Using saliency with custom nets and grayscale images

[vedantbhatia]

Hi,

Currently the saliency method is not working for images with a single channel (ex (1,244,244)). At some point the code seems to require three channels. This results in shape errors, etc.

Could you please help me run the saliency code with grayscale images? Thanks.

[sytelus]

Thanks for pointing this out! Are you using MNIST data? Also, is there a specific explainer you are interested in?

[vedantbhatia]

I am using a custom dataset. I'm interested in saliency approaches for a modified GoogLeNet architecture (with grayscale input and custom output classes).

Do you think this will be compatible anytime soon? What sort of changes will you have to make for this?

Thanks for the prompt reply.