392b75b1ac
revert for running
2023-03-10 16:56:55 +08:00
4008588343
add amp decorator and parse_batch decorator
2023-03-10 16:56:55 +08:00
d3941dc1e9
rm amp code from train and eval & use decorator for amp training
2023-03-10 16:56:55 +08:00
97935164fe
use decorator to parse batch
2023-03-10 16:56:55 +08:00
e0847f1800
Update pretrained backbone
2023-02-28 15:01:21 +08:00
7c3bb2754b
Update files according to reviews
...
https://github.com/PaddlePaddle/PaddleClas/pull/2633
2023-02-28 15:01:21 +08:00
7a31d9a97f
fix: optimizer setting
2023-02-28 15:01:21 +08:00
21e76d08b6
Modify codes based on reviews
2023-02-28 15:01:21 +08:00
c3fa6eca86
Create backbone for MetaBIN
2023-02-28 15:01:21 +08:00
adb9930317
fix name
2023-02-28 14:28:23 +08:00
5604a13fac
fix name
2023-02-28 14:28:23 +08:00
f24e1f9bcf
cae config fix
2023-02-23 16:28:50 +08:00
811b483e30
fix: set dtype in paddle.to_tensor()
2023-02-10 15:51:53 +08:00
23f5af9f2a
add field vit to foundationvit's name
2023-02-07 17:38:50 +08:00
e8bf35d1b4
fix foundationvit load fail
2023-02-07 10:46:25 +08:00
83fc10cd0f
modify uniformer link
2023-02-06 19:28:30 +08:00
a170400d20
modify init
2023-02-01 10:43:11 +08:00
79cbd7350a
Aesthetic
2023-02-01 10:43:11 +08:00
4fdcda7c60
fix bug in randaug, train_progressive and efficientnet_v2
2023-01-31 21:15:22 +08:00
7e12c73e0a
polish progressive training code
2023-01-31 21:15:22 +08:00
8b8e0431ed
fix AttrDict import path
2023-01-31 21:15:22 +08:00
3a8b56801d
feat(model): add EfficientNetV2 code and fix AttrDict BUG
2023-01-31 21:15:22 +08:00
b889101fd1
add docs and config
2023-01-31 19:57:43 +08:00
373a026a83
modify efficientnet_variant
2023-01-31 19:57:43 +08:00
54f7dd7484
watermark
2023-01-31 19:57:43 +08:00
add3ea238c
modify tipc
2023-01-20 14:08:06 +08:00
9ba23527b4
add swinv2
2023-01-20 14:08:06 +08:00
4786b2c96b
add swinv2
2023-01-20 14:08:06 +08:00
74de24f621
add swinv2
2023-01-20 14:08:06 +08:00
a5a3b0f567
Delete unnecessary files
2023-01-20 14:08:06 +08:00
272e75365b
modify swinv2
2023-01-20 14:08:06 +08:00
65e7a41a1d
add docs
2023-01-20 14:08:06 +08:00
7e616a16b5
modify init
2023-01-20 14:08:06 +08:00
f5e8df0e63
Some modifications
2023-01-20 14:08:06 +08:00
da95893bf2
supplement init
2023-01-20 14:08:06 +08:00
6994050983
add SwinTransformerV2
2023-01-20 14:08:06 +08:00
0a56120855
add swinV1 22k weights
2023-01-20 14:07:04 +08:00
3672d1f227
add swinV1 22k weights
2023-01-20 14:07:04 +08:00
5544dbaf8a
add ViT model for foundation models forward
2023-01-19 17:42:45 +08:00
216291618c
add the pretrained url
2023-01-06 20:29:07 +08:00
f6df698c4f
micro_block -> layer_type
2023-01-06 20:29:07 +08:00
81de331e00
rename to re_parameterize() for re-parameterization nets
2023-01-06 20:29:07 +08:00
d4d3d01384
support re_parameterize
2023-01-06 20:29:07 +08:00
f82871b1f8
add copyright & reference
2023-01-06 20:29:07 +08:00
963dcb9cac
support dbb module for ResNet
...
1. add DiverseBranchBlock module;
2. ResNet support dbb version using DiverseBranchBlock by setting micro_block="DiverseBranchBlock";
3. ResNet support official vb version by setting use_first_short_conv=False;
4. add ResNet18_dbb training config ResNet18_dbb.yaml.
2023-01-06 20:29:07 +08:00
d8f049ae76
增加代码规范，删除无用空格
2023-01-04 13:42:00 +08:00
c5a29dbaec
代码规范
2023-01-04 13:42:00 +08:00
f073e97d22
增加cifar100参数yaml
2023-01-04 13:42:00 +08:00
009f347d64
修改空格等代码规范
2023-01-04 13:42:00 +08:00
4db132444d
增加RecModel配合WideResNet代码以及参数转换成RecModel
2023-01-04 13:42:00 +08:00
gaotingquan
tianyi1997
zengshao0622
dolcexu
zh-hike
zhangyubo0722
HydrogenSulfate
zengshao0622