模型信息记录

本文记录一下在调试模型的过程中输出的信息:

FC-SOD

  1. 下面为run.py文件中下述打印信息的输出:
1
2
3
4
5
6
7
8
9
10
11
12
13
print("train_full_wraper: ", train_full_wraper)
print("semi_wrapper: ", semi_wrapper)
print("gt_full_wraper: ", gt_full_wraper)
print("adv_loss_func: ", adv_loss_func)
print("sal_loss_function: ", sal_loss_function)
print("config.MAX_ITER: ", config.MAX_ITER)
print("ignore_value: ", sal_dataset.ignore_value)
print("generator_optim_create_func: ", G_optim_obj)
print("discriminator_optim_create_func: ", D_optim_obj)
print("device: ", device)
print("model_saver: ", model_saver)
print("pretrained_model_path: ", config.PRETRAINED_MODEL_PATH)
print("is_use_grab: ", config.USE_GRAB)

输出信息为:

1
2
3
4
5
6
7
8
9
10
11
12
13
train_full_wraper:  {'dataloader': <data_provider.sod_dataloader.SOD_Dataloader object at 0x7f0564504f70>, 'lambda_sal': 1, 'lambda_adv': 0.01, 'start_time': None, 'mask_T': None}
semi_wrapper: {'dataloader': <data_provider.sod_dataloader.SOD_Dataloader object at 0x7f05644b3bb0>, 'lambda_sal': 1, 'lambda_adv': 0.007, 'start_time': 5000, 'mask_T': 0.2}
gt_full_wraper: {'dataloader': <data_provider.sod_dataloader.SOD_Dataloader object at 0x7f0564504fd0>, 'lambda_sal': None, 'lambda_adv': None, 'start_time': None, 'mask_T': None}
adv_loss_func: BCELossWithLogits_Mask()
sal_loss_function: BCELossWithLogits_Mask()
config.MAX_ITER: 25000
ignore_value: 255.0
generator_optim_create_func: <helper.optim.Optimizer object at 0x7f0548df3e50>
discriminator_optim_create_func: <helper.optim.Optimizer object at 0x7f0548df3d60>
device: cuda
model_saver: <helper.model_saver.ModelSaver object at 0x7f056444bb80>
pretrained_model_path: ./pretrained/resnet101COCO-41f33a49.pth
is_use_grab: False