Pytorch Lightning Integration

PyTorch Lightning helps organize PyTorch code and decouple the science code from the engineering code. It’s more of a style-guide than a framework. By organizing PyTorch code under a LightningModule, Lightning makes things like TPU, multi-GPU and 16-bit precision training (40+ other features) trivial.

Using the Comet Logger

The Pytorch lightning library includes a Comet logger that can be used to track the training runs in the Comet platform.

Using the Comet logger is as simple as making two small changes to your Pytorch Lightning code: 1. Enabling the save_hyperparameters method: By adding self.save_hyperparameters within your LightningModule's __init__ method Lightning to store all the provided arguments under the self.hparams attribute [Read more] 2. Adding the Comet Logger to your method: Pytorch Lightning includes a Comet logger (pytorch_lightning.loggers.CometLogger) that needs to be added to the fit method

For more information, please see: * Comet + Pytorch Lightning