Streamlit stop by itself after running for a while on AWS EC2

Hello friends,
I have a problem when hosting my streamlit app on an AWS EC2 instance. I was wondering if there is anyone who can help? Basically, my app allows a real-time training job running on the EC2, which takes up to 2 hours. This training process is smooth without any trouble for the first tens of epochs (around 20 min). But after a while, this training process will stop by itself before it completes. Neither did I refresh the app nor did I interrupt anything. It is just stopped by itself… And the CPU usage of the EC2 instance becomes low as if there is nothing running there.
However, I tried hosting this app on my local computer and everything is fine. The whole training process can be finished without any problem.
Please help me if anyone knows what’s going on… Thanks!