hao-ai-lab / FastVideo

A unified inference and post-training framework for accelerated video generation.

huggingface / diffusers

🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch.

X-GenGroup / Flow-Factory

A unified framework for easy fine-tuning in Flow-Matching models

bghira / SimpleTuner

A general fine-tuning kit geared toward image/video/audio diffusion models.

mflux-community / mflux

Apple MLX native implementations of state-of-the-art generative image & video models

intel / auto-round

PTQ-Level Efficiency, QAT-Level Accuracy|训练后量化的效率，量化感知训练的精度

ZeamoxWang / pdmd

Official Training and Benchmarking Implementation of PDMD: Projected Distribution Matching Distillation for Video Diffusion Models

xorbitsai / inference

Swap GPT for any LLM by changing a single line of code. Xinference lets you run open-source, speech, and multimodal models on cloud, on-prem, or your laptop — all through one unified, production-ready inference API.

vladmandic / sdnext

SD.Next: All-in-one WebUI for AI generative image and video creation, captioning and processing

huggingface / optimum

🚀 Accelerate inference and training of 🤗 Transformers, Diffusers, TIMM and Sentence Transformers with easy to use hardware optimization tools

rupeshs / fastsdcpu

Fast stable diffusion on CPU and AI PC

huggingface / finetrainers

Scalable and memory-optimized training of diffusion models

PrunaAI / pruna

Pruna is a model optimization framework built for developers, enabling you to deliver faster, more efficient models with minimal overhead.

Woolverine94 / biniou

a self-hosted webui for 30+ generative ai

Video-Rebirth / hyperflow

HyperFlow: Video Rebirth's 8-step LoRA for MiniMax-H3 on the diffusers Modular Pipeline

PipeNetwork / minimax-h3-mlx

MLX (Apple Silicon) port of MiniMax-H3 — 33B joint video+audio diffusion. Validated against the diffusers reference; AdaLN precompute drops 13B at inference.

animede / Realtime_Narration_Video

Real-time talking character on a single RTX 5090 — LTX-2.5 22B + local LLM + TTS, fully local. First video ~2.6s after Enter.

doofzoff / pagedMark

AI watermark removal built for Apple Silicon. SynthID-class pixel regeneration, visible vendor labels and provenance metadata, with every pipeline decision measured on Metal.