Add vLLM Support for GPT OSS and its mapping generator for tunix. #2803
+231
−1
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
abhinavclemson
requested review from
A9isha,
NicoGrande,
NuojCheng,
RissyRan,
SurbhiJainUSC,
aireenmei,
bvandermoon,
gagika,
gobbleturk,
hengtaoguo,
jiangjy1982,
khatwanimohit,
richjames0,
shralex,
suexu1025 and
vipannalla
as code owners
abhinavclemson
force-pushed
the
gpt-off-branch
branch
3 times, most recently
from
69d6317 to
9b0fe8f
Compare
abhinavclemson
force-pushed
the
gpt-off-branch
branch
from
9b0fe8f to
413bfab
Compare
gagika
reviewed
Dec 10, 2025
|
|
||
| # TODO: Enable multi-host sharding, if there is a mismatch in shapes. | ||
| # # MULTI-HOST case. | ||
| val = jax.device_put(val, current.sharding) |
Collaborator
There was a problem hiding this comment.
was it tested at multi-host ?
what sharding do you think it might not work?
| return current.at[..., 0::2].set(val) | ||
|
|
||
| def fuse_interleaved_up(val, tgt_param): | ||
| """Fuse Up (wi_1) with Multi-Host Sharding Support.""" |
Collaborator
There was a problem hiding this comment.
why Multi-Host is special? do you mean multi-devie or multi-host?
From Jax's perspective I think only number of devices (and sharding) matters.
| return STANDALONE_VLLM_WEIGHT_MAPPING[self.model_name].to_hf_mapping() | ||
| mapping_fn = STANDALONE_VLLM_WEIGHT_MAPPING[self.model_name].to_hf_mapping | ||
| total_num_layers = self.config["num_hidden_layers"] | ||
| print(f"total_num_layers: {total_num_layers} for model: {self.model_name}") |
Collaborator
There was a problem hiding this comment.
could you remove this print or make debug logging (e.g. logging.debug)?
| # TODO: Enable multi-host sharding, if there is a mismatch in shapes. | ||
| # # MULTI-HOST case. | ||
| val = jax.device_put(val, current.sharding) | ||
| val.block_until_ready() |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 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.
Description
Add vLLM Support for GPT OSS and its mapping generator for tunix.
Before submitting this PR, please make sure (put X in square brackets):
(https://maxtext.readthedocs.io/en/latest/development.html#adding-new-documentation-files).kets):
gemini-reviewlabel.