Deep Learning with AWS
Update: I recommend reading Nvidia’ GPU Server Application Deployment Made Easy for an alternative approach (Ansible) to setting up the NVDIA-Docker plugin. There are many tutorials on how to leverage Amazon’s supreme computing power to perform deep learning tasks. I would like to take this opportunity to contribute to that collection. I started working with Amazon’s EC2 instances for deep learning by reading some of these tutorials, including predominantly: Installing CUDA, OpenCL, and PyOpenCL on AWS EC2 Deep Learning Tutorial for Kaggle’s Facial Keypoints Detection Installing TensorFlow on AWS I would launch an instance, install the necessary dependencies and create an AMI. This would be repeated for each deep learning framework that I wanted to use (Theano, Caffe, Tensorflow etc.) ...