ufv_load#

t3toolbox.backend.ufv_operations.ufv_load(file, use_jax=False)#
def ufv_load(file, use_jax: bool = False):  # -> (*supercores, shape, masks)

Load a bv .data tuple saved by ufv_save(). The supercores follow use_jax; the masks always come back numpy (host) bool (a jax mask is a tracer under jit). The caller wraps the returned tuple into the OO class (the supercore count is fixed per class).

Parameters:

use_jax (bool)