Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -3,326 +3,213 @@ import torch
|
|
| 3 |
import torchaudio
|
| 4 |
import numpy as np
|
| 5 |
from transformers import AutoProcessor, SeamlessM4Tv2Model
|
| 6 |
-
from datetime import datetime
|
| 7 |
-
import time
|
| 8 |
|
| 9 |
-
class
|
| 10 |
-
def __init__(self
|
| 11 |
-
self.
|
| 12 |
-
|
|
|
|
|
|
|
| 13 |
self.sample_rate = self.model.config.sampling_rate
|
| 14 |
-
|
| 15 |
-
self.
|
| 16 |
-
"English
|
| 17 |
-
"Spanish
|
| 18 |
-
"French
|
| 19 |
-
"German
|
| 20 |
-
"Italian
|
| 21 |
-
"Portuguese
|
| 22 |
-
"Russian
|
| 23 |
-
"Chinese
|
| 24 |
-
"Japanese
|
| 25 |
-
"Korean
|
| 26 |
-
"Hindi (IN)": "hin",
|
| 27 |
-
"Arabic (AR)": "ara"
|
| 28 |
}
|
| 29 |
|
| 30 |
-
def
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 31 |
try:
|
| 32 |
-
if audio_path is None:
|
| 33 |
-
raise gr.Error("No audio input provided")
|
| 34 |
-
|
| 35 |
-
# Carregar e resample do áudio
|
| 36 |
audio, orig_freq = torchaudio.load(audio_path)
|
| 37 |
audio = torchaudio.functional.resample(audio, orig_freq=orig_freq, new_freq=16000)
|
| 38 |
|
| 39 |
-
|
| 40 |
inputs = self.processor(audios=audio, return_tensors="pt")
|
| 41 |
-
audio_array = self.model.generate(**inputs, tgt_lang=self.
|
| 42 |
-
|
| 43 |
-
|
| 44 |
-
raise gr.Error(f"Audio processing failed: {str(e)}")
|
| 45 |
-
|
| 46 |
-
def translate_text(self, text: str, src_lang: str, tgt_lang: str) -> tuple[int, np.ndarray]:
|
| 47 |
-
try:
|
| 48 |
-
if not text.strip():
|
| 49 |
-
raise gr.Error("No text input provided")
|
| 50 |
-
|
| 51 |
-
inputs = self.processor(text=text, src_lang=self.language_codes[src_lang], return_tensors="pt")
|
| 52 |
-
audio_array = self.model.generate(**inputs, tgt_lang=self.language_codes[tgt_lang])[0].cpu().numpy().squeeze()
|
| 53 |
-
return self.sample_rate, audio_array
|
| 54 |
except Exception as e:
|
| 55 |
-
raise gr.Error(
|
| 56 |
|
| 57 |
css = """
|
| 58 |
-
|
| 59 |
-
|
| 60 |
-
|
| 61 |
-
--accent: #ff3366;
|
| 62 |
-
--background: #000000;
|
| 63 |
-
--text: #ffffff;
|
| 64 |
-
}
|
| 65 |
-
|
| 66 |
-
#aris-interface {
|
| 67 |
-
background-color: var(--background);
|
| 68 |
-
background-image:
|
| 69 |
-
radial-gradient(circle at 20% 20%, rgba(0, 102, 204, 0.1) 0%, transparent 50%),
|
| 70 |
-
radial-gradient(circle at 80% 80%, rgba(0, 255, 255, 0.1) 0%, transparent 50%);
|
| 71 |
-
min-height: 100vh;
|
| 72 |
-
font-family: 'Courier New', monospace;
|
| 73 |
padding: 20px;
|
| 74 |
}
|
| 75 |
|
| 76 |
-
.
