├── images ├── cifar_gt.png ├── cifar_ig.png ├── mnist_gt.png ├── mnist_ig.png ├── cifar_gt │ ├── 1.png │ ├── 2.png │ ├── 3.png │ ├── 4.png │ ├── 5.png │ └── 6.png ├── cifar_ig │ ├── 1.png │ ├── 2.png │ ├── 3.png │ ├── 4.png │ ├── 5.png │ └── 6.png ├── cifar_ours.png ├── mnist_gt │ ├── 1.png │ ├── 2.png │ ├── 3.png │ ├── 4.png │ ├── 5.png │ └── 6.png ├── mnist_ig │ ├── 1.png │ ├── 2.png │ ├── 3.png │ ├── 4.png │ ├── 5.png │ └── 6.png ├── mnist_ours.png ├── cifar_ours │ ├── 1.png │ ├── 2.png │ ├── 3.png │ ├── 4.png │ ├── 5.png │ └── 6.png └── mnist_ours │ ├── 1.png │ ├── 2.png │ ├── 3.png │ ├── 4.png │ ├── 5.png │ └── 6.png ├── logs └── Experiment3 │ ├── MNIST_ig │ ├── gt_0.png │ ├── gt_1.png │ ├── gt_2.png │ ├── gt_3.png │ ├── gt_4.png │ ├── gt_5.png │ ├── gt_6.png │ ├── gt_7.png │ ├── gt_8.png │ ├── gt_9.png │ ├── gt_10.png │ ├── gt_11.png │ ├── gt_12.png │ ├── gt_13.png │ ├── gt_14.png │ ├── gt_15.png │ ├── gt_16.png │ ├── gt_17.png │ ├── gt_18.png │ ├── gt_19.png │ ├── gt_20.png │ ├── gt_21.png │ ├── gt_22.png │ ├── gt_23.png │ ├── gt_24.png │ ├── gt_25.png │ ├── gt_26.png │ ├── gt_27.png │ ├── gt_28.png │ ├── gt_29.png │ ├── gt_30.png │ ├── gt_31.png │ ├── gt_32.png │ ├── gt_33.png │ ├── gt_34.png │ ├── gt_35.png │ ├── gt_36.png │ ├── gt_37.png │ ├── gt_38.png │ ├── gt_39.png │ ├── gt_40.png │ ├── gt_41.png │ ├── gt_42.png │ ├── gt_43.png │ ├── gt_44.png │ ├── gt_45.png │ ├── gt_46.png │ ├── gt_47.png │ ├── gt_48.png │ ├── gt_49.png │ ├── rec_0.png │ ├── rec_1.png │ ├── rec_10.png │ ├── rec_11.png │ ├── rec_12.png │ ├── rec_13.png │ ├── rec_14.png │ ├── rec_15.png │ ├── rec_16.png │ ├── rec_17.png │ ├── rec_18.png │ ├── rec_19.png │ ├── rec_2.png │ ├── rec_20.png │ ├── rec_21.png │ ├── rec_22.png │ ├── rec_23.png │ ├── rec_24.png │ ├── rec_25.png │ ├── rec_26.png │ ├── rec_27.png │ ├── rec_28.png │ ├── rec_29.png │ ├── rec_3.png │ ├── rec_30.png │ ├── rec_31.png │ ├── rec_32.png │ ├── rec_33.png │ ├── rec_34.png │ ├── rec_35.png │ ├── rec_36.png │ ├── rec_37.png │ ├── rec_38.png │ ├── rec_39.png │ ├── rec_4.png │ ├── rec_40.png │ ├── rec_41.png │ ├── rec_42.png │ ├── rec_43.png │ ├── rec_44.png │ ├── rec_45.png │ ├── rec_46.png │ ├── rec_47.png │ ├── rec_48.png │ ├── rec_49.png │ ├── rec_5.png │ ├── rec_6.png │ ├── rec_7.png │ ├── rec_8.png │ ├── rec_9.png │ ├── help.sh │ └── exp1.log │ ├── MNIST_ours │ ├── gt_0.png │ ├── gt_1.png │ ├── gt_10.png │ ├── gt_11.png │ ├── gt_12.png │ ├── gt_13.png │ ├── gt_14.png │ ├── gt_15.png │ ├── gt_16.png │ ├── gt_17.png │ ├── gt_18.png │ ├── gt_19.png │ ├── gt_2.png │ ├── gt_20.png │ ├── gt_21.png │ ├── gt_22.png │ ├── gt_23.png │ ├── gt_24.png │ ├── gt_25.png │ ├── gt_26.png │ ├── gt_27.png │ ├── gt_28.png │ ├── gt_29.png │ ├── gt_3.png │ ├── gt_30.png │ ├── gt_31.png │ ├── gt_32.png │ ├── gt_33.png │ ├── gt_34.png │ ├── gt_35.png │ ├── gt_36.png │ ├── gt_37.png │ ├── gt_38.png │ ├── gt_39.png │ ├── gt_4.png │ ├── gt_40.png │ ├── gt_41.png │ ├── gt_42.png │ ├── gt_43.png │ ├── gt_44.png │ ├── gt_45.png │ ├── gt_46.png │ ├── gt_47.png │ ├── gt_48.png │ ├── gt_49.png │ ├── gt_5.png │ ├── gt_6.png │ ├── gt_7.png │ ├── gt_8.png │ ├── gt_9.png │ ├── rec_0.png │ ├── rec_1.png │ ├── rec_2.png │ ├── rec_3.png │ ├── rec_4.png │ ├── rec_5.png │ ├── rec_6.png │ ├── rec_7.png │ ├── rec_8.png │ ├── rec_9.png │ ├── rec_10.png │ ├── rec_11.png │ ├── rec_12.png │ ├── rec_13.png │ ├── rec_14.png │ ├── rec_15.png │ ├── rec_16.png │ ├── rec_17.png │ ├── rec_18.png │ ├── rec_19.png │ ├── rec_20.png │ ├── rec_21.png │ ├── rec_22.png │ ├── rec_23.png │ ├── rec_24.png │ ├── rec_25.png │ ├── rec_26.png │ ├── rec_27.png │ ├── rec_28.png │ ├── rec_29.png │ ├── rec_30.png │ ├── rec_31.png │ ├── rec_32.png │ ├── rec_33.png │ ├── rec_34.png │ ├── rec_35.png │ ├── rec_36.png │ ├── rec_37.png │ ├── rec_38.png │ ├── rec_39.png │ ├── rec_40.png │ ├── rec_41.png │ ├── rec_42.png │ ├── rec_43.png │ ├── rec_44.png │ ├── rec_45.png │ ├── rec_46.png │ ├── rec_47.png │ ├── rec_48.png │ ├── rec_49.png │ ├── help.sh │ └── exp1.log │ ├── cifar100_ig │ ├── gt_0.png │ ├── gt_1.png │ ├── gt_2.png │ ├── gt_3.png │ ├── gt_4.png │ ├── gt_5.png │ ├── gt_6.png │ ├── gt_7.png │ ├── gt_8.png │ ├── gt_9.png │ ├── gt_10.png │ ├── gt_11.png │ ├── gt_12.png │ ├── gt_13.png │ ├── gt_14.png │ ├── gt_15.png │ ├── rec_0.png │ ├── rec_1.png │ ├── rec_10.png │ ├── rec_11.png │ ├── rec_12.png │ ├── rec_13.png │ ├── rec_14.png │ ├── rec_15.png │ ├── rec_2.png │ ├── rec_3.png │ ├── rec_4.png │ ├── rec_5.png │ ├── rec_6.png │ ├── rec_7.png │ ├── rec_8.png │ ├── rec_9.png │ ├── help.sh │ └── exp1.log │ └── cifar100_ours │ ├── gt_0.png │ ├── gt_1.png │ ├── gt_2.png │ ├── gt_3.png │ ├── gt_4.png │ ├── gt_5.png │ ├── gt_6.png │ ├── gt_7.png │ ├── gt_8.png │ ├── gt_9.png │ ├── gt_10.png │ ├── gt_11.png │ ├── gt_12.png │ ├── gt_13.png │ ├── gt_14.png │ ├── gt_15.png │ ├── rec_0.png │ ├── rec_1.png │ ├── rec_10.png │ ├── rec_11.png │ ├── rec_12.png │ ├── rec_13.png │ ├── rec_14.png │ ├── rec_15.png │ ├── rec_2.png │ ├── rec_3.png │ ├── rec_4.png │ ├── rec_5.png │ ├── rec_6.png │ ├── rec_7.png │ ├── rec_8.png │ ├── rec_9.png │ ├── help.sh │ └── exp1.log ├── README_md_files └── f67ab870-3e10-11ed-9ab9-579d6cfc3a57.jpeg ├── .idea ├── .gitignore ├── inspectionProfiles │ ├── profiles_settings.xml │ └── Project_Default.xml ├── modules.xml ├── source_code.iml └── deployment.xml ├── requirements.txt ├── consts.py ├── medianfilt.py ├── cal_metrics.py ├── defense.py ├── loss.py ├── metrics.py ├── scheduler.py ├── methods.py ├── options.py └── new_logs ├── ExtremeDistribution ├── exp1.log ├── exp2.log ├── exp3.log └── exp4.log └── TrainingStage └── MNIST ├── exp5.log ├── exp6.log ├── exp4.log ├── exp1.log ├── exp2.log └── exp3.log /images/cifar_gt.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_gt.png -------------------------------------------------------------------------------- /images/cifar_ig.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_ig.png -------------------------------------------------------------------------------- /images/mnist_gt.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_gt.png -------------------------------------------------------------------------------- /images/mnist_ig.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_ig.png -------------------------------------------------------------------------------- /images/cifar_gt/1.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_gt/1.png -------------------------------------------------------------------------------- /images/cifar_gt/2.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_gt/2.png -------------------------------------------------------------------------------- /images/cifar_gt/3.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_gt/3.png -------------------------------------------------------------------------------- /images/cifar_gt/4.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_gt/4.png -------------------------------------------------------------------------------- /images/cifar_gt/5.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_gt/5.png -------------------------------------------------------------------------------- /images/cifar_gt/6.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_gt/6.png -------------------------------------------------------------------------------- /images/cifar_ig/1.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_ig/1.png -------------------------------------------------------------------------------- /images/cifar_ig/2.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_ig/2.png -------------------------------------------------------------------------------- /images/cifar_ig/3.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_ig/3.png -------------------------------------------------------------------------------- /images/cifar_ig/4.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_ig/4.png -------------------------------------------------------------------------------- /images/cifar_ig/5.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_ig/5.png -------------------------------------------------------------------------------- /images/cifar_ig/6.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_ig/6.png -------------------------------------------------------------------------------- /images/cifar_ours.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_ours.png -------------------------------------------------------------------------------- /images/mnist_gt/1.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_gt/1.png -------------------------------------------------------------------------------- /images/mnist_gt/2.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_gt/2.png -------------------------------------------------------------------------------- /images/mnist_gt/3.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_gt/3.png -------------------------------------------------------------------------------- /images/mnist_gt/4.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_gt/4.png -------------------------------------------------------------------------------- /images/mnist_gt/5.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_gt/5.png -------------------------------------------------------------------------------- /images/mnist_gt/6.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_gt/6.png -------------------------------------------------------------------------------- /images/mnist_ig/1.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_ig/1.png -------------------------------------------------------------------------------- /images/mnist_ig/2.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_ig/2.png -------------------------------------------------------------------------------- /images/mnist_ig/3.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_ig/3.png -------------------------------------------------------------------------------- /images/mnist_ig/4.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_ig/4.png -------------------------------------------------------------------------------- /images/mnist_ig/5.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_ig/5.png -------------------------------------------------------------------------------- /images/mnist_ig/6.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_ig/6.png -------------------------------------------------------------------------------- /images/mnist_ours.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_ours.png -------------------------------------------------------------------------------- /images/cifar_ours/1.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_ours/1.png -------------------------------------------------------------------------------- /images/cifar_ours/2.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_ours/2.png -------------------------------------------------------------------------------- /images/cifar_ours/3.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_ours/3.png -------------------------------------------------------------------------------- /images/cifar_ours/4.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_ours/4.png -------------------------------------------------------------------------------- /images/cifar_ours/5.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_ours/5.png -------------------------------------------------------------------------------- /images/cifar_ours/6.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/cifar_ours/6.png -------------------------------------------------------------------------------- /images/mnist_ours/1.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_ours/1.png -------------------------------------------------------------------------------- /images/mnist_ours/2.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_ours/2.png -------------------------------------------------------------------------------- /images/mnist_ours/3.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_ours/3.png -------------------------------------------------------------------------------- /images/mnist_ours/4.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_ours/4.png -------------------------------------------------------------------------------- /images/mnist_ours/5.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_ours/5.png -------------------------------------------------------------------------------- /images/mnist_ours/6.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/images/mnist_ours/6.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_0.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_0.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_1.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_1.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_2.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_2.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_3.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_3.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_4.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_4.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_5.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_5.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_6.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_6.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_7.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_7.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_8.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_8.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_9.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_9.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_10.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_10.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_11.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_11.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_12.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_12.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_13.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_13.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_14.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_14.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_15.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_15.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_16.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_16.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_17.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_17.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_18.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_18.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_19.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_19.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_20.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_20.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_21.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_21.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_22.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_22.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_23.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_23.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_24.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_24.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_25.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_25.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_26.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_26.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_27.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_27.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_28.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_28.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_29.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_29.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_30.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_30.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_31.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_31.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_32.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_32.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_33.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_33.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_34.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_34.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_35.