pytorch_ema - Tiny PyTorch library for maintaining a moving average of a collection of parameters.

Age	Commit message (Collapse)	Author
2021-11-17	CI with GitHub Actions	Linux-cpp-lisp

2021-11-17	changelog	Linux-cpp-lisp

2021-11-17	handle non-trainable parameters	Linux-cpp-lisp

2021-07-02	Merge pull request #8 from Linux-cpp-lisp/context_manager	Samuel Fadel
	Context manager & README updates
2021-07-02	Merge pull request #7 from Linux-cpp-lisp/device_and_dtype	Samuel Fadel
	Device and dtype improvements
2021-04-27	README updates	Alby M

2021-04-21	context manager for validation	Alby M

2021-04-21	Indents	Alby M

2021-04-21	Add .to()	Alby M

2021-04-21	More state_dict tests	Alby M

2021-04-21	Casting and error checking	Alby M

2021-04-21	Cast state dict to current device/dtype	Alby M

2021-04-21	State dict support (#6)	Alby M

2021-04-20	Option to keep parameters reference in `ExponentialMovingAverage` (#5)	Alby M

2021-04-20	Tests and miscellaneous (#4)	Alby M
	* linter * Add tests * Add restore test * Type hints * PyTorch dependency * One fewer temp buffer
2021-03-03	Version bump and only store params requiring grad.v0.2	Samuel Fadel

2021-03-03	Updated README.md with store/restore functionality.	Samuel Fadel

2021-03-03	Minor updates to docstrings and code organization.	Samuel Fadel
	* Docstrings using the same imperative verb form * Use list comprehension when collecting params with store()
2021-02-08	Merge pull request #2 from Zehui-Lin/master	Samuel Fadel
	Add feature: store/restore
2021-02-08	minor	Zehui Lin

2021-01-21	fix comments	Zehui Lin

2021-01-21	store and restore	Zehui Lin

2021-01-15	Add Feature: restore	zehui-lin

2020-08-20	Updated README.md with suggested install method.	Samuel Fadel

2019-05-26	Filter by requires_grad before zip() in update().	Samuel Fadel

2019-04-16	Initial commit.	Samuel Fadel