-
Notifications
You must be signed in to change notification settings - Fork 6
/
asr_inference_offline.py
58 lines (50 loc) · 2.22 KB
/
asr_inference_offline.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
import argparse
import asyncio
import functools
from asr.utils import AudioReader
from asr.wav2vec2 import Wav2Vec2ASR
parser = argparse.ArgumentParser(
description="ASR with recorded audio (offline)")
parser.add_argument("--recording", "-rec", required=True,
help="path to recording file")
parser.add_argument("--model", "-m", default=None, required=False,
help="path to local saved model")
parser.add_argument("--processor", "-t", default=None, required=False,
help="path to local saved processor")
parser.add_argument("--output", "-out", required=False,
help="path to save resultant transcriptions")
parser.add_argument("--lm", "-l", default=None, required=False,
help="Trained lm folder path with unigram and bigram files")
parser.add_argument("--device", "-d", default='cpu', nargs='?', choices=['cuda', 'cpu'], required=False,
help="device to use for inferencing")
parser.add_argument("--beam_width", "-bw", default=1, type=int, required=False,
help="beam width to use for beam search decoder during inferencing")
parser.add_argument("--pretrained_model_name", "-pwmn", default="facebook/wav2vec2-base-960h",
type=str, required=False, help="Pretrained wav2vec2 model name")
args = parser.parse_args()
asr = Wav2Vec2ASR(device=args.device,
processor_path=args.processor,
model_path=args.model,
pretrained_model_name=args.pretrained_model_name,
beam_width=args.beam_width,
lm_path=args.lm)
print("Loading Models ...")
asr.load()
print("Models Loaded ...")
async def main():
loop = asyncio.get_running_loop()
reader = AudioReader(audio_path=args.recording,
sr=16000,
dtype="float32")
inputs, sr = reader.read()
transcriptions = await asr.transcribe(inputs, loop=loop)
print(transcriptions)
if args.output:
with open(args.output, "w") as f:
f.write(transcriptions)
if __name__ == "__main__":
print("Start Transcribing...")
try:
asyncio.run(main())
except KeyboardInterrupt:
print("Exited")