rtxbae

Is saving the state by calling .state_dict() sufficient?

2

``` base_optimizer = torch.optim.Adam optimizer = SAM(model.parameters(), base_optimizer, lr=0.1) torch.save({"optz_state_dict":optimizer.state_dict()}, "state.pth") checkpoint = torch.load("state.pth") optimizer.load_state_dict(checkpoint["optz_state_dict"]) ``` By using the above code, the saved state size is more than halved compared...

Any chance for the implementation of the recent Fisher SAM?

2

Link to the paper: https://arxiv.org/abs/2206.04920 Any chance for this implementation in this module?

Is there any reason why AltStore doesn't support iPad Pro (12.9) screen resolution?

1

I was surprised to see that it doesn't support iPad Pro 12.9 screen resolution since it's an actively developed app. Is there any technical reason why it's not supported? Here's...

Is there direct download link for feature embeddings?

1

Hey, I noticed that once agreed to the NDA for the videos, we could download them via a hosting cloud. I wonder do we have this for the embeddings? I...

I'm trying to train on relatively small datasets, [mix-up](https://arxiv.org/abs/1710.09412) is one way to reduce it from overfitting, but it seems like focal loss is not designed to works with label...