 8076, lr = 0.00201949
I0428 14:08:17.203665  9322 solver.cpp:218] Iteration 8088 (2.2472 iter/s, 5.33999s/12 iters), loss = 0.108708
I0428 14:08:17.203714  9322 solver.cpp:237]     Train net output #0: loss = 0.108708 (* 1 = 0.108708 loss)
I0428 14:08:17.203724  9322 sgd_solver.cpp:105] Iteration 8088, lr = 0.00201469
I0428 14:08:18.716980  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:08:22.517697  9322 solver.cpp:218] Iteration 8100 (2.25819 iter/s, 5.314s/12 iters), loss = 0.0355746
I0428 14:08:22.517742  9322 solver.cpp:237]     Train net output #0: loss = 0.0355745 (* 1 = 0.0355745 loss)
I0428 14:08:22.517751  9322 sgd_solver.cpp:105] Iteration 8100, lr = 0.00200991
I0428 14:08:27.756765  9322 solver.cpp:218] Iteration 8112 (2.2905 iter/s, 5.23904s/12 iters), loss = 0.125979
I0428 14:08:27.756810  9322 solver.cpp:237]     Train net output #0: loss = 0.125979 (* 1 = 0.125979 loss)
I0428 14:08:27.756819  9322 sgd_solver.cpp:105] Iteration 8112, lr = 0.00200514
I0428 14:08:32.851756  9322 solver.cpp:218] Iteration 8124 (2.35527 iter/s, 5.09496s/12 iters), loss = 0.165503
I0428 14:08:32.851799  9322 solver.cpp:237]     Train net output #0: loss = 0.165503 (* 1 = 0.165503 loss)
I0428 14:08:32.851807  9322 sgd_solver.cpp:105] Iteration 8124, lr = 0.00200038
I0428 14:08:38.062384  9322 solver.cpp:218] Iteration 8136 (2.303 iter/s, 5.2106s/12 iters), loss = 0.117008
I0428 14:08:38.062429  9322 solver.cpp:237]     Train net output #0: loss = 0.117008 (* 1 = 0.117008 loss)
I0428 14:08:38.062438  9322 sgd_solver.cpp:105] Iteration 8136, lr = 0.00199563
I0428 14:08:43.259960  9322 solver.cpp:218] Iteration 8148 (2.30878 iter/s, 5.19755s/12 iters), loss = 0.0522845
I0428 14:08:43.260083  9322 solver.cpp:237]     Train net output #0: loss = 0.0522844 (* 1 = 0.0522844 loss)
I0428 14:08:43.260093  9322 sgd_solver.cpp:105] Iteration 8148, lr = 0.00199089
I0428 14:08:47.937827  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_8160.caffemodel
I0428 14:08:52.257742  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_8160.solverstate
I0428 14:08:57.700605  9322 solver.cpp:330] Iteration 8160, Testing net (#0)
I0428 14:08:57.700623  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:08:59.027653  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:09:02.672662  9322 solver.cpp:397]     Test net output #0: accuracy = 0.477941
I0428 14:09:02.672705  9322 solver.cpp:397]     Test net output #1: loss = 2.92253 (* 1 = 2.92253 loss)
I0428 14:09:02.790598  9322 solver.cpp:218] Iteration 8160 (0.61442 iter/s, 19.5306s/12 iters), loss = 0.0490991
I0428 14:09:02.790645  9322 solver.cpp:237]     Train net output #0: loss = 0.049099 (* 1 = 0.049099 loss)
I0428 14:09:02.790653  9322 sgd_solver.cpp:105] Iteration 8160, lr = 0.00198616
I0428 14:09:07.183818  9322 solver.cpp:218] Iteration 8172 (2.7315 iter/s, 4.39318s/12 iters), loss = 0.115315
I0428 14:09:07.183861  9322 solver.cpp:237]     Train net output #0: loss = 0.115315 (* 1 = 0.115315 loss)
I0428 14:09:07.183869  9322 sgd_solver.cpp:105] Iteration 8172, lr = 0.00198145
I0428 14:09:12.410184  9322 solver.cpp:218] Iteration 8184 (2.29606 iter/s, 5.22634s/12 iters), loss = 0.108867
I0428 14:09:12.410225  9322 solver.cpp:237]     Train net output #0: loss = 0.108866 (* 1 = 0.108866 loss)
I0428 14:09:12.410233  9322 sgd_solver.cpp:105] Iteration 8184, lr = 0.00197674
I0428 14:09:16.077438  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:09:17.598067  9322 solver.cpp:218] Iteration 8196 (2.3131 iter/s, 5.18785s/12 iters), loss = 0.161615
I0428 14:09:17.598114  9322 solver.cpp:237]     Train net output #0: loss = 0.161615 (* 1 = 0.161615 loss)
I0428 14:09:17.598122  9322 sgd_solver.cpp:105] Iteration 8196, lr = 0.00197205
I0428 14:09:22.537066  9322 solver.cpp:218] Iteration 8208 (2.42966 iter/s, 4.93897s/12 iters), loss = 0.0683085
I0428 14:09:22.537096  9322 solver.cpp:237]     Train net output #0: loss = 0.0683084 (* 1 = 0.0683084 loss)
I0428 14:09:22.537103  9322 sgd_solver.cpp:105] Iteration 8208, lr = 0.00196737
I0428 14:09:27.772483  9322 solver.cpp:218] Iteration 8220 (2.29209 iter/s, 5.2354s/12 iters), loss = 0.0534403
I0428 14:09:27.772524  9322 solver.cpp:237]     Train net output #0: loss = 0.0534402 (* 1 = 0.0534402 loss)
I0428 14:09:27.772533  9322 sgd_solver.cpp:105] Iteration 8220, lr = 0.0019627
I0428 14:09:32.915264  9322 solver.cpp:218] Iteration 8232 (2.33338 iter/s, 5.14275s/12 iters), loss = 0.0675903
I0428 14:09:32.915305  9322 solver.cpp:237]     Train net output #0: loss = 0.0675903 (* 1 = 0.0675903 loss)
I0428 14:09:32.915313  9322 sgd_solver.cpp:105] Iteration 8232, lr = 0.00195804
I0428 14:09:37.915796  9322 solver.cpp:218] Iteration 8244 (2.39976 iter/s, 5.0005s/12 iters), loss = 0.0901915
I0428 14:09:37.915848  9322 solver.cpp:237]     Train net output #0: loss = 0.0901914 (* 1 = 0.0901914 loss)
I0428 14:09:37.915858  9322 sgd_solver.cpp:105] Iteration 8244, lr = 0.00195339
I0428 14:09:43.130645  9322 solver.cpp:218] Iteration 8256 (2.30113 iter/s, 5.21482s/12 iters), loss = 0.0438474
I0428 14:09:43.130683  9322 solver.cpp:237]     Train net output #0: loss = 0.0438473 (* 1 = 0.0438473 loss)
I0428 14:09:43.130690  9322 sgd_solver.cpp:105] Iteration 8256, lr = 0.00194875
I0428 14:09:45.203934  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_8262.caffemodel
I0428 14:09:48.386618  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_8262.solverstate
I0428 14:09:53.261950  9322 solver.cpp:330] Iteration 8262, Testing net (#0)
I0428 14:09:53.261970  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:09:54.528872  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:09:58.178561  9322 solver.cpp:397]     Test net output #0: accuracy = 0.481618
I0428 14:09:58.178608  9322 solver.cpp:397]     Test net output #1: loss = 2.81113 (* 1 = 2.81113 loss)
I0428 14:10:00.093582  9322 solver.cpp:218] Iteration 8268 (0.707422 iter/s, 16.963s/12 iters), loss = 0.0635304
I0428 14:10:00.093626  9322 solver.cpp:237]     Train net output #0: loss = 0.0635303 (* 1 = 0.0635303 loss)
I0428 14:10:00.093634  9322 sgd_solver.cpp:105] Iteration 8268, lr = 0.00194412
I0428 14:10:05.246958  9322 solver.cpp:218] Iteration 8280 (2.32858 iter/s, 5.15335s/12 iters), loss = 0.0647979
I0428 14:10:05.246994  9322 solver.cpp:237]     Train net output #0: loss = 0.0647978 (* 1 = 0.0647978 loss)
I0428 14:10:05.247001  9322 sgd_solver.cpp:105] Iteration 8280, lr = 0.00193951
I0428 14:10:10.408849  9322 solver.cpp:218] Iteration 8292 (2.32474 iter/s, 5.16188s/12 iters), loss = 0.0738173
I0428 14:10:10.408890  9322 solver.cpp:237]     Train net output #0: loss = 0.0738172 (* 1 = 0.0738172 loss)
I0428 14:10:10.408898  9322 sgd_solver.cpp:105] Iteration 8292, lr = 0.0019349
I0428 14:10:11.100330  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:10:15.574767  9322 solver.cpp:218] Iteration 8304 (2.32293 iter/s, 5.16589s/12 iters), loss = 0.056891
I0428 14:10:15.574810  9322 solver.cpp:237]     Train net output #0: loss = 0.0568909 (* 1 = 0.0568909 loss)
I0428 14:10:15.574820  9322 sgd_solver.cpp:105] Iteration 8304, lr = 0.00193031
I0428 14:10:18.498319  9322 blocking_queue.cpp:49] Waiting for data
I0428 14:10:20.707455  9322 solver.cpp:218] Iteration 8316 (2.33797 iter/s, 5.13266s/12 iters), loss = 0.159033
I0428 14:10:20.707499  9322 solver.cpp:237]     Train net output #0: loss = 0.159033 (* 1 = 0.159033 loss)
I0428 14:10:20.707505  9322 sgd_solver.cpp:105] Iteration 8316, lr = 0.00192573
I0428 14:10:25.862455  9322 solver.cpp:218] Iteration 8328 (2.32785 iter/s, 5.15498s/12 iters), loss = 0.0554604
I0428 14:10:25.862495  9322 solver.cpp:237]     Train net output #0: loss = 0.0554603 (* 1 = 0.0554603 loss)
I0428 14:10:25.862502  9322 sgd_solver.cpp:105] Iteration 8328, lr = 0.00192115
I0428 14:10:31.033799  9322 solver.cpp:218] Iteration 8340 (2.32049 iter/s, 5.17132s/12 iters), loss = 0.0677746
I0428 14:10:31.033847  9322 solver.cpp:237]     Train net output #0: loss = 0.0677745 (* 1 = 0.0677745 loss)
I0428 14:10:31.033856  9322 sgd_solver.cpp:105] Iteration 8340, lr = 0.00191659
I0428 14:10:36.206274  9322 solver.cpp:218] Iteration 8352 (2.31999 iter/s, 5.17244s/12 iters), loss = 0.101048
I0428 14:10:36.206317  9322 solver.cpp:237]     Train net output #0: loss = 0.101048 (* 1 = 0.101048 loss)
I0428 14:10:36.206326  9322 sgd_solver.cpp:105] Iteration 8352, lr = 0.00191204
