多属性分类模型修改，给模型添加neck，使每个头输出效果最优化。
-
cdto folder where you want to download this repo -
Run `git clone
-
Install dependencies:
- pytorch>=0.4
- torchvision
- yacs
-
Prepare dataset
-
Prepare pretrained model if you don't have
（1）Resnet
from torchvision import models models.resnet50(pretrained=True)
（2）Senet
import torch.utils.model_zoo as model_zoo model_zoo.load_url('the pth you want to download (specific urls are listed in ./modeling/backbones/senet.py)')
Then it will automatically download model in
~/.torch/models/, you should set this path inconfig/defaults.pyfor all training or set in every single training config file inconfigs/or set in every single command.（3）Load your self-trained model
If you want to continue your train process based on your self-trained model, you can change the configuration
PRETRAIN_CHOICEfrom 'imagenet' to 'self' and set thePRETRAIN_PATHto your self-trained model. -
If you want to know the detailed configurations and their meaning, please refer to
config/defaults.py. If you want to set your own parameters, you can follow our method: create a new yml file, then set your own parameters. Add--config_file='config/your yml file'int the commands described below, then our code will merge your configuration. automatically.
用以下格式生成train.txt 和 val.txt即可。对应txt路径写出config文件夹对应配置文件中。
#图片路径 属性1对应索引 属性2对应索引 属性3对应索引（ps：中间都存在一个空格，图片路径中不可以有空格）
#eg:
/media/中东数据/国家分类数据/middle_att3_0419/Lebanon/white/red/Lebanon_white_00016758.jpg 3 4 3
/media/中东数据/国家分类数据/middle_att3_0419/Lebanon/white/red/Lebanon_white_00018923.jpg 3 4 3
python3 tools/train.py --config_file='config/middle_att.yml' MODEL.DEVICE_ID "('your device id'，只有一张显卡可以省去)" OUTPUT_DIR "output/middle_3att"Please replace the data path of the model and set the PRETRAIN_CHOICE as 'self' to avoid time consuming on loading ImageNet pretrained model.
python3 tools/test.py --config_file='configs/your.yml' MODEL.DEVICE_ID "('your device id')" MODEL.PRETRAIN_CHOICE "('self')" TEST.WEIGHT "('your path to trained checkpoints')"