Update handler.py

handler.py

@@ -188,7 +188,7 @@ class EndpointHandler:
         inputs = self.tokenizer(prompt, return_tensors="pt")
         input_ids = inputs["input_ids"].to(device)
 
-        max_new_tokens = 600 - len(input_ids[0]) 
+        max_new_tokens = self.model.config.max_seq_len - len(input_ids[0]) 
         try:
             if set(["up", "down"]) & set(config_data.keys()):
                 acc_embs_up1 = []