mirror of
https://github.com/alphacep/vosk-api.git
synced 2026-02-04 14:38:26 +08:00
4209f3a9fe
* methods get_model_by_name, get_model_by_lang, get_model were added into the model class * importing modules changed to using components; introduced constant MODELS_HOME_DIR; simplified code * added new model folders into init; changed samples and transcriber bin for new mode loader * changed back in cli.py lang arg to args.lang * added 3 directories instead of 1 to check for models * cli.py: added 3 args instead of 1 for model; __init__.py: changed script get_model_path for run get_model_by_name/lang inside current directory * deleted default env var * cli.py: changed arg_name; __init__.py: changed const name, changed model loading only for last directory * deleted unused method * changed by_name, by_lang methods, added download_model method * deleted env variable initialization * deleted print() * deteled unused modules * added progress_bar, added folder AppData/Local/vosk for model search * changed download_model methond; added my_hook method
28 lines
748 B
Python
Executable File
28 lines
748 B
Python
Executable File
#!/usr/bin/env python3
|
|
|
|
from vosk import Model, KaldiRecognizer
|
|
import sys
|
|
import os
|
|
import wave
|
|
|
|
wf = wave.open(sys.argv[1], "rb")
|
|
if wf.getnchannels() != 1 or wf.getsampwidth() != 2 or wf.getcomptype() != "NONE":
|
|
print ("Audio file must be WAV format mono PCM.")
|
|
exit (1)
|
|
|
|
model = Model(lang="en-us")
|
|
|
|
# You can also specify the possible word or phrase list as JSON list, the order doesn't have to be strict
|
|
rec = KaldiRecognizer(model, wf.getframerate(), '["oh one two three four five six seven eight nine zero", "[unk]"]')
|
|
|
|
while True:
|
|
data = wf.readframes(4000)
|
|
if len(data) == 0:
|
|
break
|
|
if rec.AcceptWaveform(data):
|
|
print(rec.Result())
|
|
else:
|
|
print(rec.PartialResult())
|
|
|
|
print(rec.FinalResult())
|