lm-sys / lm-sys/FastChat

Vicuna-13b-16k with vllm not repeats a single word in output

Open
#2,330 4 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 tried to use Vicuna-13b-16k with vllm worker(feature in Fastchat library). In that case, it repeats single word in output.
reproduce the error:
" python3 -m fastchat.serve.vllm_worker --model-names "gpt-3.5-turbo,text-davinci-003,text-embedding-ada-002" --model-path lmsys/vicuna-13b-v1.5-16k --num-gpus 2"

however it works when I replace "vllm_worker" to "model_worker"

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

Start by reproducing the command with fastchat.serve.vllm_worker and compare its output with model_worker using the same Vicuna model and two GPUs. Trace the vllm_worker entry point and the model-serving path to identify why output repeats a single word; done means the vllm worker produces normal, non-repeating output for this reproduction.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
backend, machine-learning
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
30/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.