AI · Technology · 2026-10-08 16:59
A tiny open-source speech recognition model just 16.9 MB in size, supporting seven languages with 11ms first-token latency on CPU, suited for resource-constrained edge devices.
Auto-generated summary for reference only. RSS