fix model size

src/se_extractor.py

@@ -9,6 +9,7 @@ from whisper_timestamped.transcribe import get_audio_tensor, get_vad_segments
 
 # Run on GPU with FP16
 model = None
+model_size = 'medium'
 def split_audio_whisper(audio_path, target_dir='processed',needs_offset=True):
     print("in whisper split")
     model = WhisperModel('medium', device="cuda:0", compute_type="float16")
@@ -156,13 +157,12 @@ def generate_voice_segments(audio_path, target_dir='processed', vad=True):
 
 
 def load_model():
-    model = WhisperModel(model_size, device="cpu", compute_type="int8")
+    model = WhisperModel('medium', device="cpu", compute_type="int8")
 
 
 def extract_segments_to_cut_audio(max_duration,audio_path,target_dir='processed'):
-    global model
-    if model is None:
-        model = WhisperModel(model_size, device="cuda", compute_type="float16")
+    
+    model = WhisperModel('medium', device="cuda:0", compute_type="float16")
     audio = AudioSegment.from_file(audio_path)
     max_len = len(audio)