veda-sparse / Veda-on-ComfyUI

# Veda Sparse Attention for ComfyUI (MiniMax-H3)
Paper · Project Page · Predictor · Training Code · 中文说明
## Introduction
**Veda** is a learned sparse-attention method for video diffusion models.
Attention dominates the cost of a long clip, and a small distilled
predictor can say in advance which tiles of the attention map carry the
result. Veda computes roughly the top 10% of them and skips the rest.
This repository packages Veda as a custom node for ComfyUI's native
MiniMax-H3 (T2VA, FL2VA and R2VA): one node, MODEL in and MODEL out.
### Highlights
- **Faster inference.** 2.9x end to end and 7.1x on attention alone
 against ComfyUI's default attention, measured on an RTX 5070. The gain
 grows with clip length.
- **Plug-and-play.** An attention override rather than a weight patch, so
 Turbo and style LoRAs, fine-tuned and quantized H3 checkpoints,
 first/last-frame and reference conditioning, and other block patches
 keep working.
- **Never silently off.** Every kernel passes a self-test on the GPU
 before it is used. When Veda cannot run, the prompt stops with the
 reason instead of quietly rendering without acceleration: a missing
 package or a node in the wrong place comes with the fix, anything else
 with a ready-to-paste report for a GitHub issue.
- **Nothing extra to install.** The sparse kernel ships with the node;
 installing from the Comfy Registry pulls in Triton (NVIDIA SM80 and
 newer, AMD RDNA3 / RDNA4 on ROCm) or MLX (Apple silicon).
## Installation
Requires ComfyUI >= 0.38.0. In ComfyUI Manager, search "Veda". Or:
 [code] 
Or clone this repository into `ComfyUI/custom_nodes/` and restart ComfyUI.
## Usage
Open *Workflow -> Browse Templates -> Veda-on-ComfyUI* and pick "Veda
MiniMax H3 T2VA" or "Veda MiniMax H3 R2VA". The missing-model dialog
offers the predictor. Write a prompt and run.
To add Veda to an existing H3 workflow, put **Veda Sparse Attention
(MiniMax H3)** on the MODEL wire after the model and any LoRA loaders,
last before the guider or sampler. Select it and press **Ctrl+B** to
bypass it and render the same seed with full attention.
Do not combine it with ComfyUI's own "Model Sparse Attention" node on H3:
that one replaces the attention blocks outright, so Veda would never be
called. The node says so when it sees both.
Other attention nodes can stay. Veda runs on top of "Patch Sage Attention
KJ" (the calls it does not take go to sage), and KJNodes' "MiniMax H3 Low
VRAM Attention" keeps working. A node that swaps out the whole H3
attention, such as "MiniMax H3 Mem Eff Sage Attention Patch", must come
*before* Veda: Veda then runs the sparse layers and leaves that node the
full-attention ones, and says so on its node. Placed after Veda it would
hide every call from Veda; the node warns about that too.
### Predictor
The node does not download anything itself. The predictor (275 MB)
reaches `models/veda` one of two ways:
* **From a template.** Open one of the Veda templates and ComfyUI offers
 the predictor in its missing-