10 Commits (c9a61b7d988ec5907c1a010a0b1bb4ab910f5264)

Author	SHA1	Message	Date
Ross Wightman	db8ad25a23	MobileNetV3 appears correct based on paper update, cleaned up comments and compacted last block def	2019-05-15 08:53:27 -07:00
Ross Wightman	20d66beead	Move RMSpropTF another step closer to Tensorflow impl ... * init square_avg with one instead of zero as per TF * match TF order of ops for square_avg accumulation * move LR scaling to momentum buffer accumulator as per TF * add decoupled weight decay flag (not in TF)	2019-05-14 18:35:43 -07:00
Ross Wightman	89147a91e6	Remove adabound optimizer, never got it working well on larger datasets	2019-05-14 18:35:43 -07:00
Ross Wightman	8fbd62a169	Exclude batchnorm and bias params from weight_decay by default	2019-04-22 17:33:22 -07:00
Ross Wightman	bc264269c9	Morph mnasnet impl into a generic mobilenet that covers Mnasnet, MobileNetV1/V2, ChamNet, FBNet, and related ... * add an alternate RMSprop opt that applies eps like TF * add bn params for passing through alternates and changing defaults to TF style	2019-04-21 15:54:28 -07:00
Ross Wightman	db1fe34d0c	Update a few comment, add some references	2019-04-12 23:16:49 -07:00
Ross Wightman	9c3859fb9c	Uniform pretrained model handling. ... * All models have 'default_cfgs' dict * load/resume/pretrained helpers factored out * pretrained load operates on state_dict based on default_cfg * test all models in validate * schedule, optim factor factored out * test time pool wrapper applied based on default_cfg	2019-04-11 21:32:16 -07:00
Ross Wightman	2295cf56c2	Add some Nvidia performance enhancements (prefetch loader, fast collate), and refactor some of training and model fact/transforms	2019-03-10 14:23:16 -07:00
Ross Wightman	9d927a389a	Add adabound, random erasing	2019-03-01 22:03:42 -08:00
Ross Wightman	5855b07ae0	Initial commit, puting some ol pieces together	2019-02-01 22:07:34 -08:00