David Novotny c3f8dad55c Move load_stats to TrainingLoop ...

Summary:
Stats are logically connected to the training loop, not to the model. Hence, moving to the training loop.

Also removing resume_epoch from OptimizerFactory in favor of a single place - ModelFactory. This removes the need for config consistency checks etc.

Reviewed By: kjchalup

Differential Revision: D38313475

fbshipit-source-id: a1d188a63e28459df381ff98ad8acdcdb14887b7

2022-08-02 15:40:53 -07:00

..

__init__.py

Updates to support Accelerate and multigpu training (#37)

2022-07-11 19:29:58 -07:00

model_factory.py

Move load_stats to TrainingLoop

2022-08-02 15:40:53 -07:00

optimizer_factory.py

Move load_stats to TrainingLoop

2022-08-02 15:40:53 -07:00

training_loop.py

Move load_stats to TrainingLoop

2022-08-02 15:40:53 -07:00

utils.py

Better seeding of random engines

2022-08-01 10:03:09 -07:00