Open snakers4 opened 3 years ago
Initial models, examples, utils for VAD only uploaded (no number detector or language classifier yet)
First readable public release
Added VAD latency and throughput metrics
Updated VAD quality
Before / after (precision / recall)
Added < 250ms compatibility
Added number detector
Language detector example, readme update + FAQ
Audiotok benchmarks added Looks like all energy based solutions are kind of similar
Added a utility to tune the VAD params properly for a domain
Some final benchmarks posted here - https://github.com/pyannote/pyannote-audio/issues/604#issue-798003383 Probably we are done with benchmarks for now
Added micro (10k params, 100x smaller) VAD models
Added micro (10k params, 100x smaller) VAD models for 8 kHz audio
improved language classifier
updated further reading section
8000 Hz
and 16000 Hz
are supported;Please see the new examples.
New get_speech_timestamps
is a simplified and unified version of the old deprecated get_speech_ts
or get_speech_ts_adaptive
methods.
speech_timestamps = get_speech_timestamps(wav, model, sampling_rate=16000)
New VADIterator
class serves as an example for streaming tasks instead of old deprecated VADiterator
and VADiteratorAdaptive
.
vad_iterator = VADIterator(model)
window_size_samples = 1536
for i in range(0, len(wav), window_size_samples):
speech_dict = vad_iterator(wav[i: i+ window_size_samples], return_seconds=True)
if speech_dict:
print(speech_dict, end=' ')
vad_iterator.reset_states()
This summarises new progress well
We finally were able to port a model to ONNX:
jit
model now can handle 8, 16, 32 and 48 kHz directly (change implemented within the model itself);onnx
model as well, but only via external wrappers (we just use each n-th sample for higher sampling rates);If you are using the VAD in a:
Do not forget to disable gradients in EACH process and / or thread. Otherwise memory may leak noticeably.
Changes:
audio_forward
method for one-line processing of a single or multiple audio without postprocessingIt is worth posting this chart:
window_size_samples
is deprecated - now the VAD only works with fixed size window;Full Changelog: https://github.com/snakers4/silero-vad/compare/v5.0...v5.1
Just a handy issue to be notified of latest changes and micro-releases (we will mostly changing the models)