TensorFlow 0.8 gives the people what they want: Distributed computing support

Google’s TensorFlow AI has been updated and now comes with a host of new offerings. TensorFlow 0.8 with distributed computing support includes “everything you need to train distributed models on your own infrastructure,” Google said in a post at the Google Research Blog. Google’s TensorFlow shortens the training process for some machine learning models. The tech can reduce training time from weeks down to hours. When TensorFlow became an open source offering, Google said the most frequently requested feature was distributed training support. The new version of TensorFlow will work in tandem with Google’s newly released Cloud Machine Learning. Cloud Machine Learning will work to “train and serve” user’s TensorFlow models via the Google Cloud Platform. Google published a distributed trainer for the Inception image classification neural network in TensorFlow…


Link to Full Article: TensorFlow 0.8 gives the people what they want: Distributed computing support