├── LICENSE ├── README.md ├── SSL_Flexmatch ├── README.md ├── __pycache__ │ ├── custom_writer.cpython-38.pyc │ ├── train_utils.cpython-38.pyc │ └── utils.cpython-38.pyc ├── config │ ├── fixmatch │ │ ├── fixmatch_cifar100_10000_0.yaml │ │ ├── fixmatch_cifar100_2500_0.yaml │ │ ├── fixmatch_cifar100_400_0.yaml │ │ ├── fixmatch_cifar10_250_0.yaml │ │ ├── fixmatch_cifar10_4000_0.yaml │ │ ├── fixmatch_cifar10_40_0.yaml │ │ ├── fixmatch_imagenet_100000_0.yaml │ │ ├── fixmatch_stl10_1000_0.yaml │ │ ├── fixmatch_stl10_250_0.yaml │ │ ├── fixmatch_stl10_40_0.yaml │ │ ├── fixmatch_svhn_1000_0.yaml │ │ ├── fixmatch_svhn_250_0.yaml │ │ └── fixmatch_svhn_40_0.yaml │ ├── flexmatch │ │ ├── flexmatch_cifar100_10000_0.yaml │ │ ├── flexmatch_cifar100_2500_0.yaml │ │ ├── flexmatch_cifar100_400_0.yaml │ │ ├── flexmatch_cifar100_400_2.yaml │ │ ├── flexmatch_cifar100_400_3.yaml │ │ ├── flexmatch_cifar10_250_0.yaml │ │ ├── flexmatch_cifar10_4000_0.yaml │ │ ├── flexmatch_cifar10_40_0.yaml │ │ ├── flexmatch_cifar10_40_1.yaml │ │ ├── flexmatch_cifar10_40_2.yaml │ │ ├── flexmatch_imagenet_100000_0.yaml │ │ ├── flexmatch_stl10_1000_0.yaml │ │ ├── flexmatch_stl10_250_0.yaml │ │ ├── flexmatch_stl10_40_0.yaml │ │ ├── flexmatch_svhn_1000_0.yaml │ │ ├── flexmatch_svhn_250_0.yaml │ │ └── flexmatch_svhn_40_0.yaml │ ├── fullysupervised │ │ ├── fullysupervised_cifar100_10000_0.yaml │ │ ├── fullysupervised_cifar100_2500_0.yaml │ │ ├── fullysupervised_cifar100_400_0.yaml │ │ ├── fullysupervised_cifar10_250_0.yaml │ │ ├── fullysupervised_cifar10_4000_0.yaml │ │ ├── fullysupervised_cifar10_40_0.yaml │ │ ├── fullysupervised_stl10_1000_0.yaml │ │ ├── fullysupervised_stl10_250_0.yaml │ │ ├── fullysupervised_stl10_40_0.yaml │ │ ├── fullysupervised_svhn_1000_0.yaml │ │ ├── fullysupervised_svhn_250_0.yaml │ │ └── fullysupervised_svhn_40_0.yaml │ ├── meanteacher │ │ ├── meanteacher_cifar100_10000_0.yaml │ │ ├── meanteacher_cifar100_2500_0.yaml │ │ ├── meanteacher_cifar100_400_0.yaml │ │ ├── meanteacher_cifar10_250_0.yaml │ │ ├── meanteacher_cifar10_4000_0.yaml │ │ ├── meanteacher_cifar10_40_0.yaml │ │ ├── meanteacher_stl10_1000_0.yaml │ │ ├── meanteacher_stl10_250_0.yaml │ │ ├── meanteacher_stl10_40_0.yaml │ │ ├── meanteacher_svhn_1000_0.yaml │ │ ├── meanteacher_svhn_250_0.yaml │ │ └── meanteacher_svhn_40_0.yaml │ ├── mixmatch │ │ ├── mixmatch_cifar100_10000_0.yaml │ │ ├── mixmatch_cifar100_2500_0.yaml │ │ ├── mixmatch_cifar100_400_0.yaml │ │ ├── mixmatch_cifar10_250_0.yaml │ │ ├── mixmatch_cifar10_4000_0.yaml │ │ ├── mixmatch_cifar10_40_0.yaml │ │ ├── mixmatch_stl10_1000_0.yaml │ │ ├── mixmatch_stl10_250_0.yaml │ │ ├── mixmatch_stl10_40_0.yaml │ │ ├── mixmatch_svhn_1000_0.yaml │ │ ├── mixmatch_svhn_250_0.yaml │ │ └── mixmatch_svhn_40_0.yaml │ ├── pimodel │ │ ├── pimodel_cifar100_10000_0.yaml │ │ ├── pimodel_cifar100_2500_0.yaml │ │ ├── pimodel_cifar100_400_0.yaml │ │ ├── pimodel_cifar10_250_0.yaml │ │ ├── pimodel_cifar10_4000_0.yaml │ │ ├── pimodel_cifar10_40_0.yaml │ │ ├── pimodel_stl10_1000_0.yaml │ │ ├── pimodel_stl10_250_0.yaml │ │ ├── pimodel_stl10_40_0.yaml │ │ ├── pimodel_svhn_1000_0.yaml │ │ ├── pimodel_svhn_250_0.yaml │ │ └── pimodel_svhn_40_0.yaml │ ├── pseudolabel │ │ ├── pseudolabel_cifar100_10000_0.yaml │ │ ├── pseudolabel_cifar100_2500_0.yaml │ │ ├── pseudolabel_cifar100_400_0.yaml │ │ ├── pseudolabel_cifar10_250_0.yaml │ │ ├── pseudolabel_cifar10_4000_0.yaml │ │ ├── pseudolabel_cifar10_40_0.yaml │ │ ├── pseudolabel_stl10_1000_0.yaml │ │ ├── pseudolabel_stl10_250_0.yaml │ │ ├── pseudolabel_stl10_40_0.yaml │ │ ├── pseudolabel_svhn_1000_0.yaml │ │ ├── pseudolabel_svhn_250_0.yaml │ │ └── pseudolabel_svhn_40_0.yaml │ ├── pseudolabel_flex │ │ ├── pseudolabel_flex_cifar100_10000_0.yaml │ │ ├── pseudolabel_flex_cifar100_2500_0.yaml │ │ ├── pseudolabel_flex_cifar100_400_0.yaml │ │ ├── pseudolabel_flex_cifar10_250_0.yaml │ │ ├── pseudolabel_flex_cifar10_4000_0.yaml │ │ ├── pseudolabel_flex_cifar10_40_0.yaml │ │ ├── pseudolabel_flex_stl10_1000_0.yaml │ │ ├── pseudolabel_flex_stl10_250_0.yaml │ │ ├── pseudolabel_flex_stl10_40_0.yaml │ │ ├── pseudolabel_flex_svhn_1000_0.yaml │ │ ├── pseudolabel_flex_svhn_250_0.yaml │ │ └── pseudolabel_flex_svhn_40_0.yaml │ ├── remixmatch │ │ ├── remixmatch_cifar100_10000_0.yaml │ │ ├── remixmatch_cifar100_2500_0.yaml │ │ ├── remixmatch_cifar100_400_0.yaml │ │ ├── remixmatch_cifar10_250_0.yaml │ │ ├── remixmatch_cifar10_4000_0.yaml │ │ ├── remixmatch_cifar10_40_0.yaml │ │ ├── remixmatch_stl10_1000_0.yaml │ │ ├── remixmatch_stl10_250_0.yaml │ │ ├── remixmatch_stl10_40_0.yaml │ │ ├── remixmatch_svhn_1000_0.yaml │ │ ├── remixmatch_svhn_250_0.yaml │ │ └── remixmatch_svhn_40_0.yaml │ ├── uda │ │ ├── uda_cifar100_10000_0.yaml │ │ ├── uda_cifar100_2500_0.yaml │ │ ├── uda_cifar100_400_0.yaml │ │ ├── uda_cifar10_250_0.yaml │ │ ├── uda_cifar10_4000_0.yaml │ │ ├── uda_cifar10_40_0.yaml │ │ ├── uda_stl10_1000_0.yaml │ │ ├── uda_stl10_250_0.yaml │ │ ├── uda_stl10_40_0.yaml │ │ ├── uda_svhn_1000_0.yaml │ │ ├── uda_svhn_250_0.yaml │ │ └── uda_svhn_40_0.yaml │ ├── uda_flex │ │ ├── uda_flex_cifar100_10000_0.yaml │ │ ├── uda_flex_cifar100_2500_0.yaml │ │ ├── uda_flex_cifar100_400_0.yaml │ │ ├── uda_flex_cifar10_250_0.yaml │ │ ├── uda_flex_cifar10_4000_0.yaml │ │ ├── uda_flex_cifar10_40_0.yaml │ │ ├── uda_flex_stl10_1000_0.yaml │ │ ├── uda_flex_stl10_250_0.yaml │ │ ├── uda_flex_stl10_40_0.yaml │ │ ├── uda_flex_svhn_1000_0.yaml │ │ ├── uda_flex_svhn_250_0.yaml │ │ └── uda_flex_svhn_40_0.yaml │ └── vat │ │ ├── vat_cifar100_10000_0.yaml │ │ ├── vat_cifar100_2500_0.yaml │ │ ├── vat_cifar100_400_0.yaml │ │ ├── vat_cifar10_250_0.yaml │ │ ├── vat_cifar10_4000_0.yaml │ │ ├── vat_cifar10_40_0.yaml │ │ ├── vat_stl10_1000_0.yaml │ │ ├── vat_stl10_250_0.yaml │ │ ├── vat_stl10_40_0.yaml │ │ ├── vat_svhn_1000_0.yaml │ │ ├── vat_svhn_250_0.yaml │ │ └── vat_svhn_40_0.yaml ├── custom_writer.py ├── datasets │ ├── DistributedProxySampler.py │ ├── __init__.py │ ├── __pycache__ │ │ ├── DistributedProxySampler.cpython-38.pyc │ │ ├── __init__.cpython-38.pyc │ │ ├── data_utils.cpython-38.pyc │ │ ├── dataset.cpython-38.pyc │ │ └── ssl_dataset.cpython-38.pyc │ ├── augmentation │ │ ├── __pycache__ │ │ │ └── randaugment.cpython-38.pyc │ │ └── randaugment.py │ ├── data_utils.py │ ├── dataset.py │ └── ssl_dataset.py ├── environment.yml ├── environment3090.yml ├── eval.py ├── figures │ ├── cf10.png │ ├── cf100.png │ ├── logo.png │ ├── stl.png │ └── svhn.png ├── fixmatch.py ├── flexmatch.py ├── flexmatch_package.jpg ├── fullysupervised.py ├── logs │ ├── 0 │ │ └── log.txt │ ├── 1 │ │ └── log.txt │ └── 2 │ │ └── log.txt ├── meanteacher.py ├── mixmatch.py ├── models │ ├── fixmatch │ │ ├── __init__.py │ │ ├── fixmatch.py │ │ └── fixmatch_utils.py │ ├── flexmatch │ │ ├── __init__.py │ │ ├── __pycache__ │ │ │ ├── __init__.cpython-38.pyc │ │ │ ├── flexmatch.cpython-38.pyc │ │ │ └── flexmatch_utils.cpython-38.pyc │ │ ├── flexmatch.py │ │ ├── flexmatch1.py │ │ └── flexmatch_utils.py │ ├── fullysupervised │ │ ├── __init__.py │ │ ├── fullysupervised.py │ │ └── fullysupervised_utils.py │ ├── meanteacher │ │ ├── __init__.py │ │ ├── meanteacher.py │ │ └── meanteacher_utils.py │ ├── mixmatch │ │ ├── __init__.py │ │ ├── mixmatch.py │ │ └── mixmatch_utils.py │ ├── nets │ │ ├── __init__.py │ │ ├── __pycache__ │ │ │ ├── __init__.cpython-38.pyc │ │ │ └── wrn.cpython-38.pyc │ │ ├── resnet50.py │ │ ├── wrn.py │ │ └── wrn_var.py │ ├── pimodel │ │ ├── __init__.py │ │ ├── pimodel.py │ │ └── pimodel_utils.py │ ├── pseudolabel │ │ ├── __init__.py │ │ ├── pseudolabel.py │ │ └── pseudolabel_utils.py │ ├── remixmatch │ │ ├── __init__.py │ │ ├── remixmatch.py │ │ └── remixmatch_utils.py │ ├── uda │ │ ├── __init__.py │ │ ├── uda.py │ │ └── uda_utils.py │ └── vat │ │ ├── __init__.py │ │ ├── vat.py │ │ └── vat_utils.py ├── my_environment.txt ├── pimodel.py ├── pseudolabel.py ├── remixmatch.py ├── scripts │ ├── average_log.py │ ├── cifar.sh │ └── config_generator.py ├── train.sh ├── train_utils.py ├── uda.py ├── utils.py └── vat.py └── UDA_GVB ├── .idea ├── .gitignore ├── clsda.iml ├── deployment.xml ├── inspectionProfiles │ ├── Project_Default.xml │ └── profiles_settings.xml ├── misc.xml ├── modules.xml ├── remote-mappings.xml └── webServers.xml ├── ReadMe.md ├── clsda ├── __init__.py ├── loader │ ├── __init__.py │ └── cls_loaders │ │ ├── __init__.py │ │ ├── builder.py │ │ ├── cls_loaders.py │ │ └── pipelines │ │ ├── __init__.py │ │ └── pipelines.py ├── loss │ ├── __init__.py │ └── loss.py ├── models │ ├── __init__.py │ └── cls_models │ │ ├── __init__.py │ │ ├── basenet.py │ │ ├── builder.py │ │ ├── gvb_network.py │ │ └── resnet.py ├── optimizers │ └── __init__.py ├── runner │ ├── __init__.py │ ├── hooks │ │ ├── __init__.py │ │ ├── cls_accuracy.py │ │ ├── cls_analysis.py │ │ ├── cls_best_accuracy_by_val.py │ │ ├── cls_classwise_accuracy.py │ │ └── training_hooks.py │ ├── trainer.py │ └── validator.py ├── schedulers │ ├── __init__.py │ ├── builder.py │ └── schedulers.py ├── trainers │ ├── __init__.py │ ├── builder.py │ └── trainer_gvb.py └── utils │ ├── __init__.py │ ├── gradcam.py │ ├── labels2wv.py │ ├── logger.py │ ├── metrics.py │ ├── spkmeans.py │ ├── utils.py │ └── writer.py ├── configs ├── _base_ │ ├── cls_datasets │ │ ├── uda_officehome │ │ │ ├── uda_officehome_A_C.py │ │ │ ├── uda_officehome_A_C_4gpu.py │ │ │ ├── uda_officehome_A_C_all_weak.py │ │ │ ├── uda_officehome_A_C_weak.py │ │ │ ├── uda_officehome_A_P.py │ │ │ ├── uda_officehome_A_P_all_weak.py │ │ │ ├── uda_officehome_A_P_weak.py │ │ │ ├── uda_officehome_A_R.py │ │ │ ├── uda_officehome_A_R_all_weak.py │ │ │ ├── uda_officehome_A_R_weak.py │ │ │ ├── uda_officehome_C_A.py │ │ │ ├── uda_officehome_C_A_all_weak.py │ │ │ ├── uda_officehome_C_A_weak.py │ │ │ ├── uda_officehome_C_P.py │ │ │ ├── uda_officehome_C_P_all_weak.py │ │ │ ├── uda_officehome_C_P_weak.py │ │ │ ├── uda_officehome_C_R.py │ │ │ ├── uda_officehome_C_R_all_weak.py │ │ │ ├── uda_officehome_C_R_weak.py │ │ │ ├── uda_officehome_P_A.py │ │ │ ├── uda_officehome_P_A_all_weak.py │ │ │ ├── uda_officehome_P_A_weak.py │ │ │ ├── uda_officehome_P_C.py │ │ │ ├── uda_officehome_P_C_all_weak.py │ │ │ ├── uda_officehome_P_C_weak.py │ │ │ ├── uda_officehome_P_R.py │ │ │ ├── uda_officehome_P_R_all_weak.py │ │ │ ├── uda_officehome_P_R_weak.py │ │ │ ├── uda_officehome_R_A.py │ │ │ ├── uda_officehome_R_A_all_weak.py │ │ │ ├── uda_officehome_R_A_weak.py │ │ │ ├── uda_officehome_R_C.py │ │ │ ├── uda_officehome_R_C_all_weak.py │ │ │ ├── uda_officehome_R_C_weak.py │ │ │ ├── uda_officehome_R_P.py │ │ │ ├── uda_officehome_R_P_all_weak.py │ │ │ └── uda_officehome_R_P_weak.py │ │ └── uda_visda │ │ │ ├── uda_visda.py │ │ │ ├── uda_visda_2gpu.py │ │ │ ├── uda_visda_2gpu_all_train_aug.py │ │ │ └── uda_visda_2gpu_train_aug.py │ └── cls_models │ │ ├── resnet_50_gvb.py │ │ └── resnet_50_gvb_visda.py └── gvb │ ├── gvb_officehome_A_C.py │ ├── gvb_officehome_A_C_fixmatch.py │ ├── gvb_officehome_A_C_fixmatch_nce.py │ ├── gvb_officehome_A_C_nce.py │ ├── gvb_officehome_A_C_nce_all_weak.py │ ├── gvb_officehome_A_C_nce_fixmatch.py │ ├── gvb_officehome_A_C_nce_weak.py │ ├── gvb_officehome_A_C_test.py │ ├── gvb_officehome_A_C_weak.py │ ├── gvb_officehome_A_P_fixmatch.py │ ├── gvb_officehome_A_P_fixmatch_nce.py │ ├── gvb_officehome_A_P_nce.py │ ├── gvb_officehome_A_P_nce_all_weak.py │ ├── gvb_officehome_A_P_nce_weak.py │ ├── gvb_officehome_A_P_weak.py │ ├── gvb_officehome_A_R_fixmatch.py │ ├── gvb_officehome_A_R_fixmatch_nce.py │ ├── gvb_officehome_A_R_nce.py │ ├── gvb_officehome_A_R_nce_all_weak.py │ ├── gvb_officehome_A_R_nce_weak.py │ ├── gvb_officehome_A_R_weak.py │ ├── gvb_officehome_C_A_fixmatch.py │ ├── gvb_officehome_C_A_fixmatch_nce.py │ ├── gvb_officehome_C_A_nce.py │ ├── gvb_officehome_C_A_nce_all_weak.py │ ├── gvb_officehome_C_A_nce_weak.py │ ├── gvb_officehome_C_A_weak.py │ ├── gvb_officehome_C_P_fixmatch.py │ ├── gvb_officehome_C_P_fixmatch_nce.py │ ├── gvb_officehome_C_P_nce.py │ ├── gvb_officehome_C_P_nce_all_weak.py │ ├── gvb_officehome_C_P_nce_weak.py │ ├── gvb_officehome_C_P_weak.py │ ├── gvb_officehome_C_R_fixmatch.py │ ├── gvb_officehome_C_R_fixmatch_nce.py │ ├── gvb_officehome_C_R_nce.py │ ├── gvb_officehome_C_R_nce_all_weak.py │ ├── gvb_officehome_C_R_nce_weak.py │ ├── gvb_officehome_C_R_weak.py │ ├── gvb_officehome_P_A_fixmatch.py │ ├── gvb_officehome_P_A_fixmatch_nce.py │ ├── gvb_officehome_P_A_nce.py │ ├── gvb_officehome_P_A_nce_all_weak.py │ ├── gvb_officehome_P_A_nce_weak.py │ ├── gvb_officehome_P_A_weak.py │ ├── gvb_officehome_P_C_fixmatch.py │ ├── gvb_officehome_P_C_fixmatch_nce.py │ ├── gvb_officehome_P_C_nce.py │ ├── gvb_officehome_P_C_nce_all_weak.py │ ├── gvb_officehome_P_C_nce_weak.py │ ├── gvb_officehome_P_C_weak.py │ ├── gvb_officehome_P_R_fixmatch.py │ ├── gvb_officehome_P_R_fixmatch_nce.py │ ├── gvb_officehome_P_R_nce.py │ ├── gvb_officehome_P_R_nce_all_weak.py │ ├── gvb_officehome_P_R_nce_weak.py │ ├── gvb_officehome_P_R_weak.py │ ├── gvb_officehome_R_A_fixmatch.py │ ├── gvb_officehome_R_A_fixmatch_nce.py │ ├── gvb_officehome_R_A_nce.py │ ├── gvb_officehome_R_A_nce_all_weak.py │ ├── gvb_officehome_R_A_nce_weak.py │ ├── gvb_officehome_R_A_weak.py │ ├── gvb_officehome_R_C_fixmatch.py │ ├── gvb_officehome_R_C_fixmatch_nce.py │ ├── gvb_officehome_R_C_nce.py │ ├── gvb_officehome_R_C_nce_all_weak.py │ ├── gvb_officehome_R_C_nce_weak.py │ ├── gvb_officehome_R_C_weak.py │ ├── gvb_officehome_R_P_fixmatch.py │ ├── gvb_officehome_R_P_fixmatch_nce.py │ ├── gvb_officehome_R_P_nce.py │ ├── gvb_officehome_R_P_nce_all_weak.py │ ├── gvb_officehome_R_P_nce_weak.py │ └── gvb_officehome_R_P_weak.py ├── data └── txt │ └── officehome │ ├── labeled_source_images_Art.txt │ ├── labeled_source_images_Clipart.txt │ ├── labeled_source_images_Product.txt │ ├── labeled_source_images_Real.txt │ ├── unlabeled_target_images_Art_0.txt │ ├── unlabeled_target_images_Clipart_0.txt │ ├── unlabeled_target_images_Product_0.txt │ ├── unlabeled_target_images_Real_0.txt │ ├── validation_target_images_Art_3.txt │ ├── validation_target_images_Clipart_3.txt │ ├── validation_target_images_Product_3.txt │ └── validation_target_images_Real_3.txt ├── experiments └── scripts │ ├── generate_random_port.py │ ├── get_visible_card_num.py │ └── uda_gvb_train_release.sh ├── train.py └── training_logs ├── OfficeHome_AC └── events.out.tfevents.1628751971.38927820282d.159.0 ├── OfficeHome_AP └── events.out.tfevents.1628752143.0a5bab1dda68.154.0 ├── OfficeHome_AR └── events.out.tfevents.1628752173.33d1f1e1bf5a.159.0 ├── OfficeHome_CA └── events.out.tfevents.1628752855.9bfb4147b9cf.159.0 ├── OfficeHome_CP └── events.out.tfevents.1628752285.66c55db0326d.159.0 ├── OfficeHome_CR └── events.out.tfevents.1628752499.791e1b1299e1.154.0 ├── OfficeHome_PA └── events.out.tfevents.1628752416.b8545536f29f.203.0 ├── OfficeHome_PC └── events.out.tfevents.1628752363.ab72c3b0dc3a.154.0 ├── OfficeHome_PR └── events.out.tfevents.1628752417.719def797de8.202.0 ├── OfficeHome_RA └── events.out.tfevents.1628752597.cddd8f788339.154.0 ├── OfficeHome_RC └── events.out.tfevents.1628748264.89588d5e003f.154.0 ├── OfficeHome_RP └── events.out.tfevents.1628752442.a76f7378e503.159.0 └── ReadMe.md /SSL_Flexmatch/README.md: -------------------------------------------------------------------------------- 1 | # TorchSSL 2 | 3 | 4 | 5 | This code is based on [FlexMatch: boosting semi-supervised learning using curriculum pseudo labeling](https://proceedings.neurips.cc/paper/2021/hash/995693c15f439e3d189b06e89d145dd5-Abstract.html). 