README: link the browser Space
Browse files
README.md
CHANGED
|
@@ -44,6 +44,8 @@ features from [WakeHuBERT tiny](https://huggingface.co/TigreGotico/wakehubert-ti
|
|
| 44 |
feature extractor distilled from HuBERT-base. A model scores the last 1.5 s of audio (75 feature frames) and
|
| 45 |
outputs a logit; its sigmoid is the probability that the window holds the wake word.
|
| 46 |
|
|
|
|
|
|
|
| 47 |
The repository holds one ONNX file per word under `models/`, and `models.json`, which lists each model's word,
|
| 48 |
featurizer, calibration, default threshold, SHA-256 and measured results. Every model carries the same facts in
|
| 49 |
its own ONNX metadata (`wake_word`, `pretrained_featurizer`, `default_threshold`, `window_frames`, `license`,
|
|
|
|
| 44 |
feature extractor distilled from HuBERT-base. A model scores the last 1.5 s of audio (75 feature frames) and
|
| 45 |
outputs a logit; its sigmoid is the probability that the window holds the wake word.
|
| 46 |
|
| 47 |
+
**Try them in your browser:** [WakeHuBERT wake-words Space](https://huggingface.co/spaces/OpenVoiceOS/wakehubert-wakewords-space) runs every model here on your microphone or an uploaded file, with a live score and the calibrated threshold. Everything runs locally in the browser.
|
| 48 |
+
|
| 49 |
The repository holds one ONNX file per word under `models/`, and `models.json`, which lists each model's word,
|
| 50 |
featurizer, calibration, default threshold, SHA-256 and measured results. Every model carries the same facts in
|
| 51 |
its own ONNX metadata (`wake_word`, `pretrained_featurizer`, `default_threshold`, `window_frames`, `license`,
|