[gpt-oss-120b] Add --enable_speculative_decode flag for Interactive scenario#2582
Open
viraatc wants to merge 1 commit into
Open
[gpt-oss-120b] Add --enable_speculative_decode flag for Interactive scenario#2582viraatc wants to merge 1 commit into
viraatc wants to merge 1 commit into
Conversation
Contributor
|
MLCommons CLA bot All contributors have signed the MLCommons CLA ✍️ ✅ |
…cenario Adds an opt-in EAGLE3 speculative decoding config path to run_server.sh, mirroring the --enable_speculative_decode flag in DeepSeek-R1. Defaults are tuned for the gpt-oss-120b Interactive scenario on GB200 using the nvidia/gpt-oss-120b-Eagle3-long-context head (num_steps=3, topk=1). Preserves backward compatibility: existing --eagle_path callers are unchanged, and the default server command still matches the previous behavior when the flag is not passed. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
viraatc
force-pushed
the
feat/viraatc-gptoss-mtp
branch
from
40d927c to
d2f48dc
Compare
Contributor
Author
|
Contributor
Author
|
Contributor
Author
|
looks like the gitleaks license is missing? repo setup issue? |
Contributor
Is this PR ready for review and approval? Also, what are the next steps to resolve the Gitleaks license issue mentioned above? @viraatc |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
--enable_speculative_decodeflag tolanguage/gpt-oss-120b/sglang/run_server.sh, mirroring the flag already present on the DeepSeek-R1 SGLang backend.nvidia/gpt-oss-120b-Eagle3-long-context,num_steps=3,topk=1. Each of these can be overridden (--speculative_draft_model_path,--speculative_num_steps,--speculative_topk).language/gpt-oss-120b/README.md.Backward compatible: the pre-existing
--eagle_pathpath takes precedence over the new defaults, and when--enable_speculative_decodeis not passed the server command is unchanged.