|
| 77 |
-
|
| 78 |
-
|
| 79 |
-
margin-bottom: 30px;
|
| 80 |
-
position: relative;
|
| 81 |
-
}
|
| 82 |
-
|
| 83 |
-
.title-container h1 {
|
| 84 |
-
font-size: 3em;
|
| 85 |
-
letter-spacing: 10px;
|
| 86 |
-
margin: 0;
|
| 87 |
-
text-shadow: 0 0 10px var(--primary);
|
| 88 |
-
}
|
| 89 |
-
|
| 90 |
-
.title-container h3 {
|
| 91 |
-
font-size: 1.2em;
|
| 92 |
-
letter-spacing: 3px;
|
| 93 |
-
opacity: 0.8;
|
| 94 |
-
margin: 5px 0;
|
| 95 |
}
|
| 96 |
|
| 97 |
-
|
| 98 |
-
|
| 99 |
-
height: 400px;
|
| 100 |
-
border: 4px solid var(--primary);
|
| 101 |
-
border-radius: 50%;
|
| 102 |
-
margin: 20px auto;
|
| 103 |
-
position: relative;
|
| 104 |
-
animation: pulse 2s infinite;
|
| 105 |
-
display: flex;
|
| 106 |
-
align-items: center;
|
| 107 |
-
justify-content: center;
|
| 108 |
-
background:
|
| 109 |
-
radial-gradient(circle at center, rgba(0, 255, 255, 0.1) 0%, transparent 70%),
|
| 110 |
-
conic-gradient(from 0deg, transparent 0%, rgba(0, 255, 255, 0.1) 50%, transparent 100%);
|
| 111 |
}
|
| 112 |
|
| 113 |
-
|
| 114 |
-
|
| 115 |
-
|
| 116 |
-
|
| 117 |
-
|
| 118 |
-
border: 1px solid rgba(0, 255, 255, 0.3);
|
| 119 |
-
animation: rotate 20s linear infinite;
|
| 120 |
}
|
| 121 |
|
| 122 |
-
|
| 123 |
-
|
| 124 |
-
|
| 125 |
}
|
| 126 |
|
| 127 |
-
|
| 128 |
-
|
| 129 |
-
70% { box-shadow: 0 0 0 20px rgba(0, 255, 255, 0); }
|
| 130 |
-
100% { box-shadow: 0 0 0 0 rgba(0, 255, 255, 0); }
|
| 131 |
}
|
| 132 |
|
| 133 |
-
.
|
| 134 |
-
|
| 135 |
-
border: 2px solid var(--primary) !important;
|
| 136 |
-
color: var(--primary) !important;
|
| 137 |
-
font-family: 'Courier New', monospace !important;
|
| 138 |
-
border-radius: 5px !important;
|
| 139 |
-
padding: 10px !important;
|
| 140 |
}
|
| 141 |
|
| 142 |
-
.
|
| 143 |
-
|
| 144 |
-
|
| 145 |
-
|
| 146 |
-
|
| 147 |
-
|
| 148 |
-
|
| 149 |
-
|
| 150 |
-
border-radius: 5px !important;
|
| 151 |
-
transition: all 0.3s ease !important;
|
| 152 |
}
|
| 153 |
|
| 154 |
-
.
|
| 155 |
-
|
| 156 |
-
|
| 157 |
-
|
| 158 |
}
|
| 159 |
|
| 160 |
-
.
|
| 161 |
-
|
| 162 |
-
|
| 163 |
-
color: var(--primary) !important;
|
| 164 |
-
padding: 15px !important;
|
| 165 |
-
border-radius: 5px !important;
|
| 166 |
-
margin: 5px !important;
|
| 167 |
-
text-align: center !important;
|
| 168 |
-
text-transform: uppercase !important;
|
| 169 |
-
letter-spacing: 1px !important;
|
| 170 |
-
transition: all 0.3s ease !important;
|
| 171 |
-
position: relative;
|
| 172 |
-
overflow: hidden;
|
| 173 |
}
|
| 174 |
|
| 175 |
-
.
