Model CoverageLarge Language Models (LLMs)

DeepSeek V4 Flash

View as Markdown

DeepSeek V4 Flash is DeepSeek’s latest fine-grained Mixture-of-Experts language model. It uses a 43-layer all-MoE backbone with 256 routed experts plus one shared expert per block, top-6 routing, and a hybrid per-layer attention zoo (SWA / CSA / HCA) selectable through compress_ratios. The first num_hash_layers blocks use a hash-clustering gate, and every block maintains hc_mult=4 Hyper-Connection streams mixed via a learned col-norm-first Sinkhorn router.

TaskText Generation (MoE)
ArchitectureDeepseekV4ForCausalLM
Parametersfine-grained MoE, 256 routed + 1 shared expert
HF Orgdeepseek-ai

Available Models

  • DeepSeek-V4-Flash

Architecture

  • DeepseekV4ForCausalLM

Example HF Models

ModelHF ID
DeepSeek V4 Flashdeepseek-ai/DeepSeek-V4-Flash

Example Recipes

RecipeDescription
deepseek_v4_flash_hellaswag.yamlSFT — DeepSeek V4 Flash on HellaSwag with pipeline parallelism

Try with NeMo AutoModel

1. Install (full instructions):

$pip install nemo-automodel

2. Clone the repo to get the example recipes:

$git clone https://github.com/NVIDIA-NeMo/Automodel.git
$cd Automodel

The full 43-layer schedule requires a multi-node run; see the recipe yaml header for ep_size / pp_size guidance. See the Launcher Guide for multi-node setup.

3. Run the recipe from inside the repo:

$automodel --nproc-per-node=8 examples/llm_finetune/deepseek_v4/deepseek_v4_flash_hellaswag.yaml

1. Pull the container and mount a checkpoint directory:

$docker run --gpus all -it --rm \
>  --shm-size=8g \
>  -v $(pwd)/checkpoints:/opt/Automodel/checkpoints \
>  nvcr.io/nvidia/nemo-automodel:26.04.00

2. Navigate to the AutoModel directory (where the recipes are):

$cd /opt/Automodel

3. Run the recipe:

$automodel --nproc-per-node=8 examples/llm_finetune/deepseek_v4/deepseek_v4_flash_hellaswag.yaml

See the Installation Guide and LLM Fine-Tuning Guide.

Fine-Tuning

See the Fine-Tune DeepSeek V4 Flash guide and the Large MoE Fine-Tuning Guide.

Hugging Face Model Cards