Brian Vaughan

It looks a little confusing and maybe could be written more clearly, but I think that's actually correct. If it was just `model`, it would be trying to FSDP-wrap the...

@mthrok does this imply torchio would be its own pypi package, so can be depended on (e.g. in torchvision) without depending on the rest of torchaudio?

cc @chauhang @HamidShojanazeri we're happy to submit a PR, but would like confirmation that aimstack integration would be welcome in this codebase. I believe we can add it in a...

looks like there are conflicts with main that will have to be resolved too

I'm glad to see this exists. I've just been relying on the directory list to see what's supported (checking https://github.com/EleutherAI/lm-evaluation-harness/tree/main/lm_eval/tasks). Is there a more thorough list somewhere within the repo?...

@Harryalways317 that looks super helpful. I think it'd be ideal if there was a way we could we could represent this info in one place (e.g. yaml or son) that...

that sounds good but I'd probably want the "task family" column to be the link. having the full desciption be a link could look awkward vs just having the one...

Looking at the table, the actual number of parameters in llama7b is 6738000000 so at half precision it makes sense that `6738000000 * 2 / (1024**3) == 12.55GB`. I usually...

what are the sizes of the files, especially the pytorch_model.bin? do we have a safetensors version? how are we loading it?

There are some suggestions here on how to fix some of these, possibly via a fully qualified import? https://stackoverflow.com/questions/68695851/mypy-cannot-find-implementation-or-library-stub-for-module @afrittoli may also have suggestions. pasting the mypy errors here, for...