Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -13,6 +13,7 @@ import re
|
|
| 13 |
#p1 = pipeline(task="automatic-speech-recognition", model="cdactvm/w2v-bert-2.0-odia_v1")
|
| 14 |
#odia_model2 = pipeline(task="automatic-speech-recognition", model="cdactvm/w2v-bert-odia_v1")
|
| 15 |
p2 = pipeline(task="automatic-speech-recognition", model="cdactvm/w2v-bert-2.0-hindi_v1")
|
|
|
|
| 16 |
punjaib_modle_155750=pipeline(task="automatic-speech-recognition", model="cdactvm/wav2vec-bert-punjabi-155750-model")
|
| 17 |
punjaib_modle_75000=pipeline(task="automatic-speech-recognition", model="cdactvm/wav2vec-bert-punjabi-75000-model")
|
| 18 |
#p3 = pipeline(task="automatic-speech-recognition", model="cdactvm/kannada_w2v-bert_model")
|
|
@@ -267,7 +268,7 @@ def process_transcription(input_sentence):
|
|
| 267 |
###########################################
|
| 268 |
|
| 269 |
def transcribe_punjabi_30000(speech):
|
| 270 |
-
text =
|
| 271 |
text = text.replace("[PAD]","")
|
| 272 |
if text is None:
|
| 273 |
return "Error: ASR returned None"
|
|
@@ -275,7 +276,7 @@ def transcribe_punjabi_30000(speech):
|
|
| 275 |
|
| 276 |
def transcribe_punjabi_eng_model_30000(speech):
|
| 277 |
trn = Transliterator(source='pan', target='eng', build_lookup=True)
|
| 278 |
-
text =
|
| 279 |
text = text.replace("[PAD]","")
|
| 280 |
if text is None:
|
| 281 |
return "Error: ASR returned None"
|
|
@@ -288,7 +289,7 @@ def transcribe_punjabi_eng_model_30000(speech):
|
|
| 288 |
return sentence
|
| 289 |
|
| 290 |
def transcribe_punjabi_75000(speech):
|
| 291 |
-
text =
|
| 292 |
text = text.replace("[PAD]","")
|
| 293 |
if text is None:
|
| 294 |
return "Error: ASR returned None"
|
|
@@ -296,7 +297,7 @@ def transcribe_punjabi_75000(speech):
|
|
| 296 |
|
| 297 |
def transcribe_punjabi_eng_model_75000(speech):
|
| 298 |
trn = Transliterator(source='pan', target='eng', build_lookup=True)
|
| 299 |
-
text =
|
| 300 |
text = text.replace("[PAD]","")
|
| 301 |
if text is None:
|
| 302 |
return "Error: ASR returned None"
|
|
@@ -308,6 +309,27 @@ def transcribe_punjabi_eng_model_75000(speech):
|
|
| 308 |
return process_transcription(processed_sentence)
|
| 309 |
return sentence
|
| 310 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 311 |
###########################################
|
| 312 |
def transcribe_odiya(speech):
|
| 313 |
text = p1(speech)["text"]
|
|
@@ -542,6 +564,10 @@ def sel_lng(lng, mic=None, file=None):
|
|
| 542 |
return transcribe_punjabi_75000(audio)
|
| 543 |
elif lng == "Punjabi_Model2_Trans":
|
| 544 |
return transcribe_punjabi_eng_model_75000(audio)
|
|
|
|
|
|
|
|
|
|
|
|
|
| 545 |
|
| 546 |
|
| 547 |
|
|
@@ -579,7 +605,7 @@ demo=gr.Interface(
|
|
| 579 |
# "Hindi","Hindi-trans",
|
| 580 |
# "Odiya","Odiya-trans","Odia_model2","Odia_trans_model2",
|
| 581 |
# "Assamese-LM","Assamese-Model2",
|
| 582 |
-
"Punjabi_Model1","Punjabi_Model1_Trans","Punjabi_Model2","Punjabi_Model2_Trans"],value="Hindi",label="Select Language"),
|
| 583 |
gr.Audio(sources=["microphone","upload"], type="filepath"),
|
| 584 |
#gr.Audio(sources="upload", type="filepath"),
|
| 585 |
#"state"
|
|
|
|
| 13 |
#p1 = pipeline(task="automatic-speech-recognition", model="cdactvm/w2v-bert-2.0-odia_v1")
|
| 14 |
#odia_model2 = pipeline(task="automatic-speech-recognition", model="cdactvm/w2v-bert-odia_v1")
|
| 15 |
p2 = pipeline(task="automatic-speech-recognition", model="cdactvm/w2v-bert-2.0-hindi_v1")
|
| 16 |
+
punjaib_modle_30000=pipeline(task="automatic-speech-recognition", model="cdactvm/wav2vec-bert-punjabi-30000-model")
|