I0428 14:10:40.851480  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_8364.caffemodel
I0428 14:10:44.646556  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_8364.solverstate
I0428 14:10:48.148938  9322 solver.cpp:330] Iteration 8364, Testing net (#0)
I0428 14:10:48.148962  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:10:49.361460  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:10:53.077359  9322 solver.cpp:397]     Test net output #0: accuracy = 0.483456
I0428 14:10:53.077400  9322 solver.cpp:397]     Test net output #1: loss = 2.92426 (* 1 = 2.92426 loss)
I0428 14:10:53.195621  9322 solver.cpp:218] Iteration 8364 (0.706323 iter/s, 16.9894s/12 iters), loss = 0.0716327
I0428 14:10:53.195668  9322 solver.cpp:237]     Train net output #0: loss = 0.0716326 (* 1 = 0.0716326 loss)
I0428 14:10:53.195677  9322 sgd_solver.cpp:105] Iteration 8364, lr = 0.0019075
I0428 14:10:57.420867  9322 solver.cpp:218] Iteration 8376 (2.84009 iter/s, 4.22521s/12 iters), loss = 0.068327
I0428 14:10:57.420907  9322 solver.cpp:237]     Train net output #0: loss = 0.068327 (* 1 = 0.068327 loss)
I0428 14:10:57.420917  9322 sgd_solver.cpp:105] Iteration 8376, lr = 0.00190297
I0428 14:11:02.549453  9322 solver.cpp:218] Iteration 8388 (2.33984 iter/s, 5.12856s/12 iters), loss = 0.0485797
I0428 14:11:02.549499  9322 solver.cpp:237]     Train net output #0: loss = 0.0485796 (* 1 = 0.0485796 loss)
I0428 14:11:02.549507  9322 sgd_solver.cpp:105] Iteration 8388, lr = 0.00189846
I0428 14:11:05.387179  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:11:07.665027  9322 solver.cpp:218] Iteration 8400 (2.34579 iter/s, 5.11554s/12 iters), loss = 0.0932951
I0428 14:11:07.665073  9322 solver.cpp:237]     Train net output #0: loss = 0.093295 (* 1 = 0.093295 loss)
I0428 14:11:07.665081  9322 sgd_solver.cpp:105] Iteration 8400, lr = 0.00189395
I0428 14:11:12.855355  9322 solver.cpp:218] Iteration 8412 (2.31201 iter/s, 5.1903s/12 iters), loss = 0.121219
I0428 14:11:12.855402  9322 solver.cpp:237]     Train net output #0: loss = 0.121219 (* 1 = 0.121219 loss)
I0428 14:11:12.855410  9322 sgd_solver.cpp:105] Iteration 8412, lr = 0.00188945
I0428 14:11:18.029501  9322 solver.cpp:218] Iteration 8424 (2.31924 iter/s, 5.17411s/12 iters), loss = 0.0719255
I0428 14:11:18.029543  9322 solver.cpp:237]     Train net output #0: loss = 0.0719255 (* 1 = 0.0719255 loss)
I0428 14:11:18.029552  9322 sgd_solver.cpp:105] Iteration 8424, lr = 0.00188497
I0428 14:11:23.126135  9322 solver.cpp:218] Iteration 8436 (2.35451 iter/s, 5.09661s/12 iters), loss = 0.0504978
I0428 14:11:23.126260  9322 solver.cpp:237]     Train net output #0: loss = 0.0504978 (* 1 = 0.0504978 loss)
I0428 14:11:23.126269  9322 sgd_solver.cpp:105] Iteration 8436, lr = 0.00188049
I0428 14:11:28.296011  9322 solver.cpp:218] Iteration 8448 (2.32119 iter/s, 5.16977s/12 iters), loss = 0.0428308
I0428 14:11:28.296056  9322 solver.cpp:237]     Train net output #0: loss = 0.0428307 (* 1 = 0.0428307 loss)
I0428 14:11:28.296063  9322 sgd_solver.cpp:105] Iteration 8448, lr = 0.00187603
I0428 14:11:33.438944  9322 solver.cpp:218] Iteration 8460 (2.33331 iter/s, 5.14291s/12 iters), loss = 0.178241
I0428 14:11:33.438985  9322 solver.cpp:237]     Train net output #0: loss = 0.178241 (* 1 = 0.178241 loss)
I0428 14:11:33.438994  9322 sgd_solver.cpp:105] Iteration 8460, lr = 0.00187157
I0428 14:11:35.532166  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_8466.caffemodel
I0428 14:11:38.885886  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_8466.solverstate
I0428 14:11:41.368249  9322 solver.cpp:330] Iteration 8466, Testing net (#0)
I0428 14:11:41.368278  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:11:42.541371  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:11:46.278873  9322 solver.cpp:397]     Test net output #0: accuracy = 0.492034
I0428 14:11:46.278909  9322 solver.cpp:397]     Test net output #1: loss = 2.82636 (* 1 = 2.82636 loss)
I0428 14:11:48.176057  9322 solver.cpp:218] Iteration 8472 (0.814269 iter/s, 14.7371s/12 iters), loss = 0.0532152
I0428 14:11:48.176100  9322 solver.cpp:237]     Train net output #0: loss = 0.0532151 (* 1 = 0.0532151 loss)
I0428 14:11:48.176108  9322 sgd_solver.cpp:105] Iteration 8472, lr = 0.00186713
I0428 14:11:53.281257  9322 solver.cpp:218] Iteration 8484 (2.35056 iter/s, 5.10516s/12 iters), loss = 0.0866411
I0428 14:11:53.281376  9322 solver.cpp:237]     Train net output #0: loss = 0.086641 (* 1 = 0.086641 loss)
I0428 14:11:53.281389  9322 sgd_solver.cpp:105] Iteration 8484, lr = 0.0018627
I0428 14:11:58.452435  9322 solver.cpp:218] Iteration 8496 (2.3206 iter/s, 5.17108s/12 iters), loss = 0.0568322
I0428 14:11:58.452478  9322 solver.cpp:237]     Train net output #0: loss = 0.0568321 (* 1 = 0.0568321 loss)
I0428 14:11:58.452487  9322 sgd_solver.cpp:105] Iteration 8496, lr = 0.00185827
I0428 14:11:58.488850  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:12:03.546584  9322 solver.cpp:218] Iteration 8508 (2.35566 iter/s, 5.09412s/12 iters), loss = 0.0371993
I0428 14:12:03.546634  9322 solver.cpp:237]     Train net output #0: loss = 0.0371992 (* 1 = 0.0371992 loss)
I0428 14:12:03.546643  9322 sgd_solver.cpp:105] Iteration 8508, lr = 0.00185386
I0428 14:12:08.708225  9322 solver.cpp:218] Iteration 8520 (2.32486 iter/s, 5.1616s/12 iters), loss = 0.137523
I0428 14:12:08.708281  9322 solver.cpp:237]     Train net output #0: loss = 0.137523 (* 1 = 0.137523 loss)
I0428 14:12:08.708293  9322 sgd_solver.cpp:105] Iteration 8520, lr = 0.00184946
I0428 14:12:13.897964  9322 solver.cpp:218] Iteration 8532 (2.31227 iter/s, 5.1897s/12 iters), loss = 0.0869639
I0428 14:12:13.898006  9322 solver.cpp:237]     Train net output #0: loss = 0.0869638 (* 1 = 0.0869638 loss)
I0428 14:12:13.898015  9322 sgd_solver.cpp:105] Iteration 8532, lr = 0.00184507
I0428 14:12:19.290947  9322 solver.cpp:218] Iteration 8544 (2.22513 iter/s, 5.39295s/12 iters), loss = 0.0838145
I0428 14:12:19.290994  9322 solver.cpp:237]     Train net output #0: loss = 0.0838144 (* 1 = 0.0838144 loss)
I0428 14:12:19.291003  9322 sgd_solver.cpp:105] Iteration 8544, lr = 0.00184069
I0428 14:12:24.466059  9322 solver.cpp:218] Iteration 8556 (2.3188 iter/s, 5.17508s/12 iters), loss = 0.042955
I0428 14:12:24.466207  9322 solver.cpp:237]     Train net output #0: loss = 0.0429549 (* 1 = 0.0429549 loss)
I0428 14:12:24.466214  9322 sgd_solver.cpp:105] Iteration 8556, lr = 0.00183632
I0428 14:12:29.041749  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_8568.caffemodel
I0428 14:12:32.201193  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_8568.solverstate
I0428 14:12:34.618237  9322 solver.cpp:330] Iteration 8568, Testing net (#0)
I0428 14:12:34.618254  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:12:35.745978  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:12:39.511515  9322 solver.cpp:397]     Test net output #0: accuracy = 0.494485
I0428 14:12:39.511552  9322 solver.cpp:397]     Test net output #1: loss = 2.88103 (* 1 = 2.88103 loss)
I0428 14:12:39.630168  9322 solver.cpp:218] Iteration 8568 (0.791346 iter/s, 15.164s/12 iters), loss = 0.0390176
I0428 14:12:39.630214  9322 solver.cpp:237]     Train net output #0: loss = 0.0390175 (* 1 = 0.0390175 loss)
I0428 14:12:39.630223  9322 sgd_solver.cpp:105] Iteration 8568, lr = 0.00183196
I0428 14:12:43.925580  9322 solver.cpp:218] Iteration 8580 (2.7937 iter/s, 4.29537s/12 iters), loss = 0.0299818
I0428 14:12:43.925621  9322 solver.cpp:237]     Train net output #0: loss = 0.0299817 (* 1 = 0.0299817 loss)
I0428 14:12:43.925630  9322 sgd_solver.cpp:105] Iteration 8580, lr = 0.00182761
I0428 14:12:49.109329  9322 solver.cpp:218] Iteration 8592 (2.31494 iter/s, 5.18372s/12 iters), loss = 0.0845987
I0428 14:12:49.109369  9322 solver.cpp:237]     Train net output #0: loss = 0.0845986 (* 1 = 0.0845986 loss)
I0428 14:12:49.109377  9322 sgd_solver.cpp:105] Iteration 8592, lr = 0.00182327
I0428 14:12:51.359884  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:12:54.303735  9322 solver.cpp:218] Iteration 8604 (2.31019 iter/s, 5.19438s/12 iters), loss = 0.0208074
I0428 14:12:54.303781  9322 solver.cpp:237]     Train net output #0: loss = 0.0208073 (* 1 = 0.0208073 loss)
I0428 14:12:54.303788  9322 sgd_solver.cpp:105] Iteration 8604, lr = 0.00181894
I0428 14:12:59.442600  9322 solver.cpp:218] Iteration 8616 (2.33516 iter/s, 5.13883s/12 iters), loss = 0.0894209
