Decide upon an approach to adding default data preprocessing functions for each dataset in AstroPile (perhaps formulate some guidelines), and implement data preprocessing for datasets already in AstroPile.
Contacts: Tom Hehir (@tom-hehir)
Participants:
Goals and deliverables
Decide upon an approach to including default preprocessing functions in AstroPile.
Add preprocessing for datasets already in AstroPile.
Write brief guidelines and some examples for dataset contributors to refer to.
Resources needed
Participants of this hack
Familiarity with dataset preprocessing in any ML library e.g. PyTorch, Keras, SKLearn [not essential].
Familiarity with applying preprocessing functions to Hugging Face datasets [not essential].
Knowledge of appropriate preprocessing for a particular dataset in AstroPile [not essential].
General
Contributions of existing preprocessing functions from previous projects of hackathon attendees.
Opportunities to seek input from hackathon attendees with expertise on a particular dataset.
Opportunity to present guidelines to hackathon attendees early in the week so that they can add preprocessing to any new datasets they contribute.
Add data preprocessing to AstroPile
Decide upon an approach to adding default data preprocessing functions for each dataset in AstroPile (perhaps formulate some guidelines), and implement data preprocessing for datasets already in AstroPile.
Contacts: Tom Hehir (@tom-hehir) Participants:
Goals and deliverables
Resources needed
Participants of this hack
General
Detailed description
[WIP -- more to follow]