E-MoE: Enhanced Mixture-of-Experts for Non-Factorized Diffusion Language Models. LM1B checkpoints: E-MoE and the MDLM / MDLM-MoE baselines.
-
E-MoE: Enhanced Mixture-of-Experts for Non-Factorized Diffusion Language Models
Paper • 2609.37533 • Published • 67 -
ArsenyIvanov/emoe-lm1b
Text Generation • 0.5B • Updated • 10 -
ArsenyIvanov/mdlm-moe-lm1b
Text Generation • 0.5B • Updated • 11 -
ArsenyIvanov/mdlm-lm1b
Text Generation • 0.1B • Updated • 10
