keras/keras_core/initializers
Matt Watson 273d3be5cd Fixed SeedGenerator in the torch backend (#434)
We were ignoring the second seed entirely, which is quite incorrect as
that is the only part of the seed we are incrementing.

I'm just summing the fixed seed and increment counter together for now.
Not sure if that's the best approach for truly random seeding, or
what approach torch is even using for there seeds, but figure this is
much better than nothing.
2023-07-10 10:43:08 -07:00
..
__init__.py Bug fixes 2023-05-30 16:29:50 -07:00
constant_initializers.py Namespace nit: operations -> ops 2023-06-28 15:36:45 -07:00
initializer.py Add missing docstrings 2023-05-26 14:11:03 -07:00
random_initializers_test.py Fixed SeedGenerator in the torch backend (#434) 2023-07-10 10:43:08 -07:00
random_initializers.py Namespace nit: operations -> ops 2023-06-28 15:36:45 -07:00