6 | 7 | ### Training 8 | 9 | Command line for training model on 4 GPU 10 | ```bash 11 | sh train.sh 12 | 13 | -------------------------------------------------------------------------------- /SSL_Flexmatch/__pycache__/custom_writer.cpython-38.pyc: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/__pycache__/custom_writer.cpython-38.pyc -------------------------------------------------------------------------------- /SSL_Flexmatch/__pycache__/train_utils.cpython-38.pyc: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/__pycache__/train_utils.cpython-38.pyc -------------------------------------------------------------------------------- /SSL_Flexmatch/__pycache__/utils.cpython-38.pyc: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/__pycache__/utils.cpython-38.pyc -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fixmatch/fixmatch_cifar100_10000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fixmatch_cifar100_10000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 10000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.001 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 8 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar100 32 | train_sampler: RandomSampler 33 | num_classes: 100 34 | num_workers: 1 35 | alg: fixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10007 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fixmatch/fixmatch_cifar100_2500_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fixmatch_cifar100_2500_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 2500 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.001 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 8 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar100 32 | train_sampler: RandomSampler 33 | num_classes: 100 34 | num_workers: 1 35 | alg: fixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10007 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fixmatch/fixmatch_cifar100_400_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fixmatch_cifar100_400_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 400 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.001 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 8 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar100 32 | train_sampler: RandomSampler 33 | num_classes: 100 34 | num_workers: 1 35 | alg: fixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10007 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fixmatch/fixmatch_cifar10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fixmatch_cifar10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: fixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10006 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fixmatch/fixmatch_cifar10_4000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fixmatch_cifar10_4000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 4000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: fixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10006 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fixmatch/fixmatch_cifar10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fixmatch_cifar10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: fixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10006 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fixmatch/fixmatch_imagenet_100000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fixmatch_imagenet_100000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 10000 10 | num_labels: 100000 11 | batch_size: 128 12 | eval_batch_size: 256 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.7 16 | ulb_loss_ratio: 1.0 17 | uratio: 1 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0003 23 | amp: False 24 | net: ResNet50 25 | net_from_name: False 26 | depth: 0 27 | widen_factor: 0 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: imagenet 32 | train_sampler: RandomSampler 33 | num_classes: 1000 34 | num_workers: 1 35 | alg: fixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10010 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fixmatch/fixmatch_stl10_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fixmatch_stl10_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNetVar 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: stl10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: fixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10009 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fixmatch/fixmatch_stl10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fixmatch_stl10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNetVar 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: stl10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: fixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10009 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fixmatch/fixmatch_stl10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fixmatch_stl10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNetVar 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: stl10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: fixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10009 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fixmatch/fixmatch_svhn_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fixmatch_svhn_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: svhn 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: fixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10008 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fixmatch/fixmatch_svhn_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fixmatch_svhn_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: svhn 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: fixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10008 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fixmatch/fixmatch_svhn_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fixmatch_svhn_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: svhn 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: fixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10008 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/flexmatch/flexmatch_cifar100_10000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: flexmatch_cifar100_10000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 10000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.001 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 8 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar100 32 | train_sampler: RandomSampler 33 | num_classes: 100 34 | num_workers: 1 35 | alg: flexmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10002 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/flexmatch/flexmatch_cifar100_2500_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: flexmatch_cifar100_2500_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 2500 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.001 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 8 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar100 32 | train_sampler: RandomSampler 33 | num_classes: 100 34 | num_workers: 1 35 | alg: flexmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10002 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/flexmatch/flexmatch_cifar100_400_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: /gdata2/lijj/Torch_ssl 2 | save_name: flexmatch_cifar100_400_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 400 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.001 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 8 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar100 32 | train_sampler: RandomSampler 33 | num_classes: 100 34 | num_workers: 1 35 | alg: flexmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10002 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/flexmatch/flexmatch_cifar100_400_2.yaml: -------------------------------------------------------------------------------- 1 | save_dir: /gdata2/lijj/Torch_ssl 2 | save_name: flexmatch_cifar100_400_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 400 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.001 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 8 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar100 32 | train_sampler: RandomSampler 33 | num_classes: 100 34 | num_workers: 1 35 | alg: flexmatch 36 | seed: 1 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10002 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/flexmatch/flexmatch_cifar100_400_3.yaml: -------------------------------------------------------------------------------- 1 | save_dir: /gdata2/lijj/Torch_ssl 2 | save_name: flexmatch_cifar100_400_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 400 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.001 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 8 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar100 32 | train_sampler: RandomSampler 33 | num_classes: 100 34 | num_workers: 1 35 | alg: flexmatch 36 | seed: 2 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10002 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/flexmatch/flexmatch_cifar10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: flexmatch_cifar10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: flexmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10001 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/flexmatch/flexmatch_cifar10_4000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: flexmatch_cifar10_4000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 4000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: flexmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10001 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/flexmatch/flexmatch_cifar10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: /gdata2/lijj/Torch_ssl 2 | save_name: flexmatch_cifar10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: flexmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10001 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/flexmatch/flexmatch_cifar10_40_1.yaml: -------------------------------------------------------------------------------- 1 | save_dir: /gdata2/lijj/Torch_ssl 2 | save_name: flexmatch_cifar10_40_1 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: flexmatch 36 | seed: 1 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10001 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/flexmatch/flexmatch_cifar10_40_2.yaml: -------------------------------------------------------------------------------- 1 | save_dir: /gdata2/lijj/Torch_ssl 2 | save_name: flexmatch_cifar10_40_2 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: flexmatch 36 | seed: 2 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10001 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/flexmatch/flexmatch_imagenet_100000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: flexmatch_imagenet_100000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 10000 10 | num_labels: 100000 11 | batch_size: 128 12 | eval_batch_size: 256 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.7 16 | ulb_loss_ratio: 1.0 17 | uratio: 1 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0003 23 | amp: False 24 | net: ResNet50 25 | net_from_name: False 26 | depth: 0 27 | widen_factor: 0 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: imagenet 32 | train_sampler: RandomSampler 33 | num_classes: 1000 34 | num_workers: 1 35 | alg: flexmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10005 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/flexmatch/flexmatch_stl10_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: flexmatch_stl10_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNetVar 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: stl10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: flexmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10004 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/flexmatch/flexmatch_stl10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: flexmatch_stl10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNetVar 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: stl10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: flexmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10004 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/flexmatch/flexmatch_stl10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: flexmatch_stl10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNetVar 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: stl10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: flexmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10004 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/flexmatch/flexmatch_svhn_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: flexmatch_svhn_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: svhn 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: flexmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10003 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/flexmatch/flexmatch_svhn_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: flexmatch_svhn_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: svhn 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: flexmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10003 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/flexmatch/flexmatch_svhn_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: flexmatch_svhn_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | hard_label: True 14 | T: 0.5 15 | p_cutoff: 0.95 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: svhn 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: flexmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10003 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fullysupervised/fullysupervised_cifar100_10000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fullysupervised_cifar100_10000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 10000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.001 18 | amp: False 19 | net: WideResNet 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 8 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: cifar100 27 | train_sampler: RandomSampler 28 | num_classes: 100 29 | num_workers: 1 30 | alg: fullysupervised 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10020 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fullysupervised/fullysupervised_cifar100_2500_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fullysupervised_cifar100_2500_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 2500 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.001 18 | amp: False 19 | net: WideResNet 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 8 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: cifar100 27 | train_sampler: RandomSampler 28 | num_classes: 100 29 | num_workers: 1 30 | alg: fullysupervised 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10020 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fullysupervised/fullysupervised_cifar100_400_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fullysupervised_cifar100_400_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 400 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.001 18 | amp: False 19 | net: WideResNet 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 8 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: cifar100 27 | train_sampler: RandomSampler 28 | num_classes: 100 29 | num_workers: 1 30 | alg: fullysupervised 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10020 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fullysupervised/fullysupervised_cifar10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fullysupervised_cifar10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.0005 