using-pytorch-lightning

page_type	languages	products	description	experimental
sample	python azurecli	azure-machine-learning	Learn how to train and log metrics with [PyTorch Lightning](https://github.com/PyTorchLightning/pytorch-lightning) and Azure ML.	issues with multinode pytorch lightning

Train with PyTorch Lightning

PyTorch Lightning is a lightweight open-source library that provides a high-level interface for PyTorch.

The model training code for this tutorial can be found in src. This tutorial goes over the steps to run PyTorch Lightning on Azure ML, and it includes the following parts:

1. train-single-node: Train single-node and single-node, multi-GPU PyTorch Lightning on Azure ML.
2. log-with-tensorboard: Use Lightning's built-in TensorBoardLogger to log metrics and leverage Azure ML's TensorBoard integration.
3. log-with-mlflow: Use Lightning's MLFlowLogger to log metrics and leverage Azure ML's MLflow integration.