File size: 9,702 Bytes
9ce3c93
 
962a236
ec35154
9ce3c93
 
 
 
 
 
962a236
9ce3c93
 
962a236
ec35154
6497d65
ec35154
 
2ba1ae3
ec35154
6497d65
ec35154
 
 
6497d65
ec35154
6497d65
9ce3c93
ec35154
2ba1ae3
9ce3c93
2cbbf1f
ec35154
990149f
2cbbf1f
 
 
 
 
 
 
 
 
 
 
990149f
 
ec35154
 
2cbbf1f
990149f
2cbbf1f
 
ec35154
 
 
 
2cbbf1f
 
 
ec35154
 
2cbbf1f
ec35154
 
 
2cbbf1f
 
 
ec35154
 
2cbbf1f
6497d65
2cbbf1f
ec35154
 
 
2cbbf1f
ec35154
 
6497d65
2cbbf1f
ec35154
 
 
 
2cbbf1f
 
 
ec35154
2cbbf1f
ec35154
 
 
 
 
2cbbf1f
 
 
 
 
ec35154
990149f
2cbbf1f
 
ec35154
 
 
2cbbf1f
 
 
 
ec35154
 
 
2cbbf1f
 
 
 
 
6497d65
990149f
2cbbf1f
6497d65
 
 
2cbbf1f
 
 
ec35154
9ce3c93
990149f
 
2cbbf1f
 
 
 
ec35154
9ce3c93
990149f
ec35154
2cbbf1f
 
 
ec35154
2cbbf1f
 
 
 
ec35154
 
 
 
2cbbf1f
 
 
ec35154
 
2cbbf1f
ec35154
 
 
2cbbf1f
 
 
 
 
 
 
 
6497d65
 
990149f
2cbbf1f
990149f
 
 
6497d65
990149f
2cbbf1f
6497d65
990149f
 
 
 
 
2cbbf1f
990149f
 
 
 
 
2cbbf1f
 
990149f
 
 
2cbbf1f
990149f
 
 
 
 
 
 
2cbbf1f
 
 
 
990149f
ec35154
 
2cbbf1f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ec35154
 
 
2cbbf1f
ec35154
 
 
 
 
 
2cbbf1f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ec35154
 
 
 
6497d65
ec35154
 
 
6497d65
990149f
ec35154
 
 
 
 
 
2cbbf1f
ec35154
 
 
 
 
 
 
 
 
 
2cbbf1f
 
ec35154
2cbbf1f
6497d65
ec35154
 
 
 
 
 
 
6497d65
ec35154
 
 
 
 
6497d65
ec35154
 
 
 
 
 
2cbbf1f
ec35154
 
 
6497d65
ec35154
 
 
 
 
 
9ce3c93
 
962a236
ec35154
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9ce3c93
 
ec35154
 
 
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
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
import gradio as gr
from transformers import pipeline
import re
import time

# Load both ASR models
whisper_asr = pipeline("automatic-speech-recognition", model="1morecupofhottea/Whisper-Code-Switching-Kh-En")
wav2vec_asr = pipeline("automatic-speech-recognition", model="1morecupofhottea/wav2vec2-Code-Switching-Kh_En")

def clean_transcript(text: str) -> str:
    # Remove <km>, </km>, <en>, </en>, etc.
    return re.sub(r"</?[^>]+>", "", text).strip()

def transcribe(audio, model_choice):
    if audio is None:
        return "Please upload or record an audio file first."
    
    try:
        if model_choice == "Whisper Model":
            result = whisper_asr(audio)
            return f"Transcription Complete:\n\n{result['text']}"
        else:
            result = wav2vec_asr(audio)
            cleaned_text = clean_transcript(result["text"])
            return f"Transcription Complete:\n\n{cleaned_text}"
    except Exception as e:
        return f"Error during transcription: {str(e)}"

def clear_all():
    return None, "Whisper Model", ""

# Custom CSS with dark theme
custom_css = """
/* 
DARK THEME COLOR PALETTE:
- Background: #000000 (Pure black - main background)
- Surface Dark: #1a1a1a (Dark gray - cards, containers)
- Surface Medium: #2d2d2d (Medium gray - inputs, components)
- Border: #404040 (Gray - borders, dividers)
- Text Primary: #ffffff (White - main text, headers)
- Text Secondary: #b0b0b0 (Light gray - secondary text)
- Accent: #4a9eff (Blue - buttons, highlights)
- Accent Hover: #357abd (Darker blue - hover states)
- Success: #28a745 (Green - success states)
- Warning: #ffc107 (Yellow - warnings)
*/