|
| 176 |
-
|
| 177 |
-
|
| 178 |
-
|
| 179 |
-
left: -100%;
|
| 180 |
-
width: 100%;
|
| 181 |
-
height: 2px;
|
| 182 |
-
background: linear-gradient(90deg, transparent, var(--primary));
|
| 183 |
-
animation: scan-line 2s linear infinite;
|
| 184 |
}
|
| 185 |
|
| 186 |
-
|
| 187 |
-
|
| 188 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 189 |
}
|
| 190 |
"""
|
| 191 |
|
| 192 |
-
def
|
| 193 |
-
translator =
|
| 194 |
-
|
| 195 |
-
|
| 196 |
-
|
| 197 |
-
|
| 198 |
-
|
| 199 |
-
|
| 200 |
-
|
| 201 |
)
|
| 202 |
-
|
| 203 |
-
|
| 204 |
-
|
| 205 |
-
|
| 206 |
-
|
| 207 |
-
|
| 208 |
-
<div class="mode-indicator">QUANTUM CORE ACTIVE</div>
|
| 209 |
-
</div>
|
| 210 |
-
''')
|
| 211 |
-
|
| 212 |
-
with gr.Column(elem_id="aris-interface"):
|
| 213 |
-
gr.HTML("""
|
| 214 |
-
<div id="status-ring">
|
| 215 |
-
<div id="outer-ring-decoration"></div>
|
| 216 |
-
<div id="inner-ring">
|
| 217 |
-
<div id="core">
|
| 218 |
-
<div>A.R.I.S.</div>
|
| 219 |
-
<div>QUANTUM CORE</div>
|
| 220 |
-
<div>v2.0.0</div>
|
| 221 |
-
<div class="system-version">NEURAL ENGINE ACTIVE</div>
|
| 222 |
-
</div>
|
| 223 |
-
</div>
|
| 224 |
-
</div>
|
| 225 |
-
""")
|
| 226 |
-
|
| 227 |
-
with gr.Row():
|
| 228 |
-
with gr.Column():
|
| 229 |
-
with gr.Tab("Text Translation"):
|
| 230 |
text_input = gr.Textbox(
|
| 231 |
-
label="
|
| 232 |
-
placeholder="Enter text
|
| 233 |
-
|
| 234 |
-
lines=3
|
| 235 |
)
|
| 236 |
with gr.Row():
|
| 237 |
-
|
| 238 |
-
choices=list(translator.
|
| 239 |
-
value="English
|
| 240 |
-
label="
|
| 241 |
-
elem_classes=["aris-textbox"]
|
| 242 |
)
|
| 243 |
-
|
| 244 |
-
choices=list(translator.
|
| 245 |
-
value="Spanish
|
| 246 |
-
label="
|
| 247 |
-
elem_classes=["aris-textbox"]
|
| 248 |
)
|
| 249 |
-
translate_btn = gr.Button("
|
| 250 |
-
|
| 251 |
-
with gr.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 252 |
audio_input = gr.Audio(
|
| 253 |
-
label="
|
| 254 |
type="filepath"
|
| 255 |
)
|
| 256 |
tgt_lang_audio = gr.Dropdown(
|
| 257 |
-
choices=list(translator.
|
| 258 |
-
value="English
|
| 259 |
-
label="
|
| 260 |
-
elem_classes=["aris-textbox"]
|
| 261 |
)
|
| 262 |
-
translate_audio_btn = gr.Button("
|
| 263 |
-
|
| 264 |
-
with gr.Column():
|
| 265 |
-
audio_output = gr.Audio(
|
| 266 |
-
label="TRANSLATION OUTPUT",
|
| 267 |
-
type="numpy"
|
| 268 |
-
)
|
| 269 |
-
|
| 270 |
-
with gr.Row():
|
| 271 |
-
with gr.Column(min_width=200):
|
| 272 |
-
gr.HTML(
|
| 273 |
-
"""
|
| 274 |
-
<div class="status-box">
|
| 275 |
-
NEURAL CORE<br>
|
| 276 |
-
<strong>OPERATIONAL</strong>
|
| 277 |
-
</div>
|
| 278 |
-
"""
|
| 279 |
-
)
|
| 280 |
-
with gr.Column(min_width=200):
|
| 281 |
-
gr.HTML(
|
| 282 |
-
"""
|
| 283 |
-
