Fast Neural Network Training with Distributed Training and Google TPUs

TensorFlow Show and Tell
, ,
1

While there is a great ground to cover in “distributed training with TPUs” I have written down a blog post which helps anyone to being with.

My latest PyImageSearch blog post covers the following details:

  • Hardwares used for DL (CPUs, GPUs and TPUs).
  • An efficient data pipeline to use TPUs (using tf.data).
  • A primer on distributed training.

Link: Fast Neural Network Training with Distributed Training and Google TPUs - PyImageSearch