Skip to content
Mar 19, 2019
Add horovodrun
Feb 21, 2019
PySpark, Apache MXNet, autotuning, TensorFlow eager execution, mixed-…
…precision & embedding improvements
Nov 28, 2018
Force allreduce of all gradients in step(), bugfixes
Oct 25, 2018
FP16 hierarchical allreduce
Oct 2, 2018
PyTorch 1.0, TF-Keras, FP16 ops on GPU
Sep 6, 2018
Parallelized hierarchical allreduce
Aug 31, 2018
Support for the upcoming PyTorch release
Jul 29, 2018
Add compatibility with PyTorch 0.4.1
Jul 14, 2018
Support for IBM PowerAI DDL & APIs to restore optimizer state
Jul 9, 2018
Critical Bugfix: PyTorch must wait for GPU data before allreduce
You can’t perform that action at this time.