search

sicara / tf-explain

Interpretability Methods for tf.keras models with Tensorflow 2.x
https://tf-explain.readthedocs.io
MIT License
1.02k stars 112 forks source link

possible support for 1D models? #160

Open arun-nemani opened 3 years ago

arun-nemani commented 3 years ago

Excellent package, are there any plans for supporting TF models that use Conv1D layers instead?

RaphaelMeudec commented 3 years ago

@arun-nemani could you provide any toy example so we can see how we can help?

arun-nemani commented 3 years ago

@RaphaelMeudec This resource is a great toy example. The usecase would be to identify zones in time series signals for highly important / discriminatory features towards a class or prediction

arun-nemani commented 3 years ago

@RaphaelMeudec Hello, I wanted to followup on this issue!

Agrover112 commented 3 years ago

@arun-nemani could you provide any toy example so we can see how we can help?

I could give another good example. In my repo here I use Flattened versions of MNIST , where it fails . https://github.com/M-68/ActivationFunctions. I was trying to add the Activations Visualizer as it would be serve as a good addition to my repo.