> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.nvidia.com/nemo/automodel/_mcp/server.

# Recipes & E2E Examples

## Docs

- [Recipes and End-to-End Examples](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/overview.md): 
- [Supervised Fine-Tuning (SFT) and Parameter-Efficient Fine-Tuning (PEFT) with NeMo AutoModel](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/sft-peft.md): Fine-tune language models with NeMo AutoModel using SFT or PEFT (LoRA), from install and YAML config through training, evaluation, publish, and vLLM deploy.
- [Function Calling with NeMo AutoModel using FunctionGemma](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/function-calling.md): 
- [Multi-Turn Agent (Tool-Calling) SFT with NeMo AutoModel](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/agent-sft.md): End-to-end multi-turn, function-calling SFT on Qwen2.5-3B with the glaive_toolcall and SynthTraces datasets
- [Knowledge Distillation with NeMo AutoModel](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/knowledge-distillation.md): 
- [Fine-Tune Large MoE LLMs](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/large-moe-fine-tuning.md): Fine-tune large Mixture-of-Experts (MoE) LLMs with NeMo AutoModel using HellaSwag recipes for MiniMax-M2.5, GLM-5, Step-3.5 Flash, and DeepSeek-V3.2.
- [DeepSeek-V4 Flash](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/deepseek-v4-flash.md): 
- [Hy3-preview](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/hy3-preview.md): 
- [Fine-Tune Nemotron-3-Ultra-550B](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/nemotron-3-ultra.md): End-to-end LoRA PEFT fine-tuning of Nemotron-3-Ultra-550B on HellaSwag using NeMo AutoModel on GB200 and H100.
- [Pretrain Megatron Core Datasets with NeMo AutoModel](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/pretraining.md): 
- [LLM Pre-Training with NeMo AutoModel](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/nanogpt-pretraining.md): 
- [Sequence Classification (SFT/PEFT) with NeMo AutoModel](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/sequence-classification.md): Train a sequence classification model with NeMo AutoModel using GLUE MRPC, RoBERTa, and optional LoRA.
- [Retrieval Fine-Tuning (Bi-Encoder and Cross-Encoder)](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/retrieval-finetune.md): Fine-tune bi-encoder and cross-encoder retrieval models in NeMo AutoModel with distributed training, low-rank adaptation (LoRA), hard-negative mining, and evaluation.
- [Fine-Tune Gemma 3 and Gemma 3n](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/gemma-3-3n.md): 
- [Fine-Tune Gemma 4 31B on CORD-v2 Receipts — End-to-End Guide](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/gemma-4.md): 
- [Fine-Tune Qwen3.5-VL](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/qwen3-5-vl.md): 
- [Fine-Tune Qwen3.8-27B](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/qwen3-8-27b.md): End-to-end full SFT and LoRA recipes for Qwen3.8-27B on MedPix-VQA.
- [Nemotron-Omni](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/nemotron-omni.md): 
- [NVIDIA Nemotron 3.5 Super VL](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/nemotron-3-5-super-vl.md): Fine-tune Nemotron 3.5 Super VL on CORD-v2 for receipt field extraction, then run inference and evaluation.
- [NVIDIA Nemotron 3.5 Super VL: Text-to-SQL with LoRA](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/nemotron-3-5-super-vl-text-to-sql.md): LoRA fine-tune Nemotron 3.5 Super VL on Spider so it answers natural-language questions with SQL for databases it has never seen, then evaluate on the Spider validation split.
- [NVIDIA Nemotron 3.5 Super VL: Continued Pretraining on FineWeb](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/nemotron-3-5-super-vl-continued-pretraining.md): Prepare FineWeb as Megatron indexed data, continue training the Super 3.5 VL language backbone, and track training and validation loss with Weights & Biases.
- [Mistral Medium 3.5 VL](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/mistral-medium-3-5.md): 
- [MiniMax-M3](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/minimax-m3.md)
- [Fine-Tune Step-3.7-Flash](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/step-3-7.md): End-to-end SFT and LoRA setup for Stepfun AI's Step-3.7-Flash 198B-A13B MoE vision-language model.
- [Fine-Tune Qwen3-Omni for ASR](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/qwen3-omni-asr.md): End-to-end ASR fine-tuning of Qwen3-Omni-30B on Hugging Face audio datasets with NeMo AutoModel.
- [Diffusion Model Fine-Tuning with NeMo AutoModel](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/diffusion-fine-tuning.md): Fine-tune text-to-image and text-to-video diffusion models with NeMo AutoModel using flow matching, from dataset prep through multi-node training and inference.
- [Fine-Tuning DiffusionGemma](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/diffusiongemma.md): Supervised fine-tuning of DiffusionGemma 26B-A4B with full SFT and LoRA on NeMo AutoModel.
- [Train an EAGLE Drafter for Speculative Decoding — End-to-End Guide](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/eagle-speculative-decoding.md): Step-by-step guide for training EAGLE-3 and EAGLE-1 speculative decoding drafters with NeMo AutoModel, from data preparation through SGLang serving.
- [Train a DFlash Drafter for Speculative Decoding](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/dflash-speculative-decoding.md): Train a DFlash block-parallel speculative-decoding drafter with NeMo AutoModel, from data through multi-GPU training, including the DFlash 2, Domino, and JetSpec variants.
- [Train a DSpark Drafter for Speculative Decoding](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/dspark-speculative-decoding.md): Train a DSpark semi-autoregressive speculative-decoding drafter with NeMo AutoModel, from data through 2-GPU FSDP2 training.
- [dLLM Fine-Tuning](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/dllm-fine-tuning.md): Fine-tune diffusion language models (dLLMs), including LLaDA, DiffusionGemma, DFlash, I-DLM, and Uno, with NeMo AutoModel.
- [Quantization-Aware Training (QAT) in NeMo Automodel](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/qat.md): 
- [Model Training on Databricks](https://docs.nvidia.com/nemo/automodel/nightly/recipes-e2e-examples/databricks.md):