AlazarM/hey-native-wakeword
Hey Native — wake-word dataset Synthetic 16 kHz mono audio for training a "Hey Native" wake-word detector. data/positive/ — utterances of "Hey Native" (label 1) data/negative/ — general speech, not the wake word (label 0) data/hard_negative/ — near-miss confusables, e.g. "hey navy", "hey maybe" (label 0) metadata.csv — file_name, label, label_name, text, source_model, mos_p808 5,000 positives · 6,000 negatives · 1,050 hard negatives.
Add 1050 hard-negative confusables (data/hard_negative) + 3-class metadata + README (part 4)
Add 1050 hard-negative confusables (data/hard_negative) + 3-class metadata + README (part 3)
Add 1050 hard-negative confusables (data/hard_negative) + 3-class metadata + README (part 2)
Add 1050 hard-negative confusables (data/hard_negative) + 3-class metadata + README
README: reflect 5000 positives + 6000 negatives
Add 6000 general-speech negatives (data/negative) + full metadata (part 12)
Add 6000 general-speech negatives (data/negative) + full metadata (part 11)
Add 6000 general-speech negatives (data/negative) + full metadata (part 10)
Add 6000 general-speech negatives (data/negative) + full metadata (part 9)
Add 6000 general-speech negatives (data/negative) + full metadata (part 8)
Add 6000 general-speech negatives (data/negative) + full metadata (part 7)
Add 6000 general-speech negatives (data/negative) + full metadata (part 6)
Add 6000 general-speech negatives (data/negative) + full metadata (part 5)
Add 6000 general-speech negatives (data/negative) + full metadata (part 4)
Add 6000 general-speech negatives (data/negative) + full metadata (part 3)
Add 6000 general-speech negatives (data/negative) + full metadata (part 2)
Add 6000 general-speech negatives (data/negative) + full metadata
Positives: 500 -> 5000 (multi-model, quality-gated, deduped) (part 11)
Positives: 500 -> 5000 (multi-model, quality-gated, deduped) (part 10)
Positives: 500 -> 5000 (multi-model, quality-gated, deduped) (part 9)
Positives: 500 -> 5000 (multi-model, quality-gated, deduped) (part 8)
Positives: 500 -> 5000 (multi-model, quality-gated, deduped) (part 7)
Positives: 500 -> 5000 (multi-model, quality-gated, deduped) (part 6)
Positives: 500 -> 5000 (multi-model, quality-gated, deduped) (part 5)
Positives: 500 -> 5000 (multi-model, quality-gated, deduped) (part 4)
Positives: 500 -> 5000 (multi-model, quality-gated, deduped) (part 3)
Positives: 500 -> 5000 (multi-model, quality-gated, deduped) (part 2)
Positives: 500 -> 5000 (multi-model, quality-gated, deduped)
Reorganize: data/positive + metadata.csv, minimal README (part 3)
Reorganize: data/positive + metadata.csv, minimal README (part 2)
Reorganize: data/positive + metadata.csv, minimal README
initial commit