Fine-tuning & Reinforcement Learning for LLMs. 🦥 Train OpenAI gpt-oss, DeepSeek, Qwen, Llama, Gemma, TTS 2x faster with 70% less VRAM.

A generative world for general-purpose robotics & embodied AI learning.

Original reference implementation of "3D Gaussian Splatting for Real-Time Radiance Field Rendering"

A TTS model capable of generating ultra-realistic dialogue in one pass.

Wan: Open and Advanced Large-Scale Video Generative Models

Video editing with Python

Efficient Triton Kernels for LLM Training

PyTorch implementation of DQN, AC, ACER, A2C, A3C, PG, DDPG, TRPO, PPO, SAC, TD3 and ....

[ECCV 2024] codes of DiffBIR: Towards Blind Image Restoration with Generative Diffusion Prior

The best OSS video generation models, created by Genmo

Imitation learning algorithms with Co-training for Mobile ALOHA: ACT, Diffusion Policy, VINN

[NeurIPS 2024] Unique3D: High-Quality and Efficient 3D Mesh Generation from a Single Image

Modularized Implementation of Deep RL Algorithms in PyTorch

SAPIEN Manipulation Skill Framework, an open source GPU parallelized robotics simulator and benchmark, led by Hillbot, Inc.

🔮 Instill Core is a full-stack AI infrastructure tool for data, model and pipeline orchestration, designed to streamline every aspect of building versatile AI-first applications

This SDK is now deprecated, use the new unified Google GenAI SDK.

[TMLR 2025] Latte: Latent Diffusion Transformer for Video Generation.

Official repository for "AM-RADIO: Reduce All Domains Into One"

Timestep Embedding Tells: It's Time to Cache for Video Diffusion Model

Cosmos-Reason1 models understand the physical common sense and generate appropriate embodied decisions in natural language through long chain-of-thought reasoning processes.

Cosmos-Transfer1 is a world-to-world transfer model designed to bridge the perceptual divide between simulated and real-world environments.

This code corresponds to simulation environments used as part of the MimicGen project.

Official repo of VLABench, a large scale benchmark designed for fairly evaluating VLA, Embodied Agent, and VLMs.

The Open-Source Repository of Taipei City Dashboard. Try it out by clicking the link below!

Reference workflow for generating large amounts of synthetic motion trajectories for robot manipulation from a few human demonstrations.

Revisiting Image Deblurring with an Efficient ConvNet - An efficient CNN performs better than Transformer

Learning Real-World Action-Video Dynamics with Heterogeneous Masked Autoregression