18 | amp: False 19 | net: WideResNet 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 2 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: cifar10 27 | train_sampler: RandomSampler 28 | num_classes: 10 29 | num_workers: 1 30 | alg: fullysupervised 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10019 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fullysupervised/fullysupervised_cifar10_4000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fullysupervised_cifar10_4000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 4000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.0005 18 | amp: False 19 | net: WideResNet 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 2 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: cifar10 27 | train_sampler: RandomSampler 28 | num_classes: 10 29 | num_workers: 1 30 | alg: fullysupervised 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10019 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fullysupervised/fullysupervised_cifar10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fullysupervised_cifar10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.0005 18 | amp: False 19 | net: WideResNet 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 2 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: cifar10 27 | train_sampler: RandomSampler 28 | num_classes: 10 29 | num_workers: 1 30 | alg: fullysupervised 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10019 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fullysupervised/fullysupervised_stl10_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fullysupervised_stl10_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.0005 18 | amp: False 19 | net: WideResNetVar 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 2 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: stl10 27 | train_sampler: RandomSampler 28 | num_classes: 10 29 | num_workers: 1 30 | alg: fullysupervised 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10022 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fullysupervised/fullysupervised_stl10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fullysupervised_stl10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.0005 18 | amp: False 19 | net: WideResNetVar 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 2 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: stl10 27 | train_sampler: RandomSampler 28 | num_classes: 10 29 | num_workers: 1 30 | alg: fullysupervised 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10022 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fullysupervised/fullysupervised_stl10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fullysupervised_stl10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.0005 18 | amp: False 19 | net: WideResNetVar 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 2 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: stl10 27 | train_sampler: RandomSampler 28 | num_classes: 10 29 | num_workers: 1 30 | alg: fullysupervised 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10022 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fullysupervised/fullysupervised_svhn_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fullysupervised_svhn_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.0005 18 | amp: False 19 | net: WideResNet 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 2 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: svhn 27 | train_sampler: RandomSampler 28 | num_classes: 10 29 | num_workers: 1 30 | alg: fullysupervised 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10021 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fullysupervised/fullysupervised_svhn_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fullysupervised_svhn_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.0005 18 | amp: False 19 | net: WideResNet 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 2 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: svhn 27 | train_sampler: RandomSampler 28 | num_classes: 10 29 | num_workers: 1 30 | alg: fullysupervised 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10021 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/fullysupervised/fullysupervised_svhn_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: fullysupervised_svhn_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.0005 18 | amp: False 19 | net: WideResNet 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 2 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: svhn 27 | train_sampler: RandomSampler 28 | num_classes: 10 29 | num_workers: 1 30 | alg: fullysupervised 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10021 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/meanteacher/meanteacher_cifar100_10000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: meanteacher_cifar100_10000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 10000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | ulb_loss_ratio: 50 15 | unsup_warm_up: 0.4 16 | ema_m: 0.999 17 | optim: SGD 18 | lr: 0.03 19 | momentum: 0.9 20 | weight_decay: 0.001 21 | amp: False 22 | net: WideResNet 23 | net_from_name: False 24 | depth: 28 25 | widen_factor: 8 26 | leaky_slope: 0.1 27 | dropout: 0.0 28 | data_dir: ./data 29 | dataset: cifar100 30 | train_sampler: RandomSampler 31 | num_classes: 100 32 | num_workers: 1 33 | alg: meanteacher 34 | seed: 0 35 | world_size: 1 36 | rank: 0 37 | multiprocessing_distributed: True 38 | dist_url: tcp://127.0.0.1:10032 39 | dist_backend: nccl 40 | gpu: None 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/meanteacher/meanteacher_cifar100_2500_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: meanteacher_cifar100_2500_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 2500 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | ulb_loss_ratio: 50 15 | unsup_warm_up: 0.4 16 | ema_m: 0.999 17 | optim: SGD 18 | lr: 0.03 19 | momentum: 0.9 20 | weight_decay: 0.001 21 | amp: False 22 | net: WideResNet 23 | net_from_name: False 24 | depth: 28 25 | widen_factor: 8 26 | leaky_slope: 0.1 27 | dropout: 0.0 28 | data_dir: ./data 29 | dataset: cifar100 30 | train_sampler: RandomSampler 31 | num_classes: 100 32 | num_workers: 1 33 | alg: meanteacher 34 | seed: 0 35 | world_size: 1 36 | rank: 0 37 | multiprocessing_distributed: True 38 | dist_url: tcp://127.0.0.1:10032 39 | dist_backend: nccl 40 | gpu: None 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/meanteacher/meanteacher_cifar100_400_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: meanteacher_cifar100_400_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 400 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | ulb_loss_ratio: 50 15 | unsup_warm_up: 0.4 16 | ema_m: 0.999 17 | optim: SGD 18 | lr: 0.03 19 | momentum: 0.9 20 | weight_decay: 0.001 21 | amp: False 22 | net: WideResNet 23 | net_from_name: False 24 | depth: 28 25 | widen_factor: 8 26 | leaky_slope: 0.1 27 | dropout: 0.0 28 | data_dir: ./data 29 | dataset: cifar100 30 | train_sampler: RandomSampler 31 | num_classes: 100 32 | num_workers: 1 33 | alg: meanteacher 34 | seed: 0 35 | world_size: 1 36 | rank: 0 37 | multiprocessing_distributed: True 38 | dist_url: tcp://127.0.0.1:10032 39 | dist_backend: nccl 40 | gpu: None 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/meanteacher/meanteacher_cifar10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: meanteacher_cifar10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | ulb_loss_ratio: 50 15 | unsup_warm_up: 0.4 16 | ema_m: 0.999 17 | optim: SGD 18 | lr: 0.03 19 | momentum: 0.9 20 | weight_decay: 0.0005 21 | amp: False 22 | net: WideResNet 23 | net_from_name: False 24 | depth: 28 25 | widen_factor: 2 26 | leaky_slope: 0.1 27 | dropout: 0.0 28 | data_dir: ./data 29 | dataset: cifar10 30 | train_sampler: RandomSampler 31 | num_classes: 10 32 | num_workers: 1 33 | alg: meanteacher 34 | seed: 0 35 | world_size: 1 36 | rank: 0 37 | multiprocessing_distributed: True 38 | dist_url: tcp://127.0.0.1:10031 39 | dist_backend: nccl 40 | gpu: None 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/meanteacher/meanteacher_cifar10_4000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: meanteacher_cifar10_4000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 4000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | ulb_loss_ratio: 50 15 | unsup_warm_up: 0.4 16 | ema_m: 0.999 17 | optim: SGD 18 | lr: 0.03 19 | momentum: 0.9 20 | weight_decay: 0.0005 21 | amp: False 22 | net: WideResNet 23 | net_from_name: False 24 | depth: 28 25 | widen_factor: 2 26 | leaky_slope: 0.1 27 | dropout: 0.0 28 | data_dir: ./data 29 | dataset: cifar10 30 | train_sampler: RandomSampler 31 | num_classes: 10 32 | num_workers: 1 33 | alg: meanteacher 34 | seed: 0 35 | world_size: 1 36 | rank: 0 37 | multiprocessing_distributed: True 38 | dist_url: tcp://127.0.0.1:10031 39 | dist_backend: nccl 40 | gpu: None 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/meanteacher/meanteacher_cifar10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: meanteacher_cifar10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | ulb_loss_ratio: 50 15 | unsup_warm_up: 0.4 16 | ema_m: 0.999 17 | optim: SGD 18 | lr: 0.03 19 | momentum: 0.9 20 | weight_decay: 0.0005 21 | amp: False 22 | net: WideResNet 23 | net_from_name: False 24 | depth: 28 25 | widen_factor: 2 26 | leaky_slope: 0.1 27 | dropout: 0.0 28 | data_dir: ./data 29 | dataset: cifar10 30 | train_sampler: RandomSampler 31 | num_classes: 10 32 | num_workers: 1 33 | alg: meanteacher 34 | seed: 0 35 | world_size: 1 36 | rank: 0 37 | multiprocessing_distributed: True 38 | dist_url: tcp://127.0.0.1:10031 39 | dist_backend: nccl 40 | gpu: None 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/meanteacher/meanteacher_stl10_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: meanteacher_stl10_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | ulb_loss_ratio: 50 15 | unsup_warm_up: 0.4 16 | ema_m: 0.999 17 | optim: SGD 18 | lr: 0.03 19 | momentum: 0.9 20 | weight_decay: 0.0005 21 | amp: False 22 | net: WideResNetVar 23 | net_from_name: False 24 | depth: 28 25 | widen_factor: 2 26 | leaky_slope: 0.1 27 | dropout: 0.0 28 | data_dir: ./data 29 | dataset: stl10 30 | train_sampler: RandomSampler 31 | num_classes: 10 32 | num_workers: 1 33 | alg: meanteacher 34 | seed: 0 35 | world_size: 1 36 | rank: 0 37 | multiprocessing_distributed: True 38 | dist_url: tcp://127.0.0.1:10034 39 | dist_backend: nccl 40 | gpu: None 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/meanteacher/meanteacher_stl10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: meanteacher_stl10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | ulb_loss_ratio: 50 15 | unsup_warm_up: 0.4 16 | ema_m: 0.999 17 | optim: SGD 18 | lr: 0.03 19 | momentum: 0.9 20 | weight_decay: 0.0005 21 | amp: False 22 | net: WideResNetVar 23 | net_from_name: False 24 | depth: 28 25 | widen_factor: 2 26 | leaky_slope: 0.1 27 | dropout: 0.0 28 | data_dir: ./data 29 | dataset: stl10 30 | train_sampler: RandomSampler 31 | num_classes: 10 32 | num_workers: 1 33 | alg: meanteacher 34 | seed: 0 35 | world_size: 1 36 | rank: 0 37 | multiprocessing_distributed: True 38 | dist_url: tcp://127.0.0.1:10034 39 | dist_backend: nccl 40 | gpu: None 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/meanteacher/meanteacher_stl10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: meanteacher_stl10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | ulb_loss_ratio: 50 15 | unsup_warm_up: 0.4 16 | ema_m: 0.999 17 | optim: SGD 18 | lr: 0.03 19 | momentum: 0.9 20 | weight_decay: 0.0005 21 | amp: False 22 | net: WideResNetVar 23 | net_from_name: False 24 | depth: 28 25 | widen_factor: 2 26 | leaky_slope: 0.1 27 | dropout: 0.0 28 | data_dir: ./data 29 | dataset: stl10 30 | train_sampler: RandomSampler 31 | num_classes: 10 32 | num_workers: 1 33 | alg: meanteacher 34 | seed: 0 35 | world_size: 1 36 | rank: 0 37 | multiprocessing_distributed: True 38 | dist_url: tcp://127.0.0.1:10034 39 | dist_backend: nccl 40 | gpu: None 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/meanteacher/meanteacher_svhn_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: meanteacher_svhn_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | ulb_loss_ratio: 50 15 | unsup_warm_up: 0.4 16 | ema_m: 0.999 17 | optim: SGD 18 | lr: 0.03 19 | momentum: 0.9 20 | weight_decay: 0.0005 21 | amp: False 22 | net: WideResNet 23 | net_from_name: False 24 | depth: 28 25 | widen_factor: 2 26 | leaky_slope: 0.1 27 | dropout: 0.0 28 | data_dir: ./data 29 | dataset: svhn 30 | train_sampler: RandomSampler 31 | num_classes: 10 32 | num_workers: 1 33 | alg: meanteacher 34 | seed: 0 35 | world_size: 1 36 | rank: 0 37 | multiprocessing_distributed: True 38 | dist_url: tcp://127.0.0.1:10033 39 | dist_backend: nccl 40 | gpu: None 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/meanteacher/meanteacher_svhn_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: meanteacher_svhn_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | ulb_loss_ratio: 50 15 | unsup_warm_up: 0.4 16 | ema_m: 0.999 17 | optim: SGD 18 | lr: 0.03 19 | momentum: 0.9 20 | weight_decay: 0.0005 21 | amp: False 22 | net: WideResNet 23 | net_from_name: False 24 | depth: 28 25 | widen_factor: 2 26 | leaky_slope: 0.1 27 | dropout: 0.0 28 | data_dir: ./data 29 | dataset: svhn 30 | train_sampler: RandomSampler 31 | num_classes: 10 32 | num_workers: 1 33 | alg: meanteacher 34 | seed: 0 35 | world_size: 1 36 | rank: 0 37 | multiprocessing_distributed: True 38 | dist_url: tcp://127.0.0.1:10033 39 | dist_backend: nccl 40 | gpu: None 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/meanteacher/meanteacher_svhn_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: meanteacher_svhn_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | ulb_loss_ratio: 50 15 | unsup_warm_up: 0.4 16 | ema_m: 0.999 17 | optim: SGD 18 | lr: 0.03 19 | momentum: 0.9 20 | weight_decay: 0.0005 21 | amp: False 22 | net: WideResNet 23 | net_from_name: False 24 | depth: 28 25 | widen_factor: 2 26 | leaky_slope: 0.1 27 | dropout: 0.0 28 | data_dir: ./data 29 | dataset: svhn 30 | train_sampler: RandomSampler 31 | num_classes: 10 32 | num_workers: 1 33 | alg: meanteacher 34 | seed: 0 35 | world_size: 1 36 | rank: 0 37 | multiprocessing_distributed: True 38 | dist_url: tcp://127.0.0.1:10033 39 | dist_backend: nccl 40 | gpu: None 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/mixmatch/mixmatch_cifar100_10000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: mixmatch_cifar100_10000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 10000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | alpha: 0.5 15 | T: 0.5 16 | ulb_loss_ratio: 100 17 | ramp_up: 0.4 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.001 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 8 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar100 32 | train_sampler: RandomSampler 33 | num_classes: 100 34 | num_workers: 1 35 | alg: mixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10028 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/mixmatch/mixmatch_cifar100_2500_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: mixmatch_cifar100_2500_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 2500 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | alpha: 0.5 15 | T: 0.5 16 | ulb_loss_ratio: 100 17 | ramp_up: 0.4 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.001 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 8 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar100 32 | train_sampler: RandomSampler 33 | num_classes: 100 34 | num_workers: 1 35 | alg: mixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10028 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/mixmatch/mixmatch_cifar100_400_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: mixmatch_cifar100_400_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 400 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | alpha: 0.5 15 | T: 0.5 16 | ulb_loss_ratio: 100 17 | ramp_up: 0.4 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.001 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 8 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar100 32 | train_sampler: RandomSampler 33 | num_classes: 100 34 | num_workers: 1 35 | alg: mixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10028 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/mixmatch/mixmatch_cifar10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: mixmatch_cifar10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | alpha: 0.5 15 | T: 0.5 16 | ulb_loss_ratio: 100 17 | ramp_up: 0.4 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: mixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10027 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/mixmatch/mixmatch_cifar10_4000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: mixmatch_cifar10_4000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 4000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | alpha: 0.5 15 | T: 0.5 16 | ulb_loss_ratio: 100 17 | ramp_up: 0.4 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: mixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10027 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/mixmatch/mixmatch_cifar10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: mixmatch_cifar10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | alpha: 0.5 15 | T: 0.5 16 | ulb_loss_ratio: 100 17 | ramp_up: 0.4 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: mixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10027 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/mixmatch/mixmatch_stl10_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: mixmatch_stl10_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | alpha: 0.5 15 | T: 0.5 16 | ulb_loss_ratio: 100 17 | ramp_up: 0.4 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNetVar 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: stl10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: mixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10030 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/mixmatch/mixmatch_stl10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: mixmatch_stl10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | alpha: 0.5 15 | T: 0.5 16 | ulb_loss_ratio: 100 17 | ramp_up: 0.4 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNetVar 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: stl10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: mixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10030 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/mixmatch/mixmatch_stl10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: mixmatch_stl10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | alpha: 0.5 15 | T: 0.5 16 | ulb_loss_ratio: 100 17 | ramp_up: 0.4 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNetVar 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: stl10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: mixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10030 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/mixmatch/mixmatch_svhn_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: mixmatch_svhn_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | alpha: 0.5 15 | T: 0.5 16 | ulb_loss_ratio: 100 17 | ramp_up: 0.4 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: svhn 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: mixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10029 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/mixmatch/mixmatch_svhn_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: mixmatch_svhn_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | alpha: 0.5 15 | T: 0.5 16 | ulb_loss_ratio: 100 17 | ramp_up: 0.4 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: svhn 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: mixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10029 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/mixmatch/mixmatch_svhn_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: mixmatch_svhn_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | uratio: 1 14 | alpha: 0.5 15 | T: 0.5 16 | ulb_loss_ratio: 100 17 | ramp_up: 0.4 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: svhn 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: mixmatch 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10029 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pimodel/pimodel_cifar100_10000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pimodel_cifar100_10000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 10000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 10 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.001 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 8 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: cifar100 29 | train_sampler: RandomSampler 30 | num_classes: 100 31 | num_workers: 1 32 | alg: pimodel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10036 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pimodel/pimodel_cifar100_2500_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pimodel_cifar100_2500_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 2500 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 10 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.001 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 8 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: cifar100 29 | train_sampler: RandomSampler 30 | num_classes: 100 31 | num_workers: 1 32 | alg: pimodel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10036 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pimodel/pimodel_cifar100_400_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pimodel_cifar100_400_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 400 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 10 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.001 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 8 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: cifar100 29 | train_sampler: RandomSampler 30 | num_classes: 100 31 | num_workers: 1 32 | alg: pimodel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10036 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pimodel/pimodel_cifar10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pimodel_cifar10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 10 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: cifar10 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pimodel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10035 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pimodel/pimodel_cifar10_4000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pimodel_cifar10_4000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 4000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 10 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: cifar10 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pimodel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10035 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pimodel/pimodel_cifar10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pimodel_cifar10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 10 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: cifar10 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pimodel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10035 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pimodel/pimodel_stl10_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pimodel_stl10_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 10 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNetVar 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: stl10 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pimodel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10038 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pimodel/pimodel_stl10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pimodel_stl10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 10 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNetVar 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: stl10 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pimodel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10038 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pimodel/pimodel_stl10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pimodel_stl10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 10 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNetVar 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: stl10 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pimodel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10038 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pimodel/pimodel_svhn_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pimodel_svhn_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 10 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: svhn 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pimodel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10037 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pimodel/pimodel_svhn_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pimodel_svhn_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 10 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: svhn 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pimodel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10037 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pimodel/pimodel_svhn_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pimodel_svhn_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 10 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: svhn 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pimodel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10037 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel/pseudolabel_cifar100_10000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_cifar100_10000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 10000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.001 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 8 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: cifar100 29 | train_sampler: RandomSampler 30 | num_classes: 100 31 | num_workers: 1 32 | alg: pseudolabel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10016 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel/pseudolabel_cifar100_2500_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_cifar100_2500_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 2500 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.001 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 8 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: cifar100 29 | train_sampler: RandomSampler 30 | num_classes: 100 31 | num_workers: 1 32 | alg: pseudolabel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10016 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel/pseudolabel_cifar100_400_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_cifar100_400_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 400 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.001 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 8 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: cifar100 29 | train_sampler: RandomSampler 30 | num_classes: 100 31 | num_workers: 1 32 | alg: pseudolabel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10016 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel/pseudolabel_cifar10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_cifar10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: cifar10 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pseudolabel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10015 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel/pseudolabel_cifar10_4000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_cifar10_4000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 4000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: cifar10 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pseudolabel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10015 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel/pseudolabel_cifar10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_cifar10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: cifar10 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pseudolabel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10015 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel/pseudolabel_stl10_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_stl10_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNetVar 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: stl10 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pseudolabel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10018 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel/pseudolabel_stl10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_stl10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNetVar 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: stl10 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pseudolabel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10018 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel/pseudolabel_stl10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_stl10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNetVar 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: stl10 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pseudolabel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10018 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel/pseudolabel_svhn_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_svhn_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: svhn 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pseudolabel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10017 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel/pseudolabel_svhn_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_svhn_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: svhn 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pseudolabel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10017 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel/pseudolabel_svhn_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_svhn_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: svhn 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pseudolabel 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:10017 38 | dist_backend: nccl 39 | gpu: None 40 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel_flex/pseudolabel_flex_cifar100_10000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_flex_cifar100_10000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 10000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.001 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 8 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: cifar100 29 | train_sampler: RandomSampler 30 | num_classes: 100 31 | num_workers: 1 32 | alg: pseudolabel_flex 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:11126 