Allow torch tensor memory to be released and resumed later

Ongoing research training transformer models at scale

UniRL is a Framework for Unified Multimodal Model Reinforcement Learning

An LLM post-training framework with vLLM for RL Scaling

Compact and Agent-Native MoE Training System

📰 Must-read papers and blogs on Speculative Decoding ⚡️

how to optimize some algorithm in cuda.

FlashInfer: Kernel Library for LLM Serving

[Experimental] Miles-diffusion is an post-training framework for large-scale diffusion model training and production workloads, forked from and co-evolving with miles.

An Open-Source Large-Scale Reinforcement Learning Project for Search Agents

A Distributed Attention Towards Linear Scalability for Ultra-Long Context, Heterogeneous Data Training

Official Implementation of EAGLE-1 (ICML'24), EAGLE-2 (EMNLP'24), and EAGLE-3 (NeurIPS'25).

A unified library for building, evaluating, and storing speculative decoding algorithms for LLM inference in vLLM

A unified framework for building, running, and training general agents at scale.

Open MinT training runtime on veRL

TokenSpeed is a speed-of-light LLM inference engine.

🚀 An open-source, hands-on curriculum bridging the gap from basic RL concepts to LLM alignment, RLVR, and advanced Agentic systems.

verl Zero-Mismatch Dense/MoE HuggingFace Rollout

Multimodal RL training framework for diffusion & omni models

On demand communication

A kernel library written in tilelang

A minimal and fully-customizable CV template for Typst.

Resume template for Typst. Mirror to https://typst.app/project/rVVa3y9vXemUKyvNKnabKV

A simple, elegant, academic style CV template for typst. Support for English and Chinese (and more).

An Asynchronous Reinforcement Learning Engine for Omni-Modal Post-Training at Scale

A project implementing various agentic RL based on the Slime post-training framework

Learning and Debugging for FSDP/FSDP2 Training

[ASPLOS'26] Taming the Long-Tail: Efficient Reasoning RL Training with Adaptive Drafter