Merge pull request #3327 from buqi-code/add-qwen-image21-multigpu
Add Qwen-Image 2.1 Multi-GPU node
This commit is contained in:
+12
-1
@@ -60650,7 +60650,18 @@
|
||||
],
|
||||
"nodename_pattern": "MiniMaxH3SP.*",
|
||||
"install_type": "git-clone",
|
||||
"description": "Multi-GPU (Ulysses sequence-parallel) inference for MiniMax-H3: drop-in UNETLoader replacement sharding the packed video+audio sequence across 2/4/7/8 GPUs, bit-identical output. Requires ComfyUI >= 0.30.0 and Linux or WSL2 (NCCL; no native Windows)."
|
||||
"description": "Multi-GPU Ulysses sequence-parallel inference for MiniMax-H3 video and audio generation across 2/4/7/8 GPUs. Requires ComfyUI commit c194dd00 or newer and Linux or WSL2 with NCCL; native Windows is not supported."
|
||||
},
|
||||
{
|
||||
"author": "buqi-code",
|
||||
"title": "buqi-qwen-image-2.1-multigpu",
|
||||
"reference": "https://github.com/buqi-code/buqi-qwen-image-2.1-multigpu",
|
||||
"files": [
|
||||
"https://github.com/buqi-code/buqi-qwen-image-2.1-multigpu"
|
||||
],
|
||||
"nodename_pattern": "QwenImage21SP.*",
|
||||
"install_type": "git-clone",
|
||||
"description": "ComfyUI-native Ulysses sequence-parallel inference for Qwen-Image 2.1 across 1/2/4/8 GPUs, including text-to-image, multi-reference editing, and distributed prefix caching. Requires ComfyUI commit c194dd00 or newer and Linux or WSL2 with NCCL; native Windows is not supported."
|
||||
},
|
||||
{
|
||||
"author": "ANe5s",
|
||||
|
||||
Reference in New Issue
Block a user