Move compile into AutoParallel API #77
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
meta-cla
bot
added
the
CLA Signed
wconstab
force-pushed
the
whc/compile
branch
2 times, most recently
from
b07c2a6 to
27583b4
Compare
|
ah crap. this change actually isn't possible to land in isolation becuase the meta info needs to be updated before compile will even run. I'll merge my meta-fixup PR into this PR. |
Update node metas after parallelization
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Requires upstream pytorch/pytorch#159814 to land first. (done)
also updates joint_with_descriptors meta info to reflect that arg/placeholder shapes have changed (inputs and params are now sharded, not global, as they were when first traced).
kicked off a verification run from this PR @ acaaf9c: https://www.internalfb.com/mlhub/pipelines/runs/mast/torchtitan-64-whc-kf1llhnr
tbm FSDP_eager:torchtitan-64-whc-p3s1bn compile_pr:torchtitan-64-whc-kf1llhnr compile_noac_from_update_post:torchtitan-64-fmassa-r4rvfnf6
hmm. this compile PR has higher memory usage than the compile_noac job we had last week.. what changed..
rerun w/ memory profiling: https://www.internalfb.com/mlhub/pipelines/runs/mast/torchtitan-64-whc-k3l2mcd