Horovod (machine learning)
| Horovod | |
|---|---|
| Developer | Uber |
| Initial release | August 9, 2017 |
| Stable release | v0.28.1
/ June 12, 2023 |
| Written in | Python, C++, CUDA |
| Platform | Linux, macOS, Windows |
| Type | Artificial intelligence ecosystem |
| License | Apache License 2.0 |
| Website | horovod |
Horovod is a free and open-source distributed deep learning training framework for TensorFlow, Keras, PyTorch and Apache MXNet.
It is designed to scale existing single-GPU training scripts to efficiently run on multiple GPUs and computer nodes with minimal code changes, using synchronous data-parallel training based on the ring-allreduce communication pattern. Horovod was initially developed at Uber and released as an open-source project in 2017, and is now hosted by the LF AI & Data Foundation, a project of the Linux Foundation.