Making a First Submission to the FLAT DATA challenge of The MOSTLY AI Prize 🏆

the-prize-flat.py

# install Synthetic Data SDK
# see also https://github.com/mostly-ai/mostlyai
#!uv pip install "mostlyai[local]"

# load training data
import pandas as pd
trn = pd.read_csv('/Users/mplatzer/github/the-prize-data/flat/flat-training.csv')

# instantiate SDK in LOCAL mode
from mostlyai.sdk import MostlyAI
mostly = MostlyAI(local=True)

# train a generator
g = mostly.train(config={
    'tables': [{
        'name': 'flat',
        'data': trn,  # your training data
        'tabular_model_configuration': {
          'max_training_time': 2,  # e.g. limit training to 2 minutes
        }
    }]
})

# create a synthetic dataset
sd = mostly.generate(g)
syn = sd.data()
syn.to_csv('your-flat-submission01.csv.gz', index=False)