NVIDIA-NeMo/Nemotron resource

Health v2 · maintenance only

Adoption not part of the score

1983 stars · 404 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

NVIDIA's developer asset hub for the Nemotron family of open LLMs, providing reproducible training recipes (pretraining, SFT, RL), usage cookbooks, datasets, and end-to-end reference examples. It includes a CLI pipeline runner that submits stage-based training jobs to Slurm clusters via NeMo-Run, plus deployment guides for TensorRT-LLM, vLLM, SGLang, and NIM.

Use cases

• fine-tune a Nemotron model with supervised fine-tuning
• run a full pretraining to RL training pipeline on a Slurm cluster
• find training recipes for hybrid Mamba-Transformer MoE models
• deploy Nemotron models with vLLM or TensorRT-LLM
• build RAG agents and multi-agent systems with Nemotron
• learn how to reproduce NVIDIA's open model training
• prepare datasets for LLM pretraining and alignment

When to choose

• you want to train, fine-tune, or align Nemotron models with proven recipes
• you need reproducible end-to-end LLM training pipelines with transparent data and weights
• you want cookbooks and examples for deploying Nemotron on GPUs or via NIM
• you're building agentic AI applications on NVIDIA open models

When to avoid

• you need a general-purpose training framework unrelated to Nemotron models
• you only want to run inference without training or deployment guidance
• you lack GPU or Slurm cluster resources for the training pipelines

Facets

learning-resource · maturity active

llm-training machine-learning rag agent-framework developer-tools large-language-models machine-learning artificial-intelligence tutorials developer-tools python cli nvidia nemotron training-recipes fine-tuning reinforcement-learning cookbooks jupyter-notebooks slurm open-models ai-agents linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="NVIDIA-NeMo/Nemotron")