How to Create a Jupyter Notebook on AWS with Deep Learning AMIs

AWS recently released the Deep Learning AMIs, which come pre-installed with Jupyter and various deep learning frameworks like MXNet, TensorFlow, Caffe, Theano, Torch, and CNTK. This significantly simplifies the process of setting up your machine learning environment on AWS! This article guides you through the steps to set up a Jupyter Notebook on AWS, with detailed instructions and helpful tips for both beginners and experienced users.

Setting Up the AWS Environment

To start, create an AWS EC2 instance using one of the Deep Learning AMIs. AWS provides two versions:

Amazon Linux version: Deep Learning AMI Ubuntu version: Deep Learning AMI Ubuntu Version

Create a Security Group for Your EC2 Instance

Before you launch your instance, create a security group that allows incoming traffic on the necessary ports.

CLI Commands:

aws ec2 create-security-group --group-name group_name --description My Jupyter Notebook Security Group

aws ec2 authorize-security-group-ingress --group-id my-security-group --protocol tcp --port 22 --cidr 0.0.0.0/0 --region my-region

aws ec2 authorize-security-group-ingress --group-id my-security-group --protocol tcp --port 443 --cidr 0.0.0.0/0 --region my-region

aws ec2 authorize-security-group-ingress --group-id my-security-group --protocol tcp --port 8888 --cidr 0.0.0.0/0 --region my-region

Connect to Your EC2 Instance

Once the instance is launched, connect to it using SSH. If you prefer command line interface (CLI), you can use the following command:

ssh -i path_to_ssh_key 

If you're using a graphical interface, you can use the AWS Management Console to connect to your instance.

Install Required Packages and Start Jupyter Notebook

After connecting, update your package manager and install Git:

sudo apt-get updatesudo apt-get install -y git

Clone the Anaconda repository and install the latest version:

git clone git_repocd jupyter_notebook/install_anaconda

To start Jupyter notebook, use the following commands:

cd notebooktmux new -s nbjupyter notebook --certfile~/.jupyter/wp_ --keyfile ~/.jupyter/wp_ --no-browser

Access Jupyter Notebook Remotely

Now, you can access Jupyter notebook remotely via your browser:

https://public_dhcp_id:8888/

Additional Resources

If you need more detailed instructions, you can refer to the following resources:

Run Jupyter Notebook Server on AWS EC2 Project Jupyter on Amazon EC2

Conclusion

Diligently following these steps should help you set up a Jupyter Notebook on AWS with ease. If you have any questions or need further assistance, feel free to comment below! Happy coding!