Does the Model retrain on same dataset for each task?

[arc-arnob]

In config, we pass the dataset as well as the horizon window and I was wondering if the model considers different horizons from the same domain as a different task.

for eg: for tasks in yaml: LTF_ECL_p96 and LTF_ECL_p336, are these two different tasks?

This could mean the model retrains on the same data for more than the expected amount of time.