0.2.1

What's Changed

• Remove extra LR in default config by @Xiao-Chenguang in #69
• Update the doc strings by @Xiao-Chenguang in #71
• Subprocess exception handling for train and test by @Xiao-Chenguang in #72

Full Changelog: 0.2.0...0.2.1

0.2.0

Milestone version update

Since we achieved both powerful simulation and simple development goals, we have released this milestone version, 0.2.0.

What's Changed

• Config torch for all platforms except macos by @Xiao-Chenguang in #57
• Add wandb into third party package explicitly to avoid wrong sort when import by @Xiao-Chenguang in #58
• Use config instead of args by @Xiao-Chenguang in #59
• Notice of CUDA multiprocessing problem on Windows by @Xiao-Chenguang in #62
• Support StateDict inplace operations by @Xiao-Chenguang in #64
• Add inplace copy to StateDict and remove unused variables. by @Xiao-Chenguang in #65
• Use fix and dynamic args group for the local training process by @Xiao-Chenguang in #66
• Use inplace add to avoid broke the reference to model parameters by @Xiao-Chenguang in #67
• New FL develop schema in easy 3 steps. by @Xiao-Chenguang in #68

Full Changelog: 0.1.12...0.2.0

0.1.12

What's Changed

• Set cudnn to deterministic when using seed in Tiny ImageNet demo by @Xiao-Chenguang in #50
• Fix wong args TEST_PROCESS to TEST_SUBPROCESS by @Xiao-Chenguang in #51
• Add download count badge by @Xiao-Chenguang in #55
• Add acknowledgement to the easydict project. by @Xiao-Chenguang in #56

Full Changelog: 0.1.11...0.1.12

0.1.11

What's Changed

• support cudnn reproductivity by @Xiao-Chenguang in #46
• use test subprocess by @Xiao-Chenguang in #47
• set start method only necessary by @Xiao-Chenguang in #48
• Add Tiny Imagenet demo by @Xiao-Chenguang in #49

Full Changelog: 0.1.10...0.1.11

0.1.10

What's Changed

• rename double underscore to single for inherent and overwrite by @Xiao-Chenguang in #38
• add dampening config to SGD by @Xiao-Chenguang in #41
• allow more args for client training by @Xiao-Chenguang in #42
• Add MFL algorithm and dome script by @Xiao-Chenguang in #43

Full Changelog: 0.1.9...0.1.10

0.1.9

What's Changed

• Fix the wrong badge by @Xiao-Chenguang in #33
• add release link by @Xiao-Chenguang in #34
• default seed to -1 by @Xiao-Chenguang in #35

Full Changelog: 0.1.8...0.1.9

0.1.8

What's Changed

• Readme badges by @Xiao-Chenguang in #31
• update feature list by @Xiao-Chenguang in #32

Full Changelog: 0.1.7...0.1.8

0.1.7

What's Changed

• add cuda examples with mp by @Xiao-Chenguang in #20
• enable test loop in subprocess by @Xiao-Chenguang in #23
• rename internal attributes and funcs with underscore by @Xiao-Chenguang in #25
• Fix wrong cid in mp training by @Xiao-Chenguang in #27
• deep copy the model passed to subprocess by @Xiao-Chenguang in #29
• Random seed by @Xiao-Chenguang in #30

Full Changelog: 0.1.6...0.1.7

0.1.6

What's Changed

• Fix mp stuck under Linux by @Xiao-Chenguang in #14
• work around wrong isoft on wandb by @Xiao-Chenguang in #16
• Support flexiable device by @Xiao-Chenguang in #17
• fix shared cuda tensor issue by @Xiao-Chenguang in #19

Full Changelog: 0.1.5...0.1.6

0.1.5

What's Changed

• add the level name to log by @Xiao-Chenguang in #5
• use static method for client train and embed server test by @Xiao-Chenguang in #6
• use dynamic log info in fedavg by @Xiao-Chenguang in #7
• fix typo: train_client by @Xiao-Chenguang in #8
• Fedprox by @Xiao-Chenguang in #9
• export core apis by @Xiao-Chenguang in #10
• Type dict by @Xiao-Chenguang in #11

Full Changelog: 0.1.4...0.1.5