I0428 14:12:59.442708  9322 solver.cpp:237]     Train net output #0: loss = 0.0894208 (* 1 = 0.0894208 loss)
I0428 14:12:59.442718  9322 sgd_solver.cpp:105] Iteration 8616, lr = 0.00181462
I0428 14:13:04.601755  9322 solver.cpp:218] Iteration 8628 (2.326 iter/s, 5.15906s/12 iters), loss = 0.088471
I0428 14:13:04.601802  9322 solver.cpp:237]     Train net output #0: loss = 0.0884709 (* 1 = 0.0884709 loss)
I0428 14:13:04.601811  9322 sgd_solver.cpp:105] Iteration 8628, lr = 0.00181031
I0428 14:13:09.797740  9322 solver.cpp:218] Iteration 8640 (2.30949 iter/s, 5.19595s/12 iters), loss = 0.0513843
I0428 14:13:09.797789  9322 solver.cpp:237]     Train net output #0: loss = 0.0513842 (* 1 = 0.0513842 loss)
I0428 14:13:09.797797  9322 sgd_solver.cpp:105] Iteration 8640, lr = 0.00180602
I0428 14:13:14.941890  9322 solver.cpp:218] Iteration 8652 (2.33276 iter/s, 5.14412s/12 iters), loss = 0.0911864
I0428 14:13:14.941936  9322 solver.cpp:237]     Train net output #0: loss = 0.0911863 (* 1 = 0.0911863 loss)
I0428 14:13:14.941943  9322 sgd_solver.cpp:105] Iteration 8652, lr = 0.00180173
I0428 14:13:20.032542  9322 solver.cpp:218] Iteration 8664 (2.35728 iter/s, 5.09061s/12 iters), loss = 0.0475032
I0428 14:13:20.032610  9322 solver.cpp:237]     Train net output #0: loss = 0.0475031 (* 1 = 0.0475031 loss)
I0428 14:13:20.032624  9322 sgd_solver.cpp:105] Iteration 8664, lr = 0.00179745
I0428 14:13:22.116165  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_8670.caffemodel
I0428 14:13:25.281335  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_8670.solverstate
I0428 14:13:27.694797  9322 solver.cpp:330] Iteration 8670, Testing net (#0)
I0428 14:13:27.694818  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:13:28.781903  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:13:32.612993  9322 solver.cpp:397]     Test net output #0: accuracy = 0.487132
I0428 14:13:32.613138  9322 solver.cpp:397]     Test net output #1: loss = 2.94321 (* 1 = 2.94321 loss)
I0428 14:13:34.528426  9322 solver.cpp:218] Iteration 8676 (0.82782 iter/s, 14.4959s/12 iters), loss = 0.0994468
I0428 14:13:34.528465  9322 solver.cpp:237]     Train net output #0: loss = 0.0994467 (* 1 = 0.0994467 loss)
I0428 14:13:34.528472  9322 sgd_solver.cpp:105] Iteration 8676, lr = 0.00179318
I0428 14:13:39.712210  9322 solver.cpp:218] Iteration 8688 (2.31492 iter/s, 5.18376s/12 iters), loss = 0.107761
I0428 14:13:39.712256  9322 solver.cpp:237]     Train net output #0: loss = 0.107761 (* 1 = 0.107761 loss)
I0428 14:13:39.712265  9322 sgd_solver.cpp:105] Iteration 8688, lr = 0.00178893
I0428 14:13:44.330305  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:13:45.081634  9322 solver.cpp:218] Iteration 8700 (2.23489 iter/s, 5.36939s/12 iters), loss = 0.0963803
I0428 14:13:45.081674  9322 solver.cpp:237]     Train net output #0: loss = 0.0963802 (* 1 = 0.0963802 loss)
I0428 14:13:45.081682  9322 sgd_solver.cpp:105] Iteration 8700, lr = 0.00178468
I0428 14:13:50.321038  9322 solver.cpp:218] Iteration 8712 (2.29035 iter/s, 5.23938s/12 iters), loss = 0.078651
I0428 14:13:50.321080  9322 solver.cpp:237]     Train net output #0: loss = 0.0786509 (* 1 = 0.0786509 loss)
I0428 14:13:50.321089  9322 sgd_solver.cpp:105] Iteration 8712, lr = 0.00178044
I0428 14:13:55.461076  9322 solver.cpp:218] Iteration 8724 (2.33463 iter/s, 5.14s/12 iters), loss = 0.0967749
I0428 14:13:55.461143  9322 solver.cpp:237]     Train net output #0: loss = 0.0967748 (* 1 = 0.0967748 loss)
I0428 14:13:55.461156  9322 sgd_solver.cpp:105] Iteration 8724, lr = 0.00177621
I0428 14:14:00.607127  9322 solver.cpp:218] Iteration 8736 (2.33191 iter/s, 5.146s/12 iters), loss = 0.11064
I0428 14:14:00.607175  9322 solver.cpp:237]     Train net output #0: loss = 0.11064 (* 1 = 0.11064 loss)
I0428 14:14:00.607183  9322 sgd_solver.cpp:105] Iteration 8736, lr = 0.001772
I0428 14:14:05.779505  9322 solver.cpp:218] Iteration 8748 (2.32003 iter/s, 5.17235s/12 iters), loss = 0.0446297
I0428 14:14:05.779623  9322 solver.cpp:237]     Train net output #0: loss = 0.0446296 (* 1 = 0.0446296 loss)
I0428 14:14:05.779633  9322 sgd_solver.cpp:105] Iteration 8748, lr = 0.00176779
I0428 14:14:10.972600  9322 solver.cpp:218] Iteration 8760 (2.31081 iter/s, 5.19298s/12 iters), loss = 0.0845893
I0428 14:14:10.972666  9322 solver.cpp:237]     Train net output #0: loss = 0.0845892 (* 1 = 0.0845892 loss)
I0428 14:14:10.972679  9322 sgd_solver.cpp:105] Iteration 8760, lr = 0.00176359
I0428 14:14:15.703177  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_8772.caffemodel
I0428 14:14:21.598486  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_8772.solverstate
I0428 14:14:24.686961  9322 solver.cpp:330] Iteration 8772, Testing net (#0)
I0428 14:14:24.686981  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:14:25.715227  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:14:29.615554  9322 solver.cpp:397]     Test net output #0: accuracy = 0.49326
I0428 14:14:29.615600  9322 solver.cpp:397]     Test net output #1: loss = 2.8836 (* 1 = 2.8836 loss)
I0428 14:14:29.733379  9322 solver.cpp:218] Iteration 8772 (0.639631 iter/s, 18.7608s/12 iters), loss = 0.10354
I0428 14:14:29.733428  9322 solver.cpp:237]     Train net output #0: loss = 0.10354 (* 1 = 0.10354 loss)
I0428 14:14:29.733435  9322 sgd_solver.cpp:105] Iteration 8772, lr = 0.00175941
I0428 14:14:34.045900  9322 solver.cpp:218] Iteration 8784 (2.78262 iter/s, 4.31248s/12 iters), loss = 0.10974
I0428 14:14:34.045944  9322 solver.cpp:237]     Train net output #0: loss = 0.109739 (* 1 = 0.109739 loss)
I0428 14:14:34.045953  9322 sgd_solver.cpp:105] Iteration 8784, lr = 0.00175523
I0428 14:14:39.159629  9322 solver.cpp:218] Iteration 8796 (2.34664 iter/s, 5.1137s/12 iters), loss = 0.0752509
I0428 14:14:39.159750  9322 solver.cpp:237]     Train net output #0: loss = 0.0752508 (* 1 = 0.0752508 loss)
I0428 14:14:39.159760  9322 sgd_solver.cpp:105] Iteration 8796, lr = 0.00175106
I0428 14:14:40.587546  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:14:44.279955  9322 solver.cpp:218] Iteration 8808 (2.34365 iter/s, 5.12022s/12 iters), loss = 0.0573072
I0428 14:14:44.280000  9322 solver.cpp:237]     Train net output #0: loss = 0.0573071 (* 1 = 0.0573071 loss)
I0428 14:14:44.280009  9322 sgd_solver.cpp:105] Iteration 8808, lr = 0.0017469
I0428 14:14:49.442102  9322 solver.cpp:218] Iteration 8820 (2.32463 iter/s, 5.16211s/12 iters), loss = 0.16776
I0428 14:14:49.442149  9322 solver.cpp:237]     Train net output #0: loss = 0.16776 (* 1 = 0.16776 loss)
I0428 14:14:49.442157  9322 sgd_solver.cpp:105] Iteration 8820, lr = 0.00174276
I0428 14:14:54.606403  9322 solver.cpp:218] Iteration 8832 (2.32366 iter/s, 5.16427s/12 iters), loss = 0.0719597
I0428 14:14:54.606449  9322 solver.cpp:237]     Train net output #0: loss = 0.0719596 (* 1 = 0.0719596 loss)
I0428 14:14:54.606457  9322 sgd_solver.cpp:105] Iteration 8832, lr = 0.00173862
I0428 14:14:59.815639  9322 solver.cpp:218] Iteration 8844 (2.30361 iter/s, 5.20921s/12 iters), loss = 0.0935177
I0428 14:14:59.815681  9322 solver.cpp:237]     Train net output #0: loss = 0.0935176 (* 1 = 0.0935176 loss)
I0428 14:14:59.815690  9322 sgd_solver.cpp:105] Iteration 8844, lr = 0.00173449
I0428 14:15:04.967474  9322 solver.cpp:218] Iteration 8856 (2.32928 iter/s, 5.15181s/12 iters), loss = 0.0932437
I0428 14:15:04.967519  9322 solver.cpp:237]     Train net output #0: loss = 0.0932436 (* 1 = 0.0932436 loss)
I0428 14:15:04.967527  9322 sgd_solver.cpp:105] Iteration 8856, lr = 0.00173037
I0428 14:15:10.222645  9322 solver.cpp:218] Iteration 8868 (2.28348 iter/s, 5.25514s/12 iters), loss = 0.0105136
I0428 14:15:10.222743  9322 solver.cpp:237]     Train net output #0: loss = 0.0105135 (* 1 = 0.0105135 loss)
I0428 14:15:10.222754  9322 sgd_solver.cpp:105] Iteration 8868, lr = 0.00172626
I0428 14:15:12.335999  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_8874.caffemodel
I0428 14:15:16.727610  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_8874.solverstate
I0428 14:15:21.021739  9322 solver.cpp:330] Iteration 8874, Testing net (#0)
I0428 14:15:21.021764  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:15:22.033361  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:15:26.006541  9322 solver.cpp:397]     Test net output #0: accuracy = 0.500613
I0428 14:15:26.006588  9322 solver.cpp:397]     Test net output #1: loss = 2.85579 (* 1 = 2.85579 loss)
