pytorch-grad-cam
pytorch-grad-cam copied to clipboard
How to add the 'targets' attribute for the SwinT?
Thanks for your great work. I ran SwinT_examle.py on my own dataset. But the results were terrible, always failing to focus on the goal. But the results are much better after I use cnn and specify targets. I think the most likely reason is that I don't set the 'targers' in SwinT.
Can you please try following the new Swin transformer section in this tutorial and let me know if it helped: https://jacobgil.github.io/pytorch-gradcam-book/HuggingFace.html