OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Models

Nemotron

Nemotron is NVIDIA-NeMo’s public developer repository of modular training steps, full training recipes, datasets, model-use and deployment cookbooks, and end-to-end examples for building with NVIDIA Nemotron models.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Nemotron — OpenViST project profile in AI Models.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use the repository when you need Nemotron-specific training materials, model-use guidance or application examples, and select the section based on whether you are composing pipeline stages, following a full recipe, deploying a model or building an application.

Who it is for

  • Developers preparing or fine-tuning Nemotron models.
  • Practitioners looking for guidance on model deployment or inference.
  • Teams exploring application examples involving RAG, agents or tool use.
  • Claude Code users who want the repository’s nemotron-customize skill to assemble a pipeline configuration.

Documented uses

  • Find modular steps for data preparation, training, evaluation and related pipeline stages.
  • Follow a complete training recipe from raw data through model training.
  • Consult model-specific cookbooks for deployment and inference.
  • Study end-to-end examples when building applications with Nemotron models.
  • Generate YAML configuration for a pipeline using the Claude Code plugin.
Before you choose

Requirements & limits

The repository is public and identifies its license as Apache 2.0. The README lists Python 3.10+ as a requirement. The documented Claude Code plugin requires installation and must be invoked from the repository root.

  • Training recipes are Python packages with configurations, scripts and evaluation; usage cookbooks are Jupyter notebooks; use-case examples are notebooks and scripts.
  • The repository distinguishes training and evaluation building blocks from model deployment and application examples.
  • Cookbooks cover different model sizes and deployment contexts, including resource-constrained environments and multi-GPU deployments.
  • The nemotron-customize plugin emits pipeline YAML and checks artifact wiring; its documented role is configuration generation, not pipeline execution.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Prepare a pipeline configuration with nemotron-customize

Use the repository’s Claude Code skill to plan a Nemotron steps pipeline and generate its YAML configuration.

Before you start

  • A Nemotron repository checkout containing pyproject.toml and src/nemotron/steps/ at its root.
  • Claude Code and the nemotron-customize plugin, installed using the repository’s documented marketplace commands.

Steps

  1. Add the NVIDIA/Nemotron marketplace in Claude Code with /plugin marketplace add NVIDIA/Nemotron.

  2. Install the plugin with /plugin install nemotron-customize@nvidia-nemotron.

  3. Start Claude Code from the Nemotron checkout root, then invoke /nemotron-customize.

  4. Use the skill to plan the step DAG and artifact wiring and emit the requested pipeline YAML configuration.

Expected outcome

A planned pipeline and YAML configuration files for the requested workflow.

Guide limitations

  • The skill must run from the repository root; invoking it from a subdirectory causes file reads to fail.
  • The plugin marketplace installs only nemotron-customize; other folders under skills are not loaded as plugins.
  • The documented skill creates configuration; this guide does not cover running the pipeline.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
14 October 2026
Original website title
NVIDIA-NeMo/Nemotron · Official homepage title · retained in its source language
How we review sourcesSuggest a correction