Peter Dudfield

Could you do something like: **Train** - use smaller images and validate the model on these smaller images as well **Test / Predict** - Split the large image into smaller...

Thanks you this helped me


Here's some groupings that might help ``` zones['z1'] = [139,288, 203,57,43, 138, 179, 262,12,6, 18, 94,42,114, 204, 108,154,115,59,239,127,29, 87] zones['z2'] = [125,225,259, 224,195, 21,282, 232,313,120,99,77,248,311,121, 39] zones['z3'] = [284,236,50, 98]...

Further feedback - Zonal information would be great as they almost map with the Constraint Boundaries. - If it was possible to select multiple zones at once, that would be...

This might also help - https://www.nationalgrideso.com/research-and-publications/electricity-ten-year-statement-etys/electricity-transmission-network/scottish-boundaries

I tried something, but its a bit manual ``` ModelFlatPydantic = sqlalchemy_to_pydantic(ModelSQL) ChildPydantic = sqlalchemy_to_pydantic(ChildSQL) class ModelPydantic(ModelFlatPydantic): child: ChildPydantic ``` Does that help?

I had a quick look here - https://docs.mapbox.com/android/maps/api/10.0.0-rc.1/-mapbox%20-maps%20-android/com.mapbox.maps/-mapbox-map/#com.mapbox.maps/MapboxMap/getFeatureState/#kotlin.String#kotlin.String?#kotlin.String#com.mapbox.maps.QueryFeatureStateCallback/PointingToDeclaration/ which is what we use at the moment to select a region, it doesnt look easy to select multiple regions

Persistence model has been added, see https://github.com/openclimatefix/predict_pv_yield/blob/master/predict_pv_yield/models/baseline/last_value.py Have taken the last historic value, but infact could take the 'current' value of PV yield using y_hat = pv_yield[:, -self.forecast_len , 0]...

I would like to add a weighted mae / mse, where first time step is weighted a lot more heavily than the last. The weights could exponentially decay, and the...