LLaDA

model
Ant Group
Experimental diffusion language model from Inclusion AI. Explores an alternative to autoregressive generation. LLaDA-MoE extends the approach with Mixture-of-Experts.
Paper (arXiv)GitHubProject PageHuggingFace (LLaDA-MoE)

Outputs 2

LLaDA

model

Diffusion language model from Inclusion AI.

Paper (arXiv)GitHub

arXiv: 2502.09992

LLaDA-MoE

model

Mixture-of-Experts variant of LLaDA.

Paper (arXiv)HuggingFace
Architecture MOE
generationarchitectureresearch

Related