v0.2.2 - Massialot

• Remove transpose convolution option in when using interpolate upsampling

v0.2.1 - Massialot

• Cleanup Makefile
• Add interpolation resize test

v0.2.0 - Massialot

• Add option to upsample with interpolation instead of transpose conv.
• Remove separate padding layer and use conv. built-in padding (breaking change from v0.1)
• Improvements to ONNX converter

v0.1.37 - Venezia

• Improve type annotations

v0.1.36 - Venezia

• Fix validation logging in MLFlow

v0.1.35 - Venezia

• Add support for deterministic validation patches

v0.1.34 - Venezia

• Add support for bandit
• Fix warning filter for dataloader
• Add support for a dataset that can prevent a known PyTorch/Python memory leak issue

v0.1.32 - Venezia

• Fix bug in prediction where data not transferred to GPU

v0.1.31 - Venezia

• Updates to support pytorch-lightning~=1.5.1

v0.1.30 - Venezia

• Add option to change label sampling probabilities
• Bump pip version in requirements_dev.txt for security