I0428 14:15:27.922961  9322 solver.cpp:218] Iteration 8880 (0.677954 iter/s, 17.7003s/12 iters), loss = 0.0914868
I0428 14:15:27.923008  9322 solver.cpp:237]     Train net output #0: loss = 0.0914867 (* 1 = 0.0914867 loss)
I0428 14:15:27.923017  9322 sgd_solver.cpp:105] Iteration 8880, lr = 0.00172217
I0428 14:15:33.064148  9322 solver.cpp:218] Iteration 8892 (2.33411 iter/s, 5.14116s/12 iters), loss = 0.0892676
I0428 14:15:33.064189  9322 solver.cpp:237]     Train net output #0: loss = 0.0892675 (* 1 = 0.0892675 loss)
I0428 14:15:33.064198  9322 sgd_solver.cpp:105] Iteration 8892, lr = 0.00171808
I0428 14:15:36.804780  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:15:38.331526  9322 solver.cpp:218] Iteration 8904 (2.27818 iter/s, 5.26735s/12 iters), loss = 0.124452
I0428 14:15:38.331566  9322 solver.cpp:237]     Train net output #0: loss = 0.124452 (* 1 = 0.124452 loss)
I0428 14:15:38.331574  9322 sgd_solver.cpp:105] Iteration 8904, lr = 0.001714
I0428 14:15:43.644461  9322 solver.cpp:218] Iteration 8916 (2.25865 iter/s, 5.3129s/12 iters), loss = 0.0909944
I0428 14:15:43.644610  9322 solver.cpp:237]     Train net output #0: loss = 0.0909943 (* 1 = 0.0909943 loss)
I0428 14:15:43.644620  9322 sgd_solver.cpp:105] Iteration 8916, lr = 0.00170993
I0428 14:15:48.919910  9322 solver.cpp:218] Iteration 8928 (2.27475 iter/s, 5.27531s/12 iters), loss = 0.0777533
I0428 14:15:48.919955  9322 solver.cpp:237]     Train net output #0: loss = 0.0777532 (* 1 = 0.0777532 loss)
I0428 14:15:48.919965  9322 sgd_solver.cpp:105] Iteration 8928, lr = 0.00170587
I0428 14:15:54.002445  9322 solver.cpp:218] Iteration 8940 (2.36104 iter/s, 5.08251s/12 iters), loss = 0.109278
I0428 14:15:54.002485  9322 solver.cpp:237]     Train net output #0: loss = 0.109278 (* 1 = 0.109278 loss)
I0428 14:15:54.002494  9322 sgd_solver.cpp:105] Iteration 8940, lr = 0.00170182
I0428 14:15:59.169520  9322 solver.cpp:218] Iteration 8952 (2.32241 iter/s, 5.16705s/12 iters), loss = 0.0775464
I0428 14:15:59.169564  9322 solver.cpp:237]     Train net output #0: loss = 0.0775463 (* 1 = 0.0775463 loss)
I0428 14:15:59.169571  9322 sgd_solver.cpp:105] Iteration 8952, lr = 0.00169778
I0428 14:16:04.329924  9322 solver.cpp:218] Iteration 8964 (2.32542 iter/s, 5.16036s/12 iters), loss = 0.0397835
I0428 14:16:04.329995  9322 solver.cpp:237]     Train net output #0: loss = 0.0397834 (* 1 = 0.0397834 loss)
I0428 14:16:04.330008  9322 sgd_solver.cpp:105] Iteration 8964, lr = 0.00169375
I0428 14:16:08.998095  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_8976.caffemodel
I0428 14:16:12.616098  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_8976.solverstate
I0428 14:16:16.300159  9322 solver.cpp:330] Iteration 8976, Testing net (#0)
I0428 14:16:16.300256  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:16:17.209735  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:16:21.164192  9322 solver.cpp:397]     Test net output #0: accuracy = 0.485294
I0428 14:16:21.164227  9322 solver.cpp:397]     Test net output #1: loss = 2.89935 (* 1 = 2.89935 loss)
I0428 14:16:21.282027  9322 solver.cpp:218] Iteration 8976 (0.707876 iter/s, 16.9521s/12 iters), loss = 0.0217791
I0428 14:16:21.282073  9322 solver.cpp:237]     Train net output #0: loss = 0.021779 (* 1 = 0.021779 loss)
I0428 14:16:21.282081  9322 sgd_solver.cpp:105] Iteration 8976, lr = 0.00168973
I0428 14:16:25.638521  9322 solver.cpp:218] Iteration 8988 (2.75453 iter/s, 4.35646s/12 iters), loss = 0.0557435
I0428 14:16:25.638563  9322 solver.cpp:237]     Train net output #0: loss = 0.0557434 (* 1 = 0.0557434 loss)
I0428 14:16:25.638571  9322 sgd_solver.cpp:105] Iteration 8988, lr = 0.00168571
I0428 14:16:28.956351  9322 blocking_queue.cpp:49] Waiting for data
I0428 14:16:30.763696  9322 solver.cpp:218] Iteration 9000 (2.3414 iter/s, 5.12514s/12 iters), loss = 0.0868037
I0428 14:16:30.763746  9322 solver.cpp:237]     Train net output #0: loss = 0.0868035 (* 1 = 0.0868035 loss)
I0428 14:16:30.763756  9322 sgd_solver.cpp:105] Iteration 9000, lr = 0.00168171
I0428 14:16:31.525399  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:16:36.202807  9322 solver.cpp:218] Iteration 9012 (2.20626 iter/s, 5.43908s/12 iters), loss = 0.038416
I0428 14:16:36.202848  9322 solver.cpp:237]     Train net output #0: loss = 0.0384159 (* 1 = 0.0384159 loss)
I0428 14:16:36.202857  9322 sgd_solver.cpp:105] Iteration 9012, lr = 0.00167772
I0428 14:16:41.528887  9322 solver.cpp:218] Iteration 9024 (2.25308 iter/s, 5.32605s/12 iters), loss = 0.201891
I0428 14:16:41.528928  9322 solver.cpp:237]     Train net output #0: loss = 0.201891 (* 1 = 0.201891 loss)
I0428 14:16:41.528937  9322 sgd_solver.cpp:105] Iteration 9024, lr = 0.00167374
I0428 14:16:46.770155  9322 solver.cpp:218] Iteration 9036 (2.28954 iter/s, 5.24124s/12 iters), loss = 0.127901
I0428 14:16:46.770326  9322 solver.cpp:237]     Train net output #0: loss = 0.127901 (* 1 = 0.127901 loss)
I0428 14:16:46.770336  9322 sgd_solver.cpp:105] Iteration 9036, lr = 0.00166976
I0428 14:16:51.927251  9322 solver.cpp:218] Iteration 9048 (2.32696 iter/s, 5.15694s/12 iters), loss = 0.0357397
I0428 14:16:51.927296  9322 solver.cpp:237]     Train net output #0: loss = 0.0357395 (* 1 = 0.0357395 loss)
I0428 14:16:51.927304  9322 sgd_solver.cpp:105] Iteration 9048, lr = 0.0016658
I0428 14:16:57.126065  9322 solver.cpp:218] Iteration 9060 (2.30823 iter/s, 5.19878s/12 iters), loss = 0.077581
I0428 14:16:57.126112  9322 solver.cpp:237]     Train net output #0: loss = 0.0775808 (* 1 = 0.0775808 loss)
I0428 14:16:57.126121  9322 sgd_solver.cpp:105] Iteration 9060, lr = 0.00166184
I0428 14:17:02.287685  9322 solver.cpp:218] Iteration 9072 (2.32487 iter/s, 5.16158s/12 iters), loss = 0.022676
I0428 14:17:02.287735  9322 solver.cpp:237]     Train net output #0: loss = 0.0226759 (* 1 = 0.0226759 loss)
I0428 14:17:02.287744  9322 sgd_solver.cpp:105] Iteration 9072, lr = 0.0016579
I0428 14:17:04.366905  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_9078.caffemodel
I0428 14:17:09.210552  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_9078.solverstate
I0428 14:17:11.673676  9322 solver.cpp:330] Iteration 9078, Testing net (#0)
I0428 14:17:11.673697  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:17:12.600219  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:17:16.619661  9322 solver.cpp:397]     Test net output #0: accuracy = 0.492647
I0428 14:17:16.619710  9322 solver.cpp:397]     Test net output #1: loss = 2.86276 (* 1 = 2.86276 loss)
I0428 14:17:18.521152  9322 solver.cpp:218] Iteration 9084 (0.739212 iter/s, 16.2335s/12 iters), loss = 0.0434509
I0428 14:17:18.521265  9322 solver.cpp:237]     Train net output #0: loss = 0.0434508 (* 1 = 0.0434508 loss)
I0428 14:17:18.521273  9322 sgd_solver.cpp:105] Iteration 9084, lr = 0.00165396
I0428 14:17:23.690965  9322 solver.cpp:218] Iteration 9096 (2.32121 iter/s, 5.16971s/12 iters), loss = 0.00730822
I0428 14:17:23.691010  9322 solver.cpp:237]     Train net output #0: loss = 0.00730809 (* 1 = 0.00730809 loss)
I0428 14:17:23.691020  9322 sgd_solver.cpp:105] Iteration 9096, lr = 0.00165003
I0428 14:17:26.770303  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:17:28.926272  9322 solver.cpp:218] Iteration 9108 (2.29215 iter/s, 5.23527s/12 iters), loss = 0.0388377
I0428 14:17:28.926321  9322 solver.cpp:237]     Train net output #0: loss = 0.0388376 (* 1 = 0.0388376 loss)
I0428 14:17:28.926328  9322 sgd_solver.cpp:105] Iteration 9108, lr = 0.00164612
I0428 14:17:34.069331  9322 solver.cpp:218] Iteration 9120 (2.33326 iter/s, 5.14303s/12 iters), loss = 0.0354728
I0428 14:17:34.069367  9322 solver.cpp:237]     Train net output #0: loss = 0.0354726 (* 1 = 0.0354726 loss)
I0428 14:17:34.069375  9322 sgd_solver.cpp:105] Iteration 9120, lr = 0.00164221
I0428 14:17:39.245906  9322 solver.cpp:218] Iteration 9132 (2.31815 iter/s, 5.17655s/12 iters), loss = 0.0474433
I0428 14:17:39.245947  9322 solver.cpp:237]     Train net output #0: loss = 0.0474432 (* 1 = 0.0474432 loss)
I0428 14:17:39.245955  9322 sgd_solver.cpp:105] Iteration 9132, lr = 0.00163831
I0428 14:17:44.406510  9322 solver.cpp:218] Iteration 9144 (2.32532 iter/s, 5.16058s/12 iters), loss = 0.0473319
I0428 14:17:44.406556  9322 solver.cpp:237]     Train net output #0: loss = 0.0473318 (* 1 = 0.0473318 loss)
I0428 14:17:44.406564  9322 sgd_solver.cpp:105] Iteration 9144, lr = 0.00163442
I0428 14:17:49.525701  9322 solver.cpp:218] Iteration 9156 (2.34414 iter/s, 5.11915s/12 iters), loss = 0.112976
