Story

arxiv_cs_cl ยท Nov 19, 2026 ยท paper

Source brief

One Schedule Does Not Fit All: Adaptive Inference Schedules for Diffusion Language Models

arxiv.orgNov 19, 2026
original source linked

In brief

Diffusion language models typically use a fixed denoising schedule across workloads. We show that the cheapest schedule that preserves quality depends on the task, model, and sampler. We propose Task-Calibrated Denois...

Feed lens
evaluation

Continue reading

Read the original at arxiv.org โ†’Open in live feedRead that dayโ€™s brief

Earlier in this thread 4 items