[ECCV 2026] Moebius: 0.2B Lightweight Image Inpainting Framework with 10B-Level Performance

Official repository for the survey "Reinforcement Learning for Visual Generation: A Model-Oriented Survey".

Lore is a next-generation, open source version control system

Official JAX code of MiniT2I.

Pipeline-parallel LLM inference across GPUs on separate machines.

Tile-Based Runtime for Ultra-Low-Latency LLM Inference

DreamReasoner-8B, a large diffusion reasoning model

Official PyTorch re-implementation of MiniT2I.

OpenAI's privacy filter NER model architecture implemented in a minimal C++/GGML runtime

Code of "Show the Signal, Hide the Noise: Spectral Forcing for Pixel-Space Diffusion"

UniRL is a Framework for Unified Multimodal Model Reinforcement Learning

ComfyUI custom nodes and workflows for Boogu-Image-0.1.

Boogu-Image-0.1 is an Apache-2.0 open-source image generation and editing model family that delivers near-closed-source performance with an order of magnitude less data.

[ICML 2026] The official implementation of "Mean Flow Distillation: Robust and Stable Distillation for Flow Matching Models".

Open Machine Learning Compiler Framework

compiler learning resources collect.

Repository for the CVPR 2026 paper MeshFlow Efficient Artistic Mesh Generation via MeshVAE and Flow-based Diffusion Transformer by Weiyu Li, Antoine Toisoul, Tom Monnier, Roman Shapovalov, Rakesh R…

Fork of SageAttention for Windows wheels and easy installation

Stable Audio LoRA Trainer of salty goodness

A super light-weight embedded code search engine CLI (AST based) that just works - saves 70% token and improves speed for coding agent 🌟 Star if you like it!

The all-in-one, open-source backend platform for agentic coding. InsForge gives your coding agent database, auth, storage, compute, hosting, and AI gateway to ship full-stack apps end-to-end.

🙌 OpenHands: AI-Driven Development

Full Transformer into a custom chip. microGPT in RTL, generating names on a Virtex-5 FPGA at ~56k tokens/second.

Our inference and training framework to run on the Cosmos Models

DeepSeek-native AI coding agent for your terminal. Engineered around prefix-cache stability — leave it running.

Code repository for "Spectral Progressive Diffusion for Efficient Image and Video Generation"

Official Implementation of SCAIL-2: Unifying Controlled Character Animation with End-to-end In-Context Conditioning

TORAX: Tokamak transport simulation in JAX

llama.cpp fork with TurboQuant WHT-rotated KV cache & weight compression + Gemma 4 MTP and Qwen 3.6 NextN speculative decoding (+30-50% throughput).

Ideogram 4: Open image model at the forefront of design