Template Repository for OCF Projects
To create new batches, modify the configs/datapipe/forecastor.yaml
file to point to the appropriate files
Also update configs/batch.yaml
and irradiance/batch_writing.py
to write to the correct places and with the correct options.
Then run python run.py --config-name batch
to create the batches.
To change what data is being created with the datapipe, the changes have to happen to the ocf_datapipes.training.pseudo_irradiance_datapipe
, which is what handles
the data processing.
To run inference, modify configs/inference_irradiance.yaml
to have the model configuration you want,
and irradiance/model_inference.py
to run the current-best model on the data that you want.
Then run python run.py --config-name inference_irradiance
to run inference. By default it uses the 4th GPU on Donatello.