38 | dist_backend: nccl 39 | gpu: None 40 | use_flex: True 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel_flex/pseudolabel_flex_cifar100_2500_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_flex_cifar100_2500_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 2500 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.001 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 8 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: cifar100 29 | train_sampler: RandomSampler 30 | num_classes: 100 31 | num_workers: 1 32 | alg: pseudolabel_flex 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:11126 38 | dist_backend: nccl 39 | gpu: None 40 | use_flex: True 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel_flex/pseudolabel_flex_cifar100_400_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_flex_cifar100_400_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 400 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.001 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 8 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: cifar100 29 | train_sampler: RandomSampler 30 | num_classes: 100 31 | num_workers: 1 32 | alg: pseudolabel_flex 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:11126 38 | dist_backend: nccl 39 | gpu: None 40 | use_flex: True 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel_flex/pseudolabel_flex_cifar10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_flex_cifar10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: cifar10 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pseudolabel_flex 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:11125 38 | dist_backend: nccl 39 | gpu: None 40 | use_flex: True 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel_flex/pseudolabel_flex_cifar10_4000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_flex_cifar10_4000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 4000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: cifar10 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pseudolabel_flex 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:11125 38 | dist_backend: nccl 39 | gpu: None 40 | use_flex: True 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel_flex/pseudolabel_flex_cifar10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_flex_cifar10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: cifar10 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pseudolabel_flex 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:11125 38 | dist_backend: nccl 39 | gpu: None 40 | use_flex: True 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel_flex/pseudolabel_flex_stl10_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_flex_stl10_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNetVar 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: stl10 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pseudolabel_flex 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:11128 38 | dist_backend: nccl 39 | gpu: None 40 | use_flex: True 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel_flex/pseudolabel_flex_stl10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_flex_stl10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNetVar 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: stl10 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pseudolabel_flex 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:11128 38 | dist_backend: nccl 39 | gpu: None 40 | use_flex: True 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel_flex/pseudolabel_flex_stl10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_flex_stl10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNetVar 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: stl10 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pseudolabel_flex 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:11128 38 | dist_backend: nccl 39 | gpu: None 40 | use_flex: True 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel_flex/pseudolabel_flex_svhn_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_flex_svhn_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: svhn 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pseudolabel_flex 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:11127 38 | dist_backend: nccl 39 | gpu: None 40 | use_flex: True 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel_flex/pseudolabel_flex_svhn_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_flex_svhn_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: svhn 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pseudolabel_flex 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:11127 38 | dist_backend: nccl 39 | gpu: None 40 | use_flex: True 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/pseudolabel_flex/pseudolabel_flex_svhn_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: pseudolabel_flex_svhn_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ulb_loss_ratio: 1.0 14 | uratio: 1 15 | ema_m: 0.999 16 | optim: SGD 17 | lr: 0.03 18 | momentum: 0.9 19 | weight_decay: 0.0005 20 | amp: False 21 | net: WideResNet 22 | net_from_name: False 23 | depth: 28 24 | widen_factor: 2 25 | leaky_slope: 0.1 26 | dropout: 0.0 27 | data_dir: ./data 28 | dataset: svhn 29 | train_sampler: RandomSampler 30 | num_classes: 10 31 | num_workers: 1 32 | alg: pseudolabel_flex 33 | seed: 0 34 | world_size: 1 35 | rank: 0 36 | multiprocessing_distributed: True 37 | dist_url: tcp://127.0.0.1:11127 38 | dist_backend: nccl 39 | gpu: None 40 | use_flex: True 41 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/uda/uda_cifar100_10000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: uda_cifar100_10000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 10000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | TSA_schedule: none 14 | T: 0.4 15 | p_cutoff: 0.8 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.001 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 8 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar100 32 | train_sampler: RandomSampler 33 | num_classes: 100 34 | num_workers: 1 35 | alg: uda 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10012 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/uda/uda_cifar100_2500_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: uda_cifar100_2500_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 2500 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | TSA_schedule: none 14 | T: 0.4 15 | p_cutoff: 0.8 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.001 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 8 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar100 32 | train_sampler: RandomSampler 33 | num_classes: 100 34 | num_workers: 1 35 | alg: uda 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10012 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/uda/uda_cifar100_400_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: uda_cifar100_400_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 400 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | TSA_schedule: none 14 | T: 0.4 15 | p_cutoff: 0.8 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.001 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 8 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar100 32 | train_sampler: RandomSampler 33 | num_classes: 100 34 | num_workers: 1 35 | alg: uda 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10012 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/uda/uda_cifar10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: uda_cifar10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | TSA_schedule: none 14 | T: 0.4 15 | p_cutoff: 0.8 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: uda 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10011 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/uda/uda_cifar10_4000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: uda_cifar10_4000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 4000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | TSA_schedule: none 14 | T: 0.4 15 | p_cutoff: 0.8 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: uda 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10011 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/uda/uda_cifar10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: uda_cifar10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | TSA_schedule: none 14 | T: 0.4 15 | p_cutoff: 0.8 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: uda 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10011 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/uda/uda_stl10_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: uda_stl10_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | TSA_schedule: none 14 | T: 0.4 15 | p_cutoff: 0.8 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNetVar 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: stl10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: uda 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10014 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/uda/uda_stl10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: uda_stl10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | TSA_schedule: none 14 | T: 0.4 15 | p_cutoff: 0.8 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNetVar 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: stl10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: uda 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10014 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/uda/uda_stl10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: uda_stl10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | TSA_schedule: none 14 | T: 0.4 15 | p_cutoff: 0.8 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNetVar 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: stl10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: uda 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10014 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/uda/uda_svhn_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: uda_svhn_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | TSA_schedule: none 14 | T: 0.4 15 | p_cutoff: 0.8 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: svhn 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: uda 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10013 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/uda/uda_svhn_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: uda_svhn_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | TSA_schedule: none 14 | T: 0.4 15 | p_cutoff: 0.8 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: svhn 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: uda 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10013 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/uda/uda_svhn_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: uda_svhn_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | TSA_schedule: none 14 | T: 0.4 15 | p_cutoff: 0.8 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: svhn 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: uda 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:10013 41 | dist_backend: nccl 42 | gpu: None 43 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/uda_flex/uda_flex_cifar10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: uda_flex_cifar10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | TSA_schedule: none 14 | T: 0.4 15 | p_cutoff: 0.8 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: uda_flex 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:11121 41 | dist_backend: nccl 42 | gpu: None 43 | use_flex: True 44 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/uda_flex/uda_flex_cifar10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: uda_flex_cifar10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | TSA_schedule: none 14 | T: 0.4 15 | p_cutoff: 0.8 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: cifar10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: uda_flex 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:11121 41 | dist_backend: nccl 42 | gpu: None 43 | use_flex: True 44 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/uda_flex/uda_flex_stl10_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: uda_flex_stl10_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | TSA_schedule: none 14 | T: 0.4 15 | p_cutoff: 0.8 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNetVar 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: stl10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: uda_flex 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:11124 41 | dist_backend: nccl 42 | gpu: None 43 | use_flex: True 44 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/uda_flex/uda_flex_stl10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: uda_flex_stl10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | TSA_schedule: none 14 | T: 0.4 15 | p_cutoff: 0.8 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNetVar 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: stl10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: uda_flex 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:11124 41 | dist_backend: nccl 42 | gpu: None 43 | use_flex: True 44 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/uda_flex/uda_flex_stl10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: uda_flex_stl10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | TSA_schedule: none 14 | T: 0.4 15 | p_cutoff: 0.8 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNetVar 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: stl10 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: uda_flex 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:11124 41 | dist_backend: nccl 42 | gpu: None 43 | use_flex: True 44 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/uda_flex/uda_flex_svhn_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: uda_flex_svhn_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | TSA_schedule: none 14 | T: 0.4 15 | p_cutoff: 0.8 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: svhn 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: uda_flex 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:11123 41 | dist_backend: nccl 42 | gpu: None 43 | use_flex: True 44 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/uda_flex/uda_flex_svhn_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: uda_flex_svhn_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | TSA_schedule: none 14 | T: 0.4 15 | p_cutoff: 0.8 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: svhn 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: uda_flex 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:11123 41 | dist_backend: nccl 42 | gpu: None 43 | use_flex: True 44 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/uda_flex/uda_flex_svhn_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: uda_flex_svhn_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | TSA_schedule: none 14 | T: 0.4 15 | p_cutoff: 0.8 16 | ulb_loss_ratio: 1.0 17 | uratio: 7 18 | ema_m: 0.999 19 | optim: SGD 20 | lr: 0.03 21 | momentum: 0.9 22 | weight_decay: 0.0005 23 | amp: False 24 | net: WideResNet 25 | net_from_name: False 26 | depth: 28 27 | widen_factor: 2 28 | leaky_slope: 0.1 29 | dropout: 0.0 30 | data_dir: ./data 31 | dataset: svhn 32 | train_sampler: RandomSampler 33 | num_classes: 10 34 | num_workers: 1 35 | alg: uda_flex 36 | seed: 0 37 | world_size: 1 38 | rank: 0 39 | multiprocessing_distributed: True 