I0428 14:17:49.525838  9322 solver.cpp:237]     Train net output #0: loss = 0.112976 (* 1 = 0.112976 loss)
I0428 14:17:49.525851  9322 sgd_solver.cpp:105] Iteration 9156, lr = 0.00163054
I0428 14:17:54.831880  9322 solver.cpp:218] Iteration 9168 (2.26156 iter/s, 5.30606s/12 iters), loss = 0.0994693
I0428 14:17:54.831930  9322 solver.cpp:237]     Train net output #0: loss = 0.0994692 (* 1 = 0.0994692 loss)
I0428 14:17:54.831939  9322 sgd_solver.cpp:105] Iteration 9168, lr = 0.00162667
I0428 14:17:59.563297  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_9180.caffemodel
I0428 14:18:03.714579  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_9180.solverstate
I0428 14:18:07.093443  9322 solver.cpp:330] Iteration 9180, Testing net (#0)
I0428 14:18:07.093462  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:18:07.982283  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:18:12.051054  9322 solver.cpp:397]     Test net output #0: accuracy = 0.485294
I0428 14:18:12.051096  9322 solver.cpp:397]     Test net output #1: loss = 2.89713 (* 1 = 2.89713 loss)
I0428 14:18:12.169224  9322 solver.cpp:218] Iteration 9180 (0.692146 iter/s, 17.3374s/12 iters), loss = 0.131615
I0428 14:18:12.169293  9322 solver.cpp:237]     Train net output #0: loss = 0.131614 (* 1 = 0.131614 loss)
I0428 14:18:12.169304  9322 sgd_solver.cpp:105] Iteration 9180, lr = 0.00162281
I0428 14:18:16.508378  9322 solver.cpp:218] Iteration 9192 (2.76556 iter/s, 4.33909s/12 iters), loss = 0.0981713
I0428 14:18:16.508436  9322 solver.cpp:237]     Train net output #0: loss = 0.0981712 (* 1 = 0.0981712 loss)
I0428 14:18:16.508450  9322 sgd_solver.cpp:105] Iteration 9192, lr = 0.00161895
I0428 14:18:21.737263  9322 solver.cpp:218] Iteration 9204 (2.29496 iter/s, 5.22884s/12 iters), loss = 0.0540963
I0428 14:18:21.737435  9322 solver.cpp:237]     Train net output #0: loss = 0.0540962 (* 1 = 0.0540962 loss)
I0428 14:18:21.737445  9322 sgd_solver.cpp:105] Iteration 9204, lr = 0.00161511
I0428 14:18:21.804399  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:18:26.911646  9322 solver.cpp:218] Iteration 9216 (2.31919 iter/s, 5.17423s/12 iters), loss = 0.0668708
I0428 14:18:26.911687  9322 solver.cpp:237]     Train net output #0: loss = 0.0668707 (* 1 = 0.0668707 loss)
I0428 14:18:26.911695  9322 sgd_solver.cpp:105] Iteration 9216, lr = 0.00161128
I0428 14:18:32.089442  9322 solver.cpp:218] Iteration 9228 (2.3176 iter/s, 5.17776s/12 iters), loss = 0.044075
I0428 14:18:32.089488  9322 solver.cpp:237]     Train net output #0: loss = 0.0440749 (* 1 = 0.0440749 loss)
I0428 14:18:32.089496  9322 sgd_solver.cpp:105] Iteration 9228, lr = 0.00160745
I0428 14:18:37.262743  9322 solver.cpp:218] Iteration 9240 (2.31961 iter/s, 5.17328s/12 iters), loss = 0.0941568
I0428 14:18:37.262778  9322 solver.cpp:237]     Train net output #0: loss = 0.0941567 (* 1 = 0.0941567 loss)
I0428 14:18:37.262786  9322 sgd_solver.cpp:105] Iteration 9240, lr = 0.00160363
I0428 14:18:42.467448  9322 solver.cpp:218] Iteration 9252 (2.30562 iter/s, 5.20468s/12 iters), loss = 0.0370444
I0428 14:18:42.467487  9322 solver.cpp:237]     Train net output #0: loss = 0.0370442 (* 1 = 0.0370442 loss)
I0428 14:18:42.467495  9322 sgd_solver.cpp:105] Iteration 9252, lr = 0.00159983
I0428 14:18:47.646488  9322 solver.cpp:218] Iteration 9264 (2.31704 iter/s, 5.17901s/12 iters), loss = 0.07367
I0428 14:18:47.646533  9322 solver.cpp:237]     Train net output #0: loss = 0.0736698 (* 1 = 0.0736698 loss)
I0428 14:18:47.646540  9322 sgd_solver.cpp:105] Iteration 9264, lr = 0.00159603
I0428 14:18:52.780472  9322 solver.cpp:218] Iteration 9276 (2.33738 iter/s, 5.13395s/12 iters), loss = 0.124997
I0428 14:18:52.780601  9322 solver.cpp:237]     Train net output #0: loss = 0.124997 (* 1 = 0.124997 loss)
I0428 14:18:52.780611  9322 sgd_solver.cpp:105] Iteration 9276, lr = 0.00159224
I0428 14:18:54.848098  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_9282.caffemodel
I0428 14:18:58.015461  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_9282.solverstate
I0428 14:19:00.464259  9322 solver.cpp:330] Iteration 9282, Testing net (#0)
I0428 14:19:00.464278  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:19:01.271991  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:19:05.368450  9322 solver.cpp:397]     Test net output #0: accuracy = 0.489583
I0428 14:19:05.368495  9322 solver.cpp:397]     Test net output #1: loss = 2.91412 (* 1 = 2.91412 loss)
I0428 14:19:07.393209  9322 solver.cpp:218] Iteration 9288 (0.821205 iter/s, 14.6127s/12 iters), loss = 0.0753415
I0428 14:19:07.393252  9322 solver.cpp:237]     Train net output #0: loss = 0.0753414 (* 1 = 0.0753414 loss)
I0428 14:19:07.393260  9322 sgd_solver.cpp:105] Iteration 9288, lr = 0.00158846
I0428 14:19:12.607339  9322 solver.cpp:218] Iteration 9300 (2.30145 iter/s, 5.2141s/12 iters), loss = 0.129334
I0428 14:19:12.607383  9322 solver.cpp:237]     Train net output #0: loss = 0.129333 (* 1 = 0.129333 loss)
I0428 14:19:12.607391  9322 sgd_solver.cpp:105] Iteration 9300, lr = 0.00158469
I0428 14:19:14.884111  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:19:17.786975  9322 solver.cpp:218] Iteration 9312 (2.31678 iter/s, 5.1796s/12 iters), loss = 0.114894
I0428 14:19:17.787020  9322 solver.cpp:237]     Train net output #0: loss = 0.114893 (* 1 = 0.114893 loss)
I0428 14:19:17.787029  9322 sgd_solver.cpp:105] Iteration 9312, lr = 0.00158092
I0428 14:19:22.810844  9322 solver.cpp:218] Iteration 9324 (2.38862 iter/s, 5.02383s/12 iters), loss = 0.0279576
I0428 14:19:22.810957  9322 solver.cpp:237]     Train net output #0: loss = 0.0279575 (* 1 = 0.0279575 loss)
I0428 14:19:22.810967  9322 sgd_solver.cpp:105] Iteration 9324, lr = 0.00157717
I0428 14:19:27.985560  9322 solver.cpp:218] Iteration 9336 (2.31901 iter/s, 5.17462s/12 iters), loss = 0.0524601
I0428 14:19:27.985605  9322 solver.cpp:237]     Train net output #0: loss = 0.05246 (* 1 = 0.05246 loss)
I0428 14:19:27.985613  9322 sgd_solver.cpp:105] Iteration 9336, lr = 0.00157343
I0428 14:19:33.095863  9322 solver.cpp:218] Iteration 9348 (2.34821 iter/s, 5.11027s/12 iters), loss = 0.091456
I0428 14:19:33.095913  9322 solver.cpp:237]     Train net output #0: loss = 0.0914559 (* 1 = 0.0914559 loss)
I0428 14:19:33.095921  9322 sgd_solver.cpp:105] Iteration 9348, lr = 0.00156969
I0428 14:19:38.365010  9322 solver.cpp:218] Iteration 9360 (2.27742 iter/s, 5.26912s/12 iters), loss = 0.0402687
I0428 14:19:38.365044  9322 solver.cpp:237]     Train net output #0: loss = 0.0402686 (* 1 = 0.0402686 loss)
I0428 14:19:38.365052  9322 sgd_solver.cpp:105] Iteration 9360, lr = 0.00156596
I0428 14:19:43.467007  9322 solver.cpp:218] Iteration 9372 (2.35203 iter/s, 5.10197s/12 iters), loss = 0.0337742
I0428 14:19:43.467051  9322 solver.cpp:237]     Train net output #0: loss = 0.0337741 (* 1 = 0.0337741 loss)
I0428 14:19:43.467058  9322 sgd_solver.cpp:105] Iteration 9372, lr = 0.00156225
I0428 14:19:48.129366  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_9384.caffemodel
I0428 14:19:51.309526  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_9384.solverstate
I0428 14:19:53.997661  9322 solver.cpp:330] Iteration 9384, Testing net (#0)
I0428 14:19:53.997725  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:19:54.763329  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:19:58.739663  9322 solver.cpp:397]     Test net output #0: accuracy = 0.499387
I0428 14:19:58.739697  9322 solver.cpp:397]     Test net output #1: loss = 2.94045 (* 1 = 2.94045 loss)
I0428 14:19:58.857056  9322 solver.cpp:218] Iteration 9384 (0.779723 iter/s, 15.3901s/12 iters), loss = 0.0249061
I0428 14:19:58.857105  9322 solver.cpp:237]     Train net output #0: loss = 0.024906 (* 1 = 0.024906 loss)
I0428 14:19:58.857115  9322 sgd_solver.cpp:105] Iteration 9384, lr = 0.00155854
I0428 14:20:03.101016  9322 solver.cpp:218] Iteration 9396 (2.82758 iter/s, 4.24391s/12 iters), loss = 0.102591
I0428 14:20:03.101063  9322 solver.cpp:237]     Train net output #0: loss = 0.102591 (* 1 = 0.102591 loss)
I0428 14:20:03.101071  9322 sgd_solver.cpp:105] Iteration 9396, lr = 0.00155484
I0428 14:20:07.600932  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:20:08.291375  9322 solver.cpp:218] Iteration 9408 (2.31199 iter/s, 5.19033s/12 iters), loss = 0.0564275
I0428 14:20:08.291420  9322 solver.cpp:237]     Train net output #0: loss = 0.0564274 (* 1 = 0.0564274 loss)
