deepspeedai / deepspeedai/DeepSpeed

Deep Speed Fast-Gen Multi-LoRA[REQUEST]

Open
#5,027 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

enhancement
Dominant language
Python
Stars
43.1k
Forks
5k
Avg merge
4d 15h
Merged PRs (30d)
112

Description

Feature Request : Multi-LoRA
Deep Speed Fast Gen sounds amazing, but I think it's lacking a feature of supporting LoRA adapters trained on Custom Dataset. So, It would be great if you provided some kind of direction of using LoRA adapters with Deepspeed fastgen, potentially with multiple LoRA adapters , efficiently switching between them or even disabling all of them off based on user's choice.

Need Some pointers and Directions on how to do it.
Even if implementing this is not your priority, it would be great, if you are able to provide pointers/direction on how to do this easily. I can make a pull request and add this feature.

Contributor guide

Open the contributing guide

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

The issue names no files, tests, or entry points. Start by locating DeepSpeed Fast Gen and its LoRA-related integration points, then clarify the design for custom-dataset adapters, switching between multiple adapters, and disabling them; done would require an agreed implementation scope or clear usage direction.

Written by the indexing model from the issue text.

Assessment

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

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.