Ross Wightman 5940cc167f Change start/end args

2025-01-30 15:44:02 -08:00

..

__init__.py

…

_optim_factory.py

Some more kron work. Figured out why some tests fail, implemented a deterministic rng state load but too slow so skipping some tests for now.

2025-01-27 21:02:26 -08:00

_param_groups.py

…

_types.py

…

adabelief.py

Fix typos

2025-01-19 13:39:40 -08:00

adafactor_bv.py

…

adafactor.py

Update README, missed small discrep in adafactor min dim update

2024-11-29 10:57:47 -08:00

adahessian.py

Fix typos

2025-01-19 13:39:40 -08:00

adamp.py

…

adamw.py

…

adan.py

Add caution to Adan. Add decouple decay option to LAMB.

2024-12-05 13:50:30 -08:00

adopt.py

…

kron.py

Change start/end args

2025-01-30 15:44:02 -08:00

lamb.py

Add caution to Adan. Add decouple decay option to LAMB.

2024-12-05 13:50:30 -08:00

laprop.py

Make LaProp weight decay match typical PyTorch 'decoupled' behaviour where it's scaled by LR

2024-11-29 16:44:43 -08:00

lars.py

…

lion.py

To be technically correct, need to check the in-place _ ver of op

2024-11-28 15:11:58 -08:00

lookahead.py

…

madgrad.py

…

mars.py

Add caution to Adan. Add decouple decay option to LAMB.

2024-12-05 13:50:30 -08:00

nadam.py

…

nadamw.py

To be technically correct, need to check the in-place _ ver of op

2024-11-28 15:11:58 -08:00

nvnovograd.py

…

optim_factory.py

…

radam.py

…

rmsprop_tf.py

…

sgdp.py

…

sgdw.py

…