GAIN icon indicating copy to clipboard operation
GAIN copied to clipboard

Pytorch implementation of paper: [Tell Me Where to Look: Guided Attention Inference Network](http://openaccess.thecvf.com/content_cvpr_2018/papers/Li_Tell_Me_Where_CVPR_2018_paper.pdf)

Results 6 GAIN issues
Sort by recently updated
recently updated
newest added

Hi, could you upload the datasets maybe somewhere ? Thanks.

What do I have to change in the code to run the model without cuda, just on cpu (is it possible?) ? I have downloaded the Apex version for only...

can u give us a tutorial that how to use this code? for example. how to use dockerfile? And which dataset can we use to train model

Hi, in forward( ) function, in gain.py if remove the code " with torch.enable_grad():" what will it happen? Does it influence the final result?

At /src/models/ gain.py 86 gradient = logits * labels_ohe 87 grad_logits = (logits * labels_ohe).sum() 88 grad_logits.backward(gradient=gradient, retain_graph=True) Line 88 shows an error: File "C:\ProgramData\Anaconda2\envs\python3\lib\site-packages\torch\autograd\__init__.py", line 90, in backward allow_unreachable=True)...

Hi, Thanks a lot for sharing the pytorch version of GAIN!! I'm trying to run the code and make some changes based on GAIN. I was wondering why you define...