Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Change 10.FSDP/0.create_conda_env.sh to pytorch 2.1.0 #228

Closed
wants to merge 1 commit into from

Conversation

pbelevich
Copy link
Collaborator

This fixes #219

I'm not aware of the context what PyTorch version should be used, but to fix ""triu_tril_cuda_template" not implemented for 'BFloat16'" we need PyTorch 2.1.0+

Explanation:

""triu_tril_cuda_template" not implemented for 'BFloat16'" is fixed in this PR and this commit

git tag --contains c51dfbf5b41b1acdad798c10ef7bfb1ed58abc85 | grep "v2" | head -n 1
v2.1.0

@sean-smith @verdimrc

@johnbensnyder
Copy link
Collaborator

The issue is, without changes to the checkpointing functions, PyTorch>2.0 fails. Created another PR that increases version to 2.2, with a fix for checkpointing.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Example 10.FSDP is broken again with dependencies issue
2 participants