<div class="status-box">
|
| 284 |
-
QUANTUM ENGINE<br>
|
| 285 |
-
<strong>ACTIVE</strong>
|
| 286 |
-
</div>
|
| 287 |
-
"""
|
| 288 |
-
)
|
| 289 |
-
|
| 290 |
-
with gr.Row():
|
| 291 |
-
with gr.Column(min_width=200):
|
| 292 |
-
gr.HTML(
|
| 293 |
-
"""
|
| 294 |
-
<div class="status-box">
|
| 295 |
-
TRANSLATION MATRIX<br>
|
| 296 |
-
<strong>CALIBRATED</strong>
|
| 297 |
-
</div>
|
| 298 |
-
"""
|
| 299 |
-
)
|
| 300 |
-
with gr.Column(min_width=200):
|
| 301 |
-
gr.HTML(
|
| 302 |
-
"""
|
| 303 |
-
<div class="status-box">
|
| 304 |
-
VOICE SYNTHESIS<br>
|
| 305 |
-
<strong>READY</strong>
|
| 306 |
-
</div>
|
| 307 |
-
"""
|
| 308 |
-
)
|
| 309 |
|
| 310 |
-
|
| 311 |
-
|
| 312 |
-
|
| 313 |
-
|
| 314 |
-
|
| 315 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 316 |
|
| 317 |
-
translate_audio_btn.click(
|
| 318 |
-
fn=translator.process_audio,
|
| 319 |
-
inputs=[audio_input, tgt_lang_audio],
|
| 320 |
-
outputs=audio_output
|
| 321 |
-
)
|
| 322 |
-
|
| 323 |
return demo
|
| 324 |
|
| 325 |
if __name__ == "__main__":
|
| 326 |
-
demo =
|
| 327 |
demo.queue()
|
| 328 |
demo.launch()
|
|
|
|
| 3 |
import torchaudio
|
| 4 |
import numpy as np
|
| 5 |
from transformers import AutoProcessor, SeamlessM4Tv2Model
|
|
|
|
|
|
|
| 6 |
|
| 7 |
+
class SeamlessTranslator:
|
| 8 |
+
def __init__(self):
|
| 9 |
+
self.model_name = "facebook/seamless-m4t-v2-large"
|
| 10 |
+
print("Loading model...")
|
| 11 |
+
self.processor = AutoProcessor.from_pretrained(self.model_name)
|
| 12 |
+
self.model = SeamlessM4Tv2Model.from_pretrained(self.model_name)
|
| 13 |
self.sample_rate = self.model.config.sampling_rate
|
| 14 |
+
|
| 15 |
+
self.languages = {
|
| 16 |
+
"English": "eng",
|
| 17 |
+
"Spanish": "spa",
|
| 18 |
+
"French": "fra",
|
| 19 |
+
"German": "deu",
|
| 20 |
+
"Italian": "ita",
|
| 21 |
+
"Portuguese": "por",
|
| 22 |
+
"Russian": "rus",
|
| 23 |
+
"Chinese": "cmn",
|
| 24 |
+
"Japanese": "jpn",
|
| 25 |
+
"Korean": "kor"
|
|
|
|
|
|
|
| 26 |
}
|
| 27 |
|
| 28 |
+
def translate_text(self, text, src_lang, tgt_lang, progress=gr.Progress()):
|
| 29 |
+
progress(0.3, desc="Processing input...")
|
| 30 |
+
try:
|
| 31 |
+
inputs = self.processor(text=text, src_lang=self.languages[src_lang], return_tensors="pt")
|
| 32 |
+
progress(0.6, desc="Generating audio...")
|
| 33 |
+
audio_array = self.model.generate(**inputs, tgt_lang=self.languages[tgt_lang])[0].cpu().numpy().squeeze()
|
| 34 |
+
progress(1.0, desc="Done!")
|
| 35 |
+
return (self.sample_rate, audio_array)
|
| 36 |
+
except Exception as e:
|
| 37 |
+
raise gr.Error(str(e))
|
| 38 |
+
|
| 39 |
+
def translate_audio(self, audio_path, tgt_lang, progress=gr.Progress()):
|
| 40 |
+
progress(0.3, desc="Loading audio...")
