sgl-project / sgl-project/SpecForge

[Question] Why Qwen3-30B-A3B eagle3 draft model accept rate is lower than Qwen3-8B?

Open
#252 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
1.2k
Forks
347
Avg merge
4d 1h
Merged PRs (30d)
41

Description

I trained eagle3 draft model on a Chinese dataset for Qwen3-8B and Qwen3-30B-A3B, but Qwen3-30B-A3B-eagle3's accept rate is quite lower than Qwen3-8B-eagle3, got any explanation and advice, thanks.

Image

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

The payload names no files, tests, or entry points. Start by comparing the Qwen3-8B and Qwen3-30B-A3B Eagle3 training configurations and acceptance-rate measurements; done means an evidence-backed explanation and concrete advice for the reported difference.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, pytorch
Domain
machine-learning, performance
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.