The motivation is to speed up prediction when running on a list of benchmark datasets.
The predict() function in Forecaster() class used to always predict 100 samples and return mean, quantiles and samples. To speed up, we only predict one sample when the uncertainty argument in Forecaster() is False, when uncertainty is not needed by the user.
The motivation is to speed up prediction when running on a list of benchmark datasets. The predict() function in Forecaster() class used to always predict 100 samples and return mean, quantiles and samples. To speed up, we only predict one sample when the uncertainty argument in Forecaster() is False, when uncertainty is not needed by the user.