Code and Dataset Preps

[liellnima]

Code:

• New repo
• Docs, Readme, etc.
• Label 1st version
• Check with Francis
• Reproducability / Runnability (tested by Julia)

Later:

• Clean up Code (comments, remove unsused scripts, ambiguous variable etc)

Dataset:

• Get data from cluster
• Talk with Francis on how to document that best
• Guide on how to reproduce stuff
• Host data somewhere (csv for GoogleEarth, and real datasets for fine-tuning)
• Weights / checkpoints of pretraining + fine-tuning + baseline weights