Westlake AGI Lab releases eDMD, a new method for few-step diffusion distillation
_akhaliq · x · 2026-11-17
A new paper, "eDMD: Endpoint Distribution Matching for Few-Step Diffusion Distillation" (arXiv 2611.17194), from Westlake AGI Lab builds on distribution matching distillation (DMD), where an auxiliary fake score model tracks the student's distribution.
- eDMD adds endpoint distribution matching to improve few-step distillation quality, aiming to compress diffusion models into very few inference steps.
- Sample outputs span photorealistic, anime and illustration styles.
- Code and paper page are open on GitHub and Hugging Face.
Related event: Westlake AGI Lab Open-Sources eDMD for Few-Step Diffusion Distillation(3 posts)→
More from Research
- Benchmark: scaffold-assisted AI writing fools Pangram detector 78% of the time — amplifiedamp · 2026-11-17
- Oxford Internet Institute PhD thesis on human-AI alignment dynamics passes viva — natashajaques · 2026-11-17
- Nature paper uses AI to decode regulatory programs behind glioblastoma plasticity — anshulkundaje · 2026-11-17
- Vault-LD: markdown wiki + YAML-LD knowledge graph as an open standard for agent memory — _AustinCalvert_ · 2026-11-17
- Personal benchmark 'I don't know' fine-tunes open models via Tinker to test if models know when they don't know — brandon_galang · 2026-11-17
- MemInStream benchmark separates real memory from mere recognition in streaming video — Zhenqi He · 2026-11-17