Track validation data when running script

[nlebovits]

When testing + validating a model, write the outputs to a .csv file stored in the cloud. Can use as a benchmark for evaluation. Should include relevant metrics (accuracy, recall, false positive rate, AUC, kappa, etc.) plus also sample size, time of evaluation, and location from which the data come (e.g., costa rica)

[nlebovits]

Can also save trained model as an asset in GCloud