Update handler.py

handler.py

@@ -8,7 +8,7 @@ class EndpointHandler:
         self.processor = AutoProcessor.from_pretrained(path)
         self.model = MusicgenForConditionalGeneration.from_pretrained(path, torch_dtype=torch.float16).to("cuda")
 
-    def __call__(self, data: Dict[str, Any]) -> bytes:
+    def __call__(self, data: Dict[str, Any]) -> Dict[str, str]:
         """
         Args:
             data (:dict:):
@@ -29,6 +29,6 @@ class EndpointHandler:
                 outputs = self.model.generate(**inputs, do_sample=False, max_new_tokens=400)
 
         # postprocess the prediction
-        prediction = outputs[0].cpu().numpy().tobytes()
+        prediction = outputs[0].cpu().numpy().tolist()
 
-        return prediction
+        return [{"generated_audio": prediction}]