40 | dist_url: tcp://127.0.0.1:11123 41 | dist_backend: nccl 42 | gpu: None 43 | use_flex: True 44 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/vat/vat_cifar100_10000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: vat_cifar100_10000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 10000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.001 18 | amp: False 19 | net: WideResNet 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 8 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: cifar100 27 | train_sampler: RandomSampler 28 | num_classes: 100 29 | num_workers: 1 30 | alg: vat 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10040 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/vat/vat_cifar100_2500_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: vat_cifar100_2500_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 2500 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.001 18 | amp: False 19 | net: WideResNet 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 8 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: cifar100 27 | train_sampler: RandomSampler 28 | num_classes: 100 29 | num_workers: 1 30 | alg: vat 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10040 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/vat/vat_cifar100_400_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: vat_cifar100_400_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 400 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.001 18 | amp: False 19 | net: WideResNet 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 8 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: cifar100 27 | train_sampler: RandomSampler 28 | num_classes: 100 29 | num_workers: 1 30 | alg: vat 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10040 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/vat/vat_cifar10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: vat_cifar10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.0005 18 | amp: False 19 | net: WideResNet 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 2 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: cifar10 27 | train_sampler: RandomSampler 28 | num_classes: 10 29 | num_workers: 1 30 | alg: vat 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10039 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/vat/vat_cifar10_4000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: vat_cifar10_4000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 4000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.0005 18 | amp: False 19 | net: WideResNet 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 2 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: cifar10 27 | train_sampler: RandomSampler 28 | num_classes: 10 29 | num_workers: 1 30 | alg: vat 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10039 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/vat/vat_cifar10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: vat_cifar10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.0005 18 | amp: False 19 | net: WideResNet 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 2 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: cifar10 27 | train_sampler: RandomSampler 28 | num_classes: 10 29 | num_workers: 1 30 | alg: vat 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10039 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/vat/vat_stl10_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: vat_stl10_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.0005 18 | amp: False 19 | net: WideResNetVar 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 2 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: stl10 27 | train_sampler: RandomSampler 28 | num_classes: 10 29 | num_workers: 1 30 | alg: vat 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10042 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/vat/vat_stl10_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: vat_stl10_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.0005 18 | amp: False 19 | net: WideResNetVar 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 2 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: stl10 27 | train_sampler: RandomSampler 28 | num_classes: 10 29 | num_workers: 1 30 | alg: vat 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10042 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/vat/vat_stl10_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: vat_stl10_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.0005 18 | amp: False 19 | net: WideResNetVar 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 2 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: stl10 27 | train_sampler: RandomSampler 28 | num_classes: 10 29 | num_workers: 1 30 | alg: vat 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10042 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/vat/vat_svhn_1000_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: vat_svhn_1000_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 1000 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.0005 18 | amp: False 19 | net: WideResNet 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 2 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: svhn 27 | train_sampler: RandomSampler 28 | num_classes: 10 29 | num_workers: 1 30 | alg: vat 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10041 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/vat/vat_svhn_250_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: vat_svhn_250_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 250 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.0005 18 | amp: False 19 | net: WideResNet 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 2 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: svhn 27 | train_sampler: RandomSampler 28 | num_classes: 10 29 | num_workers: 1 30 | alg: vat 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10041 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/config/vat/vat_svhn_40_0.yaml: -------------------------------------------------------------------------------- 1 | save_dir: ./saved_models 2 | save_name: vat_svhn_40_0 3 | resume: False 4 | load_path: None 5 | overwrite: True 6 | use_tensorboard: True 7 | epoch: 1 8 | num_train_iter: 1048576 9 | num_eval_iter: 5000 10 | num_labels: 40 11 | batch_size: 64 12 | eval_batch_size: 1024 13 | ema_m: 0.999 14 | optim: SGD 15 | lr: 0.03 16 | momentum: 0.9 17 | weight_decay: 0.0005 18 | amp: False 19 | net: WideResNet 20 | net_from_name: False 21 | depth: 28 22 | widen_factor: 2 23 | leaky_slope: 0.1 24 | dropout: 0.0 25 | data_dir: ./data 26 | dataset: svhn 27 | train_sampler: RandomSampler 28 | num_classes: 10 29 | num_workers: 1 30 | alg: vat 31 | seed: 0 32 | world_size: 1 33 | rank: 0 34 | multiprocessing_distributed: True 35 | dist_url: tcp://127.0.0.1:10041 36 | dist_backend: nccl 37 | gpu: None 38 | -------------------------------------------------------------------------------- /SSL_Flexmatch/datasets/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/datasets/__init__.py -------------------------------------------------------------------------------- /SSL_Flexmatch/datasets/__pycache__/DistributedProxySampler.cpython-38.pyc: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/datasets/__pycache__/DistributedProxySampler.cpython-38.pyc -------------------------------------------------------------------------------- /SSL_Flexmatch/datasets/__pycache__/__init__.cpython-38.pyc: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/datasets/__pycache__/__init__.cpython-38.pyc -------------------------------------------------------------------------------- /SSL_Flexmatch/datasets/__pycache__/data_utils.cpython-38.pyc: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/datasets/__pycache__/data_utils.cpython-38.pyc -------------------------------------------------------------------------------- /SSL_Flexmatch/datasets/__pycache__/dataset.cpython-38.pyc: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/datasets/__pycache__/dataset.cpython-38.pyc -------------------------------------------------------------------------------- /SSL_Flexmatch/datasets/__pycache__/ssl_dataset.cpython-38.pyc: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/datasets/__pycache__/ssl_dataset.cpython-38.pyc -------------------------------------------------------------------------------- /SSL_Flexmatch/datasets/augmentation/__pycache__/randaugment.cpython-38.pyc: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/datasets/augmentation/__pycache__/randaugment.cpython-38.pyc -------------------------------------------------------------------------------- /SSL_Flexmatch/figures/cf10.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/figures/cf10.png -------------------------------------------------------------------------------- /SSL_Flexmatch/figures/cf100.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/figures/cf100.png -------------------------------------------------------------------------------- /SSL_Flexmatch/figures/logo.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/figures/logo.png -------------------------------------------------------------------------------- /SSL_Flexmatch/figures/stl.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/figures/stl.png -------------------------------------------------------------------------------- /SSL_Flexmatch/figures/svhn.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/figures/svhn.png -------------------------------------------------------------------------------- /SSL_Flexmatch/flexmatch_package.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/flexmatch_package.jpg -------------------------------------------------------------------------------- /SSL_Flexmatch/models/fixmatch/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/models/fixmatch/__init__.py -------------------------------------------------------------------------------- /SSL_Flexmatch/models/flexmatch/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/models/flexmatch/__init__.py -------------------------------------------------------------------------------- /SSL_Flexmatch/models/flexmatch/__pycache__/__init__.cpython-38.pyc: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/models/flexmatch/__pycache__/__init__.cpython-38.pyc -------------------------------------------------------------------------------- /SSL_Flexmatch/models/flexmatch/__pycache__/flexmatch.cpython-38.pyc: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/models/flexmatch/__pycache__/flexmatch.cpython-38.pyc -------------------------------------------------------------------------------- /SSL_Flexmatch/models/flexmatch/__pycache__/flexmatch_utils.cpython-38.pyc: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/models/flexmatch/__pycache__/flexmatch_utils.cpython-38.pyc -------------------------------------------------------------------------------- /SSL_Flexmatch/models/fullysupervised/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/models/fullysupervised/__init__.py -------------------------------------------------------------------------------- /SSL_Flexmatch/models/fullysupervised/fullysupervised_utils.py: -------------------------------------------------------------------------------- 1 | import torch 2 | import torch.nn.functional as F 3 | from train_utils import ce_loss 4 | 5 | 6 | class Get_Scalar: 7 | def __init__(self, value): 8 | self.value = value 9 | 10 | def get_value(self, iter): 11 | return self.value 12 | 13 | def __call__(self, iter): 14 | return self.value 15 | 16 | 17 | -------------------------------------------------------------------------------- /SSL_Flexmatch/models/meanteacher/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/models/meanteacher/__init__.py -------------------------------------------------------------------------------- /SSL_Flexmatch/models/meanteacher/meanteacher_utils.py: -------------------------------------------------------------------------------- 1 | import torch 2 | import torch.nn.functional as F 3 | from train_utils import ce_loss 4 | 5 | 6 | class Get_Scalar: 7 | def __init__(self, value): 8 | self.value = value 9 | 10 | def get_value(self, iter): 11 | return self.value 12 | 13 | def __call__(self, iter): 14 | return self.value 15 | 16 | 17 | def consistency_loss(logits_w1, logits_w2): 18 | logits_w2 = logits_w2.detach() 19 | assert logits_w1.size() == logits_w2.size() 20 | return F.mse_loss(torch.softmax(logits_w1,dim=-1), torch.softmax(logits_w2,dim=-1), reduction='mean') 21 | 22 | 23 | -------------------------------------------------------------------------------- /SSL_Flexmatch/models/mixmatch/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/models/mixmatch/__init__.py -------------------------------------------------------------------------------- /SSL_Flexmatch/models/nets/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/models/nets/__init__.py -------------------------------------------------------------------------------- /SSL_Flexmatch/models/nets/__pycache__/__init__.cpython-38.pyc: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/models/nets/__pycache__/__init__.cpython-38.pyc -------------------------------------------------------------------------------- /SSL_Flexmatch/models/nets/__pycache__/wrn.cpython-38.pyc: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/models/nets/__pycache__/wrn.cpython-38.pyc -------------------------------------------------------------------------------- /SSL_Flexmatch/models/pimodel/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/models/pimodel/__init__.py -------------------------------------------------------------------------------- /SSL_Flexmatch/models/pimodel/pimodel_utils.py: -------------------------------------------------------------------------------- 1 | import torch 2 | import torch.nn.functional as F 3 | from train_utils import ce_loss 4 | 5 | 6 | class Get_Scalar: 7 | def __init__(self, value): 8 | self.value = value 9 | 10 | def get_value(self, iter): 11 | return self.value 12 | 13 | def __call__(self, iter): 14 | return self.value 15 | 16 | 17 | def consistency_loss(logits_w1, logits_w2): 18 | logits_w2 = logits_w2.detach() 19 | assert logits_w1.size() == logits_w2.size() 20 | return F.mse_loss(torch.softmax(logits_w1, dim=-1), torch.softmax(logits_w2, dim=-1), reduction='mean') 21 | -------------------------------------------------------------------------------- /SSL_Flexmatch/models/pseudolabel/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/models/pseudolabel/__init__.py -------------------------------------------------------------------------------- /SSL_Flexmatch/models/remixmatch/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/models/remixmatch/__init__.py -------------------------------------------------------------------------------- /SSL_Flexmatch/models/uda/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/models/uda/__init__.py -------------------------------------------------------------------------------- /SSL_Flexmatch/models/vat/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/SSL_Flexmatch/models/vat/__init__.py -------------------------------------------------------------------------------- /SSL_Flexmatch/scripts/cifar.sh: -------------------------------------------------------------------------------- 1 | cd ../ 2 | pwd 3 | source ~/.bashrc 4 | conda activate ssl 5 | 6 | 7 | declare -A alg_dist_dict 8 | alg_dist_dict=([pimodel]="10001" [pseudolabel]="10002" [meanteacher]="10003" [uda]="10004" [mixmatch]="10005" [remixmatch]="10006" [fixmatch]="10007") 9 | alg=$1 10 | num_class=$2 11 | weight_decay=$3 12 | dist_port=${alg_dist_dict[${alg}]} 13 | exp_name=${alg}_cifar${num_class} 14 | 15 | 16 | #for size in 4000 250 40; do 17 | for size in 250 4000; do 18 | python ${alg}.py --world-size 1 --rank 0 --amp False --multiprocessing-distributed True --num_labels ${size} --save_name ${exp_name}@${size}_${weight_decay} --weight_decay ${weight_decay} --dataset cifar${num_class} --num_classes ${num_class} --widen_factor 2 --overwrite True --dist-url 'tcp://127.0.0.1:'${dist_port} 19 | done 20 | 21 | -------------------------------------------------------------------------------- /SSL_Flexmatch/train.sh: -------------------------------------------------------------------------------- 1 | cd /ghome/lijj/DA/TorchSSL-main/TorchSSL-main/ 2 | python flexmatch.py --c ./config/flexmatch/flexmatch_cifar100_400_0.yaml 3 | #python flexmatch.py --c ./config/flexmatch/flexmatch_cifar10_40_0.yaml -------------------------------------------------------------------------------- /UDA_GVB/.idea/.gitignore: -------------------------------------------------------------------------------- 1 | # Default ignored files 2 | /shelf/ 3 | /workspace.xml 4 | # Editor-based HTTP Client requests 5 | /httpRequests/ 6 | # Datasource local storage ignored files 7 | /dataSources/ 8 | /dataSources.local.xml 9 | -------------------------------------------------------------------------------- /UDA_GVB/.idea/clsda.iml: -------------------------------------------------------------------------------- 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | -------------------------------------------------------------------------------- /UDA_GVB/.idea/deployment.xml: -------------------------------------------------------------------------------- 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 15 | -------------------------------------------------------------------------------- /UDA_GVB/.idea/inspectionProfiles/profiles_settings.xml: -------------------------------------------------------------------------------- 1 | 2 | 3 | 6 | -------------------------------------------------------------------------------- /UDA_GVB/.idea/modules.xml: -------------------------------------------------------------------------------- 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | -------------------------------------------------------------------------------- /UDA_GVB/.idea/remote-mappings.xml: -------------------------------------------------------------------------------- 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | -------------------------------------------------------------------------------- /UDA_GVB/.idea/webServers.xml: -------------------------------------------------------------------------------- 1 | 2 | 3 | 4 | 9 | 10 | -------------------------------------------------------------------------------- /UDA_GVB/clsda/__init__.py: -------------------------------------------------------------------------------- 1 | # 2 | # ---------------------------------------------- 3 | 4 | 5 | if __name__ == "__main__": 6 | pass 7 | -------------------------------------------------------------------------------- /UDA_GVB/clsda/loader/cls_loaders/builder.py: -------------------------------------------------------------------------------- 1 | # 2 | # ---------------------------------------------- 3 | from mmcv.utils import Registry, build_from_cfg 4 | from mmcls.datasets import PIPELINES as CLS_PIPELINES 5 | 6 | CLS_DATASETS = Registry('cls_datasets') 7 | 8 | 9 | def build_dataset(cfg, default_args=None): 10 | dataset = build_from_cfg(cfg, CLS_DATASETS, default_args) 11 | return dataset 12 | -------------------------------------------------------------------------------- /UDA_GVB/clsda/loader/cls_loaders/pipelines/__init__.py: -------------------------------------------------------------------------------- 1 | # 2 | # ---------------------------------------------- 3 | from mmcls.datasets.pipelines import Compose 4 | from .pipelines import Apply, ColorJitter 5 | from ..builder import CLS_PIPELINES 6 | 7 | 8 | if __name__ == "__main__": 9 | pass 10 | -------------------------------------------------------------------------------- /UDA_GVB/clsda/models/cls_models/builder.py: -------------------------------------------------------------------------------- 1 | # 2 | # ---------------------------------------------- 3 | import torch.nn as nn 4 | from mmcv.utils import Registry, build_from_cfg 5 | 6 | CLS_MODELS = Registry('cls_models') 7 | 8 | 9 | def build_cls_models(cfg, default_args=None): 10 | if isinstance(cfg, list): 11 | modules = [ 12 | build_from_cfg(cfg_, CLS_MODELS, default_args) for cfg_ in cfg 13 | ] 14 | return nn.Sequential(*modules) 15 | else: 16 | return build_from_cfg(cfg, CLS_MODELS, default_args) 17 | 18 | 19 | if __name__ == "__main__": 20 | pass 21 | -------------------------------------------------------------------------------- /UDA_GVB/clsda/runner/__init__.py: -------------------------------------------------------------------------------- 1 | # 2 | # ---------------------------------------------- 3 | -------------------------------------------------------------------------------- /UDA_GVB/clsda/schedulers/__init__.py: -------------------------------------------------------------------------------- 1 | from .builder import SCHEDULER, build_scheduler 2 | from .schedulers import TORCH_SCHEDULER, ConstantLR, PolynomialLR, WarmUpLR, InvLR 3 | 4 | __all__ = [ 5 | 'SCHEDULER', 'build_scheduler', 'TORCH_SCHEDULER', 'ConstantLR', 'PolynomialLR', 'WarmUpLR', 'InvLR' 6 | ] 7 | -------------------------------------------------------------------------------- /UDA_GVB/clsda/schedulers/builder.py: -------------------------------------------------------------------------------- 1 | # 2 | # ---------------------------------------------- 3 | import logging 4 | from copy import deepcopy 5 | from mmcv.utils import build_from_cfg 6 | from mmcv.utils import Registry 7 | from clsda.utils import get_root_logger 8 | 9 | SCHEDULER = Registry('scheduler') 10 | 11 | 12 | def build_scheduler(optimizer, scheduler_dict): 13 | temp_scheduler_dict = deepcopy(scheduler_dict) 14 | logger = get_root_logger() 15 | # 16 | if temp_scheduler_dict is None: 17 | logger.info('Using No LR Scheduling') 18 | temp_scheduler_dict = {'type': 'ConstantLR'} 19 | # 20 | s_type = temp_scheduler_dict['type'] 21 | logging.info('Using {} scheduler with {} params'.format(s_type, 22 | temp_scheduler_dict)) 23 | # 24 | temp_scheduler_dict['optimizer'] = optimizer 25 | return build_from_cfg(temp_scheduler_dict, SCHEDULER) 26 | -------------------------------------------------------------------------------- /UDA_GVB/clsda/trainers/__init__.py: -------------------------------------------------------------------------------- 1 | # 2 | # ---------------------------------------------- 3 | from .builder import TRAINER, VALIDATOR, build_validator, build_trainer 4 | from .trainer_gvb import TrainerGVB, ValidatorGVB 5 | 6 | __all__ = [ 7 | 'TRAINER', 'VALIDATOR', 'build_validator', 'build_trainer', 'TrainerGVB', 'ValidatorGVB', 8 | ] 9 | -------------------------------------------------------------------------------- /UDA_GVB/clsda/trainers/builder.py: -------------------------------------------------------------------------------- 1 | # 2 | # ---------------------------------------------- 3 | from mmcv.utils import Registry, build_from_cfg 4 | 5 | TRAINER = Registry('trainer') 6 | VALIDATOR = Registry('validator') 7 | 8 | 9 | def build_trainer(cfg, default_args=None): 10 | return build_from_cfg(cfg, TRAINER, default_args=default_args) 11 | 12 | 13 | def build_validator(cfg, default_args=None): 14 | return build_from_cfg(cfg, VALIDATOR, default_args=default_args) 15 | -------------------------------------------------------------------------------- /UDA_GVB/clsda/utils/__init__.py: -------------------------------------------------------------------------------- 1 | # 2 | # ---------------------------------------------- 3 | from clsda.utils.utils import * 4 | from clsda.utils.metrics import * 5 | from .logger import get_root_logger 6 | from .writer import get_root_writer 7 | -------------------------------------------------------------------------------- /UDA_GVB/clsda/utils/logger.py: -------------------------------------------------------------------------------- 1 | # 2 | # ---------------------------------------------- 3 | import logging 4 | from mmcv.utils import get_logger 5 | from mmcv.utils.logging import logger_initialized 6 | 7 | def get_root_logger(log_file=None, log_level=logging.INFO): 8 | if log_file is None: 9 | assert 'DAExp' in logger_initialized, 'logger not initialized' 10 | return get_logger('DAExp', log_file, log_level) 11 | 12 | -------------------------------------------------------------------------------- /UDA_GVB/clsda/utils/writer.py: -------------------------------------------------------------------------------- 1 | # 2 | # ---------------------------------------------- 3 | import torch.utils.tensorboard as tb 4 | from mmcv.runner import get_dist_info 5 | 6 | ROOT_TB_WRITER = [] 7 | 8 | 9 | def get_root_writer(log_dir=None): 10 | rank, _ = get_dist_info() 11 | if rank == 0: 12 | if log_dir is None: 13 | if len(ROOT_TB_WRITER) == 0: 14 | raise RuntimeError('You should initialize the tensorboard writer first') 15 | else: 16 | return ROOT_TB_WRITER[0] 17 | elif log_dir is not None: 18 | if len(ROOT_TB_WRITER) == 0: 19 | ROOT_TB_WRITER.append(tb.SummaryWriter(log_dir=log_dir)) 20 | return ROOT_TB_WRITER[0] 21 | else: 22 | raise RuntimeError('You have initialized the tensorboard writer before') 23 | else: 24 | return None 25 | -------------------------------------------------------------------------------- /UDA_GVB/configs/_base_/cls_models/resnet_50_gvb.py: -------------------------------------------------------------------------------- 1 | backbone_optimizer = dict( 2 | type='SGD', 3 | lr=0.001, 4 | weight_decay=0.001, 5 | momentum=0.9, 6 | nesterov=True, 7 | ) 8 | 9 | backbone = dict( 10 | type='GVBResNetFc', 11 | resnet_name='ResNet50', 12 | class_num=65, 13 | optimizer=backbone_optimizer, 14 | ) 15 | 16 | discriminator_optimizer = dict( 17 | type='SGD', 18 | lr=0.001, 19 | weight_decay=0.001, 20 | momentum=0.9, 21 | nesterov=True, 22 | ) 23 | 24 | discriminator = dict( 25 | type='GVBAdversarialNetwork', 26 | in_feature=65, 27 | hidden_size=1024, 28 | optimizer=discriminator_optimizer, 29 | ) 30 | 31 | scheduler = dict( 32 | type='InvLR', 33 | gamma=0.0001, 34 | power=0.75, 35 | ) 36 | 37 | models = dict( 38 | base_model=backbone, 39 | discriminator=discriminator, 40 | lr_scheduler=scheduler, 41 | ) 42 | -------------------------------------------------------------------------------- /UDA_GVB/configs/_base_/cls_models/resnet_50_gvb_visda.py: -------------------------------------------------------------------------------- 1 | backbone_optimizer = dict( 2 | type='SGD', 3 | lr=0.0003, 4 | weight_decay=0.001, 5 | momentum=0.9, 6 | nesterov=True, 7 | ) 8 | 9 | backbone = dict( 10 | type='GVBResNetFc', 11 | resnet_name='ResNet50', 12 | class_num=12, 13 | optimizer=backbone_optimizer, 14 | ) 15 | 16 | discriminator_optimizer = dict( 17 | type='SGD', 18 | lr=0.001, 19 | weight_decay=0.0003, 20 | momentum=0.9, 21 | nesterov=True, 22 | ) 23 | 24 | discriminator = dict( 25 | type='GVBAdversarialNetwork', 26 | in_feature=12, 27 | hidden_size=1024, 28 | optimizer=discriminator_optimizer, 29 | ) 30 | 31 | scheduler = dict( 32 | type='InvLR', 33 | gamma=0.0001, 34 | power=0.75, 35 | ) 36 | 37 | models = dict( 38 | base_model=backbone, 39 | discriminator=discriminator, 40 | lr_scheduler=scheduler, 41 | ) 42 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_C.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_C.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_C_fixmatch.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_C_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_C_fixmatch_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_C_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | using_NCE=True, 23 | NCE_weight=0.05, 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_C_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_C.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | ) 24 | 25 | test = dict( 26 | ) 27 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_C_nce_all_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_C_all_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_C_nce_fixmatch.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_C.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | using_fixmatch=True, 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_C_nce_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_C_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.02, 23 | 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_C_test.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_C.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 10 7 | val_interval = 50 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=200, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=False, 22 | using_fixmatch=False, 23 | ) 24 | 25 | test = dict( 26 | ) 27 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_C_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_C_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_P_fixmatch.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_P_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_P_fixmatch_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_P_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | using_NCE=True, 23 | NCE_weight=0.05, 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_P_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_C.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | ) 24 | 25 | test = dict( 26 | ) 27 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_P_nce_all_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_P_all_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_P_nce_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_P_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.02, 23 | 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_P_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_P_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_R_fixmatch.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_R_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_R_fixmatch_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_R_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | using_NCE=True, 23 | NCE_weight=0.05, 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_R_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_R.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | ) 24 | 25 | test = dict( 26 | ) 27 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_R_nce_all_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_R_all_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_R_nce_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_R_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.02, 23 | 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_A_R_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_A_R_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_C_A_fixmatch.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_C_A_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_C_A_fixmatch_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_C_A_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | using_NCE=True, 23 | NCE_weight=0.05, 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_C_A_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_C_A.