Update handler.py

handler.py

@@ -1,9 +1,10 @@
-from transformers import pipeline, AutoTokenizer
+from transformers import pipeline
+from typing import Any
 
 class EndpointHandler():
     def __init__(self, path=""):
         # create inference pipeline
-        self.pipeline = pipeline("text-to-speech", path)
+        self.pipeline = pipeline("text-to-speech", model=path)
 
     def __call__(self, data: Any) -> Any:
         inputs = data.pop("inputs", data)
@@ -14,5 +15,13 @@ class EndpointHandler():
             prediction = self.pipeline(inputs, **parameters)
         else:
             prediction = self.pipeline(inputs)
+        
         # postprocess the prediction
-        return prediction
+        audio_array = prediction['audio']
+        sampling_rate = prediction['sampling_rate']
+        
+        # If you need to return raw audio data
+        return {
+            "audio": audio_array,
+            "sampling_rate": sampling_rate
+        }