Comfy-Org / Comfy-Org/ComfyUI

[Feature request] frame by frame incremental sampler.

Open
#6,942 0 comments 0 reactions 0 assignees View on GitHub
Feature
Dominant language
Python
Stars
133k
Forks
15.7k
Avg merge
1d 7h
Merged PRs (30d)
158

Description

### Feature Idea

I just realized that with 1 frame of hunyuan video, i can create 4k frames and still have memory to spare on my 3090. (although with artifacts, but doable)

But if i want to make a longer than 1 frame video, that's where vram limitation comes in.
So my request if possible, is to add a "sampler customadvanced-incremental" which has 2 bars. samples only 1 frame at a time similar to "tiled samplers", but the "tiles" are each frame, holds onto them, samples next, which fills the second bar with total progress. And once second bar is full, it then shipps that off to the VAE decode.

Here's a hunyuan frame generated at 2480x1360
![Image](https://github.com/user-attachments/assets/974b1f4d-da6d-4d4b-8de8-a23d287b6a41)

### Existing Solutions

Large pricy industrial gpu's that costs way too much.

### Other

_No response_

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.