diff options
| -rw-r--r-- | AGENTS.md | 12 |
1 files changed, 7 insertions, 5 deletions
@@ -115,9 +115,10 @@ Everything under `/data/voice-models/` (outside the repo, kept apart from `/data/LLM-models/`): - `whisper/ggml-base.bin`, 142 MB, multilingual. SHA1 matches upstream - (`465707469ff3a37a2b9b8d8f89f2f99de7299dac`). + (`465707469ff3a37a2b9b8d8f89f2f99de7299dac`). Deleted 2026-09-26. - `piper/it_IT-paola-medium.onnx` and `piper/en_US-lessac-medium.onnx`, - 61 MB each, plus their `.onnx.json`. Both output 22050 Hz. + 61 MB each, plus their `.onnx.json`. Both output 22050 Hz. Deleted + 2026-09-26. - `samples/it.wav`, `samples/en.wav`: 10 s, 16 kHz mono mic recordings used for the STT comparison. Personal voice data, never copy into the repo. @@ -231,7 +232,7 @@ drives `arecord` on/off, clip POSTed to `/inference`. plausible Italian and scores *higher* (-0.164 vs -0.287). Moving to `ggml-small`. - `ggml-small.bin` (466 MB, SHA1 `55356645c2b361a969dfd0ef2c5a50d530afd8d5`, - matches upstream) in `/data/voice-models/whisper/`. On the Arc via + matches upstream) in `/data/voice-models/whisper/` (deleted 2026-09-26). On the Arc via whisper-server: **~0.35 s per pass** on a 10 s clip. Transcripts clean on both samples (Italian and English self-introductions, all words right). - small's detector *still* mislabels the English sample (la 0.45, it 0.31, @@ -368,8 +369,9 @@ no particular order: - **Latency:** STT ~1.2 s is the biggest share; the double encoder pass (detect + transcribe) is the target if it matters. - **Tools** (time, weather, calendar) so the model stops guessing facts. -- Piper leftovers: package unused, voices in `/data/voice-models/piper/` - (~120 MB) can go. ggml-base and ggml-small too, once turbo is settled. +- Piper leftovers: piper voices and ggml-base/small deleted 2026-09-26. + Installed `piper-tts` and `python3-pathvalidate` packages still to remove + (root). - OpenVINO in the whisper.cpp package: runs on CPU only here (no Intel compute-runtime), decision goes to `my-slackbuilds`. |