resource type	Name	urL	comment
Cloud platform	Amazon EC2	https://aws.amazon.com/ec2/	Most popular cloud platform
	Microsoft Azure	https://azure.microsoft.com/	Second-largest cloud platform
Plug-and-play cloud GPU services	FloydHub	https://www.floydhub.com/	All startups in the GPU service space; pay-by-the-hour model on top of basic monthly subscriptions
	PaperSpace	https://www.paperspace.com/
	Valohai	https://valohai.com/
	Google CloudML	https://cloud.google.com/ml-engine/	Can run your own models on Google’s hardware, including tensor processor units
	Google Colaboratory	https://colab.research.google.com/	Notebook environment with free GPUs (during 12 h)
Design services for deep learning models	Fabrik	https://github.com/Cloud-CV/Fabrik/	Model export to Keras code; no training
	IBM Data Cloud	https://datascience.ibm.com/	Model export to Keras, PyTorch, TensorFlow or Caffe
	DeepCognition	http://deepcognition.ai/	Training and evaluation included
Prebuilt images with CUDA support	Docker Hub	https://hub.docker.com/r/nvidia/cuda/	Docker images from NVIDIA with CUDA/cuDNN GPU support
	Amazon Deep Learning AMIs	https://aws.amazon.com/machine-learning/amis/	Amazon Machine Images (AMIs) with GPU support
Software libraries	Keras	https://keras.io/	More high-level than TensorFlow(general) (below) but can be integrated with it in many ways
	TensorFlow	https://www.tensorflow.org/	Developed by Google; most popular deep learning framework
	PyTorch	http://pytorch.org/	Developed by Facebook
Software libraries (specific for genomics)	DragoNN	https://kundajelab.github.io/dragonn/	Tutorials included
	Kipoi	http://kipoi.org/	Model zoo for deep learning in genomics
Educational resources	fast.ai	http://www.fast.ai/	E.g., Deep Learning for Coders 1 and 2
	Coursera	https://www.coursera.org/specializations/deep-learning/	Deep-learning-specialization course package
	Textbook	http://neuralnetworksanddeeplearning.com/	Free online textbook with example code
	Fast.ai tips on configuring a deep learning environment	https://github.com/reshamas/fastai_deeplearn_part1/blob/master/README.md#platforms-for-using-fastai-gpu-required/	Instructions for configuring deep learning frameworks for a variety of platforms; from the fast.ai course but general; the details of these procedures change quickly
	Setting up TensorFlow with GPU on Google Cloud Engine	https://medium.com/google-cloud/jupyter-tensorflownvidia-gpu-docker-google-compute-engine-4a146f085f17/	Recipe for Docker-based setup of Google Cloud instance with TensorFlow, GPU support and Jupiter Notebooks