Skip to content

launcher: add draft_model to GlobalVariables for spec-decode parents #1675

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
ChenhanYu merged 1 commit into from
Jun 10, 2026
Merged

launcher: add draft_model to GlobalVariables for spec-decode parents #1675

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 10 additions & 0 deletions tools/launcher/core.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -154,6 +154,16 @@ class GlobalVariables:
hf_data: str = None
hf_local: str = None
output_dir: str = None
# Speculative-decoding draft / assistant model path. SPEED-bench
# MTP/EAGLE3/DRAFT_TARGET/DFLASH parent YAMLs reference this via
# ``--draft_model_dir <<global_vars.draft_model>>`` on both the
# qualitative + throughput_32k tasks so the path lives in one
# place. Surfaced on OMNIML-5024: the gemma-4-E4B-it / MTP / vLLM
# parent used the indirection but the launcher rejected it with
# ``No parameter named 'draft_model' exists`` because the
# dataclass schema didn't include the key; the agent worked
# around it inline but the canonical YAML stayed broken.
draft_model: str = None


@dataclass
Expand Down
Loading