Skip to main content
Ctrl+K
NeMo Gym - Home NeMo Gym - Home

NeMo Gym

  • GitHub
NeMo Gym - Home NeMo Gym - Home

NeMo Gym

  • GitHub

Table of Contents

  • Home

About

  • Overview
  • Concepts
    • Training Approaches
    • Environment Components
    • Configuration System
    • Architecture
    • Task Verification
    • Key Terminology
  • Ecosystem

Get Started

  • Quickstart
  • Detailed Setup Guide
  • Rollout Collection

Agent Server

  • Overview

Model Server

  • Overview
  • vLLM

Resources Server

  • Overview

Data

  • Overview
  • Prepare and Validate
  • Download from Hugging Face

Environment Tutorials

  • Overview
  • Single-Step Environment
  • Multi-Step Environment
  • Stateful Environment
  • Real-World Environment
    • Generating Training Data
    • Resources Server Implementation
  • Integrate external libraries
  • Add a benchmark
  • Design a customer evaluation

Training Tutorials

  • Overview
  • NeMo RL
    • About Workplace Assistant
    • Gym Configuration
    • NeMo RL Configuration
    • Setup
    • Single Node Training
    • Multi-Node Training
  • Unsloth
  • Multi-Environment Training
  • Offline Training (SFT/DPO)

Model Recipes

  • Overview
  • Nemotron 3 Nano
  • Nemotron 3 Super

Infrastructure

  • Overview
  • Deployment Topology
  • Engineering Notes
    • Responses API
    • SWE RL Case Study
    • aiohttp vs httpx

Reference

  • Configuration
  • RL Framework Compatibility
  • CLI Commands
  • API Reference
    • nemo_gym
      • nemo_gym.config_types
      • nemo_gym.dataset_orchestrator
      • nemo_gym.package_info
      • nemo_gym.global_config
      • nemo_gym.hf_utils
      • nemo_gym.train_data_utils
      • nemo_gym.base_resources_server
      • nemo_gym.openai_utils
      • nemo_gym.server_utils
      • nemo_gym.cli
      • nemo_gym.server_status
      • nemo_gym.reward_profile
      • nemo_gym.gitlab_utils
      • nemo_gym.rollout_collection
      • nemo_gym.base_responses_api_model
      • nemo_gym.cli_setup_command
      • nemo_gym.benchmarks
      • nemo_gym.base_responses_api_agent
      • nemo_gym.profiling
  • FAQ

Troubleshooting

  • Configuration Errors

Contribute

  • Overview
  • Development Setup
  • Environments
    • New Environment
  • Integrate RL Frameworks
    • Generation Backend
    • On-Policy Corrections
    • Integration Footprint
    • Success Criteria
  • API Reference
  • nemo_gym
  • nemo_gym.profiling

nemo_gym.profiling#

Module Contents#

Classes#

Profiler

API#

class nemo_gym.profiling.Profiler(/, **data: typing.Any)[source]#

Bases: pydantic.BaseModel

name: str#

None

base_profile_dir: pathlib.Path#

None

required_str: Optional[str]#

None

model_post_init(context)[source]#
_check_for_dot_installation() → None[source]#
start() → None[source]#
stop() → None[source]#
dump() → None[source]#

previous

nemo_gym.base_responses_api_agent

next

FAQ

On this page
  • Module Contents
    • Classes
    • API
      • Profiler
        • name
        • base_profile_dir
        • required_str
        • model_post_init()
        • _check_for_dot_installation()
        • start()
        • stop()
        • dump()
NVIDIA NVIDIA
Privacy Policy | Manage My Privacy | Do Not Sell or Share My Data | Terms of Service | Accessibility | Corporate Policies | Product Security | Contact

Copyright © 2025, NVIDIA Corporation.