NeMo Gym Documentation#

NeMo Gym is a library for building reinforcement learning (RL) training environments for large language models (LLMs). NeMo Gym provides infrastructure to develop environments, scale rollout collection, and integrate seamlessly with your preferred training framework.

A training environment consists of three server components: Agents orchestrate the rollout lifecycle—calling models, executing tool calls through resources, and coordinating verification. Models provide stateless text generation using LLM inference endpoints. Resources define tasks, tool implementations, and verification logic.


Introduction to NeMo Gym#

Understand NeMo Gym’s purpose and core components before diving into tutorials.

About NeMo Gym

Motivation and benefits of NeMo Gym.

About NVIDIA NeMo Gym
Concepts

Core components, configuration, verification and RL terminology.

Understanding Concepts for NeMo Gym
Ecosystem

Understand how NeMo Gym fits within the RL environment ecosystem.

RL Environment Ecosystem

Get Started#

Install and run NeMo Gym to start collecting rollouts.

Quickstart

Install, start servers, and collect your first rollouts in one page.

Quickstart
Detailed Setup Guide

Step-by-step installation with requirements, configuration, and troubleshooting.

Detailed Setup Guide
Rollout Collection

Generate batches of scored interactions and view them with the rollout viewer.

Rollout Collection

Environment Configuration#

Configure and customize environment components and prepare datasets.

Model Server

Configure LLM inference backends including vLLM.

Model Server
Data

Prepare and validate training datasets.

Data

Environment Tutorials#

Learn how to build custom training environments for various RL scenarios.

Creating Environments

Build a complete training environment from scratch.

Creating a Training Environment
Multi-Environment Training

Run multiple training environments simultaneously for rollout collection.

Multi-Environment Training

View all environment tutorials →

Training Tutorials#

Train models using NeMo Gym with your preferred RL framework.

RL (GRPO)

Hands-on tutorials with NeMo RL, TRL, Unsloth, and more.

Training Tutorials
SFT & DPO

Transform rollouts into SFT and DPO format.

Offline Training with Rollouts (SFT/DPO) - Experimental

View all training tutorials →

Infrastructure#

Deploy NeMo Gym and plan cluster resources for training.

Deployment Topology

Production deployment patterns and configurations.

Deployment Topology

Contribute#

Contribute to NeMo Gym development.

Contribute Environments

Contribute new environments or integrate existing benchmarks.

Contribute Environments
Integrate RL Frameworks

Implement NeMo Gym integration into a new training framework.

Training Framework Integration