|
| 41 |
try:
|
|
|
|
|
|
|
|
|
|
|
|
|
| 42 |
audio, orig_freq = torchaudio.load(audio_path)
|
| 43 |
audio = torchaudio.functional.resample(audio, orig_freq=orig_freq, new_freq=16000)
|
| 44 |
|
| 45 |
+
progress(0.6, desc="Translating...")
|
| 46 |
inputs = self.processor(audios=audio, return_tensors="pt")
|
| 47 |
+
audio_array = self.model.generate(**inputs, tgt_lang=self.languages[tgt_lang])[0].cpu().numpy().squeeze()
|
| 48 |
+
progress(1.0, desc="Done!")
|
| 49 |
+
return (self.sample_rate, audio_array)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 50 |
except Exception as e:
|
| 51 |
+
raise gr.Error(str(e))
|
| 52 |
|
| 53 |
css = """
|
| 54 |
+
#component-0 {
|
| 55 |
+
max-width: 1200px;
|
| 56 |
+
margin: auto;
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 57 |
padding: 20px;
|
| 58 |
}
|
| 59 |
|
| 60 |
+
.container {
|
| 61 |
+
border-radius: 12px;
|
| 62 |
+
padding: 20px;
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 63 |
}
|
| 64 |
|
| 65 |
+
.gr-form {
|
| 66 |
+
border-color: #e5e7eb !important;
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 67 |
}
|
| 68 |
|
| 69 |
+
.gr-button {
|
| 70 |
+
border-radius: 8px !important;
|
| 71 |
+
background: linear-gradient(to right, #2563eb, #4f46e5) !important;
|
| 72 |
+
color: white !important;
|
| 73 |
+
font-weight: 600 !important;
|
|
|
|
|
|
|
| 74 |
}
|
| 75 |
|
| 76 |
+
.gr-button:hover {
|
| 77 |
+
box-shadow: 0 4px 6px -1px rgb(0 0 0 / 0.1) !important;
|
| 78 |
+
transform: translateY(-1px);
|
| 79 |
}
|
| 80 |
|
| 81 |
+
.gr-input, .gr-select {
|
| 82 |
+
border-radius: 8px !important;
|
|
|
|
|
|
|
| 83 |
}
|
| 84 |
|
| 85 |
+
.gr-panel {
|
| 86 |
+
border-radius: 12px !important;
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 87 |
}
|
| 88 |
|
| 89 |
+
.title {
|
| 90 |
+
text-align: center;
|
| 91 |
+
font-size: 2.5rem;
|
| 92 |
+
font-weight: bold;
|
| 93 |
+
margin: 1rem 0;
|
| 94 |
+
background: linear-gradient(to right, #2563eb, #4f46e5);
|
| 95 |
+
-webkit-background-clip: text;
|
| 96 |
+
-webkit-text-fill-color: transparent;
|
|
|
|
|
|
|
| 97 |
}
|
| 98 |
|
| 99 |
+
.subtitle {
|
| 100 |
+
text-align: center;
|
| 101 |
+
color: #6b7280;
|
| 102 |
+
margin-bottom: 2rem;
|
| 103 |
}
|
| 104 |
|
| 105 |
+
.tab-nav {
|
| 106 |
+
border-bottom: 2px solid #e5e7eb;
|
| 107 |
+
margin-bottom: 1rem;
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 108 |
}
|
| 109 |
|
| 110 |
+
.output-label {
|
| 111 |
+
font-weight: 600;
|
| 112 |
+
color: #374151;
|
| 113 |
+
margin-bottom: 0.5rem;
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 114 |
}
|
| 115 |
|
| 116 |
+
.footer {
|
| 117 |
+
text-align: center;
|
| 118 |
+
margin-top: 2rem;
|
| 119 |
+
padding-top: 1rem;
|
| 120 |
+
border-top: 1px solid #e5e7eb;
|
| 121 |
+
color: #6b7280;
|
| 122 |
+
font-size: 0.875rem;
|
| 123 |
}
|
| 124 |
"""
|
| 125 |
|
| 126 |
+
def create_ui():
|
| 127 |
+
translator = SeamlessTranslator()
|
| 128 |
+
|
| 129 |
+
