[PR #3037] [CLOSED] feat: register ComfyUI-LTX2-MultiGPU by THE-ANGEL-AI #2967

New issue

Closed

opened 2026-10-28 21:14:08 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-28 21:14:08 +00:00

Owner

📋 Pull Request Information

Original PR: https://github.com/Comfy-Org/ComfyUI-Manager/pull/3037
Author: @THE-ANGEL-AI
Created: 6/30/2026
Status: ❌ Closed

Base: main ← Head: main

📝 Commits (1)

aa8486f feat: register ComfyUI-LTX2-MultiGPU by THE-ANGEL-AI

📊 Changes

1 file changed (+789 additions, -774 deletions)

View changed files

📝 node_db/new/custom-node-list.json (+789 -774)

📄 Description

Register ComfyUI-LTX2-MultiGPU

Author: THE-ANGEL-AI (The Angel Studio)
Repo: https://github.com/THE-ANGEL-AI/ComfyUI-LTX2-MultiGPU
License: GPL-3.0-or-later
Nodes: 7

#	Node	Emoji
1	Splitter (DiT 2×GPU)	🔀
2	Dual CLIP Loader (Gemma 3)	📝
3	VRAM Diagnostics	🩺
4	GPU Strategy (hot-switch)	⚙️
5	VRAM Parking (DiT↔CPU)	🅿️
6	SageAttention (T4 Turbo)	⚡
7	VAE Loader (GPU select)	🖼️

Hybrid Multi-GPU split loader for LTX 2.3 GGUF on 2×T4 (Kaggle).
Replaces DisTorch2 regex splitting with explicit block-aware sharding
and forward-hook hidden-state transfer.

Requires: ComfyUI-GGUF (city96)
pip: gguf, safetensors

Entry added to node_db/new/custom-node-list.json

_{🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.}

## 📋 Pull Request Information **Original PR:** https://github.com/Comfy-Org/ComfyUI-Manager/pull/3037 **Author:** [@THE-ANGEL-AI](https://github.com/THE-ANGEL-AI) **Created:** 6/30/2026 **Status:** ❌ Closed **Base:** `main` ← **Head:** `main` --- ### 📝 Commits (1) - [`aa8486f`](https://github.com/Comfy-Org/ComfyUI-Manager/commit/aa8486f8e4a8cc8ad3f6a637d5adbb822dd1fbe2) feat: register ComfyUI-LTX2-MultiGPU by THE-ANGEL-AI ### 📊 Changes **1 file changed** (+789 additions, -774 deletions) <details> <summary>View changed files</summary> 📝 `node_db/new/custom-node-list.json` (+789 -774) </details> ### 📄 Description ## Register ComfyUI-LTX2-MultiGPU **Author:** THE-ANGEL-AI (The Angel Studio) **Repo:** https://github.com/THE-ANGEL-AI/ComfyUI-LTX2-MultiGPU **License:** GPL-3.0-or-later **Nodes:** 7 | # | Node | Emoji | |---|------|-------| | 1 | Splitter (DiT 2×GPU) | 🔀 | | 2 | Dual CLIP Loader (Gemma 3) | 📝 | | 3 | VRAM Diagnostics | 🩺 | | 4 | GPU Strategy (hot-switch) | ⚙️ | | 5 | VRAM Parking (DiT↔CPU) | 🅿️ | | 6 | SageAttention (T4 Turbo) | ⚡ | | 7 | VAE Loader (GPU select) | 🖼️ | Hybrid Multi-GPU split loader for LTX 2.3 GGUF on 2×T4 (Kaggle). Replaces DisTorch2 regex splitting with explicit block-aware sharding and forward-hook hidden-state transfer. **Requires:** ComfyUI-GGUF (city96) **pip:** gguf, safetensors --- *Entry added to node_db/new/custom-node-list.json* --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>

bsakaguchi

2026-10-28 21:14:08 +00:00