ZhouYanFeng

I have seen https://github.com/jacobgil/pytorch-grad-cam/blob/master/tutorials/Pixel%20Attribution%20for%20embeddings.ipynb But I still don't know how to use, sorry!

> This is tricky. What does model do here - does it have to parallel branches for input1 and input2 ? Or does it stack them to be one image...

> How do the spatial activations that you want to target look like ? The model target layer corresponds to one of the images, or both of them ? The...

> Same problem here, do you solve it? I have two branches with shared backbone, and one of them is used for cam generation while the model returns 2 outputs....

> > > Same problem here, do you solve it? I have two branches with shared backbone, and one of them is used for cam generation while the model returns...

How to read the patch size by ConfigurationManager? Static methods cannot read self variables. Writing like this will result in an error! ``` import torch import torch.nn as nn import...

The function **build_network_architecture** uses a static method, which cannot introduce **self.configuration_manager.patch_size**! Thank you for your reply!

> Use the `arch_init_kwargs` dictionary and add the patch size to that Sorry, I can't find ```arch_init_kwargs```, can you give me a minimum implementation? Thank you!

> You find it in the plans file as part of the configuration you are trying to run. It's called `arch_kwargs` in there. Just add the patch size there and...

We have discussed this issue in #6. Only binary classification problems have this parameter, and this parameter is printed out during the training process. If you do not have a...