FP16xINT4 LLM inference kernel that can achieve near-ideal ~4x speedups up to medium batchsizes of 16-32 tokens.

Simplifying reinforcement learning for complex game environments

Type annotations and runtime checking for shape and dtype of JAX/NumPy/PyTorch/etc. arrays. https://docs.kidger.site/jaxtyping/

Solve puzzles. Improve your pytorch.

A scalable asynchronous reinforcement learning implementation with in-flight weight updates.

Atropos is a Language Model Reinforcement Learning Environments framework for collecting and evaluating LLM trajectories through diverse environments

Scalable RL solution for advanced reasoning of language models

Tongyi Deep Research, the Leading Open-source Deep Research Agent

My learning notes/codes for ML SYS.

SkyRL: A Modular Full-stack RL Library for LLMs

A Collection of Competitive Text-Based Games for Language Model Evaluation and Reinforcement Learning

Optimize prompts, code, and more with AI-powered Reflective Text Evolution

slime is an LLM post-training framework for RL Scaling.

Async RL Training at Scale

Real-time terminal monitor for InfiniBand networks - htop for high-speed interconnects

Voyager is an interactive RGBD video generation model conditioned on camera input, and supports real-time 3D reconstruction.

Training-Ready RL Environments + Evals

Hunyuan-GameCraft: High-dynamic Interactive Game Video Generation with Hybrid History Condition

A PyTorch native platform for training generative AI models

Automatic evals for LLMs

Environments for LLM Reinforcement Learning

verl: Volcano Engine Reinforcement Learning for LLMs

A high-throughput and memory-efficient inference and serving engine for LLMs

Qwen Code is a coding agent that lives in the digital world.

Python SDK, Proxy Server (AI Gateway) to call 100+ LLM APIs in OpenAI (or native) format, with cost tracking, guardrails, loadbalancing and logging. [Bedrock, Azure, OpenAI, VertexAI, Cohere, Anthr…