lm-sys / lm-sys/FastChat

GPU memory keeps growing with more talks, but the GPU memory is not released when the tallk is closed.

Open
#3,223 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
39.5k
Forks
4.8k
PR merge metrics
No merged PRs in 30d

Description

I used the local model about 40GB, then increased it to 50GB as the coversation increased, but couldn't free the GPU when exited the conversations

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 report names no files, tests, or entry points. First reproduce the local-model conversation growth and identify the serving and conversation-close path responsible for GPU allocation; done means GPU memory is released after a conversation is exited without regressing continued local-model use.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning, performance
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.