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | ) 24 | 25 | test = dict( 26 | ) 27 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_C_A_nce_all_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_C_A_all_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_C_A_nce_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_C_A_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.02, 23 | 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_C_A_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_C_A_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_C_P_fixmatch.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_C_P_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_C_P_fixmatch_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_C_P_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | using_NCE=True, 23 | NCE_weight=0.05, 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_C_P_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_C_P.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | ) 24 | 25 | test = dict( 26 | ) 27 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_C_P_nce_all_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_C_P_all_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_C_P_nce_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_C_P_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.02, 23 | 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_C_P_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_C_P_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_C_R_fixmatch.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_C_R_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_C_R_fixmatch_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_C_R_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | using_NCE=True, 23 | NCE_weight=0.05, 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_C_R_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_C_R.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | ) 24 | 25 | test = dict( 26 | ) 27 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_C_R_nce_all_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_C_R_all_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_C_R_nce_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_C_R_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.02, 23 | 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_C_R_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_C_R_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_P_A_fixmatch.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_P_A_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_P_A_fixmatch_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_P_A_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | using_NCE=True, 23 | NCE_weight=0.05, 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_P_A_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_P_A.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | ) 24 | 25 | test = dict( 26 | ) 27 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_P_A_nce_all_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_P_A_all_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_P_A_nce_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_P_A_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.02, 23 | ) 24 | 25 | test = dict( 26 | ) 27 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_P_A_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_P_A_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_P_C_fixmatch.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_P_C_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_P_C_fixmatch_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_P_C_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | using_NCE=True, 23 | NCE_weight=0.05, 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_P_C_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_P_C.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | ) 24 | 25 | test = dict( 26 | ) 27 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_P_C_nce_all_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_P_C_all_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_P_C_nce_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_P_C_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.02, 23 | ) 24 | 25 | test = dict( 26 | ) 27 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_P_C_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_P_C_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_P_R_fixmatch.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_P_R_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_P_R_fixmatch_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_P_R_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | using_NCE=True, 23 | NCE_weight=0.05, 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_P_R_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_P_R.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | ) 24 | 25 | test = dict( 26 | ) 27 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_P_R_nce_all_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_P_R_all_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_P_R_nce_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_P_R_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.02, 23 | ) 24 | 25 | test = dict( 26 | ) 27 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_P_R_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_P_R_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_R_A_fixmatch.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_R_A_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_R_A_fixmatch_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_R_A_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | using_NCE=True, 23 | NCE_weight=0.05, 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_R_A_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_R_A.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | ) 24 | 25 | test = dict( 26 | ) 27 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_R_A_nce_all_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_R_A_all_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_R_A_nce_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_R_A_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.02, 23 | ) 24 | 25 | test = dict( 26 | ) 27 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_R_A_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_R_A_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_R_C_fixmatch.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_R_C_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_R_C_fixmatch_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_R_C_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | using_NCE=True, 23 | NCE_weight=0.05, 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_R_C_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_R_C.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | ) 24 | 25 | test = dict( 26 | ) 27 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_R_C_nce_all_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_R_C_all_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_R_C_nce_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_R_C_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.02, 23 | ) 24 | 25 | test = dict( 26 | ) 27 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_R_C_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_R_C_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_R_P_fixmatch.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_R_P_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_R_P_fixmatch_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_R_P_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_fixmatch=True, 22 | using_NCE=True, 23 | NCE_weight=0.05, 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_R_P_nce.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_R_P.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | ) 24 | 25 | test = dict( 26 | ) 27 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_R_P_nce_all_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_R_P_all_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.05, 23 | 24 | ) 25 | 26 | test = dict( 27 | ) 28 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_R_P_nce_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_R_P_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | using_NCE=True, 22 | NCE_weight=0.02, 23 | ) 24 | 25 | test = dict( 26 | ) 27 | -------------------------------------------------------------------------------- /UDA_GVB/configs/gvb/gvb_officehome_R_P_weak.py: -------------------------------------------------------------------------------- 1 | _base_ = [ 2 | '../_base_/cls_datasets/uda_officehome/uda_officehome_R_P_weak.py', 3 | '../_base_/cls_models/resnet_50_gvb.py' 4 | ] 5 | 6 | log_interval = 100 7 | val_interval = 500 8 | 9 | control = dict( 10 | log_interval=log_interval, 11 | max_iters=10000, 12 | val_interval=val_interval, 13 | cudnn_deterministic=True, 14 | save_interval=1000, 15 | max_save_num=1, 16 | seed=2, 17 | ) 18 | 19 | train = dict( 20 | log_interval=log_interval, 21 | 22 | ) 23 | 24 | test = dict( 25 | ) 26 | -------------------------------------------------------------------------------- /UDA_GVB/experiments/scripts/generate_random_port.py: -------------------------------------------------------------------------------- 1 | # 2 | # ---------------------------------------------- 3 | import random 4 | 5 | run_id = random.randint(15000, 20000) 6 | print(run_id) 7 | -------------------------------------------------------------------------------- /UDA_GVB/experiments/scripts/get_visible_card_num.py: -------------------------------------------------------------------------------- 1 | import os 2 | import torch 3 | 4 | 5 | def get_visible_card_num(): 6 | if 'CUDA_VISIBLE_DEVICES' in os.environ: 7 | cards = os.environ.get('CUDA_VISIBLE_DEVICES') 8 | cards_num = cards.count(',') + 1 9 | assert cards_num <= torch.cuda.device_count(), 'CUDA_VISIBLE_DEVICES should be lower than real number' 10 | return cards_num 11 | else: 12 | return torch.cuda.device_count() 13 | 14 | 15 | if __name__ == "__main__": 16 | print(get_visible_card_num()) 17 | -------------------------------------------------------------------------------- /UDA_GVB/experiments/scripts/uda_gvb_train_release.sh: -------------------------------------------------------------------------------- 1 | #!/usr/bin/env bash 2 | job_id=$1 3 | config_file=$2 4 | 5 | 6 | project_home='ghome/your_name/DA/UDA_test/' 7 | export HOME='/ghome/your_name/DA/UDA_test' 8 | export HOME=${HOME} 9 | echo 'HOME is '${HOME} 10 | cd ${HOME} || exit 11 | 12 | trainer_class=gvb 13 | validator_class=gvb 14 | scripts_path=$HOME'/experiments/scripts/get_visible_card_num.py' 15 | GPUS=$(python ${scripts_path}) 16 | PORT=16614 17 | 18 | python_file=./train.py 19 | # TODO: removing CUDA_LAUNCH_BLOCKING=1 20 | python -m torch.distributed.launch --nproc_per_node=$GPUS --master_port=$PORT \ 21 | ${python_file} --task_type cls --job_id ${job_id} --config ${config_file} \ 22 | --trainer ${trainer_class} --validator ${validator_class} 23 | -------------------------------------------------------------------------------- /UDA_GVB/training_logs/OfficeHome_AC/events.out.tfevents.1628751971.38927820282d.159.0: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/UDA_GVB/training_logs/OfficeHome_AC/events.out.tfevents.1628751971.38927820282d.159.0 -------------------------------------------------------------------------------- /UDA_GVB/training_logs/OfficeHome_AP/events.out.tfevents.1628752143.0a5bab1dda68.154.0: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/UDA_GVB/training_logs/OfficeHome_AP/events.out.tfevents.1628752143.0a5bab1dda68.154.0 -------------------------------------------------------------------------------- /UDA_GVB/training_logs/OfficeHome_AR/events.out.tfevents.1628752173.33d1f1e1bf5a.159.0: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/UDA_GVB/training_logs/OfficeHome_AR/events.out.tfevents.1628752173.33d1f1e1bf5a.159.0 -------------------------------------------------------------------------------- /UDA_GVB/training_logs/OfficeHome_CA/events.out.tfevents.1628752855.9bfb4147b9cf.159.0: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/UDA_GVB/training_logs/OfficeHome_CA/events.out.tfevents.1628752855.9bfb4147b9cf.159.0 -------------------------------------------------------------------------------- /UDA_GVB/training_logs/OfficeHome_CP/events.out.tfevents.1628752285.66c55db0326d.159.0: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/UDA_GVB/training_logs/OfficeHome_CP/events.out.tfevents.1628752285.66c55db0326d.159.0 -------------------------------------------------------------------------------- /UDA_GVB/training_logs/OfficeHome_CR/events.out.tfevents.1628752499.791e1b1299e1.154.0: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/UDA_GVB/training_logs/OfficeHome_CR/events.out.tfevents.1628752499.791e1b1299e1.154.0 -------------------------------------------------------------------------------- /UDA_GVB/training_logs/OfficeHome_PA/events.out.tfevents.1628752416.b8545536f29f.203.0: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/UDA_GVB/training_logs/OfficeHome_PA/events.out.tfevents.1628752416.b8545536f29f.203.0 -------------------------------------------------------------------------------- /UDA_GVB/training_logs/OfficeHome_PC/events.out.tfevents.1628752363.ab72c3b0dc3a.154.0: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/UDA_GVB/training_logs/OfficeHome_PC/events.out.tfevents.1628752363.ab72c3b0dc3a.154.0 -------------------------------------------------------------------------------- /UDA_GVB/training_logs/OfficeHome_PR/events.out.tfevents.1628752417.719def797de8.202.0: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/UDA_GVB/training_logs/OfficeHome_PR/events.out.tfevents.1628752417.719def797de8.202.0 -------------------------------------------------------------------------------- /UDA_GVB/training_logs/OfficeHome_RA/events.out.tfevents.1628752597.cddd8f788339.154.0: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/UDA_GVB/training_logs/OfficeHome_RA/events.out.tfevents.1628752597.cddd8f788339.154.0 -------------------------------------------------------------------------------- /UDA_GVB/training_logs/OfficeHome_RC/events.out.tfevents.1628748264.89588d5e003f.154.0: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/UDA_GVB/training_logs/OfficeHome_RC/events.out.tfevents.1628748264.89588d5e003f.154.0 -------------------------------------------------------------------------------- /UDA_GVB/training_logs/OfficeHome_RP/events.out.tfevents.1628752442.a76f7378e503.159.0: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ljjcoder/Probabilistic-Contrastive-Learning/0273938ebc3b26d3c84cb064da5507dad5dc955e/UDA_GVB/training_logs/OfficeHome_RP/events.out.tfevents.1628752442.a76f7378e503.159.0 -------------------------------------------------------------------------------- /UDA_GVB/training_logs/ReadMe.md: -------------------------------------------------------------------------------- 1 | ### Declaim 2 | 3 | tensorboard --logdir ./training_logs/OfficeHome_AC 4 | ##########!!!!!!!!!!!!!!!!!!NOTICE!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!#################### 5 | Note that 'best_acc' is not the final result. See the final result please refer to 'pred_acc_officehome_Clipart_unlabeled_target' 6 | 7 | --------------------------------------------------------------------------------