Harbor is a framework for running agent evaluations and creating and using RL environments.

A package to evaluate factuality of long-form generation. Original implementation of our EMNLP 2023 paper "FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generation"

Developer Asset Hub for NVIDIA Nemotron — A one-stop resource for training recipes, usage cookbooks, and full end-to-end reference examples to build with Nemotron models

Build RL environments for LLM training

A project to improve skills of large language models

Official implementation of "Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decoding"

[ICML 2024 Best Paper] Discrete Diffusion Modeling by Estimating the Ratios of the Data Distribution (https://arxiv.org/abs/2310.16834)

[ICLR2025] DiffuGPT and DiffuLLaMA: Scaling Diffusion Language Models via Adaptation from Autoregressive Models

Official PyTorch implementation for "Large Language Diffusion Models"

Trae Agent is an LLM-based agent for general purpose software engineering tasks.

Nano vLLM

A benchmark for LLMs on complicated tasks in the terminal

Scalable toolkit for efficient model reinforcement

SkyRL: A Modular Full-stack RL Library for LLMs