Skip to content

[perf] Fun-ASR-Nano fine-tuning on B200: 0.244 s to 0.068 s per step at 1 GPU, mostly from cuDNN SDPA plan builds on every new batch shape - #3705

Merged
LauraGPT merged 3 commits into
modelscope:mainfrom
TarzanZhao:perf/fun-asr-nano-training-speed
Sep 21, 2026
Merged

[perf] Fun-ASR-Nano fine-tuning on B200: 0.244 s to 0.068 s per step at 1 GPU, mostly from cuDNN SDPA plan builds on every new batch shape#3705
LauraGPT merged 3 commits into
modelscope:mainfrom
TarzanZhao:perf/fun-asr-nano-training-speed

Commits

  1. Commits on Sep 15, 2026

  2. Commits on Sep 20, 2026