How does `datasets.Dataset.map` parallelize data?

[ahans1]

Was this issue solved? I’m facing the same problem even after using multiprocess.set_start_method("spawn", force=True)