ASH ASR workbench. Moonshine LibriSpeech scores and the asr.jezzick.net report.
- Python 57.2%
- HTML 42.8%
| config | ||
| reports | ||
| scripts | ||
| HANDOFF.md | ||
| README.md | ||
asr-workbench
ASH-local speech recognition bench. The Rowan inference-lab-benchmarks harness was the starting checkpoint for the result shape. The runs, audio, and scored report are on ASH.
What this scores
Moonshine Streaming Tiny Q8 (private-moonshine-asr) on the official LibriSpeech test sets:
| Backend | Set | Corpus WER | Edits / words | RTF |
|---|---|---|---|---|
CPU 127.0.0.1:18197 |
test-clean | 4.84% | 2,546 / 52,576 | 0.0067 |
GPU1 127.0.0.1:18198 |
test-clean | 4.87% | 2,560 / 52,576 | 0.0071 |
| CPU | test-other | 12.94% | 6,773 / 52,343 | 0.0070 |
| GPU1 | test-other | 13.12% | 6,870 / 52,343 | 0.0076 |
Corpus WER is total word edits divided by total reference words. Tokenizer: lowercase letters, digits, apostrophes. This is not the official SCTK scorer.
The public page is reports/index.html, served at asr.jezzick.net. It does not include LibriSpeech audio or any private recording.
Layout
scripts/librispeech_eval.py— transcribe a LibriSpeech subset and append jsonlscripts/summarize_librispeech.py— corpus WER summaryreports/— scored jsonl, summary JSON, and the public HTMLconfig/cpu-asr.json— the CPU audio.cpp server config used for the run
LibriSpeech itself stays on ASH at /mnt/models/librispeech-20260928 and is not in this repo.