gradCAM visualization for a specific layer
I used a pretrained model from hugging face in code script.
Extract grad-CAM image, here taking dog.png as an example, you can choose another image with input image size 448 x 448
python grad_cam.py --image-path=dog.png --use-cuda
Source code: https://github.com/jacobgil/pytorch-grad-cam