Update handler.py

handler.py

@@ -8,16 +8,17 @@ class EndpointHandler():
         device = "cuda:0" if torch.cuda.is_available() else "cpu"
         torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
         model_id = "openai/whisper-large-v3"
-        self.model = AutoModelForSpeechSeq2Seq.from_pretrained(
+
+        model = AutoModelForSpeechSeq2Seq.from_pretrained(
             model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, use_safetensors=True
         )
-        self.model.to(device)
+        model.to(device)
 
         processor = AutoProcessor.from_pretrained(model_id)
 
         self.pipeline = pipeline(
             "automatic-speech-recognition",
-            model=self.model,
+            model=model,
             tokenizer=processor.tokenizer,
             feature_extractor=processor.feature_extractor,
             max_new_tokens=128,
@@ -27,6 +28,7 @@ class EndpointHandler():
             torch_dtype=torch_dtype,
             device=device,
         )
+        self.model = model
 
 
     def __call__(self, data: Dict[str, bytes]) -> Dict[str, str]: