Update services/chat_service.py

services/chat_service.py

@@ -233,7 +233,7 @@ class ChatService:
                 early_stopping=False
             ) 
 
-            input_ids = tokenizer.encode(prompt, return_tensors="pt", truncation=True, max_length=4096).to("cpu")
+            input_ids = self.tokenizer.encode(prompt, return_tensors="pt", truncation=True, max_length=4096).to("cpu")
 
             response = self.tokenizer.decode(outputs[0][input_ids.shape[-1]:], skip_special_tokens=True)