Posted On: Sep 6, 2018
Amazon SageMaker now supports version 1.10 in its pre-built TensorFlow containers. This makes it easier to run TensorFlow scripts, while taking advantage of the capabilities Amazon SageMaker offers, including a library of high-performance algorithms, managed and distributed training with automatic model tuning, one-click deployment, and managed hosting.
TensorFlow 1.10 comes with a number of performance improvements and bug fixes including support for complex number data types such as complex64 within TensorFlow's lightweight solution for mobile and embedded devices, popularly known as tf.lite. Additionally, the improved local run behavior in tf.estimator.train_and_evaluate removes the need to reload checkpoints for evaluation.
The Amazon SageMaker built-in TensorFlow container with version 1.10 is now available in the US East (N. Virginia), US East (Ohio), US West (Oregon), Europe (Ireland), Europe (Frankfurt), Asia Pacific (Tokyo), Asia Pacific (Seoul) and Asia Pacific (Sydney) AWS Regions. For additional information, please visit the documentation here.