alibaba / alibaba/ROLL

是否支持Wan2.2 5B的pipeline?

Open
#238 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
3.4k
Forks
312
Avg merge
1h 2m
Merged PRs (30d)
2

Description

你好,我看目前代码里改写的wan_video_vae.py文件是针对Wan2.2 14B的VAE的,请问是否有Wan2.2 5B里所使用的VAE的版本?此外我观察到代码中将Wan里chunk-wise + feature cache的模式替换成了整个时间轴一起进行CausalConv3D操作,注释中提及这是参考了tbstar的做法,请问这个tbstar具体是指哪篇论文呢?

Contributor guide

No contributing guide indexed for this repository

Research direction

Start by reading wan_video_vae.py and comparing the implementation described for Wan2.2 14B with the requested 5B VAE pipeline. Investigate the comment about chunk-wise feature cache, full-timeline CausalConv3D, and the referenced tbstar work; done means the 5B support question and the tbstar paper reference are clearly resolved.

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
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.