TensorFlow-Neuron (ResNet-50 Tutorial)

Steps Overview:

  1. Launch an EC2 Compilation Instance (recommended instance: z1d.4xlarge)
  2. Install TensorFlow-Neuron and Neuron-Compiler on the Compilation Instance
  3. Compile the compute-graph on the compilation-instance, and copy the artifacts into the deployment-instance
  4. Install TensorFlow-Neuron and Neuron-Runtime on Deployment Instance
  5. Deploy inferences inference on the Deployment Instance (Inf1)