microsoft / microsoft/VPTQ

Is the accuracy reported in the paper the result with finetuning applied?

Open
#203 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
681
Forks
53
PR merge metrics
No merged PRs in 30d

Description

I have the following questions.
I know you must be busy, but I would truly appreciate it if you could find the time to respond.

  1. Are the perplexity and accuracy results reported in Tables 2, 3, 4, 5, and 6 of the paper obtained after finetuning?
  2. If so, I would like to know whether it is end-to-end or layer-wise finetuning.
  3. Additionally, I am not sure how to implement the finetuning step after vector quantization in code, so I would greatly appreciate any hints or guidance on this.
  4. Finally, if you have any insight into the difference in effectiveness between LoRA finetuning and the end-to-end or layer-wise finetuning described in the paper, I would be grateful if you could share that as well.

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

Begin with the paper's Tables 2–6 and the repository's vector-quantization workflow. Determine whether the reported metrics use fine-tuning and whether it is end-to-end or layer-wise, then document the post-quantization fine-tuning guidance and the requested comparison with LoRA. No files or tests are identified in the issue.

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.