194 Commits

Author	SHA1	Message	Date
HydrogenSulfate	f2982e5e47	update code	2022-04-22 12:01:05 +08:00
HydrogenSulfate	17fd1bc2c0	refine code	2022-04-22 12:00:03 +08:00
HydrogenSulfate	aa26a8c1d8	fix bug for static graph	2022-04-21 16:31:28 +08:00
HydrogenSulfate	daf7eea23d	fix code	2022-04-21 15:43:53 +08:00
HydrogenSulfate	dfd7749828	refine hard code	2022-04-21 14:37:00 +08:00
HydrogenSulfate	41e1a86caf	add center loss	2022-04-21 00:17:54 +08:00
Wei Shengyu	e6feb68bb8	Merge pull request #1824 from TingquanGao/dev/spt_amp_eval ... fix: fp32 eval by default when enable amp	2022-04-20 14:40:46 +08:00
HydrogenSulfate	3a1276d315	train_loss_func only used in train mode	2022-04-19 19:54:48 +08:00
HydrogenSulfate	24abea151a	support for multi optimizer case	2022-04-19 14:26:42 +08:00
gaotingquan	83ed5195c3	fix: set use_fp16_test to True when AMP O2 is enabled	2022-04-18 06:14:43 +00:00
weishengyu	1789da6422	fix bug	2022-04-18 11:26:32 +08:00
gaotingquan	a35cdd2aec	uncommit: sync bn is too slow to use and convert_sync_batchnorm() is not effective for BatchNorm	2022-04-14 08:19:39 +00:00
gaotingquan	13d5e59051	fix: convert bn to sync_bn ... the running_mean and running_var of bn would not be synchronized in dist, so which leads to bug that eval loss in training is inconsistent with eval only.	2022-04-14 07:36:39 +00:00
gaotingquan	efde56ffc6	fix: only fp16 evaluation is supported when ampO2 is enabled	2022-04-13 12:14:14 +00:00
gaotingquan	474c918b27	fix: fix bug of batch_size statistics error	2022-04-13 09:19:30 +00:00
gaotingquan	c46189bad0	fix: fix bug about calc loss in dist	2022-04-12 06:56:44 +00:00
HydrogenSulfate	af90cd7c59	update center loss config and related code	2022-04-12 13:07:53 +08:00
weishengyu	9de22673df	dbg	2022-04-08 14:29:03 +08:00
gaotingquan	b761325faa	fix: fp32 eval by default when enable amp ... If you want to eval by fp16 when enable amp, please set Amp.use_fp16_test=True, False by default.	2022-04-02 19:22:10 +08:00
dongshuilong	a944603da0	fix log twice bug	2022-03-30 08:31:35 +00:00
huangqipeng	b62b98d79f	feat: support mlu device and amp of mlu	2022-03-14 15:48:26 +08:00
littletomatodonkey	f68c098a4a	fix train acc log	2022-03-09 19:58:36 +08:00
WangChen0902	7595ba6d70	add AFD (#1683) ... * add AFD	2022-02-28 19:11:50 +08:00
dongshuilong	dc6281a6d4	add benchmark for tipc	2022-02-10 08:25:52 +00:00
Tingquan Gao	42134cd8dd	fix: raise warning when using Global.class_num	2022-01-25 15:06:36 +08:00
Tingquan Gao	bb6581d21b	refactor: raise warning when gpu numbers is not 4	2022-01-25 15:06:36 +08:00
Tingquan Gao	8f0bd5b582	fix: fix vdl makedir	2022-01-25 15:06:36 +08:00
gaotingquan	10c93c55d1	fix: enable amp only in training	2022-01-25 11:58:07 +08:00
gaotingquan	7040ce8314	refactor: change params to be consistent with amp	2022-01-25 11:58:07 +08:00
zhangbo9674	cd039a7b37	add save_dtype	2022-01-10 18:19:03 +08:00
zhangbo9674	d437bb0a7e	use fp32 to eval	2022-01-10 18:19:03 +08:00
zhangbo9674	bb19c1f7a6	fix eval bug	2022-01-10 18:19:03 +08:00
zhangbo9674	b2956c1b41	refine code	2022-01-10 18:19:03 +08:00
zhangbo9674	205592a3e3	fix amp with distribute bug	2022-01-10 18:19:03 +08:00
littletomatodonkey	aea712cc87	add dist of rec model (#1574) ... * add distillation loss func and rec distillation	2022-01-05 19:25:36 +08:00
gaotingquan	6e13ff3068	fix: use hasattr() to check if collate_fn is in dataloader ... fix bug caused by PR #1596	2021-12-30 16:35:05 +08:00
gaotingquan	7da2a997e9	fix: save latest model every epoch	2021-12-27 22:04:26 +08:00
gaotingquan	5d53e9f152	fix: raise warning when setting batch_transform_ops and TopkAcc	2021-12-24 21:22:52 +08:00
zhangbo9674	28061f537c	refine optimizer init logice	2021-12-21 06:28:13 +00:00
zhangbo9674	b54ee04491	Accelerate dynamic graph amp training	2021-12-20 06:36:56 +00:00
gaotingquan	7732a69f1b	fix: fix key error in distillation	2021-12-16 18:21:08 +08:00
weishengyu	534037d145	dbg	2021-12-10 11:14:14 +08:00
weishengyu	ad1a2fd137	move slim into arch	2021-12-09 20:08:57 +08:00
weishengyu	7c6567cc6b	dbg	2021-12-09 18:08:16 +08:00
weishengyu	6c5d1ebc28	add pruner and quanter for theseus	2021-12-09 14:51:40 +08:00
cuicheng01	a3b54f15d5	fix export quant_model	2021-12-03 07:39:31 +00:00
dongshuilong	f7ccc874e2	fix dali distributed eval bug	2021-11-16 11:09:21 +08:00
stephon	7a17f72fc2	fix seed=0 bug	2021-11-01 06:16:24 +00:00
Walter	a5d0e37b02	Merge pull request #1341 from RainFrost1/googlenet_bug ... fig goooglenet distributed eval bug	2021-10-29 10:28:37 +08:00
gaotingquan	ed459a2a16	refactor: adapt to static graph in deprecating MixCELoss	2021-10-27 19:47:43 +08:00