Fellowship.AI Challenge: Cohort 19

This is the submission of the Fellowship.AI Challenge: Cohort 19.

The Challenge

Use a pretrained ResNet 50 and train on the STL10 dataset for 5 epochs and then visualize the layer activations using GRADCAM for a mislabeled image of your weakest class. What additional steps would you take to improve the model without changing the number of epochs?

ResNet 50

• main https://pytorch.org/docs/stable/torchvision/models.html#torchvision.models.resnet50
• tutorial https://pytorch.org/tutorials/beginner/transfer_learning_tutorial.html

STL10

• main https://cs.stanford.edu/~acoates/stl10/
• tutorial https://github.com/uoguelph-mlrg/Cutout
• sota https://paperswithcode.com/sota/image-classification-on-stl-10

The Results

Please see the details in notebook FellowshipAI_Challenge.ipynb