I0428 14:20:08.291429  9322 sgd_solver.cpp:105] Iteration 9408, lr = 0.00155114
I0428 14:20:13.460364  9322 solver.cpp:218] Iteration 9420 (2.32155 iter/s, 5.16896s/12 iters), loss = 0.0423121
I0428 14:20:13.460403  9322 solver.cpp:237]     Train net output #0: loss = 0.042312 (* 1 = 0.042312 loss)
I0428 14:20:13.460412  9322 sgd_solver.cpp:105] Iteration 9420, lr = 0.00154746
I0428 14:20:18.612589  9322 solver.cpp:218] Iteration 9432 (2.3291 iter/s, 5.1522s/12 iters), loss = 0.0665286
I0428 14:20:18.612637  9322 solver.cpp:237]     Train net output #0: loss = 0.0665285 (* 1 = 0.0665285 loss)
I0428 14:20:18.612645  9322 sgd_solver.cpp:105] Iteration 9432, lr = 0.00154379
I0428 14:20:23.775180  9322 solver.cpp:218] Iteration 9444 (2.32443 iter/s, 5.16255s/12 iters), loss = 0.0545826
I0428 14:20:23.775229  9322 solver.cpp:237]     Train net output #0: loss = 0.0545825 (* 1 = 0.0545825 loss)
I0428 14:20:23.775238  9322 sgd_solver.cpp:105] Iteration 9444, lr = 0.00154012
I0428 14:20:29.002130  9322 solver.cpp:218] Iteration 9456 (2.29581 iter/s, 5.22691s/12 iters), loss = 0.0638557
I0428 14:20:29.002288  9322 solver.cpp:237]     Train net output #0: loss = 0.0638556 (* 1 = 0.0638556 loss)
I0428 14:20:29.002298  9322 sgd_solver.cpp:105] Iteration 9456, lr = 0.00153647
I0428 14:20:34.167451  9322 solver.cpp:218] Iteration 9468 (2.32325 iter/s, 5.16519s/12 iters), loss = 0.0203877
I0428 14:20:34.167485  9322 solver.cpp:237]     Train net output #0: loss = 0.0203876 (* 1 = 0.0203876 loss)
I0428 14:20:34.167493  9322 sgd_solver.cpp:105] Iteration 9468, lr = 0.00153282
I0428 14:20:39.315393  9322 solver.cpp:218] Iteration 9480 (2.33104 iter/s, 5.14792s/12 iters), loss = 0.11946
I0428 14:20:39.315428  9322 solver.cpp:237]     Train net output #0: loss = 0.11946 (* 1 = 0.11946 loss)
I0428 14:20:39.315436  9322 sgd_solver.cpp:105] Iteration 9480, lr = 0.00152918
I0428 14:20:41.382989  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_9486.caffemodel
I0428 14:20:44.536306  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_9486.solverstate
I0428 14:20:46.989606  9322 solver.cpp:330] Iteration 9486, Testing net (#0)
I0428 14:20:46.989627  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:20:47.721311  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:20:51.783021  9322 solver.cpp:397]     Test net output #0: accuracy = 0.492647
I0428 14:20:51.783068  9322 solver.cpp:397]     Test net output #1: loss = 2.9372 (* 1 = 2.9372 loss)
I0428 14:20:53.701787  9322 solver.cpp:218] Iteration 9492 (0.83412 iter/s, 14.3864s/12 iters), loss = 0.0646408
I0428 14:20:53.701838  9322 solver.cpp:237]     Train net output #0: loss = 0.0646407 (* 1 = 0.0646407 loss)
I0428 14:20:53.701846  9322 sgd_solver.cpp:105] Iteration 9492, lr = 0.00152555
I0428 14:20:58.857846  9322 solver.cpp:218] Iteration 9504 (2.32738 iter/s, 5.15602s/12 iters), loss = 0.0394379
I0428 14:20:58.857889  9322 solver.cpp:237]     Train net output #0: loss = 0.0394377 (* 1 = 0.0394377 loss)
I0428 14:20:58.857897  9322 sgd_solver.cpp:105] Iteration 9504, lr = 0.00152193
I0428 14:21:00.369010  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:21:04.042263  9322 solver.cpp:218] Iteration 9516 (2.31464 iter/s, 5.18439s/12 iters), loss = 0.119343
I0428 14:21:04.042306  9322 solver.cpp:237]     Train net output #0: loss = 0.119343 (* 1 = 0.119343 loss)
I0428 14:21:04.042315  9322 sgd_solver.cpp:105] Iteration 9516, lr = 0.00151831
I0428 14:21:09.197988  9322 solver.cpp:218] Iteration 9528 (2.32752 iter/s, 5.15569s/12 iters), loss = 0.0121598
I0428 14:21:09.198029  9322 solver.cpp:237]     Train net output #0: loss = 0.0121596 (* 1 = 0.0121596 loss)
I0428 14:21:09.198036  9322 sgd_solver.cpp:105] Iteration 9528, lr = 0.00151471
I0428 14:21:14.339323  9322 solver.cpp:218] Iteration 9540 (2.33404 iter/s, 5.14131s/12 iters), loss = 0.0724027
I0428 14:21:14.339365  9322 solver.cpp:237]     Train net output #0: loss = 0.0724026 (* 1 = 0.0724026 loss)
I0428 14:21:14.339372  9322 sgd_solver.cpp:105] Iteration 9540, lr = 0.00151111
I0428 14:21:19.443126  9322 solver.cpp:218] Iteration 9552 (2.3512 iter/s, 5.10378s/12 iters), loss = 0.0499856
I0428 14:21:19.443171  9322 solver.cpp:237]     Train net output #0: loss = 0.0499855 (* 1 = 0.0499855 loss)
I0428 14:21:19.443179  9322 sgd_solver.cpp:105] Iteration 9552, lr = 0.00150752
I0428 14:21:24.470710  9322 solver.cpp:218] Iteration 9564 (2.38685 iter/s, 5.02754s/12 iters), loss = 0.0705835
I0428 14:21:24.470758  9322 solver.cpp:237]     Train net output #0: loss = 0.0705833 (* 1 = 0.0705833 loss)
I0428 14:21:24.470767  9322 sgd_solver.cpp:105] Iteration 9564, lr = 0.00150395
I0428 14:21:29.632521  9322 solver.cpp:218] Iteration 9576 (2.32478 iter/s, 5.16177s/12 iters), loss = 0.0748352
I0428 14:21:29.632583  9322 solver.cpp:237]     Train net output #0: loss = 0.0748351 (* 1 = 0.0748351 loss)
I0428 14:21:29.632597  9322 sgd_solver.cpp:105] Iteration 9576, lr = 0.00150037
I0428 14:21:34.243830  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_9588.caffemodel
I0428 14:21:42.196723  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_9588.solverstate
I0428 14:21:44.614946  9322 solver.cpp:330] Iteration 9588, Testing net (#0)
I0428 14:21:44.614964  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:21:45.289284  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:21:49.453230  9322 solver.cpp:397]     Test net output #0: accuracy = 0.491422
I0428 14:21:49.453277  9322 solver.cpp:397]     Test net output #1: loss = 2.88545 (* 1 = 2.88545 loss)
I0428 14:21:49.567797  9322 solver.cpp:218] Iteration 9588 (0.601947 iter/s, 19.9353s/12 iters), loss = 0.0215247
I0428 14:21:49.567843  9322 solver.cpp:237]     Train net output #0: loss = 0.0215246 (* 1 = 0.0215246 loss)
I0428 14:21:49.567852  9322 sgd_solver.cpp:105] Iteration 9588, lr = 0.00149681
I0428 14:21:53.872558  9322 solver.cpp:218] Iteration 9600 (2.78764 iter/s, 4.30472s/12 iters), loss = 0.0762832
I0428 14:21:53.872597  9322 solver.cpp:237]     Train net output #0: loss = 0.0762831 (* 1 = 0.0762831 loss)
I0428 14:21:53.872606  9322 sgd_solver.cpp:105] Iteration 9600, lr = 0.00149326
I0428 14:21:57.598305  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:21:59.055158  9322 solver.cpp:218] Iteration 9612 (2.31545 iter/s, 5.18257s/12 iters), loss = 0.0459653
I0428 14:21:59.055205  9322 solver.cpp:237]     Train net output #0: loss = 0.0459652 (* 1 = 0.0459652 loss)
I0428 14:21:59.055214  9322 sgd_solver.cpp:105] Iteration 9612, lr = 0.00148971
I0428 14:22:04.212762  9322 solver.cpp:218] Iteration 9624 (2.32668 iter/s, 5.15757s/12 iters), loss = 0.0391272
I0428 14:22:04.212803  9322 solver.cpp:237]     Train net output #0: loss = 0.0391271 (* 1 = 0.0391271 loss)
I0428 14:22:04.212811  9322 sgd_solver.cpp:105] Iteration 9624, lr = 0.00148618
I0428 14:22:09.366714  9322 solver.cpp:218] Iteration 9636 (2.32832 iter/s, 5.15392s/12 iters), loss = 0.0474972
I0428 14:22:09.366812  9322 solver.cpp:237]     Train net output #0: loss = 0.0474971 (* 1 = 0.0474971 loss)
I0428 14:22:09.366822  9322 sgd_solver.cpp:105] Iteration 9636, lr = 0.00148265
I0428 14:22:14.523628  9322 solver.cpp:218] Iteration 9648 (2.32701 iter/s, 5.15683s/12 iters), loss = 0.0569964
I0428 14:22:14.523670  9322 solver.cpp:237]     Train net output #0: loss = 0.0569962 (* 1 = 0.0569962 loss)
I0428 14:22:14.523679  9322 sgd_solver.cpp:105] Iteration 9648, lr = 0.00147913
I0428 14:22:19.687187  9322 solver.cpp:218] Iteration 9660 (2.32399 iter/s, 5.16353s/12 iters), loss = 0.121732
I0428 14:22:19.687227  9322 solver.cpp:237]     Train net output #0: loss = 0.121732 (* 1 = 0.121732 loss)
I0428 14:22:19.687234  9322 sgd_solver.cpp:105] Iteration 9660, lr = 0.00147562
I0428 14:22:24.871019  9322 solver.cpp:218] Iteration 9672 (2.31491 iter/s, 5.18379s/12 iters), loss = 0.0572133
I0428 14:22:24.871086  9322 solver.cpp:237]     Train net output #0: loss = 0.0572132 (* 1 = 0.0572132 loss)
I0428 14:22:24.871102  9322 sgd_solver.cpp:105] Iteration 9672, lr = 0.00147211
I0428 14:22:29.965934  9322 solver.cpp:218] Iteration 9684 (2.35531 iter/s, 5.09486s/12 iters), loss = 0.0510739
I0428 14:22:29.965981  9322 solver.cpp:237]     Train net output #0: loss = 0.0510738 (* 1 = 0.0510738 loss)
I0428 14:22:29.965989  9322 sgd_solver.cpp:105] Iteration 9684, lr = 0.00146862
I0428 14:22:32.047780  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_9690.caffemodel
