National University of Singapore

Trends

• Efficiency methods cut the compute and memory models need

Items

1. Apr 9

   DMax speeds up diffusion LLM decoding

   NUS recasts masked-diffusion decoding as progressive self-refinement so the model can correct its own predictions, raising tokens per forward pass on GSM8K from 2.04 to 5.47 with no accuracy loss and reaching 1,338 tokens/s on two H200s.