Mistral-Small-4-119B-2603 | NVIDIA NeMo AutoModel

Mistral-Small-4-119B-2603 has a checked-in NeMo AutoModel recipe for text generation. The Hugging Face configuration declares the Mistral3ForConditionalGeneration architecture.

Fine-Tune Mistral-Small-4-119B-2603

Follow the installation instructions, then run the recipe from the repository root:

automodel examples/llm_benchmark/mistral/mistral_small_4_te_deepep.yaml --nproc-per-node 8

Use the Slurm launcher guide for the multi-node run.

Choose a Workflow

Goal	Start Here
Run the primary recipe for this model	Use the recipe configuration.
Prepare your environment	Follow the installation instructions.

Configuration

Setting	Configuration
Hardware	-
Strategy	FSDP2; tp_size=1, pp_size=1, cp_size=1, ep_size=64
Nodes	8
Features	Activation checkpointing (activation_checkpointing=true), Transformer Engine, HybridEP (dispatcher=hybridep)
Advanced	attention=te, linear=te, experts=torch_mm, dispatcher=hybridep

Model Reference

Model Architecture

Property	Value
Task	Text generation
Architecture	`Mistral3ForConditionalGeneration`

Available Models

mistralai/Mistral-Small-4-119B-2603

LLM fine-tuning guide