@apresland Your approach looks interesting by using a GradCAM method for cancer detection on histopathologic images.
I tried to use it on my dataset and as a model I used the following
from keras.applications.resnet import ResNet50, preprocess_input
model = ResNet50(weights = 'imagenet')
For the last cell to plot the prediction show_gradCAMs(model, gradCAM, guidedBP, img_path) I get this errors
----> new_img = overlay_gradCAM(img, cam3, preds[0])
----> cam3 = np.uint8(255cam3prob)
ValueError: operands could not be broadcast together with shapes (1034,1034,3) (1000,)
Any ideas how how to fix this shapes issue to be able to plot the prediction? Thanks in advance!