/* Global Styles */
.gradio-container {
    background: #000000 !important;
    font-family: 'Georgia', 'Times New Roman', serif !important;
    color: #ffffff !important;
    min-height: 100vh !important;
}

/* Header Styling */
.header-section {
    background: #1a1a1a;
    border: 2px solid #404040;
    border-radius: 12px;
    padding: 30px;
    margin-bottom: 25px;
    box-shadow: 0 4px 12px rgba(255, 255, 255, 0.1);
}

.header-section h1 {
    color: #ffffff !important;
    font-size: 2.4em !important;
    font-weight: 700 !important;
    text-align: center;
    margin-bottom: 15px;
    border-bottom: 3px solid #4a9eff;
    padding-bottom: 15px;
    text-shadow: 0 2px 4px rgba(0, 0, 0, 0.5);
}

.header-section p {
    color: #b0b0b0;
    font-size: 1.1em;
    text-align: center;
    margin: 8px 0;
    line-height: 1.6;
}

/* Main Content Cards */
.input-card, .output-card {
    background: #1a1a1a;
    border: 2px solid #404040;
    border-radius: 12px;
    padding: 25px;
    box-shadow: 0 4px 12px rgba(255, 255, 255, 0.05);
    margin-bottom: 20px;
}

/* Button Styling */
.primary-button {
    background: linear-gradient(135deg, #4a9eff, #357abd) !important;
    border: 2px solid #4a9eff !important;
    border-radius: 8px !important;
    padding: 14px 28px !important;
    font-size: 1.1em !important;
    font-weight: 600 !important;
    color: #ffffff !important;
    transition: all 0.3s ease !important;
    box-shadow: 0 4px 12px rgba(74, 158, 255, 0.3) !important;
}

.primary-button:hover {
    background: linear-gradient(135deg, #357abd, #2a5d8f) !important;
    border-color: #357abd !important;
    transform: translateY(-2px) !important;
    box-shadow: 0 6px 16px rgba(74, 158, 255, 0.4) !important;
}

.secondary-button {
    background: #2d2d2d !important;
    border: 2px solid #404040 !important;
    border-radius: 8px !important;
    padding: 12px 24px !important;
    font-size: 1em !important;
    font-weight: 500 !important;
    color: #ffffff !important;
    transition: all 0.3s ease !important;
}

.secondary-button:hover {
    background: #404040 !important;
    border-color: #555555 !important;
    transform: translateY(-1px) !important;
}

/* Component Styling */
.audio-component, .dropdown-component {
    border-radius: 8px !important;
    border: 2px solid #404040 !important;
    background: #2d2d2d !important;
    color: #ffffff !important;
}

/* Text Styling */
.output-text {
    background: #2d2d2d !important;
    border-radius: 8px !important;
    border: 2px solid #404040 !important;
    padding: 20px !important;
    font-size: 1.05em !important;
    line-height: 1.7 !important;
    font-family: 'Consolas', 'Courier New', monospace !important;
    color: #ffffff !important;
}

/* Features Section */
.features-section {
    background: #1a1a1a;
    border: 2px solid #404040;
    border-radius: 12px;
    padding: 25px;
    margin-top: 25px;
    box-shadow: 0 4px 12px rgba(255, 255, 255, 0.05);
}

.feature-item {
    margin-bottom: 15px;
    color: #b0b0b0;
    font-size: 1.05em;
    line-height: 1.6;
    padding: 10px;
    background: #2d2d2d;
    border-radius: 6px;
    border-left: 4px solid #4a9eff;
}

.feature-item strong {
    color: #ffffff;
}

/* Typography */
h3 {
    font-family: 'Georgia', 'Times New Roman', serif !important;
    color: #ffffff !important;
    font-weight: 600 !important;
    margin-bottom: 15px !important;
}

/* Component Overrides */
label {
    color: #ffffff !important;
    font-weight: 500 !important;
    font-family: 'Georgia', 'Times New Roman', serif !important;
}

