PyTorch Lightning

PyTorch Lightning is a very light-weight structure for PyTorch — it’s more of a style guide than a framework. But once you structure your code, we give you free GPU, TPU, 16-bit precision support and much more!

Figure

Lightning is just structured PyTorch

Metrics

This release has a major new package inside lightning, a multi-GPU metrics package!

import pytorch_lightning.metrics.functional as plm

pred = torch.tensor([0, 1, 2, 3])target = torch.tensor([0, 1, 2, 2])

# many popular classification metrics and moreplm.accuracy(pred, target)plm.auc(pred, target)plm.auroc(pred, target)plm.average_precision(pred, target)plm.confusion_matrix(pred, target)plm.dice_score(pred, target)plm.f1_score(pred, target)…# 20+ metrics

view raw

metrics.py hosted with ❤ by GitHub

#overviews #gpu #python #pytorch

GPU Metrics Library  in New PyTorch Release
2.50 GEEK