Update app.py

app.py

@@ -38,9 +38,9 @@ def create_chat_template_messages(history, prompt):
 @spaces.GPU
 def generate_responses(prompt, history):
     messages = create_chat_template_messages(history, prompt)
-    wrapped_prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_special_tokens=True, add_generation_prompt=True)
+    wrapped_prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_special_tokens=False, add_generation_prompt=True)
 
-    inputs = tokenizer.encode(wrapped_prompt, add_special_tokens=False, return_tensors="pt").to("cuda")
+    inputs = tokenizer.encode(wrapped_prompt, add_special_tokens=True, return_tensors="pt").to("cuda")
 
     async def custom_sampler_task():
         generated_list = []