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_35.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_36.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_36.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_37.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_37.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_38.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_38.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_39.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_39.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_40.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_40.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_41.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_41.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_42.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_42.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_43.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_43.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_44.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_44.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_45.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_45.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_46.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_46.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_47.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_47.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_48.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_48.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/gt_49.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/gt_49.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_0.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_0.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_1.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_1.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_10.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_10.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_11.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_11.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_12.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_12.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_13.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_13.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_14.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_14.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_15.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_15.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_16.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_16.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_17.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_17.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_18.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_18.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_19.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_19.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_2.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_2.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_20.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_20.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_21.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_21.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_22.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_22.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_23.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_23.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_24.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_24.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_25.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_25.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_26.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_26.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_27.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_27.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_28.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_28.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_29.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_29.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_3.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_3.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_30.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_30.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_31.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_31.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_32.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_32.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_33.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_33.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_34.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_34.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_35.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_35.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_36.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_36.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_37.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_37.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_38.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_38.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_39.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_39.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_4.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_4.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_40.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_40.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_41.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_41.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_42.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_42.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_43.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_43.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_44.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_44.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_45.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_45.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_46.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_46.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_47.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_47.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_48.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_48.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_49.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_49.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_5.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_5.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_6.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_6.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_7.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_7.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_8.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_8.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/rec_9.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ig/rec_9.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_0.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_0.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_1.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_1.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_10.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_10.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_11.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_11.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_12.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_12.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_13.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_13.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_14.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_14.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_15.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_15.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_16.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_16.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_17.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_17.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_18.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_18.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_19.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_19.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_2.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_2.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_20.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_20.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_21.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_21.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_22.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_22.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_23.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_23.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_24.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_24.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_25.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_25.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_26.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_26.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_27.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_27.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_28.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_28.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_29.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_29.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_3.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_3.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_30.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_30.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_31.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_31.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_32.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_32.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_33.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_33.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_34.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_34.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_35.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_35.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_36.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_36.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_37.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_37.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_38.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_38.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_39.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_39.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_4.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_4.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_40.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_40.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_41.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_41.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_42.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_42.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_43.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_43.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_44.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_44.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_45.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_45.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_46.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_46.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_47.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_47.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_48.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_48.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_49.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_49.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_5.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_5.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_6.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_6.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_7.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_7.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_8.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_8.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/gt_9.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/gt_9.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_0.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_0.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_1.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_1.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_2.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_2.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_3.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_3.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_4.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_4.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_5.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_5.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_6.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_6.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_7.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_7.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_8.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_8.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_9.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_9.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/gt_0.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/gt_0.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/gt_1.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/gt_1.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/gt_2.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/gt_2.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/gt_3.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/gt_3.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/gt_4.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/gt_4.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/gt_5.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/gt_5.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/gt_6.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/gt_6.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/gt_7.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/gt_7.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/gt_8.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/gt_8.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/gt_9.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/gt_9.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_10.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_10.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_11.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_11.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_12.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_12.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_13.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_13.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_14.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_14.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_15.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_15.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_16.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_16.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_17.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_17.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_18.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_18.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_19.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_19.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_20.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_20.