| 17 |
punjaib_modle_155750=pipeline(task="automatic-speech-recognition", model="cdactvm/wav2vec-bert-punjabi-155750-model")
|
| 18 |
punjaib_modle_75000=pipeline(task="automatic-speech-recognition", model="cdactvm/wav2vec-bert-punjabi-75000-model")
|
| 19 |
#p3 = pipeline(task="automatic-speech-recognition", model="cdactvm/kannada_w2v-bert_model")
|
|
|
|
| 268 |
###########################################
|
| 269 |
|
| 270 |
def transcribe_punjabi_30000(speech):
|
| 271 |
+
text = punjaib_modle_30000(speech)["text"]
|
| 272 |
text = text.replace("[PAD]","")
|
| 273 |
if text is None:
|
| 274 |
return "Error: ASR returned None"
|
|
|
|
| 276 |
|
| 277 |
def transcribe_punjabi_eng_model_30000(speech):
|
| 278 |
trn = Transliterator(source='pan', target='eng', build_lookup=True)
|
| 279 |
+
text = punjaib_modle_30000(speech)["text"]
|
| 280 |
text = text.replace("[PAD]","")
|
| 281 |
if text is None:
|
| 282 |
return "Error: ASR returned None"
|
|
|
|
| 289 |
return sentence
|
| 290 |
|
| 291 |
def transcribe_punjabi_75000(speech):
|
| 292 |
+
text = punjaib_modle_75000(speech)["text"]
|
| 293 |
text = text.replace("[PAD]","")
|
| 294 |
if text is None:
|
| 295 |
return "Error: ASR returned None"
|
|
|
|
| 297 |
|
| 298 |
def transcribe_punjabi_eng_model_75000(speech):
|
| 299 |
trn = Transliterator(source='pan', target='eng', build_lookup=True)
|
| 300 |
+
text = punjaib_modle_75000(speech)["text"]
|
| 301 |
text = text.replace("[PAD]","")
|
| 302 |
if text is None:
|
| 303 |
return "Error: ASR returned None"
|
|
|
|
| 309 |
return process_transcription(processed_sentence)
|
| 310 |
return sentence
|
| 311 |
|
| 312 |
+
def transcribe_punjabi_155750(speech):
|
| 313 |
+
text = punjaib_modle_155750(speech)["text"]
|
| 314 |
+
text = text.replace("[PAD]","")
|
| 315 |
+
if text is None:
|
| 316 |
+
return "Error: ASR returned None"
|
| 317 |
+
return text
|
| 318 |
+
|
| 319 |
+
def transcribe_punjabi_eng_model_155750(speech):
|
| 320 |
+
trn = Transliterator(source='pan', target='eng', build_lookup=True)
|
| 321 |
+
text = punjaib_modle_155750(speech)["text"]
|
| 322 |
+
text = text.replace("[PAD]","")
|
| 323 |
+
if text is None:
|
| 324 |
+
return "Error: ASR returned None"
|
| 325 |
+
sentence = trn.transform(text)
|
| 326 |
+
if sentence is None:
|
| 327 |
+
return "Error: Transliteration returned None"
|
| 328 |
+
replaced_words = replace_words(sentence)
|
| 329 |
+
processed_sentence = process_doubles(replaced_words)
|
| 330 |
+
return process_transcription(processed_sentence)
|
| 331 |
+
return sentence
|
| 332 |
+
|
| 333 |
###########################################
|
| 334 |
def transcribe_odiya(speech):
|
| 335 |
text = p1(speech)["text"]
|
|
|
|
| 564 |
return transcribe_punjabi_75000(audio)
|
| 565 |
elif lng == "Punjabi_Model2_Trans":
|
| 566 |
return transcribe_punjabi_eng_model_75000(audio)
|
| 567 |
+
elif lng == "Punjabi_Model3":
|
| 568 |
+
return transcribe_punjabi_155750(audio)
|
| 569 |
+
elif lng == "Punjabi_Model3_Trans":
|
| 570 |
+
return transcribe_punjabi_eng_model_155750(audio)
|
| 571 |
|
| 572 |
|
| 573 |
|
|
|
|
| 605 |
# "Hindi","Hindi-trans",
|
| 606 |
# "Odiya","Odiya-trans","Odia_model2","Odia_trans_model2",
|
| 607 |
# "Assamese-LM","Assamese-Model2",
|
| 608 |
+
"Punjabi_Model1","Punjabi_Model1_Trans","Punjabi_Model2","Punjabi_Model2_Trans","Punjabi_Model3","Punjabi_Model3_Trans"],value="Hindi",label="Select Language"),
|
| 609 |
gr.Audio(sources=["microphone","upload"], type="filepath"),
|
| 610 |
#gr.Audio(sources="upload", type="filepath"),
|
| 611 |
#"state"
|