Update app.py

app.py

@@ -9,7 +9,7 @@ import nbimporter
 from transformers import pipeline
 from transformers import AutoProcessor
 from pyctcdecode import build_ctcdecoder
-#from transformers import Wav2Vec2ProcessorWithLM
+from transformers import Wav2Vec2ProcessorWithLM
 from text2int import text_to_int
 from isNumber import is_number
 from processDoubles import process_doubles
@@ -17,6 +17,20 @@ from replaceWords import replace_words
 
 transcriber_hindi_new = pipeline(task="automatic-speech-recognition", model="cdactvm/w2v-bert-2.0-hindi_v1")
 transcriber_hindi_old = pipeline(task="automatic-speech-recognition", model="cdactvm/w2v-bert-2.0-hindi_old")
+processor = AutoProcessor.from_pretrained("cdactvm/w2v-bert-2.0-hindi_v1")
+vocab_dict = processor.tokenizer.get_vocab()
+sorted_vocab_dict = {k.lower(): v for k, v in sorted(vocab_dict.items(), key=lambda item: item[1])}
+decoder = build_ctcdecoder(
+    labels=list(sorted_vocab_dict.keys()),
+    kenlm_model_path="lm.binary",
+    )
+processor_with_lm = Wav2Vec2ProcessorWithLM(
+    feature_extractor=processor.feature_extractor,
+    tokenizer=processor.tokenizer,
+    decoder=decoder
+    )
+processor.feature_extractor._processor_class = "Wav2Vec2ProcessorWithLM"
+transcriber_hindi_lm = pipeline("automatic-speech-recognition", model="cdactvm/w2v-bert-2.0-hindi_v1", tokenizer=processor_with_lm, feature_extractor=processor_with_lm.feature_extractor, decoder=processor_with_lm.decoder)
 
 
 def transcribe_hindi_new(audio):
@@ -27,6 +41,15 @@ def transcribe_hindi_new(audio):
     replaced_words = replace_words(processd_doubles)
     converted_text=text_to_int(replaced_words)
     return converted_text
+    
+def transcribe_hindi_lm(audio):
+    # # Process the audio file
+    transcript = transcriber_hindi_lm(audio)
+    text_value = transcript['text']
+    processd_doubles=process_doubles(text_value)
+    replaced_words = replace_words(processd_doubles)
+    converted_text=text_to_int(replaced_words)
+    return converted_text
 
 def transcribe_hindi_old(audio):
     # # Process the audio file
@@ -50,6 +73,8 @@ def sel_lng(lng, mic=None, file=None):
         return transcribe_hindi_old(audio)
     elif lng == "model_2":
         return transcribe_hindi_new(audio)
+    elif lng== "model_3":
+        return transcribe_hindi_lm(audio)
         
 # demo=gr.Interface(
 #     transcribe,
@@ -68,7 +93,7 @@ demo=gr.Interface(
       
     inputs=[
         gr.Dropdown([
-            "model_1","model_2"],label="Select Language"),
+            "model_1","model_2","model_3"],label="Select Model"),
         gr.Audio(sources=["microphone","upload"], type="filepath"),
     ],
     outputs=[