lm-sys / lm-sys/FastChat

【bug】KeyError: 'factor' when use Triplex

Open
#3,470 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

command is
`
python -m fastchat.serve.model_worker --model-names Triplex --model-path /data/cache/Triplex --controller-address http://127.0.0.1:21001 --worker-address http://127.0.0.1:8082 --host 0.0.0.0 --port 8082

`
this is the log
`
2024-08-08 14:43:27 | ERROR | stderr | ^MLoading checkpoint shards: 0%| | 0/2 [00:00
2024-08-08 14:43:32 | ERROR | stderr | args, worker = create_model_worker()
2024-08-08 14:43:32 | ERROR | stderr | File "/data/env/FastChat/fastchat/serve/model_worker.py", line 386, in create_model_worker
2024-08-08 14:43:32 | ERROR | stderr | worker = ModelWorker(
2024-08-08 14:43:32 | ERROR | stderr | File "/data/env/FastChat/fastchat/serve/model_worker.py", line 96, in __init__
2024-08-08 14:43:32 | ERROR | stderr | self.context_len = get_context_length(self.model.config)
2024-08-08 14:43:32 | ERROR | stderr | File "/data/env/FastChat/fastchat/utils.py", line 367, in get_context_length
2024-08-08 14:43:32 | ERROR | stderr | rope_scaling_factor = config.rope_scaling["factor"]
2024-08-08 14:43:32 | ERROR | stderr | KeyError: 'factor'
`

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

Reproduce the issue with the provided model_worker command, then start at get_context_length in fastchat/utils.py and follow its call from ModelWorker in fastchat/serve/model_worker.py. Done means Triplex can initialize without the KeyError and the worker proceeds past model loading.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
ai, backend
Issue type
Bug
Difficulty
3/5
Estimated time
1-2 days
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
58/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.