Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -14,6 +14,7 @@ p1 = pipeline(task="automatic-speech-recognition", model="cdactvm/w2v-bert-2.0-o
|
|
| 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 |
#p3 = pipeline(task="automatic-speech-recognition", model="cdactvm/kannada_w2v-bert_model")
|
| 18 |
#p4 = pipeline(task="automatic-speech-recognition", model="cdactvm/telugu_w2v-bert_model")
|
| 19 |
#p5 = pipeline(task="automatic-speech-recognition", model="Sajjo/w2v-bert-2.0-bangala-gpu-CV16.0_v2")
|
|
@@ -286,6 +287,27 @@ def transcribe_punjabi_eng_model_30000(speech):
|
|
| 286 |
return process_transcription(processed_sentence)
|
| 287 |
return sentence
|
| 288 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 289 |
###########################################
|
| 290 |
def transcribe_odiya(speech):
|
| 291 |
text = p1(speech)["text"]
|
|
@@ -512,10 +534,14 @@ def sel_lng(lng, mic=None, file=None):
|
|
| 512 |
return transcribe_odiya_model2(audio)
|
| 513 |
elif lng == "Odia_trans_model2":
|
| 514 |
return transcribe_odiya_eng_model2(audio)
|
| 515 |
-
elif lng == "
|
| 516 |
return transcribe_punjabi_30000(audio)
|
| 517 |
-
elif lng == "
|
| 518 |
return transcribe_punjabi_eng_model_30000(audio)
|
|
|
|
|
|
|
|
|
|
|
|
|
| 519 |
|
| 520 |
|
| 521 |
|
|
@@ -549,7 +575,11 @@ demo=gr.Interface(
|
|
| 549 |
inputs=[
|
| 550 |
|
| 551 |
#gr.Dropdown(["Hindi","Hindi-trans","Odiya","Odiya-trans","Kannada","Kannada-trans","Telugu","Telugu-trans","Bangala","Bangala-trans"],value="Hindi",label="Select Language"),
|
| 552 |
-
gr.Dropdown([
|
|
|
|
|
|
|
|
|
|
|
|
|
| 553 |
gr.Audio(sources=["microphone","upload"], type="filepath"),
|
| 554 |
#gr.Audio(sources="upload", type="filepath"),
|
| 555 |
#"state"
|
|
|
|
| 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_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")
|
| 19 |
#p4 = pipeline(task="automatic-speech-recognition", model="cdactvm/telugu_w2v-bert_model")
|
| 20 |
#p5 = pipeline(task="automatic-speech-recognition", model="Sajjo/w2v-bert-2.0-bangala-gpu-CV16.0_v2")
|
|
|
|
| 287 |
return process_transcription(processed_sentence)
|
| 288 |
return sentence
|
| 289 |
|
| 290 |
+
def transcribe_punjabi_75000(speech):
|
| 291 |
+
text = punjaib_modle_30000(speech)["text"]
|
| 292 |
+
text = text.replace("[PAD]","")
|
| 293 |
+
if text is None:
|
| 294 |
+
return "Error: ASR returned None"
|
| 295 |
+
return text
|
| 296 |
+
|
| 297 |
+
def transcribe_punjabi_eng_model_75000(speech):
|
| 298 |
+
trn = Transliterator(source='pan', target='eng', build_lookup=True)
|
| 299 |
+
text = punjaib_modle_30000(speech)["text"]
|
| 300 |
+
text = text.replace("[PAD]","")
|
| 301 |
+
if text is None:
|
| 302 |
+
return "Error: ASR returned None"
|
| 303 |
+
sentence = trn.transform(text)
|
| 304 |
+
if sentence is None:
|
| 305 |
+
return "Error: Transliteration returned None"
|
| 306 |
+
replaced_words = replace_words(sentence)
|
| 307 |
+
processed_sentence = process_doubles(replaced_words)
|
| 308 |
+
return process_transcription(processed_sentence)
|
| 309 |
+
return sentence
|
| 310 |
+
|
| 311 |
###########################################
|
| 312 |
def transcribe_odiya(speech):
|
| 313 |
text = p1(speech)["text"]
|
|
|
|
| 534 |
return transcribe_odiya_model2(audio)
|
| 535 |
elif lng == "Odia_trans_model2":
|
| 536 |
return transcribe_odiya_eng_model2(audio)
|
| 537 |
+
elif lng == "punjabi_30000":
|
| 538 |
return transcribe_punjabi_30000(audio)
|
| 539 |
+
elif lng == "punjabi_trans_30000":
|
| 540 |
return transcribe_punjabi_eng_model_30000(audio)
|
| 541 |
+
elif lng == "punjabi_75000":
|
| 542 |
+
return transcribe_punjabi_75000(audio)
|
| 543 |
+
elif lng == "punjabi_trans_75000":
|
| 544 |
+
return transcribe_punjabi_eng_model_75000(audio)
|
| 545 |
|
| 546 |
|
| 547 |
|
|
|
|
| 575 |
inputs=[
|
| 576 |
|
| 577 |
#gr.Dropdown(["Hindi","Hindi-trans","Odiya","Odiya-trans","Kannada","Kannada-trans","Telugu","Telugu-trans","Bangala","Bangala-trans"],value="Hindi",label="Select Language"),
|
| 578 |
+
gr.Dropdown([
|
| 579 |
+
# "Hindi","Hindi-trans",
|
| 580 |
+
# "Odiya","Odiya-trans","Odia_model2","Odia_trans_model2",
|
| 581 |
+
# "Assamese-LM","Assamese-Model2",
|
| 582 |
+
"punjabi_30000","punjabi_trans_30000","punjabi_75000","punjabi_trans_75000"],value="Hindi",label="Select Language"),
|
| 583 |
gr.Audio(sources=["microphone","upload"], type="filepath"),
|
| 584 |
#gr.Audio(sources="upload", type="filepath"),
|
| 585 |
#"state"
|