Design of noise, interpolation schedules, and sources in generative dynamics, with enhanced numerical performance

The codebase of our paper "Improving the Training of Rectified Flows", NeurIPS 2024

State-of-the-Art Embeddings, Retrieval, and Reranking

Emergent Hierarchical Reasoning in LLMs/VLMs through Reinforcement Learning [ICLR26]

Code for Fine-grained Uncertainty Quantification for LLMs from Semantic Similarities (NeurIPS'24)

a neuroscience based model simulating basal ganglia for mouse maze pathfinding with q-learning

Minimal reproduction of DeepSeek R1-Zero

An Open-source RL System from ByteDance Seed and Tsinghua AIR

The Entropy Mechanism of Reinforcement Learning for Large Language Model Reasoning.

Train transformer language models with reinforcement learning.

[NeurIPS 2025] TTRL: Test-Time Reinforcement Learning

Lighteval is your all-in-one toolkit for evaluating LLMs across multiple backends

verl/HybridFlow: A Flexible and Efficient RL Post-Training Framework

PyTorch implementation of Advantage Actor Critic (A2C), Proximal Policy Optimization (PPO), Scalable trust-region method for deep reinforcement learning using Kronecker-factored approximation (ACKT…

Code for NeurIPS'24 paper 'Grokked Transformers are Implicit Reasoners: A Mechanistic Journey to the Edge of Generalization'

Simple RL training for reasoning

Code for the paper "The Impact of Positional Encoding on Length Generalization in Transformers", NeurIPS 2023

An Easy-to-use, Scalable and High-performance Agentic RL Framework based on Ray (PPO & DAPO & REINFORCE++ & VLM & TIS & vLLM & Ray & Async RL)

L1: Controlling How Long A Reasoning Model Thinks With Reinforcement Learning

s1: Simple test-time scaling

Cliff walking reinforcement learning example, with a variety of RL algorithms

LLM training in simple, raw C/CUDA

The simplest, fastest repository for training/finetuning medium-sized GPTs.