Update app.py

app.py

@@ -21,7 +21,6 @@ tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=False)
 model = transformers.AutoModelForCausalLM.from_pretrained(model_name, device_map="auto", load_in_8bit=True)
 
 class OrcaChatBot:
-    # Code below from [microsoft/ari9dam](https://huggingface.co/spaces/ari9dam/Orca-2-13B)  
     def __init__(self, model, tokenizer, system_message="You are Orca, an AI language model created by Microsoft. You are a cautious assistant. You carefully follow instructions. You are helpful and harmless and you follow ethical guidelines and promote positive behavior."):
         self.model = model
         self.tokenizer = tokenizer
@@ -38,6 +37,8 @@ class OrcaChatBot:
         for role, message in self.conversation_history:
             if message.strip():
                 prompt += f"<|im_start|>{role}\n{message}<|im_end|>\n"
+                if role == "assistant":
+                    prompt += f"{self.tokenizer.eos_token}\n"  # Add EOS token after assistant's messages
         prompt += "<|im_start|> assistant\n"
         return prompt