Tutorials | NVIDIA NeMo AutoModel

Tutorials walk through a specific model or task, including its data, training configuration, and launch steps. Choose a tutorial below or browse the model groups in the sidebar.

For instructions on adapting a training workflow across models, use Training Recipes. For support details and available configurations, use Model Coverage.

Large Language Models

Tutorial	Focus
Function Calling with FunctionGemma	Fine-tune for function calling
Multi-Turn Agent SFT	Train on multi-turn tool-calling conversations
NanoGPT Pretraining	Pretrain a GPT model
DeepSeek-V4 Flash	Fine-tune DeepSeek-V4 Flash
Hy3-preview	Fine-tune Hy3-preview
Nemotron-3-Ultra-550B	Fine-tune Nemotron-3-Ultra
DiffusionGemma	Fine-tune a diffusion language model

Vision and Multimodal Models

Tutorial	Focus
Gemma 3 / 3n	Fine-tune Gemma 3 and Gemma 3n
Gemma 4 31B	Fine-tune Gemma 4
Qwen3.5-VL	Fine-tune Qwen3.5 vision-language models
Qwen3.8-27B	Run full SFT or LoRA on MedPix-VQA
Nemotron-Omni	Fine-tune Nemotron-Omni
Nemotron 3.5 Super VL	Fine-tune Nemotron 3.5 Super VL
Nemotron 3.5 Super VL: Text-to-SQL	Adapt the model with LoRA
Nemotron 3.5 Super VL: Continued Pretraining	Continue pretraining on FineWeb
Mistral Medium 3.5 VL	Fine-tune Mistral Medium 3.5
MiniMax-M3	Fine-tune MiniMax-M3 with full SFT or LoRA
Step-3.7-Flash	Fine-tune Step-3.7-Flash

Audio

Tutorial	Focus
ASR with Qwen3-Omni	Fine-tune for automatic speech recognition

Choose Where to Run

After choosing a tutorial, use Run Jobs for local workstations, Slurm, NeMo Run, SkyPilot, Kubernetes, and Databricks.