Update app.py

app.py

@@ -1,11 +1,13 @@
-from faster_whisper import WhisperModel
+#from faster_whisper import WhisperModel
+import whisper
 import pandas as pd
 import gradio as gr
 import psutil
 import time
 import whisperx
 
-model = WhisperModel('large-v2', device="cuda", compute_type="float16")
+#model = WhisperModel('large-v2', device="cuda", compute_type="float16")
+model = whisper.load_model('large-v2')
 
 def speech_to_text(mic=None, file=None, lang=None, task='transcribe'):
     if mic is not None:
@@ -17,11 +19,13 @@ def speech_to_text(mic=None, file=None, lang=None, task='transcribe'):
     print(lang, task)
     
     time_start = time.time()
-    segments, info = model.transcribe(audio, task=task, language=lang, beam_size=5)
+    #segments, info = model.transcribe(audio, task=task, language=lang, beam_size=5)
+    results = model.transcribe(audio, task=task, language=lang, beam_size=5)
     print("Detected language '%s' with probability %f" % (info.language, info.language_probability))
 
     #  Decode audio to Text
-    objects = [s._asdict() for s in segments]
+    #objects = [s._asdict() for s in segments]
+    objects = results["segments"]
     print(objects)
     time_end = time.time()
     time_diff = time_end - time_start