Closed ShawnHymel closed 3 years ago
Hi @ShawnHymel .
You're using original Keras
, aren't you.
Unfortunately, as README.md
said, tf-keras-vis is for debugging tf.keras models in Tensorflow2.0+, not original Keras.
I recommend you to use tf.keras
or , if you're using tensorflow1.x
, you can use keras-vis!
Thanks!
Ah, makes sense, thank you. The network was trained with Edge Impulse, which appears to use Keras (not tf.keras), so I don't think I'll be able to use tf-keras-vis for this particular project. Thank you for putting together a great visualization library, though!
I have the following model created and trained with Keras:
When I print out the layers, I get the following:
I try to run the following:
I get the following error:
From what I can tell, Gradcam is attempting to look for
tensorflow.python.keras.layers.convolutional.Conv
layers as opposed tokeras.layers.convolutional.Conv
layers (these are apparently different types of objects from what I can tell).Even if I call:
The
cams
variable comes back as all 0s. This does not seem right. Any help would be appreciated.