I0428 14:22:40.243791  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_9690.solverstate
I0428 14:22:44.509840  9322 solver.cpp:330] Iteration 9690, Testing net (#0)
I0428 14:22:44.509860  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:22:45.138140  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:22:48.351895  9322 blocking_queue.cpp:49] Waiting for data
I0428 14:22:49.460705  9322 solver.cpp:397]     Test net output #0: accuracy = 0.494485
I0428 14:22:49.460752  9322 solver.cpp:397]     Test net output #1: loss = 2.89876 (* 1 = 2.89876 loss)
I0428 14:22:51.375849  9322 solver.cpp:218] Iteration 9696 (0.560487 iter/s, 21.41s/12 iters), loss = 0.0539013
I0428 14:22:51.375892  9322 solver.cpp:237]     Train net output #0: loss = 0.0539012 (* 1 = 0.0539012 loss)
I0428 14:22:51.375900  9322 sgd_solver.cpp:105] Iteration 9696, lr = 0.00146513
I0428 14:22:56.475981  9322 solver.cpp:218] Iteration 9708 (2.3529 iter/s, 5.1001s/12 iters), loss = 0.0584662
I0428 14:22:56.476024  9322 solver.cpp:237]     Train net output #0: loss = 0.0584661 (* 1 = 0.0584661 loss)
I0428 14:22:56.476032  9322 sgd_solver.cpp:105] Iteration 9708, lr = 0.00146165
I0428 14:22:57.224159  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:23:01.657042  9322 solver.cpp:218] Iteration 9720 (2.31614 iter/s, 5.18103s/12 iters), loss = 0.0164752
I0428 14:23:01.657083  9322 solver.cpp:237]     Train net output #0: loss = 0.0164751 (* 1 = 0.0164751 loss)
I0428 14:23:01.657090  9322 sgd_solver.cpp:105] Iteration 9720, lr = 0.00145818
I0428 14:23:06.731601  9322 solver.cpp:218] Iteration 9732 (2.36475 iter/s, 5.07452s/12 iters), loss = 0.0633759
I0428 14:23:06.731645  9322 solver.cpp:237]     Train net output #0: loss = 0.0633758 (* 1 = 0.0633758 loss)
I0428 14:23:06.731654  9322 sgd_solver.cpp:105] Iteration 9732, lr = 0.00145472
I0428 14:23:11.889432  9322 solver.cpp:218] Iteration 9744 (2.32657 iter/s, 5.1578s/12 iters), loss = 0.0319977
I0428 14:23:11.889528  9322 solver.cpp:237]     Train net output #0: loss = 0.0319976 (* 1 = 0.0319976 loss)
I0428 14:23:11.889537  9322 sgd_solver.cpp:105] Iteration 9744, lr = 0.00145127
I0428 14:23:17.061067  9322 solver.cpp:218] Iteration 9756 (2.32039 iter/s, 5.17155s/12 iters), loss = 0.0363971
I0428 14:23:17.061110  9322 solver.cpp:237]     Train net output #0: loss = 0.036397 (* 1 = 0.036397 loss)
I0428 14:23:17.061117  9322 sgd_solver.cpp:105] Iteration 9756, lr = 0.00144782
I0428 14:23:22.251427  9322 solver.cpp:218] Iteration 9768 (2.31199 iter/s, 5.19032s/12 iters), loss = 0.058974
I0428 14:23:22.251474  9322 solver.cpp:237]     Train net output #0: loss = 0.0589739 (* 1 = 0.0589739 loss)
I0428 14:23:22.251482  9322 sgd_solver.cpp:105] Iteration 9768, lr = 0.00144438
I0428 14:23:27.504967  9322 solver.cpp:218] Iteration 9780 (2.28419 iter/s, 5.2535s/12 iters), loss = 0.0299645
I0428 14:23:27.505009  9322 solver.cpp:237]     Train net output #0: loss = 0.0299644 (* 1 = 0.0299644 loss)
I0428 14:23:27.505017  9322 sgd_solver.cpp:105] Iteration 9780, lr = 0.00144095
I0428 14:23:32.196776  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_9792.caffemodel
I0428 14:23:37.219085  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_9792.solverstate
I0428 14:23:44.112547  9322 solver.cpp:330] Iteration 9792, Testing net (#0)
I0428 14:23:44.112635  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:23:44.714329  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:23:49.116392  9322 solver.cpp:397]     Test net output #0: accuracy = 0.49326
I0428 14:23:49.116432  9322 solver.cpp:397]     Test net output #1: loss = 2.98802 (* 1 = 2.98802 loss)
I0428 14:23:49.234582  9322 solver.cpp:218] Iteration 9792 (0.55224 iter/s, 21.7297s/12 iters), loss = 0.0455093
I0428 14:23:49.234630  9322 solver.cpp:237]     Train net output #0: loss = 0.0455092 (* 1 = 0.0455092 loss)
I0428 14:23:49.234638  9322 sgd_solver.cpp:105] Iteration 9792, lr = 0.00143753
I0428 14:23:53.544396  9322 solver.cpp:218] Iteration 9804 (2.78437 iter/s, 4.30977s/12 iters), loss = 0.0208632
I0428 14:23:53.544443  9322 solver.cpp:237]     Train net output #0: loss = 0.0208631 (* 1 = 0.0208631 loss)
I0428 14:23:53.544451  9322 sgd_solver.cpp:105] Iteration 9804, lr = 0.00143412
I0428 14:23:56.568280  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:23:58.679795  9322 solver.cpp:218] Iteration 9816 (2.33674 iter/s, 5.13536s/12 iters), loss = 0.00896484
I0428 14:23:58.679836  9322 solver.cpp:237]     Train net output #0: loss = 0.00896476 (* 1 = 0.00896476 loss)
I0428 14:23:58.679844  9322 sgd_solver.cpp:105] Iteration 9816, lr = 0.00143072
I0428 14:24:03.761569  9322 solver.cpp:218] Iteration 9828 (2.36139 iter/s, 5.08174s/12 iters), loss = 0.0600401
I0428 14:24:03.761613  9322 solver.cpp:237]     Train net output #0: loss = 0.0600401 (* 1 = 0.0600401 loss)
I0428 14:24:03.761622  9322 sgd_solver.cpp:105] Iteration 9828, lr = 0.00142732
I0428 14:24:09.023072  9322 solver.cpp:218] Iteration 9840 (2.28074 iter/s, 5.26146s/12 iters), loss = 0.104419
I0428 14:24:09.023133  9322 solver.cpp:237]     Train net output #0: loss = 0.104419 (* 1 = 0.104419 loss)
I0428 14:24:09.023144  9322 sgd_solver.cpp:105] Iteration 9840, lr = 0.00142393
I0428 14:24:14.200614  9322 solver.cpp:218] Iteration 9852 (2.31772 iter/s, 5.17749s/12 iters), loss = 0.0343477
I0428 14:24:14.200770  9322 solver.cpp:237]     Train net output #0: loss = 0.0343476 (* 1 = 0.0343476 loss)
I0428 14:24:14.200780  9322 sgd_solver.cpp:105] Iteration 9852, lr = 0.00142055
I0428 14:24:19.287729  9322 solver.cpp:218] Iteration 9864 (2.35897 iter/s, 5.08697s/12 iters), loss = 0.0738347
I0428 14:24:19.287775  9322 solver.cpp:237]     Train net output #0: loss = 0.0738346 (* 1 = 0.0738346 loss)
I0428 14:24:19.287783  9322 sgd_solver.cpp:105] Iteration 9864, lr = 0.00141718
I0428 14:24:24.444267  9322 solver.cpp:218] Iteration 9876 (2.32716 iter/s, 5.1565s/12 iters), loss = 0.0657524
I0428 14:24:24.444312  9322 solver.cpp:237]     Train net output #0: loss = 0.0657523 (* 1 = 0.0657523 loss)
I0428 14:24:24.444320  9322 sgd_solver.cpp:105] Iteration 9876, lr = 0.00141381
I0428 14:24:29.593106  9322 solver.cpp:218] Iteration 9888 (2.33064 iter/s, 5.14881s/12 iters), loss = 0.0354316
I0428 14:24:29.593143  9322 solver.cpp:237]     Train net output #0: loss = 0.0354315 (* 1 = 0.0354315 loss)
I0428 14:24:29.593150  9322 sgd_solver.cpp:105] Iteration 9888, lr = 0.00141045
I0428 14:24:31.684638  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_9894.caffemodel
I0428 14:24:35.847018  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_9894.solverstate
I0428 14:24:40.032850  9322 solver.cpp:330] Iteration 9894, Testing net (#0)
I0428 14:24:40.032871  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:24:40.583052  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:24:45.061103  9322 solver.cpp:397]     Test net output #0: accuracy = 0.495098
I0428 14:24:45.061278  9322 solver.cpp:397]     Test net output #1: loss = 2.91862 (* 1 = 2.91862 loss)
I0428 14:24:46.999459  9322 solver.cpp:218] Iteration 9900 (0.689402 iter/s, 17.4064s/12 iters), loss = 0.070581
I0428 14:24:46.999506  9322 solver.cpp:237]     Train net output #0: loss = 0.070581 (* 1 = 0.070581 loss)
I0428 14:24:46.999514  9322 sgd_solver.cpp:105] Iteration 9900, lr = 0.00140711
I0428 14:24:52.510139  9322 solver.cpp:218] Iteration 9912 (2.1776 iter/s, 5.51064s/12 iters), loss = 0.0693374
I0428 14:24:52.510179  9322 solver.cpp:237]     Train net output #0: loss = 0.0693373 (* 1 = 0.0693373 loss)
I0428 14:24:52.510188  9322 sgd_solver.cpp:105] Iteration 9912, lr = 0.00140377
I0428 14:24:52.606848  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:24:57.669296  9322 solver.cpp:218] Iteration 9924 (2.32598 iter/s, 5.15912s/12 iters), loss = 0.0592776
I0428 14:24:57.669337  9322 solver.cpp:237]     Train net output #0: loss = 0.0592775 (* 1 = 0.0592775 loss)
I0428 14:24:57.669345  9322 sgd_solver.cpp:105] Iteration 9924, lr = 0.00140043
I0428 14:25:02.851572  9322 solver.cpp:218] Iteration 9936 (2.3156 iter/s, 5.18224s/12 iters), loss = 0.127541
I0428 14:25:02.851615  9322 solver.cpp:237]     Train net output #0: loss = 0.127541 (* 1 = 0.127541 loss)
I0428 14:25:02.851624  9322 sgd_solver.cpp:105] Iteration 9936, lr = 0.00139711
I0428 14:25:08.013774  9322 solver.cpp:218] Iteration 9948 (2.32461 iter/s, 5.16216s/12 iters), loss = 0.0592621
