Tensorflow related content.
When Deep Learning with GPUs, use a Cluster Manager TLDR; If you are employing a team of Data Scientists or working with Deep Learning, a cluster manager that can share GPUs between your team will Read more
Goodbye Horovod, Hello CollectiveAllReduce Hopsworks is replacing Horovod with Keras/TensorFlow’s new CollectiveAllReduceStrategy tl;dr Distributed Deep Learning is producing state-of-the-art results in problems from NLP to machine translation to image classification. Hopsworks continually pushes the state-of-the-art Read more
Distributed Filesystems for Deep Learning More training data gives predictable gains in prediction accuracy tl;dr When you train deep learning models with lots of high quality training data, you can beat state-of-the-art prediction Read more
When the Resource Scheduler is not enough – GPUs also need application-aware scheduling TLDR; This article describes how we use dynamic executors in PySpark to ensure GPUs are only allocated to executors only when they Read more