ERA-SESSION12 PyTorch Lightning & Gradio
HF Link: https://huggingface.co/spaces/RaviNaik/ERA-Session12
Tasks:
- ✔️ Move your S10 assignment to Lightning first and then to Spaces such that:
- (You have retrained your model on Lightning)
- You are using Gradio
- ✔️ Your spaces app has these features:
- ask the user whether he/she wants to see GradCAM images and how many, and from which layer, allow opacity change as well
- ask whether he/she wants to view misclassified images, and how many
- allow users to upload new images, as well as provide 10 example images
- ask how many top classes are to be shown (make sure the user cannot enter more than 10)
- Add the full details on what your App is doing to Spaces README
- ✔️ Then:
- Submit the Spaces App Link
- Submit the Spaces README link (Space must not have a training code)
- Submit the GitHub Link where Lightning Code can be found along with detailed README with log, loss function graphs, and 10 misclassified images
Model Summary
| Name | Type | Params
------------------------------------------------
0 | prep | Sequential | 1.9 K
1 | layer1 | Sequential | 369 K
2 | layer2 | Sequential | 295 K
3 | layer3 | Sequential | 5.9 M
4 | pool | MaxPool2d | 0
5 | fc | Linear | 5.1 K
6 | softmax | Softmax | 0
7 | accuracy | MulticlassAccuracy | 0
------------------------------------------------
6.6 M Trainable params
0 Non-trainable params
6.6 M Total params
26.292 Total estimated model params size (MB)LR Finder
Accuracy
Test Accuracy
Tensorboard Plots
Misclassified Images
GradCAM Images
HuggingFace Interface
