deepspeedai / deepspeedai/DeepSpeed

[BUG] DeepSpeed Hybrid Engine Does not Work for Mistral-7B

Open
#4,954 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

bug deepspeed-chat
Dominant language
Python
Stars
43.1k
Forks
5k
Avg merge
4d 15h
Merged PRs (30d)
112

Description

For mistral-7b models, it seem that hybrid engine does not work in generation. It can be seen the generation time reported below:

|E2E latency=84.05s |Gather latency=0.00s (0.00%) |Generate time=0.00s (0.00%) |Training time=17.70s (21.06%) |Others=66.34 (78.94%)|CurSamplesPerSec=1.14 |AvgSamplesPerSec=2.28

After investigation, the following codes do not work

https://github.com/microsoft/DeepSpeed/blob/master/deepspeed/runtime/hybrid_engine.py#L357-L359

Contributor guide

Open the contributing guide

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

Start with deepspeed/runtime/hybrid_engine.py at lines 357-359 and reproduce generation with a Mistral-7B model. Trace why the hybrid engine does not perform generation there, then verify that generation completes and the reported Generate time is no longer 0.00s.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.