mit-han-lab/distrifuser
[CVPR 2024 Highlight] DistriFusion: Distributed Parallel Inference for High-Resolution Diffusion Models
Employs synchronous cross-patch communication at the first diffusion step, then hides subsequent communication overhead through asynchronous activation reuse across the inference pipeline. Integrates seamlessly with Hugging Face Diffusers API and is already adopted in NVIDIA TensorRT-LLM and ColossalAI frameworks, supporting SDXL, Stable Diffusion 1.4/2, and arbitrary resolutions with minimal code changes.
724 stars and 27 monthly downloads. No commits in the last 6 months. Available on PyPI.
Stars
724
Forks
34
Language
Python
License
MIT
Category
Last pushed
Dec 02, 2024
Monthly downloads
27
Commits (30d)
0
Dependencies
4
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/mit-han-lab/distrifuser"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
PRIS-CV/DemoFusion
Let us democratise high-resolution generation! (CVPR 2024)
thu-ml/Causal-Forcing
Official codebase for "Causal Forcing: Autoregressive Diffusion Distillation Done Right for...
Tencent-Hunyuan/HunyuanPortrait
[CVPR-2025] The official code of HunyuanPortrait: Implicit Condition Control for Enhanced...
Shilin-LU/TF-ICON
[ICCV 2023] "TF-ICON: Diffusion-Based Training-Free Cross-Domain Image Composition" (Official...
giuvecchio/matfuse
MatFuse: Controllable Material Generation with Diffusion Models (CVPR2024)