TypeError: can't convert cuda:0 device type tensor to numpy. Use Tensor.cpu() to copy the tensor to host memory first.

[mikbuch]

Current versions of modules (January 2024).

[mikbuch]

In utils add:

    heatmap = cv2.applyColorMap(np.uint8(255 * mask.cpu().squeeze()), cv2.COLORMAP_JET)

Line 15

[mikbuch]

Then it will work

[mikbuch]

I re-created the repository, indent to modify it further: https://github.com/mikbuch/gradcam-pytorch-pet. Thank you Konny for creating this very useful example.