I downloaded the Bench2Drive-Full dataset, but I noticed a discrepancy in the size. The downloaded dataset is 2.7TB, whereas the expected size is 4TB. However, the number of files (9371) matches the expected count.

Could this size difference be due to compression, and will the dataset expand to 4TB upon extraction? To ensure data integrity, could you please provide an MD5 checksum list for the files? This will allow me to verify that all files have been downloaded correctly.