Model CoverageLarge Language Models (LLMs)

Qwen3 MoE

View as Markdown

Qwen3 MoE is the Mixture-of-Experts variant of the Qwen3 series from Alibaba Cloud, activating a small fraction of parameters per token for efficient large-scale training.

TaskText Generation (MoE)
ArchitectureQwen3MoeForCausalLM
Parameters30B – 235B total
HF OrgQwen

Available Models

  • Qwen3-30B-A3B: 30B total parameters, 3B activated per token
  • Qwen3-235B-A22B: 235B total parameters, 22B activated per token

Architecture

  • Qwen3MoeForCausalLM

Example HF Models

ModelHF ID
Qwen3 30B A3BQwen/Qwen3-30B-A3B
Qwen3 235B A22BQwen/Qwen3-235B-A22B

Example Recipes

RecipeDescription
qwen3_moe_30b_te_deepep.yamlSFT — Qwen3 MoE 30B with TE + DeepEP
qwen3_moe_30b_lora.yamlLoRA — Qwen3 MoE 30B

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

3. Run the recipe from inside the repo:

$automodel --nproc-per-node=8 examples/llm_finetune/qwen/qwen3_moe_30b_te_deepep.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/qwen/qwen3_moe_30b_te_deepep.yaml

See the Installation Guide and LLM Fine-Tuning Guide.

Fine-Tuning

See the LLM Fine-Tuning Guide and the Large MoE Fine-Tuning Guide.

Hugging Face Model Cards