I'm curious why it's common for data to be repartitioned to 1 partition
when saving ml models:
This shows up in most ml models I've seen (Word2Vec
Am I missing some benefit of repartitioning like this?
Senior Software Engineer