::placeholder {
    color: #808080 !important;
    opacity: 0.8;
}

*:focus {
    outline: 2px solid #4a9eff !important;
    outline-offset: 2px !important;
}

/* Gradio Specific Overrides */
.gradio-textbox textarea,
.gradio-dropdown,
.gradio-audio {
    border: 2px solid #404040 !important;
    border-radius: 8px !important;
    background: #2d2d2d !important;
    color: #ffffff !important;
    font-family: inherit !important;
}

.gradio-textbox textarea:focus,
.gradio-dropdown:focus,
.gradio-audio:focus {
    border-color: #4a9eff !important;
    box-shadow: 0 0 0 3px rgba(74, 158, 255, 0.2) !important;
}

/* Scrollbar Styling */
::-webkit-scrollbar {
    width: 8px;
}

::-webkit-scrollbar-track {
    background: #1a1a1a;
}

::-webkit-scrollbar-thumb {
    background: #404040;
    border-radius: 4px;
}

::-webkit-scrollbar-thumb:hover {
    background: #4a9eff;
}

/* Selection Styling */
::selection {
    background: #4a9eff;
    color: #ffffff;
}

/* Responsive Design */
@media (max-width: 768px) {
    .header-section h1 {
        font-size: 2em !important;
    }
    
    .input-card, .output-card {
        padding: 20px;
        margin-bottom: 15px;
    }
    
    .primary-button, .secondary-button {
        padding: 12px 20px !important;
        font-size: 0.95em !important;
    }
}

/* Animation for better UX */
@keyframes fadeIn {
    from { opacity: 0; transform: translateY(10px); }
    to { opacity: 1; transform: translateY(0); }
}

.input-card, .output-card, .features-section, .header-section {
    animation: fadeIn 0.5s ease-out;
}
"""

# Create the main interface
with gr.Blocks(css=custom_css, title="CS-ASR | Code-Switching Speech Recognition") as demo:
    # Header Section
    with gr.Column(elem_classes="header-section"):
        gr.HTML("""
            <h1>Code-Switching ASR Platform</h1>
            <p>Speech Recognition for Khmer-English Code-Switching</p>
        """)
    
    # Main Content
    with gr.Row():
        # Input Section
        with gr.Column(scale=1, elem_classes="input-card"):
            gr.HTML("<h3 style='color: #ffffff; margin-bottom: 20px; font-weight: 600;'>Audio Input</h3>")
            
            audio_input = gr.Audio(
                sources=["microphone", "upload"], 
                type="filepath", 
                label="Record or Upload Audio",
                elem_classes="audio-component"
            )
            
            model_selector = gr.Dropdown(
                choices=[
                    "Whisper Model", 
                    "Wav2Vec2 Model"
                ], 
                value="Whisper Model", 
                label="Select AI Model",
                elem_classes="dropdown-component",
                info="Choose the model that best fits your needs"
            )
            
            # Action Buttons
            with gr.Row():
                transcribe_button = gr.Button(
                    "Start Transcription", 
                    variant="primary",
                    elem_classes="primary-button",
                    scale=2
                )
                clear_button = gr.Button(
                    "Clear All", 
                    elem_classes="secondary-button",
                    scale=1
                )
        
        # Output Section  
        with gr.Column(scale=1, elem_classes="output-card"):
            gr.HTML("<h3 style='color: #ffffff; margin-bottom: 20px; font-weight: 600;'>Transcription Result</h3>")
            
            output_text = gr.Textbox(
                label="Your Transcription Will Appear Here",
                placeholder="Upload an audio file and click 'Start Transcription' to begin processing.",
                lines=12,
                elem_classes="output-text",
                interactive=False
            )
    
    # Event Handlers
    transcribe_button.click(
        fn=transcribe,
        inputs=[audio_input, model_selector],
        outputs=output_text,
        show_progress=True
    )
    
    clear_button.click(
        fn=clear_all,
        outputs=[audio_input, model_selector, output_text]
    )
    
    # Auto-transcribe when audio is uploaded (optional)
    audio_input.change(
        fn=lambda audio, model: transcribe(audio, model) if audio is not None else "",
        inputs=[audio_input, model_selector],
        outputs=output_text,
        show_progress=True
    )

# Launch with custom configuration
if __name__ == "__main__":
    demo.launch()