I’m noticing some differences between the local_evaluation.py and phase 1 submission. In particular, local_evaluation.py uses a zero-based index for hour (0, 1, … 23) and it seems that the submission run uses a one-based index for hour (1, … 24).
Also, my local_evaluation.py score is not the same as the score that I get on a submission. Shouldn’t they be exactly the same during Phase 1?
@mt1 thank you for your observation. I have fixed both to have one-based index. Please, pull the latest version of the starter-kit to get the updated data files with 1-24 hours.
I believe the reason for the discrepancy will be the index issue you mentioned. Also, make sure you are using the latest required version of CityLearn (see here).
@kingsley_nweye thanks for the fix. I’ll make sure I’m using the latest required version of CityLearn.