with gr.Blocks(css=css, title="A.R.I.S. Translator") as demo:
|
| 130 |
+
gr.HTML(
|
| 131 |
+
"""
|
| 132 |
+
<div class="title">A.R.I.S. Translator</div>
|
| 133 |
+
<div class="subtitle">Advanced Real-time Interpretation System</div>
|
| 134 |
+
"""
|
| 135 |
)
|
| 136 |
+
|
| 137 |
+
with gr.Tabs() as tabs:
|
| 138 |
+
# Text to Speech Tab
|
| 139 |
+
with gr.Tab("Text Translation", id=1):
|
| 140 |
+
with gr.Row():
|
| 141 |
+
with gr.Column():
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 142 |
text_input = gr.Textbox(
|
| 143 |
+
label="Text to Translate",
|
| 144 |
+
placeholder="Enter your text here...",
|
| 145 |
+
lines=5
|
|
|
|
| 146 |
)
|
| 147 |
with gr.Row():
|
| 148 |
+
src_lang = gr.Dropdown(
|
| 149 |
+
choices=list(translator.languages.keys()),
|
| 150 |
+
value="English",
|
| 151 |
+
label="Source Language"
|
|
|
|
| 152 |
)
|
| 153 |
+
tgt_lang = gr.Dropdown(
|
| 154 |
+
choices=list(translator.languages.keys()),
|
| 155 |
+
value="Spanish",
|
| 156 |
+
label="Target Language"
|
|
|
|
| 157 |
)
|
| 158 |
+
translate_btn = gr.Button("Translate", variant="primary")
|
| 159 |
+
|
| 160 |
+
with gr.Column():
|
| 161 |
+
gr.HTML('<div class="output-label">Translation Output</div>')
|
| 162 |
+
audio_output = gr.Audio(
|
| 163 |
+
label="Translated Audio",
|
| 164 |
+
type="numpy"
|
| 165 |
+
)
|
| 166 |
+
|
| 167 |
+
# Audio to Speech Tab
|
| 168 |
+
with gr.Tab("Audio Translation", id=2):
|
| 169 |
+
with gr.Row():
|
| 170 |
+
with gr.Column():
|
| 171 |
audio_input = gr.Audio(
|
| 172 |
+
label="Upload Audio",
|
| 173 |
type="filepath"
|
| 174 |
)
|
| 175 |
tgt_lang_audio = gr.Dropdown(
|
| 176 |
+
choices=list(translator.languages.keys()),
|
| 177 |
+
value="English",
|
| 178 |
+
label="Target Language"
|
|
|
|
| 179 |
)
|
| 180 |
+
translate_audio_btn = gr.Button("Translate Audio", variant="primary")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 181 |
|
| 182 |
+
with gr.Column():
|
| 183 |
+
gr.HTML('<div class="output-label">Translation Output</div>')
|
| 184 |
+
audio_output_from_audio = gr.Audio(
|
| 185 |
+
label="Translated Audio",
|
| 186 |
+
type="numpy"
|
| 187 |
+
)
|
| 188 |
+
|
| 189 |
+
gr.HTML(
|
| 190 |
+
"""
|
| 191 |
+
<div class="footer">
|
| 192 |
+
Powered by Meta's SeamlessM4T model | Built with Gradio
|
| 193 |
+
</div>
|
| 194 |
+
"""
|
| 195 |
+
)
|
| 196 |
+
|
| 197 |
+
# Event handlers
|
| 198 |
+
translate_btn.click(
|
| 199 |
+
fn=translator.translate_text,
|
| 200 |
+
inputs=[text_input, src_lang, tgt_lang],
|
| 201 |
+
outputs=audio_output
|
| 202 |
+
)
|
| 203 |
+
|
| 204 |
+
translate_audio_btn.click(
|
| 205 |
+
fn=translator.translate_audio,
|
| 206 |
+
inputs=[audio_input, tgt_lang_audio],
|
| 207 |
+
outputs=audio_output_from_audio
|
| 208 |
+
)
|
| 209 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 210 |
return demo
|
| 211 |
|
| 212 |
if __name__ == "__main__":
|
| 213 |
+
demo = create_ui()
|
| 214 |
demo.queue()
|
| 215 |
demo.launch()
|