aigc-apps / aigc-apps/VideoX-Fun

Drop Position Embedding When Training CogVideoX-Fun-V1.1-5b-Pose?

Open
#176 1 comment 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

Thank you for your wonderful work @bubbliiiing. When I load model from CogVideoX-Fun-V1.1-5b-Pose, I found that patch_embed.pos_embedding is missed and use_learned_positional_embeddings is False. Do you drop drop position embedding when training CogVideoX-Fun-V1.1-5b-Pose to support multi-resolution?

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 by tracing the CogVideoX-Fun-V1.1-5b-Pose model configuration and checkpoint loading around patch_embed.pos_embedding and use_learned_positional_embeddings. Determine whether the missing embedding is intentional for multi-resolution training; the issue provides no file or test, so completion would require a maintainer-confirmed explanation or a defined corrective change.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning
Issue type
Bug
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.