LLaDA2.1-mini

LLaDA 2.1 is a mixture-of-experts diffusion language model supported through its Hugging Face remote-code implementation.

Fine-Tune LLaDA2.1-mini

From the repository root, run:

uv run automodel --nproc-per-node=8 examples/dllm_sft/llada2_sft.yaml

Model Reference

Model Architecture

Example Model and Recipes

See the dLLM Fine-Tuning Guide.