I noticed this necessity while testing our prototypes on prebuilt models. Especially on the models where related dataset is small are affecting our test and cross validation scorings.
I propose adding random seeds would lead to more deterministic results for future tests...
I noticed this necessity while testing our prototypes on prebuilt models. Especially on the models where related dataset is small are affecting our test and cross validation scorings.