dmlcloud.seed

dmlcloud.seed(seed=None, group=None)

Share’s the seed from the root rank to all ranks in the group and seeds the random number generators.

The following libraries are seeded: - random - numpy - torch - tensorflow (if installed and imported)

Different ranks will be seeded differently, so that they do not generate the same random numbers.

Parameters:

• seed (int | None) – The seed to share. If None, a random seed is generated. Default is None. Value must be within the inclusive range [-0x8000_0000_0000_0000, 0xffff_ffff_ffff_ffff]. Negative inputs are remapped to positive values with the formula 0xffff_ffff_ffff_ffff + seed.

• group (ProcessGroup) – The process group to work on. If None, the default process group will be used. Default is None.

Return type:

int

Returns:

The original seed that was provided the root rank. 64-bit integer.

Raises:

RuntimeError – If seed is not within the inclusive range [-0x8000_0000_0000_0000, 0xffff_ffff_ffff_ffff].