- fix: prepare_data_per_node is a flag and was incorrectly used as a replacement for prepare_data.
3.8.2
- fix: points not dropped case in subsampling when the subtile contains only one point - fix: type error in edge case when dropping points in DropPointsByClass (when there is only one remaining point)
3.8.1
- fix: propagate input las format to output las (in particular epsg which comes either from input or config)
3.8.0
- dev: log confusion matrices to Comet after each epoch. - fix: do not mix the two way to log IoUs to avoid known lightning [Common Pitfalls](https://lightning.ai/docs/torchmetrics/stable/pages/lightning.html#common-pitfalls).
3.7.1
- fix: edge case when saving predictions under Classification channel, without saving entropy.
3.7.0
- Update all versions of Pytorch, Pytorch Lightning, and Pytorch Geometric. Changes are retrocompatible for models trained with older versions (with adjustment to the configuration file). - Refactor logging of single-class IoUs to go from num_classes+1 torchmetrics instances to only 1.