[TIL] Deploy AI Model Developed w/ Flask API Using AWS EC2 & Connect to Nest.js Project

Using AI Model for Emergency Level Prediction: EC2 Server Scale Up

t2.micro (free tier)

🚫 Killed: Indicates that the process was terminated by the operating system. ⇒ pip install tensorflow was killed by the OS. It was interrupted due to insufficient memory on the AWS EC2 free tier instance.

t3.small

🚫 Similar to t2.micro, it was killed.

t3.medium

✅ pip install tensorflow successfully installed TensorFlow. In the case of the Flask API, it does not require GPU since it loads a pre-trained model for predicting emergency levels. Therefore, a CPU EC2 instance is sufficient.

• Additional commands executed on the EC2 instance / Installed modules:

  • sudo apt update

  • sudo apt install python3-pip

  • sudo apt-get install python3-venv

  • sudo apt-get install default-jdk ⇒ JDK installation for using konlpy

  • sudo pip install konlpy

  • pip install intel-tensorflow==2.12.0 (Installed to address version-related errors)

    • pip install -U keras_applications==1.0.6 --no-deps pip install -U keras_preprocessing==1.0.5 --no-deps

Connecting to the EC2 Server Instance

1. Port forwarding

   sudo iptables -t nat -A PREROUTING -i eth0 -p tcp --dport 80 -j REDIRECT --to-port 5000

   

   My apologies for the mistake... I kept accessing the t3.small IP address instead of t3.medium. It was a silly mistake, not a troubleshooting issue.

2. Running the Flask server in the background

   nohup python -u app.py & ⇒ Run in the background

   tail -f nohup.out ⇒ Check the logs

   lsof -i :5000 ⇒ Check the PID

   sudo kill -9 [PID] ⇒ Terminate the running process

   Flask - Running in the Background with nohup