Skip to content

Which models does AutoTP support for training #7279

Answered by delock
KENTHD8mg asked this question in Q&A
Discussion options

You must be logged in to vote

AutoTP inference just support models in this list. AutoTP training is a subset of this list.

AutoTP uses pattern matching to look for modules in the model for tensor parallel, models out of this list may need to match with new pattern. Some models out of this list also have special architecture that are not compatible with AutoTP.

If you verified AutoTP training works for certain model, you may also submit a PR to update the list in the tutorial to mark it as training supported. This will help DeepSpeed community, thanks!

Replies: 2 comments 1 reply

Comment options

You must be logged in to vote
0 replies
Comment options

You must be logged in to vote
1 reply
@KENTHD8mg
Comment options

Answer selected by KENTHD8mg
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Category
Q&A
Labels
None yet
2 participants