lm-sys / lm-sys/FastChat

Unable to match the generation of demo portal with local hosting the 7b model

Open
#1,484 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

  • I am trying to generate with zero shot prompts in a local gpu setup.
  • But with the same parameters as shown in the demo portal (i.e. temp=0.7, top_p=1.0), am not able the match the generations even in multiple trials.
  • Neither was i able to find anywhere, any guide to good generations parameters.

Any suggestions or resource would be highly appreciated.

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

No file or test is named. Start by reproducing the local 7B generation with temp=0.7 and top_p=1.0, then compare it with the demo portal across multiple trials. Done would be a clear resource explaining the relevant generation parameters and how to reproduce the demo behavior locally.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning
Issue type
Documentation
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.