Update app.py

app.py

@@ -22,31 +22,6 @@ tokenizer.padding_side = "left"
 if tokenizer.pad_token is None:
     tokenizer.pad_token = tokenizer.eos_token
 
-def generate_response(messages):
-    # Convert list of dicts (messages) into the required format by the tokenizer
-    # messages should be a list of {"role": "user"/"assistant", "content": "<text>"}
-    
-    message = tokenizer.apply_chat_template(
-        messages, tokenize=False, add_generation_prompt=True
-    )
-    
-    # Tokenize inputs
-    tokenized_inputs = tokenizer(message, return_tensors="pt", padding=True, truncation=True, max_length=512).to(model.device)
-    
-    # Generate response
-    outputs = model.generate(
-        tokenized_inputs["input_ids"],
-        attention_mask=tokenized_inputs["attention_mask"],
-        max_new_tokens=2048,
-        temperature=0.3,
-        eos_token_id=tokenizer.eos_token_id,
-        pad_token_id=tokenizer.eos_token_id
-    )
-    
-    # Decode and return output
-    output_text = tokenizer.decode(outputs[0], skip_special_tokens=True)
-    return output_text
-
 def generate_response(messages):
     # Convert list of dicts (messages) into the required format by the tokenizer
     # messages should be a list of {"role": "user"/"assistant", "content": "<text>"}
@@ -68,7 +43,6 @@ def generate_response(messages):
         eos_token_id=[
             tokenizer.eos_token_id,
             tokenizer.convert_tokens_to_ids("<|eot_id|>"),
-            tokenizer.convert_tokens_to_ids("<|assistant|>"),
         ],
         pad_token_id=tokenizer.eos_token_id
     )