Update app.py

app.py

@@ -37,7 +37,8 @@ def generate(prompt, history):
             max_new_tokens=512,
         )
     
-    return tokenizer.decode(outputs[0], skip_special_tokens=True)
+    generated_tokens = outputs[0, inputs.input_ids.shape[1]:]
+    return tokenizer.decode(generated_tokens, skip_special_tokens=True)
 
 
 demo = gr.ChatInterface(fn=generate, type="messages", examples=["hello", "hola", "merhaba"], title="NanoChat")