This repository has been archived by the owner on Feb 19, 2024. It is now read-only.
/
transcribe_args.py
150 lines (131 loc) · 4.26 KB
/
transcribe_args.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
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
import argparse
from pathlib import Path
import hashlib
import dataclasses
import sys
from typing import Union, Dict
from modal.gpu import STRING_TO_GPU_CONFIG
@dataclasses.dataclass
class TranscribeConfig:
filename: Union[str, None]
video_url: Union[str, None]
url: Union[str, None]
out: Union[str, None]
model: str
min_segment_len: float
min_silence_len: float
force: Union[bool, None] = None
gpu: Union[str, None] = None
initial_prompt: Union[str, None] = None
def merge(self, other: Dict):
return dataclasses.replace(self, **other)
def identifier(self):
possible = [self.filename, self.video_url, self.url]
only_one = list(map(bool, possible)).count(True) == 1
if not only_one:
raise ValueError("Specify exactly one of: filename, video_url, or url")
if self.filename:
file = Path(self.filename)
source = file.name
elif self.url:
source = self.url
elif self.video_url:
source = self.video_url
else:
raise ValueError("Must specify either filename or url")
return source, hashlib.md5(source.encode("utf-8")).hexdigest()
@dataclasses.dataclass
class WhisperModel:
name: str
params: str
all_models = {
"tiny.en": WhisperModel(name="tiny.en", params="39M"),
"tiny": WhisperModel(name="tiny", params="39M"),
"base.en": WhisperModel(name="base.en", params="74M"),
"base": WhisperModel(name="base", params="74M"),
"small.en": WhisperModel(name="small.en", params="244M"),
"small": WhisperModel(name="small", params="244M"),
"medium.en": WhisperModel(name="medium.en", params="769M"),
"medium": WhisperModel(name="medium", params="769M"),
"large": WhisperModel(name="large", params="1550M"),
"large-v1": WhisperModel(name="large-v1", params="1550M"),
"large-v2": WhisperModel(name="large-v2", params="1550M"),
}
DEFAULT_MODEL = all_models["base.en"]
DEFAULT_ARGS = TranscribeConfig(
url=None,
video_url=None,
filename=None,
out=None,
model=DEFAULT_MODEL.name,
min_segment_len=5,
min_silence_len=2,
)
WEB_DEFAULT_ARGS = TranscribeConfig(
url=None,
video_url=None,
filename=None,
out=None,
model=all_models["base"].name,
min_segment_len=3.2,
min_silence_len=1.5,
)
local_serve = sys.argv[1] == "serve" and sys.argv[2] == "api.py"
from_cli = sys.argv[0] == "fan_transcribe.py"
def cfg():
parser = argparse.ArgumentParser()
parser.add_argument("-i", "--filename", help="a local file to transcribe")
parser.add_argument(
"-u",
"--url",
help=f"optional remote url of an audio file to transcribe",
)
parser.add_argument(
"-v",
"--video_url",
help=f"optional remote url of a video to transcribe (supports most video streaming sites)",
)
parser.add_argument(
"-o",
"--out",
help="optional output directory for transcription results. defaults to ./transcripts/ NB: unless you suffix this arg with .json, it will be interpreted as a directory",
)
parser.add_argument(
"-m",
"--model",
help=f"model to use for transcription. defaults to {DEFAULT_MODEL.name}. model options: [{', '.join(all_models.keys())}]",
default=DEFAULT_MODEL.name,
)
parser.add_argument(
"-g",
"--gpu",
help=f"optional GPU to use for transcription. defaults to None. GPU options: [{', '.join(STRING_TO_GPU_CONFIG.keys())}]",
default=None,
)
parser.add_argument(
"-sg",
"--min_segment_len",
help=f"minimum segment length (in seconds) for fan out. defaults to 5.0",
default=5.0,
type=float,
)
parser.add_argument(
"-sl",
"--min_silence_len",
help=f"minimum silence length (in seconds) to split on for segment generation. defaults to 2.0",
default=2.0,
type=float,
)
parser.add_argument(
"-f",
"--force",
action="store_true",
help="re-run a job identifier even if it's already processed",
)
return parser.parse_args()
if from_cli:
args: TranscribeConfig = TranscribeConfig(**vars(cfg()))
elif local_serve:
args = WEB_DEFAULT_ARGS
else:
args = DEFAULT_ARGS