1321 Commits

Author	SHA1	Message	Date
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
Wei Shengyu	d6d5efe055	Merge pull request #1616 from JamesLim-sy/change_some_settings_for_clas_train_performance ... Change some settings for clas model`s trainning performance	2022-01-25 11:29:27 +08:00
Tingquan Gao	ce166b6ced	docs: fix invalid links	2022-01-17 21:14:38 +08:00
gaotingquan	6963c25804	fix: fix bug ... bug is that error rasied by calling update_res() when sublayer does not have _save_sub_res_hook() because nn.Sequential and nn.LayerList is not TheseusLayer.	2022-01-14 14:58:28 +08:00
gaotingquan	a6a76f3d36	fix: args error	2022-01-14 14:58:28 +08:00
gaotingquan	43fce425ac	fix: change the returned result from dict to list	2022-01-14 14:58:28 +08:00
gaotingquan	3b9f629220	feat: support specify return_stages to update_res()	2022-01-14 14:58:28 +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
gaotingquan	1a9b1782ef	fix: fix min_lr & warmup_lr	2022-01-07 11:52:30 +08:00
gaotingquan	56c72a85a1	fix: fix lr	2022-01-07 11:52:30 +08:00
gaotingquan	025feb6de9	fix: fix -1 in dims	2022-01-07 11:52:30 +08:00
gaotingquan	7a0ed6f08d	rename	2022-01-07 11:52:30 +08:00
gaotingquan	ee819393ac	fix: fix -1 in dims and ignore swapdim in static	2022-01-07 11:52:30 +08:00
gaotingquan	00fb3f7519	feat: support PVTV2	2022-01-07 11:52:30 +08:00
gaotingquan	a53b636829	feat: support PVTV2	2022-01-07 11:52:30 +08:00
littletomatodonkey	e0a6e5bf38	fox dist err (#1621) ... * fox dist err * fix init * fix init	2022-01-06 09:46:52 +08:00
JamesLim-sy	f015860645	drop useless settings	2022-01-05 13:15:19 +00:00
littletomatodonkey	aea712cc87	add dist of rec model (#1574) ... * add distillation loss func and rec distillation	2022-01-05 19:25:36 +08:00
JamesLim-sy	24a51ae064	first	2022-01-04 07:20:31 +00:00
Bin Lu	304bafb735	Merge pull request #1610 from TingquanGao/dev/fix_theseus ... fix: upgrade replace_sub() to upgrade_sublayer()	2021-12-31 12:15:18 +08:00
gaotingquan	d6aeba6f56	fix: upgrade replace_sub() to upgrade_sublayer() ... supplement to PR #1581	2021-12-30 10:59:20 +00:00
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	cabc7a0f83	fix: replace reg exp with layer name	2021-12-28 15:24:50 +08:00
gaotingquan	71cb728b07	fix: rename to upgrade_sublayer()	2021-12-28 15:24:50 +08:00
gaotingquan	1e696ac22b	fix: remove unnecessary register_hook() call & pre-commit	2021-12-28 15:24:50 +08:00
gaotingquan	a86c4b290e	add license	2021-12-28 15:24:50 +08:00
gaotingquan	721ac0bf61	refactor: simplify code ... 1. remove WrapLayer and wrap_theseus; 2. support call update_res() one more; 3. optim parse_pattern_str() to return list of layer parsed.	2021-12-28 15:24:50 +08:00
gaotingquan	56911b573b	refactor: rename replace_sub() func to ?	2021-12-28 15:24:50 +08:00
gaotingquan	0f126b75da	refactor: strengthen parse_pattern_str() func	2021-12-28 15:24:50 +08:00
gaotingquan	8d0b0d4b0a	refactor: extract _parse_pattern_str() func	2021-12-28 15:24:50 +08:00
gaotingquan	18dec0744a	fix: fix problems commented in reviewing	2021-12-28 15:24:50 +08:00
gaotingquan	41296972a4	fix: fix comments	2021-12-28 15:24:50 +08:00
gaotingquan	f8ee6c0f86	fix: fix result returned by stop_after	2021-12-28 15:24:50 +08:00
gaotingquan	cf205e1379	fix: fix result returned by _find_layers_handle	2021-12-28 15:24:50 +08:00
gaotingquan	b0ae3a1210	fix: unify the pattern of layer's name	2021-12-28 15:24:50 +08:00
gaotingquan	7da2a997e9	fix: save latest model every epoch	2021-12-27 22:04:26 +08:00
weishengyu	377a7b4b2f	dbg	2021-12-27 20:35:42 +08:00
weishengyu	9a1d35ba3b	add license info	2021-12-27 20:16:56 +08:00
weishengyu	87f508e9f9	dbg	2021-12-27 19:37:00 +08:00
gaotingquan	5d53e9f152	fix: raise warning when setting batch_transform_ops and TopkAcc	2021-12-24 21:22:52 +08:00
weishengyu	b662ed34ac	dbg	2021-12-24 11:15:27 +08:00
weishengyu	18d99b012b	move gallery layer into extractor	2021-12-24 11:06:42 +08:00
weishengyu	25edd1c0d8	refactor code	2021-12-23 20:51:15 +08:00