Conversation
* support launching custom trainer; * init trainer components through .build() (pytorch#2386); * move data to GPU by micro-batch; * remove rescale_accumulated_loss (pytorch#2206).
pytorch-bot
Bot
added
ciflow/fake-pg
|
Hi @ysa2215! Thank you for your pull request and welcome to our community. Action RequiredIn order to merge any pull request (code, docs, etc.), we require contributors to sign our Contributor License Agreement, and we don't seem to have one on file for you. ProcessIn order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA. Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with If you have received this in error or have any questions, please contact us at cla@meta.com. Thanks! |
|
The following ciflow label(s) have been added but CI has not been triggered yet because the workflows are awaiting approval:
Once a maintainer approves the workflows (scroll to the bottom of the PR page), the corresponding CI jobs will be triggered automatically. Please ping one of the reviewers if you do not have access to approve and run workflows. |
tianyu-l
left a comment
tianyu-l
left a comment
There was a problem hiding this comment.
SDPA should no longer support GQAttention anymore. Would it not error out?
Also, need to rebase.
|
seems targeting at a wrong main branch, close this PR and reopen for related main branch |
3 participants
Summary
When
GQAttentionuses the SDPA backend, passNoneto rotary embedding instead of forwarding explicitpositions.Explicit positions remain unchanged for the flex and varlen backends.
Tests