Update handler.py

handler.py

@@ -1,4 +1,4 @@
-from typing import  Dict
+from typing import  Dict, Any
 import librosa
 import numpy as np
 import torch
@@ -52,14 +52,16 @@ class EndpointHandler():
         self.vocoder = SpeechT5HifiGan.from_pretrained("microsoft/speecht5_hifigan")
 
 
-    def __call__(self, data: Dict[str]) -> Dict[str, str]:
-        """
+    def __call__(self, data: Dict[str, Any]) -> bytes:
+        """_summary_
+
         Args:
-            data (:obj:):
-                includes the deserialized audio file as bytes
-        Return:
-            A :obj:`dict`:. base64 encoded image
+            data (Dict[str, Any]): _description_
+
+        Returns:
+            bytes: _description_
         """
+
         # process input
 
         if len(text.strip()) == 0:
@@ -75,4 +77,4 @@ class EndpointHandler():
         speaker_embedding = torch.tensor(speaker_embedding)
         speech = self.model.generate_speech(input_ids.to('cuda'), speaker_embedding.to('cuda'), vocoder=vocoder.to('cuda'))
         speech = nr.reduce_noise(y=speech.to('cpu'), sr=16000)
-        return (16000, speech)
+        return speech.tobytes()