[BUG] Bug report - Predicted forecasts frequency not same as training data.

[vikram-raju]

Describe the bug
A model trained with week-level data is forecasting steps ahead at day granularity.

To Reproduce
I put a reproducible example here - https://stats.stackexchange.com/questions/598669/forecasts-are-at-different-levels-using-htsregressor-hierarchical-time-series

Expected behavior
Forecasted output to be same as week-level.

Desktop (please complete the following information):

• OS: [e.g. Monterey]
• scikit-hts version: [e.g. 0.5.12]
• Python version: [e.g. 3.9.12]