File size: 5,217 Bytes
5ce1fe8
 
 
7e1376c
bb7ea32
5ce1fe8
 
 
 
2267fac
5ce1fe8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2267fac
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5ce1fe8
2267fac
 
 
 
 
 
 
 
 
5ce1fe8
 
e7f863b
 
 
 
 
 
 
 
 
 
5ce1fe8
 
 
7e1376c
5ce1fe8
7e1376c
 
 
 
 
5ce1fe8
 
e7f863b
5ce1fe8
 
 
 
e7f863b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5ce1fe8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f1ed46a
 
 
 
 
 
 
 
 
 
 
 
 
efb9bdb
 
 
 
 
 
bb7ea32
 
 
 
 
5ce1fe8
 
 
 
 
 
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
161
162
163
164
165
166
167
168
169
170
#!/usr/bin/python3
# -*- coding: utf-8 -*-
import argparse
from collections import defaultdict
import platform

import gradio as gr

from examples import examples
from models import model_map
from project_settings import project_path


def get_args():
    parser = argparse.ArgumentParser()
    parser.add_argument(
        "--examples_dir",
        default=(project_path / "data/examples").as_posix(),
        type=str
    )
    parser.add_argument(
        "--trained_model_dir",
        default=(project_path / "trained_models").as_posix(),
        type=str
    )
    args = parser.parse_args()
    return args


def update_model_dropdown(language: str):
    if language not in model_map.keys():
        raise ValueError(f"Unsupported language: {language}")

    choices = model_map[language]
    choices = [c["repo_id"] for c in choices]
    return gr.Dropdown(
        choices=choices,
        value=choices[0],
        interactive=True,
    )


def build_html_output(s: str, style: str = "result_item_success"):
    return f"""
    <div class='result'>
        <div class='result_item {style}'>
          {s}
        </div>
    </div>
    """


def process_uploaded_file(language: str,
                          repo_id: str,
                          decoding_method: str,
                          num_active_paths: int,
                          add_punctuation: str,
                          in_filename: str,
                          ):
    return "Dummy", build_html_output("Dummy")


# css style is copied from
# https://huggingface.co/spaces/alphacep/asr/blob/main/app.py#L113
css = """
.result {display:flex;flex-direction:column}
.result_item {padding:15px;margin-bottom:8px;border-radius:15px;width:100%}
.result_item_success {background-color:mediumaquamarine;color:white;align-self:start}
.result_item_error {background-color:#ff7070;color:white;align-self:start}
"""


def main():
    title = "# Automatic Speech Recognition with Next-gen Kaldi"

    language_choices = list(model_map.keys())

    language_to_models = defaultdict(list)
    for k, v in model_map.items():
        for m in v:
            repo_id = m["repo_id"]
            language_to_models[k].append(repo_id)

    # blocks
    with gr.Blocks(css=css) as blocks:
        gr.Markdown(value=title)

        with gr.Tabs():
            with gr.TabItem("Upload from disk"):
                language_radio = gr.Radio(
                    label="Language",
                    choices=language_choices,
                    value=language_choices[0],
                )
                model_dropdown = gr.Dropdown(
                    choices=language_to_models[language_choices[0]],
                    label="Select a model",
                    value=language_to_models[language_choices[0]][0],
                )
                decoding_method_radio = gr.Radio(
                    label="Decoding method",
                    choices=["greedy_search", "modified_beam_search"],
                    value="greedy_search",
                )
                num_active_paths_slider = gr.Slider(
                    minimum=1,
                    value=4,
                    step=1,
                    label="Number of active paths for modified_beam_search",
                )
                punct_radio = gr.Radio(
                    label="Whether to add punctuation (Only for Chinese and English)",
                    choices=["Yes", "No"],
                    value="Yes",
                )

                uploaded_file = gr.Audio(
                    sources=["upload"],
                    type="filepath",
                    label="Upload from disk",
                )
                upload_button = gr.Button("Submit for recognition")
                uploaded_output = gr.Textbox(label="Recognized speech from uploaded file")
                uploaded_html_info = gr.HTML(label="Info")

                gr.Examples(
                    examples=examples,
                    inputs=[
                        language_radio,
                        model_dropdown,
                        decoding_method_radio,
                        num_active_paths_slider,
                        punct_radio,
                        uploaded_file,
                    ],
                    outputs=[uploaded_output, uploaded_html_info],
                    fn=process_uploaded_file,
                )

            upload_button.click(
                process_uploaded_file,
                inputs=[
                    language_radio,
                    model_dropdown,
                    decoding_method_radio,
                    num_active_paths_slider,
                    punct_radio,
                    uploaded_file,
                ],
                outputs=[uploaded_output, uploaded_html_info],
            )

        language_radio.change(
            update_model_dropdown,
            inputs=language_radio,
            outputs=model_dropdown,
        )

    blocks.queue().launch(
        share=False if platform.system() == "Windows" else False,
        server_name="127.0.0.1" if platform.system() == "Windows" else "0.0.0.0",
        server_port=7860
    )

    return


if __name__ == "__main__":
    main()