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_21.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_21.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_22.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_22.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_23.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_23.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_24.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_24.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_25.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_25.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_26.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_26.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_27.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_27.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_28.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_28.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_29.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_29.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_30.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_30.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_31.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_31.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_32.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_32.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_33.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_33.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_34.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_34.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_35.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_35.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_36.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_36.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_37.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_37.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_38.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_38.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_39.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_39.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_40.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_40.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_41.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_41.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_42.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_42.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_43.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_43.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_44.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_44.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_45.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_45.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_46.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_46.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_47.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_47.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_48.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_48.png -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/rec_49.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/MNIST_ours/rec_49.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/gt_10.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/gt_10.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/gt_11.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/gt_11.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/gt_12.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/gt_12.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/gt_13.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/gt_13.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/gt_14.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/gt_14.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/gt_15.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/gt_15.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/rec_0.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/rec_0.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/rec_1.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/rec_1.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/rec_10.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/rec_10.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/rec_11.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/rec_11.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/rec_12.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/rec_12.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/rec_13.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/rec_13.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/rec_14.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/rec_14.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/rec_15.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/rec_15.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/rec_2.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/rec_2.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/rec_3.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/rec_3.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/rec_4.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/rec_4.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/rec_5.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/rec_5.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/rec_6.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/rec_6.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/rec_7.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/rec_7.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/rec_8.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/rec_8.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/rec_9.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ig/rec_9.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/gt_0.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/gt_0.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/gt_1.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/gt_1.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/gt_2.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/gt_2.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/gt_3.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/gt_3.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/gt_4.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/gt_4.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/gt_5.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/gt_5.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/gt_6.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/gt_6.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/gt_7.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/gt_7.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/gt_8.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/gt_8.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/gt_9.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/gt_9.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/gt_10.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/gt_10.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/gt_11.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/gt_11.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/gt_12.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/gt_12.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/gt_13.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/gt_13.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/gt_14.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/gt_14.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/gt_15.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/gt_15.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/rec_0.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/rec_0.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/rec_1.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/rec_1.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/rec_10.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/rec_10.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/rec_11.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/rec_11.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/rec_12.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/rec_12.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/rec_13.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/rec_13.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/rec_14.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/rec_14.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/rec_15.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/rec_15.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/rec_2.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/rec_2.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/rec_3.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/rec_3.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/rec_4.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/rec_4.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/rec_5.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/rec_5.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/rec_6.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/rec_6.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/rec_7.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/rec_7.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/rec_8.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/rec_8.png -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/rec_9.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/logs/Experiment3/cifar100_ours/rec_9.png -------------------------------------------------------------------------------- /README_md_files/f67ab870-3e10-11ed-9ab9-579d6cfc3a57.jpeg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/mankasto/iLRG/HEAD/README_md_files/f67ab870-3e10-11ed-9ab9-579d6cfc3a57.jpeg -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/help.sh: -------------------------------------------------------------------------------- 1 | python3 main.py --model dnn --num_images 50 --dataset MNIST_GRAY --num_classes 10 --num_tries 1 --rec_img --restarts 3 --signed --boxed --tv 1e-4 --save_image 2 | 3 | -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/help.sh: -------------------------------------------------------------------------------- 1 | python3 main.py --model dnn --num_images 50 --dataset MNIST_GRAY --num_classes 10 --num_tries 1 --rec_img --restarts 3 --signed --boxed --tv 1e-4 --save_image --fix_labels 2 | 3 | -------------------------------------------------------------------------------- /.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 | -------------------------------------------------------------------------------- /.idea/inspectionProfiles/profiles_settings.xml: -------------------------------------------------------------------------------- 1 | 2 | 3 | 6 | -------------------------------------------------------------------------------- /requirements.txt: -------------------------------------------------------------------------------- 1 | cvxopt==1.3.0 2 | lpips==0.1.4 3 | numpy~=1.22.3 4 | opencv_python==4.5.5.64 5 | scikit_learn==1.1.2 6 | torch~=1.7.0+cu110 7 | torchvision~=0.8.1+cu110 8 | 9 | sklearn~=0.0 10 | scikit-learn~=1.1.1 11 | scipy~=1.8.0 12 | opencv-python~=4.5.5.64 -------------------------------------------------------------------------------- /.idea/modules.xml: -------------------------------------------------------------------------------- 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/help.sh: -------------------------------------------------------------------------------- 1 | python3 main.py --model resnet18 --num_images 16 --dataset CIFAR100 --num_classes 100 --num_tries 1 --rec_img --restarts 1 --signed --boxed --tv 1e-4 --save_image --max_iterations 24000 --tv 8e-3 --fix_labels --seed 6666 --distribution random2 --num_target_cls 12 --exp_name cifar100_12cls 2 | 3 | -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/help.sh: -------------------------------------------------------------------------------- 1 | python3 main.py --model resnet18 --num_images 16 --dataset CIFAR100 --num_classes 100 --num_tries 1 --rec_img --restarts 1 --signed --boxed --tv 1e-4 --save_image --max_iterations 24000 --tv 8e-3 --fix_labels --seed 8888 --distribution random2 --num_target_cls 10 --exp_name cifar100_10cls 2 | 3 | -------------------------------------------------------------------------------- /.idea/source_code.iml: -------------------------------------------------------------------------------- 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 12 | -------------------------------------------------------------------------------- /consts.py: -------------------------------------------------------------------------------- 1 | """Setup constants.""" 2 | cifar10_mean = [0.4914672374725342, 0.4822617471218109, 0.4467701315879822] 3 | cifar10_std = [0.24703224003314972, 0.24348513782024384, 0.26158785820007324] 4 | cifar100_mean = [0.5071598291397095, 0.4866936206817627, 0.44120192527770996] 5 | cifar100_std = [0.2673342823982239, 0.2564384639263153, 0.2761504650115967] 6 | mnist_mean = (0.13066373765468597,) 7 | mnist_std = (0.30810782313346863,) 8 | imagenet_mean = [0.485, 0.456, 0.406] 9 | imagenet_std = [0.229, 0.224, 0.225] 10 | -------------------------------------------------------------------------------- /.idea/deployment.xml: -------------------------------------------------------------------------------- 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ig/exp1.log: -------------------------------------------------------------------------------- 1 | running experiments at 2022-09-04-20:56:54 2 | log_dir: exp_res 3 | cpu: False 4 | num_tries: 1 5 | num_images: 50 6 | seed: 12 7 | alpha: 1 8 | simplified: False 9 | compare: False 10 | analysis: False 11 | ablation1: False 12 | ablation2: False 13 | ablation3: False 14 | ablation4: False 15 | data_path: data 16 | dataset: MNIST_GRAY 17 | split: train 18 | distribution: random 19 | start_id: 0 20 | num_classes: 10 21 | num_uniform_cls: 32 22 | model: dnn 23 | trained_model: False 24 | iter_train: False 25 | epochs: 10 26 | iters: 1000 27 | batch_size: 128 28 | optimizer: SGD 29 | lr: 0.1 30 | scheduler: linear 31 | weight_decay: 0.0005 32 | warmup: False 33 | epoch_interval: 10 34 | iter_interval: 100 35 | mid_save: False 36 | model_path: models 37 | dryrun: False 38 | batchnorm: False 39 | dropout: False 40 | silu: False 41 | dim: 300 42 | n_hidden: 1 43 | defense: False 44 | defense_method: dp 45 | noise_std: 0.001 46 | clip_bound: 4 47 | sparse_ratio: 10 48 | prune_ratio: 10 49 | rec_img: True 50 | fix_labels: False 51 | gt_labels: False 52 | optim: ig 53 | restarts: 3 54 | cost_fn: sim 55 | indices: def 56 | weights: equal 57 | rec_lr: None 58 | rec_optimizer: adam 59 | signed: True 60 | boxed: True 61 | scoring_choice: loss 62 | init: randn 63 | tv: 0.0001 64 | l2: 1e-06 65 | max_iterations: 8000 66 | loss_thresh: 0.0001 67 | save_image: True 68 | image_path: images 69 | table_path: tables 70 | 71 | Start Experiment 1 72 | *************************************************************** 73 | Ground-truth Labels: 0,1,2,3,4,5,6,7,8,9 74 | Ground-truth Num of Instances: 7,2,2,3,6,9,1,4,7,9 75 | Our Recovered Labels: 0,1,2,3,4,5,6,7,8,9 | LeAcc: 1.000 76 | Our Recovered Num of Instances: 7,2,2,3,6,9,1,4,7,9 | LnAcc: 1.000 | IRec: 1.000 77 | 78 | Start reconstructing images 79 | End reconstructing images 80 | --------------------------------------------------------------- 81 | Mean Ours LeAcc: 1.000 | Mean Ours LnAcc: 1.000 | Mean Ours IRec: 1.000 82 | -------------------------------------------------------------------------------- /logs/Experiment3/MNIST_ours/exp1.log: -------------------------------------------------------------------------------- 1 | running experiments at 2022-09-04-21:00:57 2 | log_dir: exp_res 3 | cpu: False 4 | num_tries: 1 5 | num_images: 50 6 | seed: 12 7 | alpha: 1 8 | simplified: False 9 | compare: False 10 | analysis: False 11 | ablation1: False 12 | ablation2: False 13 | ablation3: False 14 | ablation4: False 15 | data_path: data 16 | dataset: MNIST_GRAY 17 | split: train 18 | distribution: random 19 | start_id: 0 20 | num_classes: 10 21 | num_uniform_cls: 32 22 | model: dnn 23 | trained_model: False 24 | iter_train: False 25 | epochs: 10 26 | iters: 1000 27 | batch_size: 128 28 | optimizer: SGD 29 | lr: 0.1 30 | scheduler: linear 31 | weight_decay: 0.0005 32 | warmup: False 33 | epoch_interval: 10 34 | iter_interval: 100 35 | mid_save: False 36 | model_path: models 37 | dryrun: False 38 | batchnorm: False 39 | dropout: False 40 | silu: False 41 | dim: 300 42 | n_hidden: 1 43 | defense: False 44 | defense_method: dp 45 | noise_std: 0.001 46 | clip_bound: 4 47 | sparse_ratio: 10 48 | prune_ratio: 10 49 | rec_img: True 50 | fix_labels: True 51 | gt_labels: False 52 | optim: ig 53 | restarts: 3 54 | cost_fn: sim 55 | indices: def 56 | weights: equal 57 | rec_lr: None 58 | rec_optimizer: adam 59 | signed: True 60 | boxed: True 61 | scoring_choice: loss 62 | init: randn 63 | tv: 0.0001 64 | l2: 1e-06 65 | max_iterations: 8000 66 | loss_thresh: 0.0001 67 | save_image: True 68 | image_path: images 69 | table_path: tables 70 | 71 | Start Experiment 1 72 | *************************************************************** 73 | Ground-truth Labels: 0,1,2,3,4,5,6,7,8,9 74 | Ground-truth Num of Instances: 7,2,2,3,6,9,1,4,7,9 75 | Our Recovered Labels: 0,1,2,3,4,5,6,7,8,9 | LeAcc: 1.000 76 | Our Recovered Num of Instances: 7,2,2,3,6,9,1,4,7,9 | LnAcc: 1.000 | IRec: 1.000 77 | 78 | Start reconstructing images 79 | End reconstructing images 80 | --------------------------------------------------------------- 81 | Mean Ours LeAcc: 1.000 | Mean Ours LnAcc: 1.000 | Mean Ours IRec: 1.000 82 | -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ours/exp1.log: -------------------------------------------------------------------------------- 1 | running experiments at 2022-09-18-23:15:59 2 | exp_name: cifar100_10cls 3 | cpu: False 4 | num_tries: 1 5 | num_images: 16 6 | seed: 8888 7 | alpha: 1 8 | simplified: False 9 | compare: False 10 | analysis: False 11 | data_path: data 12 | dataset: CIFAR100 13 | split: train 14 | distribution: random2 15 | start_id: 0 16 | num_classes: 100 17 | num_uniform_cls: 32 18 | num_target_cls: 10 19 | max_size: 32 20 | model: resnet18 21 | trained_model: False 22 | iter_train: False 23 | iters: 1000 24 | epochs: 10 25 | batch_size: 128 26 | lr: 0.1 27 | optimizer: SGD 28 | scheduler: linear 29 | weight_decay: 0.0005 30 | warmup: False 31 | epoch_interval: 10 32 | iter_interval: 100 33 | mid_save: False 34 | model_path: models 35 | dryrun: False 36 | batchnorm: False 37 | dropout: False 38 | silu: False 39 | leaky_relu: False 40 | n_dim: 300 41 | n_hidden: 1 42 | defense: False 43 | defense_method: dp 44 | noise_std: 0.001 45 | clip_bound: 4 46 | sparse_ratio: 10 47 | prune_ratio: 10 48 | rec_img: True 49 | fix_labels: True 50 | gt_labels: False 51 | optim: ig 52 | restarts: 1 53 | cost_fn: sim 54 | indices: def 55 | weights: equal 56 | rec_lr: None 57 | rec_optimizer: adam 58 | signed: True 59 | boxed: True 60 | scoring_choice: loss 61 | init: randn 62 | tv: 0.008 63 | l2: 1e-06 64 | max_iterations: 24000 65 | loss_thresh: 0.0001 66 | save_image: True 67 | log_dir: logs/cifar100_10cls 68 | image_dir: images/cifar100_10cls 69 | 70 | Start Experiment 1 71 | start_id: 0 72 | *************************************************************** 73 | Ground-truth Labels: 8,9,11,16,22,34,45,53,57,86 74 | Ground-truth Num of Instances: 4,2,1,1,1,1,1,3,1,1 75 | Our Recovered Labels: 8,9,11,16,22,34,45,53,57,86 | LeAcc: 1.000 76 | Our Recovered Num of Instances: 4,2,1,1,1,1,1,3,1,1 | LnAcc: 1.000 | IRec: 1.000 77 | 78 | Start reconstructing images 79 | End reconstructing images 80 | --------------------------------------------------------------- 81 | Mean Ours LeAcc: 1.000 | Mean Ours LnAcc: 1.000 | Mean Ours IRec: 1.000 82 | -------------------------------------------------------------------------------- /logs/Experiment3/cifar100_ig/exp1.log: -------------------------------------------------------------------------------- 1 | running experiments at 2022-09-19-09:08:59 2 | exp_name: cifar100_10cls_ig 3 | cpu: False 4 | num_tries: 1 5 | num_images: 16 6 | seed: 8888 7 | alpha: 1 8 | simplified: False 9 | compare: False 10 | analysis: False 11 | data_path: data 12 | dataset: CIFAR100 13 | split: train 14 | distribution: random2 15 | start_id: 0 16 | num_classes: 100 17 | num_uniform_cls: 32 18 | num_target_cls: 10 19 | max_size: 32 20 | model: resnet18 21 | trained_model: False 22 | iter_train: False 23 | iters: 1000 24 | epochs: 10 25 | batch_size: 128 26 | lr: 0.1 27 | optimizer: SGD 28 | scheduler: linear 29 | weight_decay: 0.0005 30 | warmup: False 31 | epoch_interval: 10 32 | iter_interval: 100 33 | mid_save: False 34 | model_path: models 35 | dryrun: False 36 | batchnorm: False 37 | dropout: False 38 | silu: False 39 | leaky_relu: False 40 | n_dim: 300 41 | n_hidden: 1 42 | defense: False 43 | defense_method: dp 44 | noise_std: 0.001 45 | clip_bound: 4 46 | sparse_ratio: 10 47 | prune_ratio: 10 48 | rec_img: True 49 | fix_labels: False 50 | gt_labels: False 51 | optim: ig 52 | restarts: 1 53 | cost_fn: sim 54 | indices: def 55 | weights: equal 56 | rec_lr: None 57 | rec_optimizer: adam 58 | signed: True 59 | boxed: True 60 | scoring_choice: loss 61 | init: randn 62 | tv: 0.008 63 | l2: 1e-06 64 | max_iterations: 24000 65 | loss_thresh: 0.0001 66 | save_image: True 67 | log_dir: logs/cifar100_10cls_ig 68 | image_dir: images/cifar100_10cls_ig 69 | 70 | Start Experiment 1 71 | start_id: 0 72 | *************************************************************** 73 | Ground-truth Labels: 8,9,11,16,22,34,45,53,57,86 74 | Ground-truth Num of Instances: 4,2,1,1,1,1,1,3,1,1 75 | Our Recovered Labels: 8,9,11,16,22,34,45,53,57,86 | LeAcc: 1.000 76 | Our Recovered Num of Instances: 4,2,1,1,1,1,1,3,1,1 | LnAcc: 1.000 | IRec: 1.000 77 | 78 | Start reconstructing images 79 | End reconstructing images 80 | --------------------------------------------------------------- 81 | Mean Ours LeAcc: 1.000 | Mean Ours LnAcc: 1.000 | Mean Ours IRec: 1.000 82 | -------------------------------------------------------------------------------- /medianfilt.py: -------------------------------------------------------------------------------- 1 | """This is code for median pooling from https://gist.github.com/rwightman. 2 | 3 | https://gist.github.com/rwightman/f2d3849281624be7c0f11c85c87c1598 4 | """ 5 | import torch.nn as nn 6 | import torch.nn.functional as F 7 | from torch.nn.modules.utils import _pair, _quadruple 8 | 9 | 10 | class MedianPool2d(nn.Module): 11 | """Median pool (usable as median filter when stride=1) module. 12 | 13 | Args: 14 | kernel_size: size of pooling kernel, int or 2-tuple 15 | stride: pool stride, int or 2-tuple 16 | padding: pool padding, int or 4-tuple (l, r, t, b) as in pytorch F.pad 17 | same: override padding and enforce same padding, boolean 18 | """ 19 | 20 | def __init__(self, kernel_size=3, stride=1, padding=0, same=True): 21 | """Initialize with kernel_size, stride, padding.""" 22 | super().__init__() 23 | self.k = _pair(kernel_size) 24 | self.stride = _pair(stride) 25 | self.padding = _quadruple(padding) # convert to l, r, t, b 26 | self.same = same 27 | 28 | def _padding(self, x): 29 | if self.same: 30 | ih, iw = x.size()[2:] 31 | if ih % self.stride[0] == 0: 32 | ph = max(self.k[0] - self.stride[0], 0) 33 | else: 34 | ph = max(self.k[0] - (ih % self.stride[0]), 0) 35 | if iw % self.stride[1] == 0: 36 | pw = max(self.k[1] - self.stride[1], 0) 37 | else: 38 | pw = max(self.k[1] - (iw % self.stride[1]), 0) 39 | pl = pw // 2 40 | pr = pw - pl 41 | pt = ph // 2 42 | pb = ph - pt 43 | padding = (pl, pr, pt, pb) 44 | else: 45 | padding = self.padding 46 | return padding 47 | 48 | def forward(self, x): 49 | # using existing pytorch functions and tensor ops so that we get autograd, 50 | # would likely be more efficient to implement from scratch at C/Cuda level 51 | x = F.pad(x, self._padding(x), mode='reflect') 52 | x = x.unfold(2, self.k[0], self.stride[0]).unfold(3, self.k[1], self.stride[1]) 53 | x = x.contiguous().view(x.size()[:4] + (-1,)).median(dim=-1)[0] 54 | return x 55 | -------------------------------------------------------------------------------- /cal_metrics.py: -------------------------------------------------------------------------------- 1 | import os 2 | import cv2 3 | import lpips 4 | import torch 5 | import numpy as np 6 | 7 | 8 | def get_images(images_list, img_size=28): 9 | images = [] 10 | for image_path in images_list: 11 | image = cv2.imread(image_path) 12 | image = cv2.resize(image, (img_size, img_size)) 13 | image = image.astype(np.float32) / 255 14 | image = torch.from_numpy(image) 15 | image = image.permute(2, 0, 1) 16 | images.append(image) 17 | images = torch.stack(images).to(device) 18 | return images 19 | 20 | 21 | def psnr(img_batch, ref_batch, batched=False, factor=1.0): 22 | """Standard PSNR.""" 23 | 24 | def get_psnr(img_in, img_ref): 25 | mse = ((img_in - img_ref) ** 2).mean() 26 | if mse > 0 and torch.isfinite(mse): 27 | return 10 * torch.log10(factor ** 2 / mse) 28 | elif not torch.isfinite(mse): 29 | return img_batch.new_tensor(float('nan')) 30 | else: 31 | return img_batch.new_tensor(float('inf')) 32 | 33 | if batched: 34 | psnr = get_psnr(img_batch.detach(), ref_batch) 35 | else: 36 | [B, C, m, n] = img_batch.shape 37 | psnrs = [] 38 | for sample in range(B): 39 | psnrs.append(get_psnr(img_batch.detach()[sample, :, :, :], ref_batch[sample, :, :, :])) 40 | psnr = torch.stack(psnrs, dim=0).mean() 41 | 42 | return psnr.item() 43 | 44 | 45 | if __name__ == '__main__': 46 | device = 'cuda:0' 47 | dataset = 'mnist' 48 | lpips_loss = lpips.LPIPS(net='vgg', spatial=False).to(device) 49 | x = [os.path.join(f'images/{dataset}_gt', _) for _ in os.listdir(f'images/{dataset}_gt')] 50 | y = [os.path.join(f'images/{dataset}_ig', _) for _ in os.listdir(f'images/{dataset}_ig')] 51 | z = [os.path.join(f'images/{dataset}_ours', _) for _ in os.listdir(f'images/{dataset}_ours')] 52 | x = get_images(x, 28 if dataset == 'mnist' else 32) 53 | y = get_images(y, 28 if dataset == 'mnist' else 32) 54 | z = get_images(z, 28 if dataset == 'mnist' else 32) 55 | with torch.no_grad(): 56 | lpips_score = lpips_loss(y, x).squeeze().mean() 57 | lpips_score2 = lpips_loss(z, x).squeeze().mean() 58 | psnr1 = psnr(y, x) 59 | psnr2 = psnr(z, x) 60 | print('LPIPS: IG-%.2f; Ours-%.2f' % (lpips_score.item(), lpips_score2.item())) 61 | print('PSNR: IG-%.2f; Ours-%.2f' % (psnr1, psnr2)) 62 | -------------------------------------------------------------------------------- /defense.py: -------------------------------------------------------------------------------- 1 | """ 2 | Defense methods. code from https://github.com/zhuohangli/GGL 3 | Including: 4 | - Additive noise 5 | - Gradient clipping 6 | - Gradient compression 7 | - Representation perturbation 8 | """ 9 | import numpy as np 10 | import torch 11 | 12 | params = {'dp': 'noise_std', 'clip': 'clip_bound', 'sparse': 'sparse_ratio', 'perturb': 'prune_rate'} 13 | 14 | 15 | def additive_noise(input_gradient, std=0.1): 16 | """ 17 | Additive noise mechanism for differential privacy 18 | """ 19 | gradient = [grad + torch.normal(torch.zeros_like(grad), std * torch.ones_like(grad)) for grad in input_gradient] 20 | return gradient 21 | 22 | 23 | def gradient_clipping(input_gradient, bound=4): 24 | """ 25 | Gradient clipping (clip by norm) 26 | """ 27 | max_norm = float(bound) 28 | norm_type = 2.0 # np.inf 29 | device = input_gradient[0].device 30 | 31 | if norm_type == np.inf: 32 | norms = [g.abs().max().to(device) for g in input_gradient] 33 | total_norm = norms[0] if len(norms) == 1 else torch.max(torch.stack(norms)) 34 | else: 35 | total_norm = torch.norm(torch.stack([torch.norm(g, norm_type).to(device) for g in input_gradient]), norm_type) 36 | clip_coef = max_norm / (total_norm + 1e-6) 37 | clip_coef_clamped = torch.clamp(clip_coef, max=1.0) 38 | 39 | gradient = [g.mul_(clip_coef_clamped.to(device)) for g in input_gradient] 40 | return gradient 41 | 42 | 43 | def gradient_compression(input_gradient, percentage=10): 44 | """ 45 | Prune by percentage 46 | """ 47 | device = input_gradient[0].device 48 | gradient = [None] * len(input_gradient) 49 | for i in range(len(input_gradient)): 50 | grad_tensor = input_gradient[i].clone().cpu().numpy() 51 | flattened_weights = np.abs(grad_tensor.flatten()) 52 | thresh = np.percentile(flattened_weights, percentage) 53 | grad_tensor = np.where(abs(grad_tensor) < thresh, 0, grad_tensor) 54 | gradient[i] = torch.Tensor(grad_tensor).to(device) 55 | return gradient 56 | 57 | 58 | def perturb_representation(input_gradient, model, ground_truth, pruning_rate=10): 59 | """ 60 | Defense proposed in the Soteria paper. 61 | param: 62 | - input_gradient: the input_gradient 63 | - model: the ResNet-18 model 64 | - ground_truth: the benign image (for learning perturbed representation) 65 | - pruning_rate: the prune percentage 66 | Note: This implementation only works for ResNet-18 67 | """ 68 | device = input_gradient[0].device 69 | 70 | gt_data = ground_truth.clone() 71 | gt_data.requires_grad = True 72 | 73 | out, feature_graph = model(gt_data) 74 | 75 | deviation_target = torch.zeros_like(feature_graph) 76 | deviation_x_norm = torch.zeros_like(feature_graph) 77 | for f in range(deviation_x_norm.size(1)): 78 | deviation_target[:, f] = 1 79 | feature_graph.backward(deviation_target, retain_graph=True) 80 | deviation_f1_x = gt_data.grad.data 81 | deviation_x_norm[:, f] = torch.norm(deviation_f1_x.view(deviation_f1_x.size(0), -1), dim=1) / ( 82 | (feature_graph.data[:, f]) + 1e-10) 83 | model.zero_grad() 84 | gt_data.grad.data.zero_() 85 | deviation_target[:, f] = 0 86 | 87 | # prune r_i corresponding to smallest ||dr_i/dX||/||r_i|| 88 | deviation_x_norm_sum = deviation_x_norm.sum(axis=0) 89 | thresh = np.percentile(deviation_x_norm_sum.flatten().cpu().numpy(), pruning_rate) 90 | mask = np.where(abs(deviation_x_norm_sum.cpu()) < thresh, 0, 1).astype(np.float32) 91 | 92 | print('Soteria mask: ', sum(mask)) 93 | 94 | gradient = [grad for grad in input_gradient] 95 | # apply mask 96 | gradient[-2] = gradient[-2] * torch.Tensor(mask).to(device) 97 | return gradient 98 | -------------------------------------------------------------------------------- /loss.py: -------------------------------------------------------------------------------- 1 | """Define various loss functions and bundle them with appropriate metrics.""" 2 | 3 | import torch 4 | import numpy as np 5 | 6 | 7 | class Loss: 8 | """Abstract class, containing necessary methods. 9 | 10 | Abstract class to collect information about the 'higher-level' loss function, used to train an energy-based model 11 | containing the evaluation of the loss function, its gradients w.r.t. to first and second argument and evaluations 12 | of the actual metric that is targeted. 13 | 14 | """ 15 | 16 | def __init__(self): 17 | """Init.""" 18 | pass 19 | 20 | def __call__(self, reference, argmin): 21 | """Return l(x, y).""" 22 | raise NotImplementedError() 23 | return value, name, format 24 | 25 | def metric(self, reference, argmin): 26 | """The actually sought metric.""" 27 | raise NotImplementedError() 28 | return value, name, format 29 | 30 | 31 | class PSNR(Loss): 32 | """A classical MSE target. 33 | 34 | The minimized criterion is MSE Loss, the actual metric is average PSNR. 35 | """ 36 | 37 | def __init__(self): 38 | """Init with torch MSE.""" 39 | self.loss_fn = torch.nn.MSELoss(size_average=None, reduce=None, reduction='mean') 40 | 41 | def __call__(self, x=None, y=None): 42 | """Return l(x, y).""" 43 | name = 'MSE' 44 | pf_fmt = '.6f' 45 | if x is None: 46 | return name, pf_fmt 47 | else: 48 | value = 0.5 * self.loss_fn(x, y) 49 | return value, name, pf_fmt 50 | 51 | def metric(self, x=None, y=None): 52 | """The actually sought metric.""" 53 | name = 'avg PSNR' 54 | pf_fmt = '.3f' 55 | if x is None: 56 | return name, pf_fmt 57 | else: 58 | value = self.psnr_compute(x, y) 59 | return value, name, pf_fmt 60 | 61 | @staticmethod 62 | def psnr_compute(img_batch, ref_batch, batched=False, factor=1.0): 63 | """Standard PSNR.""" 64 | 65 | def get_psnr(img_in, img_ref): 66 | mse = ((img_in - img_ref) ** 2).mean() 67 | if mse > 0 and torch.isfinite(mse): 68 | return (10 * torch.log10(factor ** 2 / mse)).item() 69 | elif not torch.isfinite(mse): 70 | return float('nan') 71 | else: 72 | return float('inf') 73 | 74 | if batched: 75 | psnr = get_psnr(img_batch.detach(), ref_batch) 76 | else: 77 | [B, C, m, n] = img_batch.shape 78 | psnrs = [] 79 | for sample in range(B): 80 | psnrs.append(get_psnr(img_batch.detach()[sample, :, :, :], ref_batch[sample, :, :, :])) 81 | psnr = np.mean(psnrs) 82 | 83 | return psnr 84 | 85 | 86 | class Classification(Loss): 87 | """A classical NLL loss for classification. Evaluation has the softmax baked in. 88 | 89 | The minimized criterion is cross entropy, the actual metric is total accuracy. 