PrototypeCommon
Saganaki22/ComfyUI-sol-attn
NVIDIA Sol-Attn for ComfyUI / Triton kernel on SM89 - SM121, with zero-copy MiniMax H3 nodes: memory-efficient attention, scheduled tau with graph preview, and feed-forward chunking. Measured 1.14–1.44× vs SageAttention and −37% MLP peak VRAM on H3
- Owner
- Saganaki22 (personal account)
- Stars
- Stars: 159
- Forks
- Forks: 15
- Language
- Python
- License
- Apache-2.0
- Created
- 2026
- Topics
- attentionblackwellcomfyuicomfyui-custom-nodesdiffusion-modelsmemory-opmemory-optimizationminimax-h3sageattentionsol-attn
- Tier
- Prototype (100 to 249 ★) What this tier means
More repos by Saganaki22
All 14 repos by Saganaki22- Saganaki22/ComfyUI-OmniVoice-TTSStars: 563PythonAlpha
- Saganaki22/AgebypassXStars: 379JavaScriptAlpha
- Saganaki22/ComfyUI-FishAudioS2Stars: 309PythonAlpha
- Saganaki22/ComfyUI-VDN-H3Stars: 245PythonPrototype
- Saganaki22/Pixal3D-ComfyUIStars: 226PythonPrototype
- Saganaki22/ComfyUI-VoxCPM2Stars: 203PythonPrototype
Nearby in Python
All Python repos- paradigma-inc/limite-violettoStars: 159PythonPrototype
- RosettaCommons/chaperoneStars: 159PythonPrototype
- aicodeking/deepseek-stackStars: 159PythonPrototype
- jchris/couchappStars: 159PythonPrototype
- cournape/BentoStars: 159PythonPrototype
- matthiask/pdfdocumentStars: 159PythonPrototype