lm-sys / lm-sys/FastChat

TypeError: unsupported operand type(s) for |: 'dict' and 'dict'

Open
#2,848 1 comment 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

My code use to work with fschat==0.2.33 and vllm==0.2.2, but when I migrated to new versions (fschat==0.2.34 and vllm==0.2.6) fastchat is throwing the following error ,

logs from fastchat

`| ERROR | stderr |   File "/usr/local/lib/python3.8/dist-packages/fastchat/serve/vllm_worker.py", line 150, in generate_stream

| ERROR | stderr |     yield (json.dumps(ret | {"finish_reason": None}) + "\0").encode()

| ERROR | stderr | TypeError: unsupported operand type(s) for |: 'dict' and 'dict'

logs from vLLM code

Received request 24bc821beced4a0abf40ac7342134701

Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 1 reqs, Swapped: 0 reqs, Pending: 0 reqs, GPU KV cache usage: 1.9%, CPU KV cache usage: 0.0%

Finished request 24bc821beced4a0abf40ac7342134701.

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 at fastchat/serve/vllm_worker.py line 150 and inspect the generate_stream path that combines ret with the finish-reason field. Reproduce the failure with Python 3.8 and the reported fschat and vllm versions, then verify the behavior with the supported versions. Done means the request completes without this TypeError or the required Python compatibility is clearly identified.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
ai, backend
Issue type
Bug
Difficulty
2/5
Estimated time
1-3 hours
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.