[RL] add post process method for low precision rl weight update - #24657
Open
xiuhu17 wants to merge 1 commit into
Open
[RL] add post process method for low precision rl weight update#24657xiuhu17 wants to merge 1 commit into
xiuhu17 wants to merge 1 commit into
Conversation
Brings the post_process_weights API set from pp_weight_update onto this branch: PostProcessWeightsReqInput/Output, scheduler dispatch, tp_worker + model_runner handlers (Marlin repack, restore-before-load, optional post_load_weights for DeepSeek MLA), tokenizer-side communicator wiring and async pause-aware locking, Engine.post_process_weights, and the /post_process_weights HTTP endpoint. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Contributor
|
Warning You have reached your daily quota limit. Please wait up to 24 hours and I will start processing your requests again! |
xiuhu17
marked this pull request as ready for review
May 12, 2026 18:35
xiuhu17
requested review from
CatherineSue,
Fridge003,
JustinTong0323,
Ying1123,
hnyls2002,
ispobock,
merrymercy,
slin1237 and
xiezhq-hermann
as code owners
May 12, 2026 18:35
xiuhu17
changed the title
yueming-yuan
changed the title
fzyzcjy
reviewed
May 14, 2026
Comment on lines
+545
to
+552
| async with self.is_pause_cond: | ||
| is_paused = self.is_pause | ||
| if is_paused: | ||
| results = await self.post_process_weights_communicator(obj) | ||
|
|
||
| if not is_paused: | ||
| async with self.model_update_lock.writer_lock: | ||
| results = await self.post_process_weights_communicator(obj) |
Collaborator
There was a problem hiding this comment.
nit: is this a bit improveable
Contributor
Author
There was a problem hiding this comment.
Thanks for the review. Do you think we should combine the logic of post_process_weights into update_weights_from_xxx?
| logger.error(f"IPC weight update failed: {e}") | ||
| return False, str(e) | ||
|
|
||
| def post_process_weights(self, recv_req): |
Collaborator
There was a problem hiding this comment.
todo: not sure whether this is the best impl
Contributor
Author
There was a problem hiding this comment.
post_process_weights only calls once after all update_weights_from_xxx finished and once before. I am not sure whether there is a better way to improve the design. Do you have any suggestion?
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
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.
upstream post_process_weights from sglang-miles branch, needed by nemorl team mxfp8 rl training