v2.0.0

Major improvement to the original code, specifically when it comes to user inputs.

Airtight Input Validation with Pydantic

• This package greatly simplifies the input validation process by using Pythons Typing module
• Deeper logic between configs is enforced to ensure no errors slip through
• Simplified configs by removing the complex branching from before associated with cross validation
• Flowchart illustrating the logic behind training_set, validation_set, and auto_split
• Changed the names of many configs to more reasonable input

EfficientNetB7 Advanced Finetuning

• Fixed a major bug where the backbone could not be frozen
• Added a feature to use random weights or ImageNet pre-trained weights
• Implemented block-level unfreezing to more deeply finetune the backbone

Other than these changes, the code works functionally the same.

Initial Release

This is the first "official" release for this project.