90 | """ 91 | 92 | def __init__(self): 93 | """Init with torch MSE.""" 94 | self.loss_fn = torch.nn.CrossEntropyLoss(weight=None, size_average=None, ignore_index=-100, 95 | reduce=None, reduction='mean') 96 | 97 | def __call__(self, x=None, y=None): 98 | """Return l(x, y).""" 99 | name = 'CrossEntropy' 100 | pf_fmt = '1.5f' 101 | if x is None: 102 | return name, pf_fmt 103 | else: 104 | value = self.loss_fn(x, y) 105 | return value, name, pf_fmt 106 | 107 | def metric(self, x=None, y=None): 108 | """The actually sought metric.""" 109 | name = 'Accuracy' 110 | pf_fmt = '6.2%' 111 | if x is None: 112 | return name, pf_fmt 113 | else: 114 | value = (x.data.argmax(dim=1) == y).sum().float() / y.shape[0] 115 | return value.detach(), name, pf_fmt 116 | -------------------------------------------------------------------------------- /metrics.py: -------------------------------------------------------------------------------- 1 | """This is code based on https://sudomake.ai/inception-score-explained/.""" 2 | import torch 3 | import torchvision 4 | 5 | from collections import defaultdict 6 | 7 | 8 | class InceptionScore(torch.nn.Module): 9 | """Class that manages and returns the inception score of images.""" 10 | 11 | def __init__(self, batch_size=32, setup=dict(device=torch.device('cpu'), dtype=torch.float)): 12 | """Initialize with setup and target inception batch size.""" 13 | super().__init__() 14 | self.preprocessing = torch.nn.Upsample(size=(299, 299), mode='bilinear', align_corners=False) 15 | self.model = torchvision.models.inception_v3(pretrained=True).to(**setup) 16 | self.model.eval() 17 | self.batch_size = batch_size 18 | 19 | def forward(self, image_batch): 20 | """Image batch should have dimensions BCHW and should be normalized. 21 | 22 | B should be divisible by self.batch_size. 23 | """ 24 | B, C, H, W = image_batch.shape 25 | batches = B // self.batch_size 26 | scores = [] 27 | for batch in range(batches): 28 | input = self.preprocessing(image_batch[batch * self.batch_size: (batch + 1) * self.batch_size]) 29 | scores.append(self.model(input)) 30 | prob_yx = torch.nn.functional.softmax(torch.cat(scores, 0), dim=1) 31 | entropy = torch.where(prob_yx > 0, -prob_yx * prob_yx.log(), torch.zeros_like(prob_yx)) 32 | return entropy.sum() 33 | 34 | 35 | def psnr(img_batch, ref_batch, batched=False, factor=1.0): 36 | """Standard PSNR.""" 37 | 38 | def get_psnr(img_in, img_ref): 39 | mse = ((img_in - img_ref) ** 2).mean() 40 | if mse > 0 and torch.isfinite(mse): 41 | return 10 * torch.log10(factor ** 2 / mse) 42 | elif not torch.isfinite(mse): 43 | return img_batch.new_tensor(float('nan')) 44 | else: 45 | return img_batch.new_tensor(float('inf')) 46 | 47 | if batched: 48 | psnr = get_psnr(img_batch.detach(), ref_batch) 49 | else: 50 | [B, C, m, n] = img_batch.shape 51 | psnrs = [] 52 | for sample in range(B): 53 | psnrs.append(get_psnr(img_batch.detach()[sample, :, :, :], ref_batch[sample, :, :, :])) 54 | psnr = torch.stack(psnrs, dim=0).mean() 55 | 56 | return psnr.item() 57 | 58 | 59 | def total_variation(x): 60 | """Anisotropic TV.""" 61 | dx = torch.mean(torch.abs(x[:, :, :, :-1] - x[:, :, :, 1:])) 62 | dy = torch.mean(torch.abs(x[:, :, :-1, :] - x[:, :, 1:, :])) 63 | return dx + dy 64 | 65 | 66 | def group_consistency(x, group_x): 67 | mean_group_x = sum(group_x) / len(group_x) 68 | return torch.norm(x - mean_group_x, p=3) 69 | 70 | 71 | def activation_errors(model, x1, x2): 72 | """Compute activation-level error metrics for every module in the network.""" 73 | model.eval() 74 | 75 | device = next(model.parameters()).device 76 | 77 | hooks = [] 78 | data = defaultdict(dict) 79 | inputs = torch.cat((x1, x2), dim=0) 80 | separator = x1.shape[0] 81 | 82 | def check_activations(self, input, output): 83 | module_name = str(*[name for name, mod in model.named_modules() if self is mod]) 84 | try: 85 | layer_inputs = input[0].detach() 86 | residual = (layer_inputs[:separator] - layer_inputs[separator:]).pow(2) 87 | se_error = residual.sum() 88 | mse_error = residual.mean() 89 | sim = torch.nn.functional.cosine_similarity(layer_inputs[:separator].flatten(), 90 | layer_inputs[separator:].flatten(), 91 | dim=0, eps=1e-8).detach() 92 | data['se'][module_name] = se_error.item() 93 | data['mse'][module_name] = mse_error.item() 94 | data['sim'][module_name] = sim.item() 95 | except (KeyboardInterrupt, SystemExit): 96 | raise 97 | except AttributeError: 98 | pass 99 | 100 | for name, module in model.named_modules(): 101 | hooks.append(module.register_forward_hook(check_activations)) 102 | 103 | try: 104 | outputs = model(inputs.to(device)) 105 | for hook in hooks: 106 | hook.remove() 107 | except Exception as e: 108 | for hook in hooks: 109 | hook.remove() 110 | raise 111 | 112 | return data 113 | -------------------------------------------------------------------------------- /scheduler.py: -------------------------------------------------------------------------------- 1 | """This file is part of https://github.com/ildoonet/pytorch-gradual-warmup-lr. 2 | 3 | MIT License 4 | 5 | Copyright (c) 2019 Ildoo Kim 6 | 7 | Permission is hereby granted, free of charge, to any person obtaining a copy 8 | of this software and associated documentation files (the "Software"), to deal 9 | in the Software without restriction, including without limitation the rights 10 | to use, copy, modify, merge, publish, distribute, sublicense, and/or sell 11 | copies of the Software, and to permit persons to whom the Software is 12 | furnished to do so, subject to the following conditions: 13 | 14 | The above copyright notice and this permission notice shall be included in all 15 | copies or substantial portions of the Software. 16 | 17 | THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR 18 | IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 19 | FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE 20 | AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER 21 | LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, 22 | OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE 23 | SOFTWARE. 24 | 25 | """ 26 | from torch.optim.lr_scheduler import _LRScheduler 27 | from torch.optim.lr_scheduler import ReduceLROnPlateau 28 | 29 | 30 | class GradualWarmupScheduler(_LRScheduler): 31 | """Gradually warm-up(increasing) learning rate in optimizer. 32 | 33 | Proposed in 'Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour'. 34 | 35 | Args: 36 | optimizer (Optimizer): Wrapped optimizer. 37 | multiplier: target learning rate = base lr * multiplier 38 | total_epoch: target learning rate is reached at total_epoch, gradually 39 | after_scheduler: after target_epoch, use this scheduler(eg. ReduceLROnPlateau) 40 | 41 | """ 42 | 43 | def __init__(self, optimizer, multiplier, total_epoch, after_scheduler=None): 44 | """Initialize the warm-up start. 45 | 46 | Usage: 47 | 48 | scheduler_normal = torch.optim.lr_scheduler.MultiStepLR(optimizer) 49 | scheduler_warmup = GradualWarmupScheduler(optimizer, multiplier=8, total_epoch=10, after_scheduler=scheduler_normal) 50 | """ 51 | self.multiplier = multiplier 52 | if self.multiplier < 1.: 53 | raise ValueError('multiplier should be greater thant or equal to 1.') 54 | self.total_epoch = total_epoch 55 | self.after_scheduler = after_scheduler 56 | self.finished = False 57 | super().__init__(optimizer) 58 | 59 | def get_lr(self): 60 | if self.last_epoch > self.total_epoch: 61 | if self.after_scheduler: 62 | if not self.finished: 63 | self.after_scheduler.base_lrs = [base_lr * self.multiplier for base_lr in self.base_lrs] 64 | self.finished = True 65 | return self.after_scheduler.get_lr() 66 | return [base_lr * self.multiplier for base_lr in self.base_lrs] 67 | 68 | return [base_lr * ((self.multiplier - 1.) * self.last_epoch / self.total_epoch + 1.) for base_lr in 69 | self.base_lrs] 70 | 71 | def step_ReduceLROnPlateau(self, metrics, epoch=None): 72 | if epoch is None: 73 | epoch = self.last_epoch + 1 74 | self.last_epoch = epoch if epoch != 0 else 1 # ReduceLROnPlateau is called at the end of epoch, whereas others are called at beginning 75 | if self.last_epoch <= self.total_epoch: 76 | warmup_lr = [base_lr * ((self.multiplier - 1.) * self.last_epoch / self.total_epoch + 1.) for base_lr in 77 | self.base_lrs] 78 | for param_group, lr in zip(self.optimizer.param_groups, warmup_lr): 79 | param_group['lr'] = lr 80 | else: 81 | if epoch is None: 82 | self.after_scheduler.step(metrics, None) 83 | else: 84 | self.after_scheduler.step(metrics, epoch - self.total_epoch) 85 | 86 | def step(self, epoch=None, metrics=None): 87 | if type(self.after_scheduler) != ReduceLROnPlateau: 88 | if self.finished and self.after_scheduler: 89 | if epoch is None: 90 | self.after_scheduler.step(None) 91 | else: 92 | self.after_scheduler.step(epoch - self.total_epoch) 93 | else: 94 | return super(GradualWarmupScheduler, self).step(epoch) 95 | else: 96 | self.step_ReduceLROnPlateau(metrics, epoch) 97 | -------------------------------------------------------------------------------- /.idea/inspectionProfiles/Project_Default.xml: -------------------------------------------------------------------------------- 1 | 2 | 3 | 99 | -------------------------------------------------------------------------------- /methods.py: -------------------------------------------------------------------------------- 1 | import torch 2 | import numpy as np 3 | 4 | 5 | def solve_lp(A, b, c): 6 | from cvxopt import matrix, solvers 7 | solvers.options['show_progress'] = False 8 | np.random.seed(None) 9 | for _ in range(1): 10 | A, b, c = matrix(A), matrix(b), matrix(c) 11 | sol = solvers.lp(c, A, b) 12 | x = sol['x'] 13 | if x is not None: 14 | ret = A * x 15 | if ret[0] < -0.1 and np.max(ret[1:]) < 1e-2 and np.count_nonzero(np.array(ret[1:]) <= 0) > 0.5 * len(ret): 16 | return True 17 | return False 18 | 19 | 20 | def solve_perceptron(X, y, fit_intercept=True, max_iter=1000, tol=1e-3, eta0=1.): 21 | from sklearn.linear_model import Perceptron 22 | clf = Perceptron(fit_intercept=fit_intercept, max_iter=max_iter, tol=tol, eta0=eta0) 23 | clf.fit(X, y) 24 | if not fit_intercept: 25 | pass 26 | if clf.score(X, y) > 0.9: 27 | return True 28 | return False 29 | 30 | 31 | def svd_infer(A, num_classes=1000, gt_k=None, epsilon=1e-8): 32 | m, n = np.shape(A) 33 | B, s, C = np.linalg.svd(A, full_matrices=False) 34 | pred_k = np.linalg.matrix_rank(A) 35 | k = min(gt_k, pred_k) 36 | C = C[:k, :].astype(np.double) 37 | # Find x: x @ C has only one positive element 38 | # Filter possible labels using perceptron algorithm 39 | bow = [] 40 | for i in range(n): 41 | if i in bow: 42 | continue 43 | indices = [j for j in range(n) if j != i] 44 | np.random.shuffle(indices) 45 | if solve_perceptron( 46 | X=np.concatenate([C[:, i:i + 1], C[:, indices[:num_classes - 1]]], 1).transpose(), 47 | y=np.array([1 if j == 0 else -1 for j in range(num_classes)]), 48 | fit_intercept=True, 49 | max_iter=1000, 50 | tol=1e-3 51 | ): 52 | bow.append(i) 53 | # Get the final set with linear programming 54 | ret_bow = [] 55 | for i in bow: 56 | if i in ret_bow: 57 | continue 58 | indices = [j for j in range(n) if j != i] 59 | D = np.concatenate([C[:, i:i + 1], C[:, indices]], 1) 60 | indices2 = np.argsort(np.linalg.norm(D[:, 1:], axis=0)) 61 | A = np.concatenate([D[:, 0:1], -D[:, 1 + indices2]], 1).transpose() 62 | if solve_lp( 63 | A=A, 64 | b=np.array([-epsilon] + [0] * len(indices2)), 65 | c=np.array(C[:, i:i + 1]) 66 | ): 67 | ret_bow.append(i) 68 | return ret_bow 69 | 70 | 71 | # Extended GradientInversion, w_grad: m * n. 72 | def gi_infer(w_grad): 73 | ret = (torch.min(w_grad, dim=0)[0] < 0).nonzero(as_tuple=False).squeeze().numpy() 74 | ret = [ret] if ret.shape == () else ret 75 | return ret 76 | 77 | 78 | # Extended iDLG 79 | def idlg_infer(w_grad): 80 | ret = (torch.sum(w_grad, dim=0) < 0).nonzero(as_tuple=False).squeeze().numpy() 81 | ret = [ret] if ret.shape == () else ret 82 | return ret 83 | 84 | 85 | # Recover embeddings 86 | def get_emb(grad_w, grad_b, exp_thre=10): 87 | # Split scientific count notation 88 | sc_grad_b = '%e' % grad_b 89 | sc_grad_w = ['%e' % w for w in grad_w] 90 | real_b, exp_b = float(sc_grad_b.split('e')[0]), int(sc_grad_b.split('e')[1]) 91 | real_w, exp_w = np.array([float(sc_w.split('e')[0]) for sc_w in sc_grad_w]), \ 92 | np.array([int(sc_w.split('e')[1]) for sc_w in sc_grad_w]) 93 | # Deal with 0 case 94 | if real_b == 0.: 95 | real_b = 1 96 | exp_b = -64 97 | # Deal with exponent value 98 | exp = exp_w - exp_b 99 | exp = np.where(exp > exp_thre, exp_thre, exp) 100 | exp = np.where(exp < -1 * exp_thre, -1 * exp_thre, exp) 101 | 102 | def get_exp(x): 103 | return 10 ** x if x >= 0 else 1. / 10 ** (-x) 104 | 105 | exp = np.array(list(map(get_exp, exp))) 106 | # Calculate recovered average embeddings for batch_i (samples of class i) 107 | res = (1. / real_b) * real_w * exp 108 | res = torch.from_numpy(res).to(torch.float32) 109 | return res 110 | 111 | 112 | # Recover Labels 113 | def iLRG(probs, grad_b, n_classes, n_images): 114 | # Solve linear equations to recover labels 115 | coefs, values = [], [] 116 | # Add the first equation: k1+k2+...+kc=K 117 | coefs.append([1 for _ in range(n_classes)]) 118 | values.append(n_images) 119 | # Add the following equations 120 | for i in range(n_classes): 121 | coef = [] 122 | for j in range(n_classes): 123 | if j != i: 124 | coef.append(probs[j][i].item()) 125 | else: 126 | coef.append(probs[j][i].item() - 1) 127 | coefs.append(coef) 128 | values.append(n_images * grad_b[i]) 129 | # Convert into numpy ndarray 130 | coefs = np.array(coefs) 131 | values = np.array(values) 132 | # Solve with Moore-Penrose pseudoinverse 133 | res_float = np.linalg.pinv(coefs).dot(values) 134 | # Filter negative values 135 | res = np.where(res_float > 0, res_float, 0) 136 | # Round values 137 | res = np.round(res).astype(int) 138 | res = np.where(res <= n_images, res, 0) 139 | err = res - res_float 140 | num_mod = np.sum(res) - n_images 141 | if num_mod > 0: 142 | inds = np.argsort(-err) 143 | mod_inds = inds[:num_mod] 144 | mod_res = res.copy() 145 | mod_res[mod_inds] -= 1 146 | elif num_mod < 0: 147 | inds = np.argsort(err) 148 | mod_inds = inds[:num_mod] 149 | mod_res = res.copy() 150 | mod_res[mod_inds] += 1 151 | else: 152 | mod_res = res 153 | 154 | return res, mod_res 155 | 156 | 157 | # Have Known about which labels exist 158 | def sim_iLRG(probs, grad_b, exist_labels, n_images): 159 | # Solve linear equations to recover labels 160 | coefs, values = [], [] 161 | # Add the first equation: k1+k2+...+kc=K 162 | coefs.append([1 for _ in range(len(exist_labels))]) 163 | values.append(n_images) 164 | # Add the following equations 165 | for i in exist_labels: 166 | coef = [] 167 | for j in exist_labels: 168 | if j != i: 169 | coef.append(probs[j][i].item()) 170 | else: 171 | coef.append(probs[j][i].item() - 1) 172 | coefs.append(coef) 173 | values.append(n_images * grad_b[i]) 174 | # Convert into numpy ndarray 175 | coefs = np.array(coefs) 176 | values = np.array(values) 177 | # Solve with Moore-Penrose pseudoinverse 178 | res_float = np.linalg.pinv(coefs).dot(values) 179 | # Filter negative values 180 | res = np.where(res_float > 0, res_float, 0) 181 | # Round values 182 | res = np.round(res).astype(int) 183 | res = np.where(res <= n_images, res, 0) 184 | err = res - res_float 185 | num_mod = np.sum(res) - n_images 186 | if num_mod > 0: 187 | inds = np.argsort(-err) 188 | mod_inds = inds[:num_mod] 189 | mod_res = res.copy() 190 | mod_res[mod_inds] -= 1 191 | elif num_mod < 0: 192 | inds = np.argsort(err) 193 | mod_inds = inds[:num_mod] 194 | mod_res = res.copy() 195 | mod_res[mod_inds] += 1 196 | else: 197 | mod_res = res 198 | 199 | return res, mod_res 200 | -------------------------------------------------------------------------------- /options.py: -------------------------------------------------------------------------------- 1 | """Parser options.""" 2 | 3 | import argparse 4 | 5 | 6 | def options(): 7 | """Construct the central argument parser, filled with useful defaults.""" 