How AI describes Horovod

Horovod is an open-source distributed deep learning training framework that supports PyTorch, TensorFlow, Keras, and Apache MXNet. Originally developed by Uber, it enables fast, scalable model training across hundreds of GPUs with only a few lines of Python code, delivering high scaling efficiency. It can be deployed on-premise or in the cloud (AWS, Azure, Databricks), can run on Apache Spark, and provides portable, framework-agnostic training across multiple frameworks as part of the LF AI Foundation.