aigc-apps / aigc-apps/VideoX-Fun

finetune control model on depth

Open
#123 0 comments 1 reaction 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
2.3k
Forks
188
Avg merge
2d 22h
Merged PRs (30d)
3

Description

The Training control model Readme recommends using DWPose for pose control training. I'd like to instead train a model on depthmaps from DepthCrafter. Will I run into any issues doing this if I use 1.1-2B or 1.1-5B Control models as the base? The Readme uses 1.1-2b-Pose but I assume I could swap for the Control model? Do I need to provide multiple control inputs then?

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 Training control model Readme and compare its 1.1-2b-Pose setup with the 1.1-2B and 1.1-5B Control models. Check whether DepthCrafter depthmaps and multiple control inputs are supported, then document the compatibility requirements and any changes needed for a reproducible training setup.

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.