0.1.3

Added

• Improved README for better user experience
• Scripts for baseline experiments of the CVPR 2026 CT Diagnosis Competition

Fixed

• Python version specifier in pyprojects.toml (#39)
• SPECTRE model configuration errors (#40)
• Vision Transformer intermediates and positional embedding handling

0.1.2

Fixed

• Fix token concatenation in VisionTransformer when "pos_embed"=None
• Multiple minor bug fixes

0.1.1

Added

• Decoders for segmentation, including a new SEoMT
• 3D version of Upsample Anything for feature upsampling

Fixed

• Minor code improvements

0.1.0

Added

• Public alpha source code under the MIT license
• SPECTRE foundation model weights for volumetric CT under the CC-BY-NC-SA license
• Self-supervised & cross-modal pretraining workflows
• Preprocessing and augmentation pipelines
• Example scripts for training, evaluation, and zero-shot retrieval