Update app.py

app.py

@@ -21,11 +21,8 @@ YT_LENGTH_LIMIT_S = 4800  # 1 hour 20 minutes
 
 device = 0 if torch.cuda.is_available() else "cpu"
 
-# Load model with bitsandbytes quantization
-bnb_config = BitsAndBytesConfig(load_in_4bit=True)
-
 # Load the model
-model = WhisperForConditionalGeneration.from_pretrained(MODEL_NAME, config=bnb_config)
+model = WhisperForConditionalGeneration.from_pretrained(MODEL_NAME, device_map="auto")
 
 # bnb_config = bnb.QuantizationConfig(bits=4)
 pipe = pipeline(task="automatic-speech-recognition", model=model, chunk_length_s=30, device=device)