| .. | ||
| architecture-investigations | ||
| data-aug-investigations | ||
| default-split | ||
| lr-investigations | ||
| split-investigations | ||
| allimages.txt | ||
| confusions.ipynb | ||
| labels.txt | ||
| README.md | ||
Stanford Cars
The majority of the work was completed on the Stanford cars dataset. This folder contains the different investigations made using the data.
- architecture-investigations
- How does altering the structure of AlexNet affect performance?
- lr-investigations
- How does affecting the learning rate, both the value itself and the schedule, affect performance?
- split-investigations
- How does the proportions of training/validation/test data affect performance?
The Cars dataset contains 16,185 images of 196 classes of cars. The data is split into 8,144 training images and 8,041 testing images, where each class has been split roughly in a 50-50 split. Classes are typically at the level of Make, Model, Year, e.g. 2012 Tesla Model S or 2012 BMW M3 coupe.