Issue #, if available:

Description of changes:

• Make __version__ available via from fev import __version__
• Silence warnings from datasets when calling ds.map(..., num_proc=num_proc) with num_proc > len(ds).
• Enable multivariate dataset support in GluonTSAdapter.
• Speed up fev.utils.validate_time_series_dataset for large datasets (down from 8s to 0.8s for m5_with_covariates dataset).

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.