deepspeedai / deepspeedai/DeepSpeedExamples

"enable_mixed_precision_lora" just available for fp16 tensor?

Open
#768 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
6.8k
Forks
1.1k
Avg merge
2d 16h
Merged PRs (30d)
1

Description

Hi, as shown in zero3 of deepspeed(version 0.11.1), i.e., "assert tensor.dtype == torch.float16, f"quantize_dstensor() expects tensor.dtype == torch.float16, got {tensor.dtype}", does this mean "enable_mixed_precision_lora" just available for fp16 tensor(i.e., dtype of fp16 training)?

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 with the ZeRO-3 implementation in DeepSpeed 0.11.1 and inspect quantize_dstensor(), especially the assertion requiring torch.float16. Determine whether enable_mixed_precision_lora supports other tensor dtypes; done should establish the supported behavior and include appropriate coverage or documentation.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.