TensorFlow 2.15 with Keras 3.0 Deep Learning Stack

Launching the AMI

1. Navigate to AWS Marketplace and subscribe to the TensorFlow 2.15 + Keras 3.0 Deep Learning Environment
2. Click "Continue to Subscribe" then "Continue to Configuration"
3. Choose your preferred region and delivery method (typically 64-bit x86)
4. Click "Continue to Launch" and select "Launch through EC2"
5. Choose your instance type based on workload requirements:
   • Development: t3.medium or t3.large
   • GPU workloads: g4dn.xlarge or p3.2xlarge
   • Production: c5.2xlarge or larger
6. Configure instance details, add storage (minimum 20GB recommended), and add tags if needed
7. Configure security group to allow SSH (port 22) and Jupyter (port 8888)
8. Review and launch the instance using your existing key pair or create a new one

Initial Access and Setup

1. Once instance is running, connect via SSH: ssh -i your-key.pem ubuntu@your-instance-ip
2. The environment is pre-configured and ready for immediate use
3. All packages are installed in the system Python environment
4. No additional setup or configuration required

Using Jupyter Notebook

1. Start Jupyter Lab with: jupyter lab --ip=0.0.0.0 --port=8888 --no-browser --allow-root
2. Note the access token displayed in the terminal output
3. Open your web browser and navigate to: http://your-instance-ip:8888 
4. Enter the token when prompted to access Jupyter Lab
5. For persistent Jupyter sessions, consider using screen or tmux

Development Workflow

1. Create your Python scripts or Jupyter notebooks in the home directory
2. Import TensorFlow and Keras as follows: import tensorflow as tf import keras
3. Verify installation with: print(f"TensorFlow version: {tf.version}") print(f"Keras version: {keras.version}")
4. Use pre-installed libraries: numpy, pandas, matplotlib, scikit-learn, opencv
5. Additional packages can be installed using pip

GPU Configuration (Optional)

1. For GPU instances, ensure you select GPU-optimized instance types
2. GPU drivers are pre-installed and configured
3. TensorFlow will automatically detect and use available GPUs
4. Verify GPU detection with: tf.config.list_physical_devices('GPU')

Best Practices

• Regularly update packages using: pip install --upgrade package-name
• Use virtual environments for project-specific dependencies
• Monitor instance performance through AWS CloudWatch
• Set up regular backups of your important work
• Use EBS snapshots for persistent storage needs
• Configure security groups to restrict access to necessary ports only

Stopping and Terminating

1. Stop instance when not in use to save costs
2. Backup important data before terminating instances
3. Terminate instance through EC2 console when no longer needed
4. Remember that instance storage is ephemeral and data will be lost on termination

Support and Resources

• Check the documentation for common issues and solutions
• Review AWS documentation for EC2 instance management
• Monitor instan