Enable AI models for video production in the browser

[CVPR 2025] Official Implementation of MotionPro: A Precise Motion Controller for Image-to-Video Generation

MAGI-1: Autoregressive Video Generation at Scale

xDiT: A Scalable Inference Engine for Diffusion Transformers (DiTs) on multi-GPU Clusters

(WIP) Parallel inference for black-forest-labs' FLUX model.

xDiT: A Scalable Inference Engine for Diffusion Transformers (DiTs) with Massive Parallelism

[CVPR 2024 Highlight] DistriFusion: Distributed Parallel Inference for High-Resolution Diffusion Models

Official PyTorch Implementation of "Optimal Stepsize for Diffusion Sampling".

[CVPR 2025] Multi-focal Conditioned Latent Diffusion for Person Image Synthesis

LBM: Latent Bridge Matching for Fast Image-to-Image Translation ✨ (ICCV 2025 Highlight)

Enhance-A-Video: Better Generated Video for Free

[CVPR 2025 Highlight] GEN3C: 3D-Informed World-Consistent Video Generation with Precise Camera Control

[CVPR 2025] Official repo for ART:Anonymous Region Transformer for Variable Multi-Layer Transparent Image Generation

[ICML2025] SpargeAttention: A training-free sparse attention that accelerates any model inference.

Official implementation of S3Diff

Minimalistic large language model 3D-parallelism training

Original reference implementation of "EVER: Exact Volumetric Ellipsoid Rendering for Real-time View Synthesis"

Inference-time scaling of diffusion-based image and video generation models.

Code release for "LLMs can see and hear without any training"

An open-source implementation of Regional Adaptive Sampling (RAS), a novel diffusion model sampling strategy that introduces regional variability in sampling steps

[CVPR2025 Highlight] Video Generation Foundation Models: https://saiyan-world.github.io/goku/

FlashVideo: Flowing Fidelity to Detail for Efficient High-Resolution Video Generation

Official code of "LayerTracer: Cognitive-Aligned Layered SVG Synthesis via Diffusion Transformer"

Simple go utility to download HuggingFace Models and Datasets

Official repo for paper "Structured 3D Latents for Scalable and Versatile 3D Generation" (CVPR'25 Spotlight).

https://wavespeed.ai/ Context parallel attention that accelerates DiT model inference with dynamic caching

Rembg is a tool to remove images background