I0428 14:25:08.013816  9322 solver.cpp:237]     Train net output #0: loss = 0.0592621 (* 1 = 0.0592621 loss)
I0428 14:25:08.013825  9322 sgd_solver.cpp:105] Iteration 9948, lr = 0.00139379
I0428 14:25:13.276082  9322 solver.cpp:218] Iteration 9960 (2.28038 iter/s, 5.26228s/12 iters), loss = 0.080014
I0428 14:25:13.276125  9322 solver.cpp:237]     Train net output #0: loss = 0.0800139 (* 1 = 0.0800139 loss)
I0428 14:25:13.276134  9322 sgd_solver.cpp:105] Iteration 9960, lr = 0.00139048
I0428 14:25:18.430008  9322 solver.cpp:218] Iteration 9972 (2.32834 iter/s, 5.15389s/12 iters), loss = 0.0666687
I0428 14:25:18.430174  9322 solver.cpp:237]     Train net output #0: loss = 0.0666686 (* 1 = 0.0666686 loss)
I0428 14:25:18.430184  9322 sgd_solver.cpp:105] Iteration 9972, lr = 0.00138718
I0428 14:25:23.501209  9322 solver.cpp:218] Iteration 9984 (2.36638 iter/s, 5.07105s/12 iters), loss = 0.0257069
I0428 14:25:23.501256  9322 solver.cpp:237]     Train net output #0: loss = 0.0257068 (* 1 = 0.0257068 loss)
I0428 14:25:23.501264  9322 sgd_solver.cpp:105] Iteration 9984, lr = 0.00138389
I0428 14:25:28.089037  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_9996.caffemodel
I0428 14:25:31.288383  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_9996.solverstate
I0428 14:25:33.892529  9322 solver.cpp:330] Iteration 9996, Testing net (#0)
I0428 14:25:33.892551  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:25:34.427245  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:25:38.731986  9322 solver.cpp:397]     Test net output #0: accuracy = 0.491422
I0428 14:25:38.732030  9322 solver.cpp:397]     Test net output #1: loss = 2.9385 (* 1 = 2.9385 loss)
I0428 14:25:38.850203  9322 solver.cpp:218] Iteration 9996 (0.78181 iter/s, 15.349s/12 iters), loss = 0.031215
I0428 14:25:38.850248  9322 solver.cpp:237]     Train net output #0: loss = 0.0312149 (* 1 = 0.0312149 loss)
I0428 14:25:38.850255  9322 sgd_solver.cpp:105] Iteration 9996, lr = 0.0013806
I0428 14:25:43.175163  9322 solver.cpp:218] Iteration 10008 (2.77462 iter/s, 4.32492s/12 iters), loss = 0.022359
I0428 14:25:43.175211  9322 solver.cpp:237]     Train net output #0: loss = 0.0223589 (* 1 = 0.0223589 loss)
I0428 14:25:43.175220  9322 sgd_solver.cpp:105] Iteration 10008, lr = 0.00137732
I0428 14:25:45.488844  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:25:48.362543  9322 solver.cpp:218] Iteration 10020 (2.31332 iter/s, 5.18734s/12 iters), loss = 0.0341819
I0428 14:25:48.362592  9322 solver.cpp:237]     Train net output #0: loss = 0.0341818 (* 1 = 0.0341818 loss)
I0428 14:25:48.362607  9322 sgd_solver.cpp:105] Iteration 10020, lr = 0.00137405
I0428 14:25:53.521039  9322 solver.cpp:218] Iteration 10032 (2.32628 iter/s, 5.15846s/12 iters), loss = 0.0308506
I0428 14:25:53.521143  9322 solver.cpp:237]     Train net output #0: loss = 0.0308506 (* 1 = 0.0308506 loss)
I0428 14:25:53.521153  9322 sgd_solver.cpp:105] Iteration 10032, lr = 0.00137079
I0428 14:25:58.678196  9322 solver.cpp:218] Iteration 10044 (2.32691 iter/s, 5.15706s/12 iters), loss = 0.0533245
I0428 14:25:58.678242  9322 solver.cpp:237]     Train net output #0: loss = 0.0533245 (* 1 = 0.0533245 loss)
I0428 14:25:58.678251  9322 sgd_solver.cpp:105] Iteration 10044, lr = 0.00136754
I0428 14:26:03.909813  9322 solver.cpp:218] Iteration 10056 (2.29376 iter/s, 5.23158s/12 iters), loss = 0.0306067
I0428 14:26:03.909859  9322 solver.cpp:237]     Train net output #0: loss = 0.0306066 (* 1 = 0.0306066 loss)
I0428 14:26:03.909869  9322 sgd_solver.cpp:105] Iteration 10056, lr = 0.00136429
I0428 14:26:09.047837  9322 solver.cpp:218] Iteration 10068 (2.33555 iter/s, 5.13799s/12 iters), loss = 0.00516411
I0428 14:26:09.047884  9322 solver.cpp:237]     Train net output #0: loss = 0.00516404 (* 1 = 0.00516404 loss)
I0428 14:26:09.047892  9322 sgd_solver.cpp:105] Iteration 10068, lr = 0.00136105
I0428 14:26:14.214857  9322 solver.cpp:218] Iteration 10080 (2.32244 iter/s, 5.16699s/12 iters), loss = 0.0161679
I0428 14:26:14.214895  9322 solver.cpp:237]     Train net output #0: loss = 0.0161678 (* 1 = 0.0161678 loss)
I0428 14:26:14.214903  9322 sgd_solver.cpp:105] Iteration 10080, lr = 0.00135782
I0428 14:26:19.295394  9322 solver.cpp:218] Iteration 10092 (2.36197 iter/s, 5.08051s/12 iters), loss = 0.0475429
I0428 14:26:19.295444  9322 solver.cpp:237]     Train net output #0: loss = 0.0475428 (* 1 = 0.0475428 loss)
I0428 14:26:19.295454  9322 sgd_solver.cpp:105] Iteration 10092, lr = 0.0013546
I0428 14:26:21.378954  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_10098.caffemodel
I0428 14:26:24.601096  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_10098.solverstate
I0428 14:26:27.025008  9322 solver.cpp:330] Iteration 10098, Testing net (#0)
I0428 14:26:27.025028  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:26:27.491577  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:26:31.779502  9322 solver.cpp:397]     Test net output #0: accuracy = 0.49326
I0428 14:26:31.779532  9322 solver.cpp:397]     Test net output #1: loss = 2.90607 (* 1 = 2.90607 loss)
I0428 14:26:33.662940  9322 solver.cpp:218] Iteration 10104 (0.835215 iter/s, 14.3676s/12 iters), loss = 0.0207934
I0428 14:26:33.662983  9322 solver.cpp:237]     Train net output #0: loss = 0.0207934 (* 1 = 0.0207934 loss)
I0428 14:26:33.662992  9322 sgd_solver.cpp:105] Iteration 10104, lr = 0.00135138
I0428 14:26:38.154765  9333 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:26:38.814141  9322 solver.cpp:218] Iteration 10116 (2.32957 iter/s, 5.15116s/12 iters), loss = 0.160328
I0428 14:26:38.814188  9322 solver.cpp:237]     Train net output #0: loss = 0.160328 (* 1 = 0.160328 loss)
I0428 14:26:38.814196  9322 sgd_solver.cpp:105] Iteration 10116, lr = 0.00134817
I0428 14:26:44.015514  9322 solver.cpp:218] Iteration 10128 (2.3071 iter/s, 5.20133s/12 iters), loss = 0.0816453
I0428 14:26:44.015565  9322 solver.cpp:237]     Train net output #0: loss = 0.0816452 (* 1 = 0.0816452 loss)
I0428 14:26:44.015573  9322 sgd_solver.cpp:105] Iteration 10128, lr = 0.00134497
I0428 14:26:49.111169  9322 solver.cpp:218] Iteration 10140 (2.35497 iter/s, 5.09561s/12 iters), loss = 0.0626995
I0428 14:26:49.111217  9322 solver.cpp:237]     Train net output #0: loss = 0.0626994 (* 1 = 0.0626994 loss)
I0428 14:26:49.111224  9322 sgd_solver.cpp:105] Iteration 10140, lr = 0.00134178
I0428 14:26:54.321084  9322 solver.cpp:218] Iteration 10152 (2.30332 iter/s, 5.20988s/12 iters), loss = 0.0400823
I0428 14:26:54.321125  9322 solver.cpp:237]     Train net output #0: loss = 0.0400822 (* 1 = 0.0400822 loss)
I0428 14:26:54.321135  9322 sgd_solver.cpp:105] Iteration 10152, lr = 0.00133859
I0428 14:26:59.428161  9322 solver.cpp:218] Iteration 10164 (2.3497 iter/s, 5.10704s/12 iters), loss = 0.0624579
I0428 14:26:59.428294  9322 solver.cpp:237]     Train net output #0: loss = 0.0624578 (* 1 = 0.0624578 loss)
I0428 14:26:59.428303  9322 sgd_solver.cpp:105] Iteration 10164, lr = 0.00133541
I0428 14:27:04.597035  9322 solver.cpp:218] Iteration 10176 (2.32165 iter/s, 5.16875s/12 iters), loss = 0.0284454
I0428 14:27:04.597086  9322 solver.cpp:237]     Train net output #0: loss = 0.0284453 (* 1 = 0.0284453 loss)
I0428 14:27:04.597095  9322 sgd_solver.cpp:105] Iteration 10176, lr = 0.00133224
I0428 14:27:09.753175  9322 solver.cpp:218] Iteration 10188 (2.32734 iter/s, 5.1561s/12 iters), loss = 0.0162305
I0428 14:27:09.753216  9322 solver.cpp:237]     Train net output #0: loss = 0.0162304 (* 1 = 0.0162304 loss)
I0428 14:27:09.753224  9322 sgd_solver.cpp:105] Iteration 10188, lr = 0.00132908
I0428 14:27:14.358839  9322 solver.cpp:447] Snapshotting to binary proto file snapshot_iter_10200.caffemodel
I0428 14:27:21.512439  9322 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_10200.solverstate
I0428 14:27:24.957962  9322 solver.cpp:310] Iteration 10200, loss = 0.0510467
I0428 14:27:24.957985  9322 solver.cpp:330] Iteration 10200, Testing net (#0)
I0428 14:27:24.957989  9322 net.cpp:676] Ignoring source layer train-data
I0428 14:27:25.365595  9345 data_layer.cpp:73] Restarting data prefetching from start.
I0428 14:27:29.906177  9322 solver.cpp:397]     Test net output #0: accuracy = 0.482843
I0428 14:27:29.906359  9322 solver.cpp:397]     Test net output #1: loss = 2.92473 (* 1 = 2.92473 loss)
I0428 14:27:29.906374  9322 solver.cpp:315] Optimization Done.
I0428 14:27:29.906391  9322 caffe.cpp:259] Optimization Done.