File size: 3,821 Bytes
05a2178
 
 
 
 
 
7ce6041
 
05a2178
 
 
7ce6041
 
 
05a2178
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7ce6041
 
 
 
 
 
 
05a2178
 
 
 
 
 
 
 
 
 
 
 
 
 
7ce6041
 
 
 
 
 
 
 
05a2178
 
 
 
 
 
 
 
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
151
152
153
154
155
156
157
158
159
160
from io import StringIO
import gradio as gr

from utils import write_vtt
import whisper

import ffmpeg

#import os
#os.system("pip install git+https://github.com/openai/whisper.git")

# Limitations (set to -1 to disable)
INPUT_AUDIO_MAX_DURATION = 60 # seconds

LANGUAGES = [
    "English",
    "Chinese",
    "German",
    "Spanish",
    "Russian",
    "Korean",
    "French",
    "Japanese",
    "Portuguese",
    "Turkish",
    "Polish",
    "Catalan",
    "Dutch",
    "Arabic",
    "Swedish",
    "Italian",
    "Indonesian",
    "Hindi",
    "Finnish",
    "Vietnamese",
    "Hebrew",
    "Ukrainian",
    "Greek",
    "Malay",
    "Czech",
    "Romanian",
    "Danish",
    "Hungarian",
    "Tamil",
    "Norwegian",
    "Thai",
    "Urdu",
    "Croatian",
    "Bulgarian",
    "Lithuanian",
    "Latin",
    "Maori",
    "Malayalam",
    "Welsh",
    "Slovak",
    "Telugu",
    "Persian",
    "Latvian",
    "Bengali",
    "Serbian",
    "Azerbaijani",
    "Slovenian",
    "Kannada",
    "Estonian",
    "Macedonian",
    "Breton",
    "Basque",
    "Icelandic",
    "Armenian",
    "Nepali",
    "Mongolian",
    "Bosnian",
    "Kazakh",
    "Albanian",
    "Swahili",
    "Galician",
    "Marathi",
    "Punjabi",
    "Sinhala",
    "Khmer",
    "Shona",
    "Yoruba",
    "Somali",
    "Afrikaans",
    "Occitan",
    "Georgian",
    "Belarusian",
    "Tajik",
    "Sindhi",
    "Gujarati",
    "Amharic",
    "Yiddish",
    "Lao",
    "Uzbek",
    "Faroese",
    "Haitian Creole",
    "Pashto",
    "Turkmen",
    "Nynorsk",
    "Maltese",
    "Sanskrit",
    "Luxembourgish",
    "Myanmar",
    "Tibetan",
    "Tagalog",
    "Malagasy",
    "Assamese",
    "Tatar",
    "Hawaiian",
    "Lingala",
    "Hausa",
    "Bashkir",
    "Javanese",
    "Sundanese"
]

model_cache = dict()

def greet(modelName, languageName, uploadFile, microphoneData, task):
    source = uploadFile if uploadFile is not None else microphoneData
    selectedLanguage = languageName.lower() if len(languageName) > 0 else None
    selectedModel = modelName if modelName is not None else "base"

    if INPUT_AUDIO_MAX_DURATION > 0:
        # Calculate audio length
        audioDuration = ffmpeg.probe(source)["format"]["duration"]
        
        if float(audioDuration) > INPUT_AUDIO_MAX_DURATION:
            return ("[ERROR]: Maximum audio file length is " + str(INPUT_AUDIO_MAX_DURATION) + "s, file was " + str(audioDuration) + "s"), "[ERROR]"

    model = model_cache.get(selectedModel, None)
    
    if not model:
        model = whisper.load_model(selectedModel)
        model_cache[selectedModel] = model

    result = model.transcribe(source, language=selectedLanguage, task=task)

    segmentStream = StringIO()
    write_vtt(result["segments"], file=segmentStream)
    segmentStream.seek(0)

    return result["text"], segmentStream.read()

ui_description = "Whisper is a general-purpose speech recognition model. It is trained on a large dataset of diverse " 
ui_description += " audio and is also a multi-task model that can perform multilingual speech recognition "
ui_description += " as well as speech translation and language identification. "

if INPUT_AUDIO_MAX_DURATION > 0:
    ui_description += "\n\n" + "Max audio file length: " + str(INPUT_AUDIO_MAX_DURATION) + " s"

demo = gr.Interface(fn=greet, description=ui_description, inputs=[
    gr.Dropdown(choices=["tiny", "base", "small", "medium", "large"], value="medium", label="Model"),
    gr.Dropdown(choices=sorted(LANGUAGES), label="Language"),
    gr.Audio(source="upload", type="filepath", label="Upload Audio"), 
    gr.Audio(source="microphone", type="filepath", label="Microphone Input"),
    gr.Dropdown(choices=["transcribe", "translate"], label="Task"),
], outputs=[gr.Text(label="Transcription"), gr.Text(label="Segments")])

demo.launch()