TensorFlow 1.8.0-rc0 釋出,bug 修復版本

類別: IT
TensorFlow

TensorFlow 1.8.0-rc0 釋出,此版本包括很多效能改進和bug修復。主要特性和改進包括:

主要特性和改進

  • Can now pass tf.contrib.distribute.MirroredStrategy()to tf.estimator.RunConfig()to run an Estimator model on multiple GPUs on one machine.

  • Add tf.contrib.data.prefetch_to_device(), which supports prefetching to GPU memory.

  • Added Gradient Boosted Trees as pre-made Estimators: BoostedTreesClassifier, BoostedTreesRegressor.

  • Add 3rd generation pipeline config for Cloud TPUs which improves performance and usability.

  • tf.contrib.bayesflowis moving out to it's own repo.

  • Added tf.contrib.{proto,rpc}to allow generic proto parsing and RPC communication.

Bug 修復和其他改變

  • tf.data:

    • Add tf.contrib.data.prefetch_to_device, which enables prefetching dataset elements to GPU memory.

    • Add tf.contrib.data.AUTOTUNE, which allows the tf.data runtime to automatically tune the prefetch buffer sizes based on your system and environment.

    • Add tf.contrib.data.make_csv_datasetfor building datasets of CSV files.

  • Eager Execution:

    • With eager execution Datasets can now be used as standard python iterators (for batch in dataset:). Both Dataset.__iter__()and Dataset.make_one_shot_iterator()can now be used to create iterators when eager execution is enabled.

    • Automatic device placement has been enabled (i.e., use a GPU if available automatically, without requiring an explicit with tf.device(“/gpu:0”)) (Fixes #14133)

    • tf.GradientTapehas moved out of contrib.

完整內容請檢視釋出主頁

下載地址:

  • Source code(zip)

  • Source code(tar.gz)

TensorFlow 1.8.0-rc0 釋出,bug 修復版本原文請看這裡