shsshs

experiencing the same issue. Found this as well: https://github.com/tensorflow/tensorflow/issues/783

Any news on this?

If you check the code it actually never tries to load imagenet weights for most of the densenets. Does anyone know if you can train these networks directly on ~50k...

Is there a quickfix?

Thanks a lot! Saliency works with this (at least it gives meaningfull results). CAM gives me only one 224x224 Matrix for my 224x224x3 image.

Nope it wasn't. Looks fine for CAM as well, thx again.

for me visualize_cam is fine (224,224,3) to (224,224,3). But this command `jet_heatmap = np.uint8(cm.jet(grads)[..., :3] * 255)` from [https://github.com/raghakot/keras-vis/blob/master/examples/vggnet/attention.ipynb](url) gives me (224,224,3,3).