Update handler.py

handler.py

@@ -62,14 +62,15 @@ class EndpointHandler:
 
     def __call__(self, data: Any) -> List[List[Dict[str, float]]]:
         embedding = data.pop("embedding", None)
+        ada_embedding = torch.tensor(embedding).unsqueeze(0)
         max_length=200
         with torch.no_grad():
-            outputs = self.model(ada_embedding=embedding, decoded_tkns=None)
+            outputs = self.model(ada_embedding=ada_embedding, decoded_tkns=None)
         decoded_tkns = outputs.logits.argmax(dim=-1)
     
         for _ in range(max_length):
             with torch.no_grad():
-                outputs = self.model(ada_embedding=embedding, decoded_tkns=decoded_tkns)
+                outputs = self.model(ada_embedding=ada_embedding, decoded_tkns=decoded_tkns)
             
             # Get the most likely next token, sampled from top k
             logits = outputs.logits[:, -1]