Update app.py

app.py

@@ -15,7 +15,9 @@ def main():
     [report](https://api.wandb.ai/links/dmeltzer/7an677es) for details on our training procedure.\
     \n\nThe two fine-tuned models (BART-Large and FLAN-T5-Base) are hosted on AWS using a combination of AWS Sagemaker, Lambda, and API gateway.\
     GPT-3.5 is called using the OpenAI API and the FLAN-T5-XXL model is hosted by HuggingFace and is called with their Inference API.\
-    \n \n **Disclaimer**: You may recieve an error message when calling the FLAN-T5-XXL model since the Inference API takes around 20 seconds to load the model.\
+    \n \n **Disclaimer**: When first running this application it may take around 30-60 seconds for the first two responses to load because of the cold start problem with AWS Lambda.
+    \ You may recieve also an error message when calling the FLAN-T5-XXL model since the Inference API takes around 20 seconds to load the model.\
+    Both issues should disappear on any subsequent calls to the application and we apologize for the delay.
     ")
     
     AWS_checkpoints = {}