Duplicate run

Config copied from existing run — adjust as needed and launch.

Run name

Optional — a descriptive label for this run

Folder

Optional — group this run with others in a folder

BigQuery

Table ID *

project.dataset.table

Client project

Leave blank to use default from credentials

Start date

YYYY-MM-DD — optional, defaults to earliest row

End date

YYYY-MM-DD — optional, defaults to latest row

Row limit

Optional — leave blank for full table

WHERE clause

Optional SQL filter

Chains

Adapt steps

Burn-in

Keep (samples)

Parallel iterations

Prior samples

Run budget optimization after training

Optimization start date

Start of the spend window used to compute the optimization budget

Optimization end date

End of the spend window. Leave blank to fall back to lookback weeks.

Google ColabLocal (this machine)Cloud Run + GCS

Dispatches to Cloud Run job gg-mmm-app. Outputs land in GCS.