FieldAccessors are now supported by python pickle. This is important for error prune usage of field accessors in pyTorch Datasets and DataLoaders. Adding LayeredDataset that combines layers across channels for loading.
1.0.7
Easier structure for creating metadata
1.0.6
Small bug fixes for add_layer method (UINT32 is now working) Gridtracers are now documented and Sampling-Line-Tracing hybrid algorithm is now implemented.
1.0.5
pyTorch Integration now works more error prune when using pytorch DataLoaders with multiple workers.
1.0.4
... and removing support for pickle for field accessors and metadata