lllyasviel / lllyasviel/FramePack

Considerations regarding accessibility of FramePack on devices with very low GPU VRAM (e.g. 2 GB VRAM and lower) or supporting only CPU and how to accelerate CPU inference

Open
#752 2 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
17.3k
Forks
1.7k
PR merge metrics
No merged PRs in 30d

Description

As of now, only CUDA works. I only have an ASUS VivoBook X512DA with an AMD Ryzen 5 3500U CPU with an integrated Radeon Vega 8 Mobile Graphics GPU, which unfortunately has 2048 MB VRAM only, and this is, in my opinion, insufficient for me. I can't afford to get a newer device because I lack a substantial source of revenue, and my autism is preventing me from having a job. I don't know what to do, CPU inference is extremely slow, and I don't know if I can make it faster. I feel like I'm out of options, and others may be in the same position as I do.

Could there be some way for me to help?

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

The issue names no files, tests, or entry points. Start by locating the CPU inference path and the CUDA/VRAM requirements, then determine whether a concrete low-VRAM or CPU-acceleration change is feasible. Done would require an agreed implementation target and validation on hardware around 2 GB VRAM or CPU-only, neither of which the issue specifies.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning, performance
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.