v0.10.4: Use trick shown here https://pytorch.org/tutorials/recipes/recipes/tu…

…ning_guide.html#use-parameter-grad-none-instead-of-model-zero-grad-or-optimizer-zero-grad

v0.10.3

Option to catch occasional CUDA OOMs to allow for more robust training.

Fix loss reporting and AMP

v0.10.2

Fix reported training loss when using AMP

Autosaving of optimizer state

v0.10.1

Fixes.

Multi-GPU training

Merge pull request #13 from microsoft/dev/miallama/distributedtrainer

A multi-GPU distributed trainer

v0.9.2: Merge pull request #8 from microsoft/dev/miallama/pna-layer

Add PNA layer.

v0.9.1

Update setup.py

GNNs now allow edge attributes beyond edge types

Merge pull request #9 from microsoft/dev/miallama/edge-features

Allow using arbitrary edge features in GNNs

v0.8.7

Update pre-commit.yml

v0.8.6: Merge pull request #7 from microsoft/mallamanis/simplify-saving

Simplify model saving, when parent folder doesn't exist.