Update app.py

app.py

@@ -16,17 +16,14 @@ import contextlib
 
 from sklearn.cluster import AgglomerativeClustering
 import numpy as np
-
-from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
+optput=""
 
 model = whisper.load_model("large-v2")
 embedding_model = PretrainedSpeakerEmbedding( 
     "speechbrain/spkrec-ecapa-voxceleb",
     device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
-    )
+)
 
-output=""
-######################################################
 def audio_to_text(audio, num_speakers):
   path, error = convert_to_wav(audio)
   if error is not None:
@@ -36,7 +33,6 @@ def audio_to_text(audio, num_speakers):
   if duration > 4 * 60 * 60:
     return "Audio duration too long"
 
-
   result = model.transcribe(path)
   segments = result["segments"]
 
@@ -48,7 +44,6 @@ def audio_to_text(audio, num_speakers):
     add_speaker_labels(segments, embeddings, num_speakers)
   global output
   output = get_output(segments)
-  #short_sum=text_to_short_summary(str(output))
   return output
 
 def convert_to_wav(path):
@@ -100,10 +95,7 @@ def get_output(segments):
         output += '\n\n'
       output += segment["speaker"] + ' ' + str(time(segment["start"])) + '\n\n'
     output += segment["text"][1:] + ' '
- 
-    return output
-
-#########################################################################
+  return output
 
 
 def text_to_short_summary():