8 | parser = argparse.ArgumentParser(description='Instance-wise Batch Label Restoration and Image Reconstruction') 9 | 10 | # Basic settings 11 | parser.add_argument('--exp_name', default='Experiment', type=str) 12 | parser.add_argument('--cpu', action='store_true', help='Use cpu') 13 | parser.add_argument('--num_tries', default=50, type=int, help='Repetition times of an experiment') 14 | parser.add_argument('--num_images', default=24, type=int, 15 | help='How many images should be recovered from the given gradient / Restoration batchsize') 16 | parser.add_argument('--seed', default=12, type=int, help='Random seed') 17 | parser.add_argument('--alpha', default=1, type=float, help='Factor for scaling outputs') 18 | parser.add_argument('--simplified', action='store_true', 19 | help='Use simplified method, given class-wise label existences') 20 | parser.add_argument('--compare', action='store_true', 21 | help='Compare our method with others') 22 | parser.add_argument('--estimate', action='store_true', 23 | help='Use 1/n to estimate probabilities') 24 | parser.add_argument('--analysis', action='store_true', 25 | help='Error analysis about four approximations and recovered embeddings & probs, etc') 26 | parser.add_argument('--ratio', default=0.00, type=float, help='Filter ratio to compute mean values') 27 | 28 | # Data settings 29 | parser.add_argument('--data_path', default='data', type=str) 30 | parser.add_argument('--dataset', default='CIFAR100', type=str) 31 | parser.add_argument('--split', default='train', type=str, help='Part of splitted dataset, default train') 32 | parser.add_argument('--distribution', default='random', type=str, help='Data distribution of a training batch,' 33 | 'random, extreme, balanced, unique') 34 | parser.add_argument('--start_id', default=0, type=int, 35 | help='The beginning image id for collecting data with extreme and balanced distribution') 36 | parser.add_argument('--num_classes', default=100, type=int) 37 | parser.add_argument('--num_uniform_cls', default=32, type=int, 38 | help='Num of classes for collecting data with balanced distribution') 39 | parser.add_argument('--num_target_cls', default=5, type=int, 40 | help='Num of classes for collecting data with random2 distribution') 41 | parser.add_argument('--max_size', default=32, type=int, 42 | help='Max batch size for ImageNet') 43 | 44 | # Model settings 45 | parser.add_argument('--model', default='lenet5', type=str, help='model name.') 46 | parser.add_argument('--trained_model', action='store_true', help='Use a trained model.') 47 | ## Training settings 48 | parser.add_argument('--iter_train', action='store_true', 49 | help='Train model with iterations setting instead of epochs') 50 | parser.add_argument('--iters', default=1000, type=int, 51 | help='If using a trained model, how many iterations was it trained?') 52 | parser.add_argument('--epochs', default=10, type=int, 53 | help='If using a trained model, how many epochs was it trained?') 54 | parser.add_argument('--batch_size', default=128, type=int, 55 | help='Batchsize for training, so is validation') 56 | parser.add_argument('--lr', default=0.1, type=float, help='Recommend 0.001 for adam series and 0.1 for sgd') 57 | parser.add_argument('--optimizer', default='SGD', type=str, help='AdamW, SGD, linear') 58 | parser.add_argument('--scheduler', default='linear', type=str, help='linear') 59 | parser.add_argument('--weight_decay', default=5e-4, help='Usually 5e-4') 60 | parser.add_argument('--warmup', action='store_true', help='Use warmup scheduler') 61 | parser.add_argument('--epoch_interval', default=10, type=int, 62 | help='How many epochs to validate or save models') 63 | parser.add_argument('--iter_interval', default=100, type=int, 64 | help='How many iterations to validate or save models') 65 | parser.add_argument('--mid_save', action='store_true', help='Save middle trained models') 66 | parser.add_argument('--model_path', default='models', type=str) 67 | parser.add_argument('--dryrun', action='store_true', help='Run everything for just one step to test functionality') 68 | ## End training settings 69 | parser.add_argument('--batchnorm', action='store_true', help='Use batchnorm for lenet5 model') 70 | parser.add_argument('--dropout', action='store_true', help='Use dropout for vgg16 model') 71 | parser.add_argument('--silu', action='store_true', help='Use silu activation, may occur negative values') 72 | parser.add_argument('--leaky_relu', action='store_true', 73 | help='Use leaky relu activation, may occur negative values') 74 | parser.add_argument('--n_dim', default=300, type=int, 75 | help='Dimension of embedding (the input of classification layer)') 76 | parser.add_argument('--n_hidden', default=1, type=int, 77 | help='Num of hidden layers') 78 | 79 | # Defense settings 80 | parser.add_argument('--defense', action='store_true', help='Defense against the attack') 81 | parser.add_argument('--defense_method', default='dp', type=str, 82 | help='dp(additive noise) or clip or sparse or perturb(soteria)') 83 | parser.add_argument('--noise_std', default=0.001, type=float) 84 | parser.add_argument('--clip_bound', default=4, type=int) 85 | parser.add_argument('--sparse_ratio', default=10, type=int) 86 | parser.add_argument('--prune_ratio', default=10, type=int) 87 | 88 | # Rec images settings 89 | parser.add_argument('--rec_img', action='store_true', help='Reconstruct images based on our attack, here IG') 90 | parser.add_argument('--fix_labels', action='store_true', help='Fix labels') 91 | parser.add_argument('--gt_labels', action='store_true', help='Fix labels with the gt') 92 | parser.add_argument('--optim', default='ig', type=str, help='IG or DLG') 93 | parser.add_argument('--restarts', default=1, type=int, help='How many restarts to run') 94 | parser.add_argument('--cost_fn', default='sim', type=str, help='Choice of cost function') 95 | parser.add_argument('--indices', default='def', type=str, help='Choice of indices from the parameter list') 96 | parser.add_argument('--weights', default='equal', type=str, help='Weigh the parameter list differently') 97 | parser.add_argument('--rec_lr', default=None, type=float, help='Learning rate for reconstruction') 98 | parser.add_argument('--rec_optimizer', default='adam', type=str, help='Optimizer for reconstruction') 99 | parser.add_argument('--signed', action='store_true', help='Use signed gradients, recommend true') 100 | parser.add_argument('--boxed', action='store_true', help='Use box constraints, recommend true') 101 | parser.add_argument('--scoring_choice', default='loss', type=str, 102 | help='How to find the best image between all restarts') 103 | parser.add_argument('--init', default='randn', type=str, help='Choice of image initialization') 104 | parser.add_argument('--tv', default=1e-6, type=float, help='Weight of TV penalty') 105 | parser.add_argument('--l2', default=1e-6, type=float, help='Weight of l2 norm') 106 | parser.add_argument('--max_iterations', default=8000, type=int, help='Max iterations of reconstruction') 107 | parser.add_argument('--loss_thresh', default=1e-4, type=float, help='Loss threshold for early stopping') 108 | 109 | # Files and folders: 110 | parser.add_argument('--save_image', action='store_true', help='Save the output to a file.') 111 | # parser.add_argument('--image_dir', default='images/Experiment', type=str) 112 | return parser 113 | -------------------------------------------------------------------------------- /new_logs/ExtremeDistribution/exp1.log: -------------------------------------------------------------------------------- 1 | running experiments at 2022-11-13-15:29:16 2 | exp_name: extreme_distribution 3 | cpu: False 4 | num_tries: 20 5 | num_images: 24 6 | seed: 12 7 | alpha: 1 8 | simplified: False 9 | compare: False 10 | estimate: False 11 | analysis: False 12 | ratio: 0.0 13 | data_path: data 14 | dataset: CIFAR100 15 | split: train 16 | distribution: custom_imbalanced 17 | start_id: 0 18 | num_classes: 100 19 | num_uniform_cls: 32 20 | num_target_cls: 5 21 | max_size: 32 22 | model: vgg16 23 | trained_model: False 24 | iter_train: False 25 | iters: 1000 26 | epochs: 10 27 | batch_size: 128 28 | lr: 0.1 29 | optimizer: SGD 30 | scheduler: linear 31 | weight_decay: 0.0005 32 | warmup: False 33 | epoch_interval: 10 34 | iter_interval: 100 35 | mid_save: False 36 | model_path: models 37 | dryrun: False 38 | batchnorm: False 39 | dropout: False 40 | silu: False 41 | leaky_relu: False 42 | n_dim: 300 43 | n_hidden: 1 44 | defense: False 45 | defense_method: dp 46 | noise_std: 0.001 47 | clip_bound: 4 48 | sparse_ratio: 10 49 | prune_ratio: 10 50 | rec_img: False 51 | fix_labels: False 52 | gt_labels: False 53 | optim: ig 54 | restarts: 1 55 | cost_fn: sim 56 | indices: def 57 | weights: equal 58 | rec_lr: None 59 | rec_optimizer: adam 60 | signed: False 61 | boxed: False 62 | scoring_choice: loss 63 | init: randn 64 | tv: 1e-06 65 | l2: 1e-06 66 | max_iterations: 8000 67 | loss_thresh: 0.0001 68 | save_image: False 69 | log_dir: logs/extreme_distribution 70 | image_dir: images/extreme_distribution 71 | 72 | Start Experiment 1 73 | start_id: 0 74 | 0/24, Acc for this batch: 0.000 75 | *************************************************************** 76 | Ground-truth Labels: 0,18,92 77 | Ground-truth Num of Instances: 1,22,1 78 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 79 | Our Recovered Num of Instances: 1,22,1 | LnAcc: 1.000 | IRec: 1.000 80 | Start Experiment 2 81 | start_id: 2558 82 | 0/24, Acc for this batch: 0.000 83 | *************************************************************** 84 | Ground-truth Labels: 0,18,92 85 | Ground-truth Num of Instances: 1,22,1 86 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 87 | Our Recovered Num of Instances: 1,22,1 | LnAcc: 1.000 | IRec: 1.000 88 | Start Experiment 3 89 | start_id: 4787 90 | 0/24, Acc for this batch: 0.000 91 | *************************************************************** 92 | Ground-truth Labels: 0,18,92 93 | Ground-truth Num of Instances: 1,22,1 94 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 95 | Our Recovered Num of Instances: 1,22,1 | LnAcc: 1.000 | IRec: 1.000 96 | Start Experiment 4 97 | start_id: 7650 98 | 0/24, Acc for this batch: 0.000 99 | *************************************************************** 100 | Ground-truth Labels: 0,18,92 101 | Ground-truth Num of Instances: 1,22,1 102 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 103 | Our Recovered Num of Instances: 1,22,1 | LnAcc: 1.000 | IRec: 1.000 104 | Start Experiment 5 105 | start_id: 9719 106 | 0/24, Acc for this batch: 0.000 107 | *************************************************************** 108 | Ground-truth Labels: 0,18,92 109 | Ground-truth Num of Instances: 1,22,1 110 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 111 | Our Recovered Num of Instances: 1,22,1 | LnAcc: 1.000 | IRec: 1.000 112 | Start Experiment 6 113 | start_id: 11524 114 | 0/24, Acc for this batch: 0.000 115 | *************************************************************** 116 | Ground-truth Labels: 0,18,92 117 | Ground-truth Num of Instances: 1,22,1 118 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 119 | Our Recovered Num of Instances: 1,22,1 | LnAcc: 1.000 | IRec: 1.000 120 | Start Experiment 7 121 | start_id: 13404 122 | 0/24, Acc for this batch: 0.000 123 | *************************************************************** 124 | Ground-truth Labels: 0,18,92 125 | Ground-truth Num of Instances: 1,22,1 126 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 127 | Our Recovered Num of Instances: 1,22,1 | LnAcc: 1.000 | IRec: 1.000 128 | Start Experiment 8 129 | start_id: 15175 130 | 0/24, Acc for this batch: 0.000 131 | *************************************************************** 132 | Ground-truth Labels: 0,18,92 133 | Ground-truth Num of Instances: 1,22,1 134 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 135 | Our Recovered Num of Instances: 1,22,1 | LnAcc: 1.000 | IRec: 1.000 136 | Start Experiment 9 137 | start_id: 18032 138 | 0/24, Acc for this batch: 0.000 139 | *************************************************************** 140 | Ground-truth Labels: 0,18,92 141 | Ground-truth Num of Instances: 1,22,1 142 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 143 | Our Recovered Num of Instances: 1,22,1 | LnAcc: 1.000 | IRec: 1.000 144 | Start Experiment 10 145 | start_id: 20144 146 | 0/24, Acc for this batch: 0.000 147 | *************************************************************** 148 | Ground-truth Labels: 0,18,92 149 | Ground-truth Num of Instances: 1,22,1 150 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 151 | Our Recovered Num of Instances: 1,22,1 | LnAcc: 1.000 | IRec: 1.000 152 | Start Experiment 11 153 | start_id: 21878 154 | 0/24, Acc for this batch: 0.000 155 | *************************************************************** 156 | Ground-truth Labels: 0,18,92 157 | Ground-truth Num of Instances: 1,22,1 158 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 159 | Our Recovered Num of Instances: 1,22,1 | LnAcc: 1.000 | IRec: 1.000 160 | Start Experiment 12 161 | start_id: 24329 162 | 0/24, Acc for this batch: 0.000 163 | *************************************************************** 164 | Ground-truth Labels: 0,18,92 165 | Ground-truth Num of Instances: 1,22,1 166 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 167 | Our Recovered Num of Instances: 1,22,1 | LnAcc: 1.000 | IRec: 1.000 168 | Start Experiment 13 169 | start_id: 27170 170 | 0/24, Acc for this batch: 0.000 171 | *************************************************************** 172 | Ground-truth Labels: 0,18,92 173 | Ground-truth Num of Instances: 1,22,1 174 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 175 | Our Recovered Num of Instances: 1,22,1 | LnAcc: 1.000 | IRec: 1.000 176 | Start Experiment 14 177 | start_id: 29035 178 | 0/24, Acc for this batch: 0.000 179 | *************************************************************** 180 | Ground-truth Labels: 0,18,92 181 | Ground-truth Num of Instances: 1,22,1 182 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 183 | Our Recovered Num of Instances: 1,22,1 | LnAcc: 1.000 | IRec: 1.000 184 | Start Experiment 15 185 | start_id: 31531 186 | 0/24, Acc for this batch: 0.000 187 | *************************************************************** 188 | Ground-truth Labels: 0,18,92 189 | Ground-truth Num of Instances: 1,22,1 190 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 191 | Our Recovered Num of Instances: 1,22,1 | LnAcc: 1.000 | IRec: 1.000 192 | Start Experiment 16 193 | start_id: 32828 194 | 0/24, Acc for this batch: 0.000 195 | *************************************************************** 196 | Ground-truth Labels: 0,18,92 197 | Ground-truth Num of Instances: 1,22,1 198 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 199 | Our Recovered Num of Instances: 1,22,1 | LnAcc: 1.000 | IRec: 1.000 200 | Start Experiment 17 201 | start_id: 35248 202 | 0/24, Acc for this batch: 0.000 203 | *************************************************************** 204 | Ground-truth Labels: 0,18,92 205 | Ground-truth Num of Instances: 1,22,1 206 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 207 | Our Recovered Num of Instances: 1,22,1 | LnAcc: 1.000 | IRec: 1.000 208 | Start Experiment 18 209 | start_id: 36762 210 | 0/24, Acc for this batch: 0.000 211 | *************************************************************** 212 | Ground-truth Labels: 0,18,92 213 | Ground-truth Num of Instances: 1,22,1 214 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 215 | Our Recovered Num of Instances: 1,22,1 | LnAcc: 1.000 | IRec: 1.000 216 | Start Experiment 19 217 | start_id: 38499 218 | 0/24, Acc for this batch: 0.000 219 | *************************************************************** 220 | Ground-truth Labels: 0,18,92 221 | Ground-truth Num of Instances: 1,22,1 222 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 223 | Our Recovered Num of Instances: 1,22,1 | LnAcc: 1.000 | IRec: 1.000 224 | Start Experiment 20 225 | start_id: 40499 226 | 0/24, Acc for this batch: 0.000 227 | *************************************************************** 228 | Ground-truth Labels: 0,18,92 229 | Ground-truth Num of Instances: 1,22,1 230 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 231 | Our Recovered Num of Instances: 1,22,1 | LnAcc: 1.000 | IRec: 1.000 232 | --------------------------------------------------------------- 233 | Mean Ours LeAcc: 1.000 | Mean Ours LnAcc: 1.000 | Mean Ours IRec: 1.000 234 | Mean Rec Instances: Class 0-1, Class 18-22, Class 92-1 235 | -------------------------------------------------------------------------------- /new_logs/ExtremeDistribution/exp2.log: -------------------------------------------------------------------------------- 1 | running experiments at 2022-11-13-15:34:12 2 | exp_name: extreme_distribution 3 | cpu: False 4 | num_tries: 20 5 | num_images: 72 6 | seed: 12 7 | alpha: 1 8 | simplified: False 9 | compare: False 10 | estimate: False 11 | analysis: False 12 | ratio: 0.0 13 | data_path: data 14 | dataset: CIFAR100 15 | split: train 16 | distribution: custom_imbalanced 17 | start_id: 0 18 | num_classes: 100 19 | num_uniform_cls: 32 20 | num_target_cls: 5 21 | max_size: 32 22 | model: vgg16 23 | trained_model: False 24 | iter_train: False 25 | iters: 1000 26 | epochs: 10 27 | batch_size: 128 28 | lr: 0.1 29 | optimizer: SGD 30 | scheduler: linear 31 | weight_decay: 0.0005 32 | warmup: False 33 | epoch_interval: 10 34 | iter_interval: 100 35 | mid_save: False 36 | model_path: models 37 | dryrun: False 38 | batchnorm: False 39 | dropout: False 40 | silu: False 41 | leaky_relu: False 42 | n_dim: 300 43 | n_hidden: 1 44 | defense: False 45 | defense_method: dp 46 | noise_std: 0.001 47 | clip_bound: 4 48 | sparse_ratio: 10 49 | prune_ratio: 10 50 | rec_img: False 51 | fix_labels: False 52 | gt_labels: False 53 | optim: ig 54 | restarts: 1 55 | cost_fn: sim 56 | indices: def 57 | weights: equal 58 | rec_lr: None 59 | rec_optimizer: adam 60 | signed: False 61 | boxed: False 62 | scoring_choice: loss 63 | init: randn 64 | tv: 1e-06 65 | l2: 1e-06 66 | max_iterations: 8000 67 | loss_thresh: 0.0001 68 | save_image: False 69 | log_dir: logs/extreme_distribution 70 | image_dir: images/extreme_distribution 71 | 72 | Start Experiment 1 73 | start_id: 0 74 | 0/72, Acc for this batch: 0.000 75 | *************************************************************** 76 | Ground-truth Labels: 0,18,92 77 | Ground-truth Num of Instances: 1,70,1 78 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 79 | Our Recovered Num of Instances: 1,70,1 | LnAcc: 1.000 | IRec: 1.000 80 | Start Experiment 2 81 | start_id: 8219 82 | 0/72, Acc for this batch: 0.000 83 | *************************************************************** 84 | Ground-truth Labels: 0,18,92 85 | Ground-truth Num of Instances: 1,70,1 86 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 87 | Our Recovered Num of Instances: 1,70,1 | LnAcc: 1.000 | IRec: 1.000 88 | Start Experiment 3 89 | start_id: 14103 90 | 0/72, Acc for this batch: 0.000 91 | *************************************************************** 92 | Ground-truth Labels: 0,18,92 93 | Ground-truth Num of Instances: 1,70,1 94 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 95 | Our Recovered Num of Instances: 1,70,1 | LnAcc: 1.000 | IRec: 1.000 96 | Start Experiment 4 97 | start_id: 21107 98 | 0/72, Acc for this batch: 0.000 99 | *************************************************************** 100 | Ground-truth Labels: 0,18,92 101 | Ground-truth Num of Instances: 1,70,1 102 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 103 | Our Recovered Num of Instances: 1,70,1 | LnAcc: 1.000 | IRec: 1.000 104 | Start Experiment 5 105 | start_id: 28379 106 | 0/72, Acc for this batch: 0.000 107 | *************************************************************** 108 | Ground-truth Labels: 0,18,92 109 | Ground-truth Num of Instances: 1,70,1 110 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 111 | Our Recovered Num of Instances: 1,70,1 | LnAcc: 1.000 | IRec: 1.000 112 | Start Experiment 6 113 | start_id: 35005 114 | 0/72, Acc for this batch: 0.000 115 | *************************************************************** 116 | Ground-truth Labels: 0,18,92 117 | Ground-truth Num of Instances: 1,70,1 118 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 119 | Our Recovered Num of Instances: 1,70,1 | LnAcc: 1.000 | IRec: 1.000 120 | Start Experiment 7 121 | start_id: 40753 122 | 0/72, Acc for this batch: 0.000 123 | *************************************************************** 124 | Ground-truth Labels: 0,18,92 125 | Ground-truth Num of Instances: 1,70,1 126 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 127 | Our Recovered Num of Instances: 1,70,1 | LnAcc: 1.000 | IRec: 1.000 128 | Start Experiment 8 129 | start_id: 48233 130 | 0/72, Acc for this batch: 0.000 131 | *************************************************************** 132 | Ground-truth Labels: 0,18,92 133 | Ground-truth Num of Instances: 1,70,1 134 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 135 | Our Recovered Num of Instances: 1,70,1 | LnAcc: 1.000 | IRec: 1.000 136 | Start Experiment 9 137 | start_id: 6746 138 | 0/72, Acc for this batch: 0.000 139 | *************************************************************** 140 | Ground-truth Labels: 0,18,92 141 | Ground-truth Num of Instances: 1,70,1 142 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 143 | Our Recovered Num of Instances: 1,70,1 | LnAcc: 1.000 | IRec: 1.000 144 | Start Experiment 10 145 | start_id: 13278 146 | 0/72, Acc for this batch: 0.000 147 | *************************************************************** 148 | Ground-truth Labels: 0,18,92 149 | Ground-truth Num of Instances: 1,70,1 150 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 151 | Our Recovered Num of Instances: 1,70,1 | LnAcc: 1.000 | IRec: 1.000 152 | Start Experiment 11 153 | start_id: 20262 154 | 0/72, Acc for this batch: 0.000 155 | *************************************************************** 156 | Ground-truth Labels: 0,18,92 157 | Ground-truth Num of Instances: 1,70,1 158 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 159 | Our Recovered Num of Instances: 1,70,1 | LnAcc: 1.000 | IRec: 1.000 160 | Start Experiment 12 161 | start_id: 27699 162 | 0/72, Acc for this batch: 0.000 163 | *************************************************************** 164 | Ground-truth Labels: 0,18,92 165 | Ground-truth Num of Instances: 1,70,1 166 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 167 | Our Recovered Num of Instances: 1,70,1 | LnAcc: 1.000 | IRec: 1.000 168 | Start Experiment 13 169 | start_id: 33607 170 | 0/72, Acc for this batch: 0.000 171 | *************************************************************** 172 | Ground-truth Labels: 0,18,92 173 | Ground-truth Num of Instances: 1,70,1 174 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 175 | Our Recovered Num of Instances: 1,70,1 | LnAcc: 1.000 | IRec: 1.000 176 | Start Experiment 14 177 | start_id: 39522 178 | 0/72, Acc for this batch: 0.000 179 | *************************************************************** 180 | Ground-truth Labels: 0,18,92 181 | Ground-truth Num of Instances: 1,70,1 182 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 183 | Our Recovered Num of Instances: 1,70,1 | LnAcc: 1.000 | IRec: 1.000 184 | Start Experiment 15 185 | start_id: 46885 186 | 0/72, Acc for this batch: 0.000 187 | *************************************************************** 188 | Ground-truth Labels: 0,18,92 189 | Ground-truth Num of Instances: 1,70,1 190 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 191 | Our Recovered Num of Instances: 1,70,1 | LnAcc: 1.000 | IRec: 1.000 192 | Start Experiment 16 193 | start_id: 5471 194 | 0/72, Acc for this batch: 0.000 195 | *************************************************************** 196 | Ground-truth Labels: 0,18,92 197 | Ground-truth Num of Instances: 1,70,1 198 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 199 | Our Recovered Num of Instances: 1,70,1 | LnAcc: 1.000 | IRec: 1.000 200 | Start Experiment 17 201 | start_id: 12620 202 | 0/72, Acc for this batch: 0.000 203 | *************************************************************** 204 | Ground-truth Labels: 0,18,92 205 | Ground-truth Num of Instances: 1,70,1 206 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 207 | Our Recovered Num of Instances: 1,70,1 | LnAcc: 1.000 | IRec: 1.000 208 | Start Experiment 18 209 | start_id: 19493 210 | 0/72, Acc for this batch: 0.000 211 | *************************************************************** 212 | Ground-truth Labels: 0,18,92 213 | Ground-truth Num of Instances: 1,70,1 214 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 215 | Our Recovered Num of Instances: 1,70,1 | LnAcc: 1.000 | IRec: 1.000 216 | Start Experiment 19 217 | start_id: 26757 218 | 0/72, Acc for this batch: 0.000 219 | *************************************************************** 220 | Ground-truth Labels: 0,18,92 221 | Ground-truth Num of Instances: 1,70,1 222 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 223 | Our Recovered Num of Instances: 1,70,1 | LnAcc: 1.000 | IRec: 1.000 224 | Start Experiment 20 225 | start_id: 32828 226 | 0/72, Acc for this batch: 0.000 227 | *************************************************************** 228 | Ground-truth Labels: 0,18,92 229 | Ground-truth Num of Instances: 1,70,1 230 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 231 | Our Recovered Num of Instances: 1,70,1 | LnAcc: 1.000 | IRec: 1.000 232 | --------------------------------------------------------------- 233 | Mean Ours LeAcc: 1.000 | Mean Ours LnAcc: 1.000 | Mean Ours IRec: 1.000 234 | Mean Rec Instances: Class 0-1, Class 18-70, Class 92-1 235 | -------------------------------------------------------------------------------- /new_logs/TrainingStage/MNIST/exp5.log: -------------------------------------------------------------------------------- 1 | running experiments at 2022-11-17-09:19:51 2 | exp_name: train_stage2 3 | cpu: False 4 | num_tries: 20 5 | num_images: 8 6 | seed: 12 7 | alpha: 1 8 | simplified: False 9 | compare: False 10 | estimate: False 11 | analysis: False 12 | ratio: 0.0 13 | data_path: data 14 | dataset: MNIST_GRAY 15 | split: train 16 | distribution: random 17 | start_id: 0 18 | num_classes: 100 19 | num_uniform_cls: 32 20 | num_target_cls: 5 21 | max_size: 32 22 | model: dnn 23 | trained_model: True 24 | iter_train: True 25 | iters: 600 26 | epochs: 10 27 | batch_size: 128 28 | lr: 0.1 29 | optimizer: SGD 30 | scheduler: linear 31 | weight_decay: 0.0005 32 | warmup: False 33 | epoch_interval: 10 34 | iter_interval: 100 35 | mid_save: False 36 | model_path: models 37 | dryrun: False 38 | batchnorm: False 39 | dropout: False 40 | silu: False 41 | leaky_relu: False 42 | n_dim: 300 43 | n_hidden: 1 44 | defense: False 45 | defense_method: dp 46 | noise_std: 0.001 47 | clip_bound: 4 48 | sparse_ratio: 10 49 | prune_ratio: 10 50 | rec_img: False 51 | fix_labels: False 52 | gt_labels: False 53 | optim: ig 54 | restarts: 1 55 | cost_fn: sim 56 | indices: def 57 | weights: equal 58 | rec_lr: None 59 | rec_optimizer: adam 60 | signed: False 61 | boxed: False 62 | scoring_choice: loss 63 | init: randn 64 | tv: 1e-06 65 | l2: 1e-06 66 | max_iterations: 8000 67 | loss_thresh: 0.0001 68 | save_image: False 69 | log_dir: logs/train_stage2 70 | image_dir: images/train_stage2 71 | 72 | Model loaded from file dnn_MNIST_GRAY_Iter600.pth. 73 | Start Experiment 1 74 | start_id: 0 75 | 8/8, Acc for this batch: 1.000 76 | *************************************************************** 77 | Ground-truth Labels: 0,3,4,5,8 78 | Ground-truth Num of Instances: 3,1,2,1,1 79 | Our Recovered Labels: 0,4,5,6 | LeAcc: 0.700 80 | Our Recovered Num of Instances: 2,3,2,1 | LnAcc: 0.400 | IRec: 0.625 81 | Start Experiment 2 82 | start_id: 11879 83 | 8/8, Acc for this batch: 1.000 84 | *************************************************************** 85 | Ground-truth Labels: 0,1,2,5,7,9 86 | Ground-truth Num of Instances: 1,1,1,2,1,2 87 | Our Recovered Labels: 0 | LeAcc: 0.500 88 | Our Recovered Num of Instances: 1 | LnAcc: 0.500 | IRec: 0.125 89 | Start Experiment 3 90 | start_id: 41168 91 | 7/8, Acc for this batch: 0.875 92 | *************************************************************** 93 | Ground-truth Labels: 0,4,5,7,9 94 | Ground-truth Num of Instances: 1,2,1,1,3 95 | Our Recovered Labels: 0,4,9 | LeAcc: 0.800 96 | Our Recovered Num of Instances: 1,6,2 | LnAcc: 0.600 | IRec: 0.625 97 | Start Experiment 4 98 | start_id: 12852 99 | 8/8, Acc for this batch: 1.000 100 | *************************************************************** 101 | Ground-truth Labels: 0,1,4,5,8,9 102 | Ground-truth Num of Instances: 1,1,1,2,2,1 103 | Our Recovered Labels: 4,5,8 | LeAcc: 0.700 104 | Our Recovered Num of Instances: 1,3,2 | LnAcc: 0.600 | IRec: 0.625 105 | Start Experiment 5 106 | start_id: 49967 107 | 8/8, Acc for this batch: 1.000 108 | *************************************************************** 109 | Ground-truth Labels: 0,2,3,5,9 110 | Ground-truth Num of Instances: 1,1,2,1,3 111 | Our Recovered Labels: 3,5,9 | LeAcc: 0.800 112 | Our Recovered Num of Instances: 2,1,5 | LnAcc: 0.700 | IRec: 0.750 113 | Start Experiment 6 114 | start_id: 13121 115 | 8/8, Acc for this batch: 1.000 116 | *************************************************************** 117 | Ground-truth Labels: 4,5,6,7,8 118 | Ground-truth Num of Instances: 1,1,1,1,4 119 | Our Recovered Labels: 7 | LeAcc: 0.600 120 | Our Recovered Num of Instances: 1 | LnAcc: 0.600 | IRec: 0.125 121 | Start Experiment 7 122 | start_id: 58442 123 | 7/8, Acc for this batch: 0.875 124 | *************************************************************** 125 | Ground-truth Labels: 1,2,3,4,5,6,7 126 | Ground-truth Num of Instances: 1,1,1,1,1,1,2 127 | Our Recovered Labels: 2,3,4,5,7 | LeAcc: 0.800 128 | Our Recovered Num of Instances: 2,1,1,2,2 | LnAcc: 0.600 | IRec: 0.750 129 | Start Experiment 8 130 | start_id: 17366 131 | 7/8, Acc for this batch: 0.875 132 | *************************************************************** 133 | Ground-truth Labels: 0,1,6,8,9 134 | Ground-truth Num of Instances: 3,1,2,1,1 135 | Our Recovered Labels: 0,6,9 | LeAcc: 0.800 136 | Our Recovered Num of Instances: 8,1,1 | LnAcc: 0.600 | IRec: 0.625 137 | Start Experiment 9 138 | start_id: 35251 139 | 8/8, Acc for this batch: 1.000 140 | *************************************************************** 141 | Ground-truth Labels: 2,3,5,6,7,9 142 | Ground-truth Num of Instances: 2,1,1,1,2,1 143 | Our Recovered Labels: 2,5,7 | LeAcc: 0.700 144 | Our Recovered Num of Instances: 2,1,6 | LnAcc: 0.600 | IRec: 0.625 145 | Start Experiment 10 146 | start_id: 39317 147 | 8/8, Acc for this batch: 1.000 148 | *************************************************************** 149 | Ground-truth Labels: 1,2,3,5,6,7,9 150 | Ground-truth Num of Instances: 1,2,1,1,1,1,1 151 | Our Recovered Labels: 2,4,7,9 | LeAcc: 0.500 152 | Our Recovered Num of Instances: 3,1,1,2 | LnAcc: 0.300 | IRec: 0.500 153 | Start Experiment 11 154 | start_id: 32817 155 | 8/8, Acc for this batch: 1.000 156 | *************************************************************** 157 | Ground-truth Labels: 1,2,4,7,8,9 158 | Ground-truth Num of Instances: 1,1,1,3,1,1 159 | Our Recovered Labels: | LeAcc: 0.400 160 | Our Recovered Num of Instances: | LnAcc: 0.400 | IRec: 0.000 161 | Start Experiment 12 162 | start_id: 12441 163 | 7/8, Acc for this batch: 0.875 164 | *************************************************************** 165 | Ground-truth Labels: 2,4,7,8,9 166 | Ground-truth Num of Instances: 2,1,2,2,1 167 | Our Recovered Labels: 2,6,8,9 | LeAcc: 0.700 168 | Our Recovered Num of Instances: 1,1,4,2 | LnAcc: 0.400 | IRec: 0.500 169 | Start Experiment 13 170 | start_id: 33651 171 | 7/8, Acc for this batch: 0.875 172 | *************************************************************** 173 | Ground-truth Labels: 2,3,5,6,7,8 174 | Ground-truth Num of Instances: 1,2,1,1,2,1 175 | Our Recovered Labels: 0,7,8 | LeAcc: 0.500 176 | Our Recovered Num of Instances: 4,1,1 | LnAcc: 0.400 | IRec: 0.250 177 | Start Experiment 14 178 | start_id: 17000 179 | 7/8, Acc for this batch: 0.875 180 | *************************************************************** 181 | Ground-truth Labels: 0,1,3,6,8,9 182 | Ground-truth Num of Instances: 2,2,1,1,1,1 183 | Our Recovered Labels: 0,3,9 | LeAcc: 0.700 184 | Our Recovered Num of Instances: 8,1,1 | LnAcc: 0.600 | IRec: 0.500 185 | Start Experiment 15 186 | start_id: 43072 187 | 8/8, Acc for this batch: 1.000 188 | *************************************************************** 189 | Ground-truth Labels: 0,1,3,4,5,8 190 | Ground-truth Num of Instances: 3,1,1,1,1,1 191 | Our Recovered Labels: | LeAcc: 0.400 192 | Our Recovered Num of Instances: | LnAcc: 0.400 | IRec: 0.000 193 | Start Experiment 16 194 | start_id: 32079 195 | 6/8, Acc for this batch: 0.750 196 | *************************************************************** 197 | Ground-truth Labels: 1,3,5,6,7,8,9 198 | Ground-truth Num of Instances: 1,1,1,2,1,1,1 199 | Our Recovered Labels: 9 | LeAcc: 0.400 200 | Our Recovered Num of Instances: 1 | LnAcc: 0.400 | IRec: 0.125 201 | Start Experiment 17 202 | start_id: 59661 203 | 6/8, Acc for this batch: 0.750 204 | *************************************************************** 205 | Ground-truth Labels: 3,4,5,6,8,9 206 | Ground-truth Num of Instances: 1,2,1,1,1,2 207 | Our Recovered Labels: 3,5,6,7,9 | LeAcc: 0.700 208 | Our Recovered Num of Instances: 1,2,2,2,2 | LnAcc: 0.500 | IRec: 0.625 209 | Start Experiment 18 210 | start_id: 44664 211 | 7/8, Acc for this batch: 0.875 212 | *************************************************************** 213 | Ground-truth Labels: 1,5,6,7,8,9 214 | Ground-truth Num of Instances: 1,1,2,1,1,2 215 | Our Recovered Labels: 5,6,7,8,9 | LeAcc: 0.900 216 | Our Recovered Num of Instances: 1,3,1,1,2 | LnAcc: 0.800 | IRec: 0.875 217 | Start Experiment 19 218 | start_id: 57949 219 | 7/8, Acc for this batch: 0.875 220 | *************************************************************** 221 | Ground-truth Labels: 0,2,3,5,8,9 222 | Ground-truth Num of Instances: 1,2,2,1,1,1 223 | Our Recovered Labels: 2,9 | LeAcc: 0.600 224 | Our Recovered Num of Instances: 1,1 | LnAcc: 0.500 | IRec: 0.250 225 | Start Experiment 20 226 | start_id: 27278 227 | 8/8, Acc for this batch: 1.000 228 | *************************************************************** 229 | Ground-truth Labels: 0,1,3,5,6,7,8 230 | Ground-truth Num of Instances: 1,1,1,1,1,2,1 231 | Our Recovered Labels: 0,5 | LeAcc: 0.500 232 | Our Recovered Num of Instances: 7,1 | LnAcc: 0.400 | IRec: 0.250 233 | --------------------------------------------------------------- 234 | Mean Ours LeAcc: 0.635 | Mean Ours LnAcc: 0.515 | Mean Ours IRec: 0.438 235 | -------------------------------------------------------------------------------- /new_logs/TrainingStage/MNIST/exp6.log: -------------------------------------------------------------------------------- 1 | running experiments at 2022-11-17-09:20:03 2 | exp_name: train_stage2 3 | cpu: False 4 | num_tries: 20 5 | num_images: 8 6 | seed: 12 7 | alpha: 1 8 | simplified: False 9 | compare: False 10 | estimate: False 11 | analysis: False 12 | ratio: 0.0 13 | data_path: data 14 | dataset: MNIST_GRAY 15 | split: train 16 | distribution: random 17 | start_id: 0 18 | num_classes: 100 19 | num_uniform_cls: 32 20 | num_target_cls: 5 21 | max_size: 32 22 | model: dnn 23 | trained_model: True 24 | iter_train: True 25 | iters: 500 26 | epochs: 10 27 | batch_size: 128 28 | lr: 0.1 29 | optimizer: SGD 30 | scheduler: linear 31 | weight_decay: 0.0005 32 | warmup: False 33 | epoch_interval: 10 34 | iter_interval: 100 35 | mid_save: False 36 | model_path: models 37 | dryrun: False 38 | batchnorm: False 39 | dropout: False 40 | silu: False 41 | leaky_relu: False 42 | n_dim: 300 43 | n_hidden: 1 44 | defense: False 45 | defense_method: dp 46 | noise_std: 0.001 47 | clip_bound: 4 48 | sparse_ratio: 10 49 | prune_ratio: 10 50 | rec_img: False 51 | fix_labels: False 52 | gt_labels: False 53 | optim: ig 54 | restarts: 1 55 | cost_fn: sim 56 | indices: def 57 | weights: equal 58 | rec_lr: None 59 | rec_optimizer: adam 60 | signed: False 61 | boxed: False 62 | scoring_choice: loss 63 | init: randn 64 | tv: 1e-06 65 | l2: 1e-06 66 | max_iterations: 8000 67 | loss_thresh: 0.0001 68 | save_image: False 69 | log_dir: logs/train_stage2 70 | image_dir: images/train_stage2 71 | 72 | Model loaded from file dnn_MNIST_GRAY_Iter500.pth. 73 | Start Experiment 1 74 | start_id: 0 75 | 8/8, Acc for this batch: 1.000 76 | *************************************************************** 77 | Ground-truth Labels: 0,3,4,5,8 78 | Ground-truth Num of Instances: 3,1,2,1,1 79 | Our Recovered Labels: 0,4,5 | LeAcc: 0.800 80 | Our Recovered Num of Instances: 2,2,4 | LnAcc: 0.600 | IRec: 0.625 81 | Start Experiment 2 82 | start_id: 11879 83 | 8/8, Acc for this batch: 1.000 84 | *************************************************************** 85 | Ground-truth Labels: 0,1,2,5,7,9 86 | Ground-truth Num of Instances: 1,1,1,2,1,2 87 | Our Recovered Labels: 0 | LeAcc: 0.500 88 | Our Recovered Num of Instances: 1 | LnAcc: 0.500 | IRec: 0.125 89 | Start Experiment 3 90 | start_id: 41168 91 | 8/8, Acc for this batch: 1.000 92 | *************************************************************** 93 | Ground-truth Labels: 0,4,5,7,9 94 | Ground-truth Num of Instances: 1,2,1,1,3 95 | Our Recovered Labels: 0,4 | LeAcc: 0.700 96 | Our Recovered Num of Instances: 1,7 | LnAcc: 0.600 | IRec: 0.375 97 | Start Experiment 4 98 | start_id: 12852 99 | 8/8, Acc for this batch: 1.000 100 | *************************************************************** 101 | Ground-truth Labels: 0,1,4,5,8,9 102 | Ground-truth Num of Instances: 1,1,1,2,2,1 103 | Our Recovered Labels: 5,8,9 | LeAcc: 0.700 104 | Our Recovered Num of Instances: 2,3,7 | LnAcc: 0.500 | IRec: 0.625 105 | Start Experiment 5 106 | start_id: 49967 107 | 8/8, Acc for this batch: 1.000 108 | *************************************************************** 109 | Ground-truth Labels: 0,2,3,5,9 110 | Ground-truth Num of Instances: 1,1,2,1,3 111 | Our Recovered Labels: 3,5,9 | LeAcc: 0.800 112 | Our Recovered Num of Instances: 3,1,4 | LnAcc: 0.600 | IRec: 0.750 113 | Start Experiment 6 114 | start_id: 13121 115 | 8/8, Acc for this batch: 1.000 116 | *************************************************************** 117 | Ground-truth Labels: 4,5,6,7,8 118 | Ground-truth Num of Instances: 1,1,1,1,4 119 | Our Recovered Labels: 7,8 | LeAcc: 0.700 120 | Our Recovered Num of Instances: 3,8 | LnAcc: 0.500 | IRec: 0.625 121 | Start Experiment 7 122 | start_id: 58442 123 | 6/8, Acc for this batch: 0.750 124 | *************************************************************** 125 | Ground-truth Labels: 1,2,3,4,5,6,7 126 | Ground-truth Num of Instances: 1,1,1,1,1,1,2 127 | Our Recovered Labels: 2,3,7 | LeAcc: 0.600 128 | Our Recovered Num of Instances: 7,1,2 | LnAcc: 0.500 | IRec: 0.500 129 | Start Experiment 8 130 | start_id: 17366 131 | 7/8, Acc for this batch: 0.875 132 | *************************************************************** 133 | Ground-truth Labels: 0,1,6,8,9 134 | Ground-truth Num of Instances: 3,1,2,1,1 135 | Our Recovered Labels: 6,9 | LeAcc: 0.700 136 | Our Recovered Num of Instances: 1,1 | LnAcc: 0.600 | IRec: 0.250 137 | Start Experiment 9 138 | start_id: 35251 139 | 6/8, Acc for this batch: 0.750 140 | *************************************************************** 141 | Ground-truth Labels: 2,3,5,6,7,9 142 | Ground-truth Num of Instances: 2,1,1,1,2,1 143 | Our Recovered Labels: 2,5,7 | LeAcc: 0.700 144 | Our Recovered Num of Instances: 1,7,3 | LnAcc: 0.400 | IRec: 0.500 145 | Start Experiment 10 146 | start_id: 39317 147 | 8/8, Acc for this batch: 1.000 148 | *************************************************************** 149 | Ground-truth Labels: 1,2,3,5,6,7,9 150 | Ground-truth Num of Instances: 1,2,1,1,1,1,1 151 | Our Recovered Labels: 2,4,7,9 | LeAcc: 0.500 152 | Our Recovered Num of Instances: 3,1,1,2 | LnAcc: 0.300 | IRec: 0.500 153 | Start Experiment 11 154 | start_id: 32817 155 | 8/8, Acc for this batch: 1.000 156 | *************************************************************** 157 | Ground-truth Labels: 1,2,4,7,8,9 158 | Ground-truth Num of Instances: 1,1,1,3,1,1 159 | Our Recovered Labels: | LeAcc: 0.400 160 | Our Recovered Num of Instances: | LnAcc: 0.400 | IRec: 0.000 161 | Start Experiment 12 162 | start_id: 12441 163 | 7/8, Acc for this batch: 0.875 164 | *************************************************************** 165 | Ground-truth Labels: 2,4,7,8,9 166 | Ground-truth Num of Instances: 2,1,2,2,1 167 | Our Recovered Labels: 2,6,8,9 | LeAcc: 0.700 168 | Our Recovered Num of Instances: 1,1,4,2 | LnAcc: 0.400 | IRec: 0.500 169 | Start Experiment 13 170 | start_id: 33651 171 | 5/8, Acc for this batch: 0.625 172 | *************************************************************** 173 | Ground-truth Labels: 2,3,5,6,7,8 174 | Ground-truth Num of Instances: 1,2,1,1,2,1 175 | Our Recovered Labels: 0,3,5,7,8 | LeAcc: 0.700 176 | Our Recovered Num of Instances: 1,4,1,1,1 | LnAcc: 0.500 | IRec: 0.625 177 | Start Experiment 14 178 | start_id: 17000 179 | 7/8, Acc for this batch: 0.875 180 | *************************************************************** 181 | Ground-truth Labels: 0,1,3,6,8,9 182 | Ground-truth Num of Instances: 2,2,1,1,1,1 183 | Our Recovered Labels: 9 | LeAcc: 0.500 184 | Our Recovered Num of Instances: 1 | LnAcc: 0.500 | IRec: 0.125 185 | Start Experiment 15 186 | start_id: 43072 187 | 7/8, Acc for this batch: 0.875 188 | *************************************************************** 189 | Ground-truth Labels: 0,1,3,4,5,8 190 | Ground-truth Num of Instances: 3,1,1,1,1,1 191 | Our Recovered Labels: 1,3,4,5,8 | LeAcc: 0.900 192 | Our Recovered Num of Instances: 3,2,2,1,1 | LnAcc: 0.600 | IRec: 0.625 193 | Start Experiment 16 194 | start_id: 32079 195 | 6/8, Acc for this batch: 0.750 196 | *************************************************************** 197 | Ground-truth Labels: 1,3,5,6,7,8,9 198 | Ground-truth Num of Instances: 1,1,1,2,1,1,1 199 | Our Recovered Labels: 3,9 | LeAcc: 0.500 200 | Our Recovered Num of Instances: 1,1 | LnAcc: 0.500 | IRec: 0.250 201 | Start Experiment 17 202 | start_id: 59661 203 | 5/8, Acc for this batch: 0.625 204 | *************************************************************** 205 | Ground-truth Labels: 3,4,5,6,8,9 206 | Ground-truth Num of Instances: 1,2,1,1,1,2 207 | Our Recovered Labels: 3,5,6,7,9 | LeAcc: 0.700 208 | Our Recovered Num of Instances: 1,2,2,1,2 | LnAcc: 0.500 | IRec: 0.625 209 | Start Experiment 18 210 | start_id: 44664 211 | 6/8, Acc for this batch: 0.750 212 | *************************************************************** 213 | Ground-truth Labels: 1,5,6,7,8,9 214 | Ground-truth Num of Instances: 1,1,2,1,1,2 215 | Our Recovered Labels: 6,8,9 | LeAcc: 0.700 216 | Our Recovered Num of Instances: 4,2,2 | LnAcc: 0.500 | IRec: 0.625 217 | Start Experiment 19 218 | start_id: 57949 219 | 8/8, Acc for this batch: 1.000 220 | *************************************************************** 221 | Ground-truth Labels: 0,2,3,5,8,9 222 | Ground-truth Num of Instances: 1,2,2,1,1,1 223 | Our Recovered Labels: 2,3,5,9 | LeAcc: 0.800 224 | Our Recovered Num of Instances: 1,5,1,1 | LnAcc: 0.600 | IRec: 0.625 225 | Start Experiment 20 226 | start_id: 27278 227 | 8/8, Acc for this batch: 1.000 228 | *************************************************************** 229 | Ground-truth Labels: 0,1,3,5,6,7,8 230 | Ground-truth Num of Instances: 1,1,1,1,1,2,1 231 | Our Recovered Labels: 0,3 | LeAcc: 0.500 232 | Our Recovered Num of Instances: 3,5 | LnAcc: 0.300 | IRec: 0.250 233 | --------------------------------------------------------------- 234 | Mean Ours LeAcc: 0.655 | Mean Ours LnAcc: 0.495 | Mean Ours IRec: 0.456 235 | -------------------------------------------------------------------------------- /new_logs/TrainingStage/MNIST/exp4.log: -------------------------------------------------------------------------------- 1 | running experiments at 2022-11-17-09:15:55 2 | exp_name: train_stage2 3 | cpu: False 4 | num_tries: 20 5 | num_images: 8 6 | seed: 12 7 | alpha: 1 8 | simplified: False 9 | compare: False 10 | estimate: False 11 | analysis: False 12 | ratio: 0.0 13 | data_path: data 14 | dataset: MNIST_GRAY 15 | split: train 16 | distribution: random 17 | start_id: 0 18 | num_classes: 100 19 | num_uniform_cls: 32 20 | num_target_cls: 5 21 | max_size: 32 22 | model: dnn 23 | trained_model: True 24 | iter_train: True 25 | iters: 700 26 | epochs: 10 27 | batch_size: 128 28 | lr: 0.1 29 | optimizer: SGD 30 | scheduler: linear 31 | weight_decay: 0.0005 32 | warmup: False 33 | epoch_interval: 10 34 | iter_interval: 100 35 | mid_save: False 36 | model_path: models 37 | dryrun: False 38 | batchnorm: False 39 | dropout: False 40 | silu: False 41 | leaky_relu: False 42 | n_dim: 300 43 | n_hidden: 1 44 | defense: False 45 | defense_method: dp 46 | noise_std: 0.001 47 | clip_bound: 4 48 | sparse_ratio: 10 49 | prune_ratio: 10 50 | rec_img: False 51 | fix_labels: False 52 | gt_labels: False 53 | optim: ig 54 | restarts: 1 55 | cost_fn: sim 56 | indices: def 57 | weights: equal 58 | rec_lr: None 59 | rec_optimizer: adam 60 | signed: False 61 | boxed: False 62 | scoring_choice: loss 63 | init: randn 64 | tv: 1e-06 65 | l2: 1e-06 66 | max_iterations: 8000 67 | loss_thresh: 0.0001 68 | save_image: False 69 | log_dir: logs/train_stage2 70 | image_dir: images/train_stage2 71 | 72 | Model loaded from file dnn_MNIST_GRAY_Iter700.pth. 73 | Start Experiment 1 74 | start_id: 0 75 | 8/8, Acc for this batch: 1.000 76 | *************************************************************** 77 | Ground-truth Labels: 0,3,4,5,8 78 | Ground-truth Num of Instances: 3,1,2,1,1 79 | Our Recovered Labels: 0,4,5 | LeAcc: 0.800 80 | Our Recovered Num of Instances: 1,4,2 | LnAcc: 0.500 | IRec: 0.500 81 | Start Experiment 2 82 | start_id: 11879 83 | 8/8, Acc for this batch: 1.000 84 | *************************************************************** 85 | Ground-truth Labels: 0,1,2,5,7,9 86 | Ground-truth Num of Instances: 1,1,1,2,1,2 87 | Our Recovered Labels: 0,1 | LeAcc: 0.600 88 | Our Recovered Num of Instances: 1,2 | LnAcc: 0.500 | IRec: 0.250 89 | Start Experiment 3 90 | start_id: 41168 91 | 7/8, Acc for this batch: 0.875 92 | *************************************************************** 93 | Ground-truth Labels: 0,4,5,7,9 94 | Ground-truth Num of Instances: 1,2,1,1,3 95 | Our Recovered Labels: 0,4,7 | LeAcc: 0.800 96 | Our Recovered Num of Instances: 1,3,2 | LnAcc: 0.600 | IRec: 0.500 97 | Start Experiment 4 98 | start_id: 12852 99 | 8/8, Acc for this batch: 1.000 100 | *************************************************************** 101 | Ground-truth Labels: 0,1,4,5,8,9 102 | Ground-truth Num of Instances: 1,1,1,2,2,1 103 | Our Recovered Labels: 3,4,5,8 | LeAcc: 0.600 104 | Our Recovered Num of Instances: 1,1,3,3 | LnAcc: 0.400 | IRec: 0.625 105 | Start Experiment 5 106 | start_id: 49967 107 | 8/8, Acc for this batch: 1.000 108 | *************************************************************** 109 | Ground-truth Labels: 0,2,3,5,9 110 | Ground-truth Num of Instances: 1,1,2,1,3 111 | Our Recovered Labels: 3,5,9 | LeAcc: 0.800 112 | Our Recovered Num of Instances: 2,1,5 | LnAcc: 0.700 | IRec: 0.750 113 | Start Experiment 6 114 | start_id: 13121 115 | 8/8, Acc for this batch: 1.000 116 | *************************************************************** 117 | Ground-truth Labels: 4,5,6,7,8 118 | Ground-truth Num of Instances: 1,1,1,1,4 119 | Our Recovered Labels: 7,8 | LeAcc: 0.700 120 | Our Recovered Num of Instances: 2,8 | LnAcc: 0.500 | IRec: 0.625 121 | Start Experiment 7 122 | start_id: 58442 123 | 6/8, Acc for this batch: 0.750 124 | *************************************************************** 125 | Ground-truth Labels: 1,2,3,4,5,6,7 126 | Ground-truth Num of Instances: 1,1,1,1,1,1,2 127 | Our Recovered Labels: 2,3,7 | LeAcc: 0.600 128 | Our Recovered Num of Instances: 2,1,1 | LnAcc: 0.400 | IRec: 0.375 129 | Start Experiment 8 130 | start_id: 17366 131 | 7/8, Acc for this batch: 0.875 132 | *************************************************************** 133 | Ground-truth Labels: 0,1,6,8,9 134 | Ground-truth Num of Instances: 3,1,2,1,1 135 | Our Recovered Labels: 2,6,8,9 | LeAcc: 0.700 136 | Our Recovered Num of Instances: 5,1,1,1 | LnAcc: 0.600 | IRec: 0.375 137 | Start Experiment 9 138 | start_id: 35251 139 | 7/8, Acc for this batch: 0.875 140 | *************************************************************** 141 | Ground-truth Labels: 2,3,5,6,7,9 142 | Ground-truth Num of Instances: 2,1,1,1,2,1 143 | Our Recovered Labels: 2,3,5,7 | LeAcc: 0.800 144 | Our Recovered Num of Instances: 2,1,3,2 | LnAcc: 0.700 | IRec: 0.750 145 | Start Experiment 10 146 | start_id: 39317 147 | 8/8, Acc for this batch: 1.000 148 | *************************************************************** 149 | Ground-truth Labels: 1,2,3,5,6,7,9 150 | Ground-truth Num of Instances: 1,2,1,1,1,1,1 151 | Our Recovered Labels: 2,4,7,9 | LeAcc: 0.500 152 | Our Recovered Num of Instances: 3,2,2,2 | LnAcc: 0.200 | IRec: 0.500 153 | Start Experiment 11 154 | start_id: 32817 155 | 8/8, Acc for this batch: 1.000 156 | *************************************************************** 157 | Ground-truth Labels: 1,2,4,7,8,9 158 | Ground-truth Num of Instances: 1,1,1,3,1,1 159 | Our Recovered Labels: 7,8,9 | LeAcc: 0.700 160 | Our Recovered Num of Instances: 1,1,5 | LnAcc: 0.500 | IRec: 0.375 161 | Start Experiment 12 162 | start_id: 12441 163 | 7/8, Acc for this batch: 0.875 164 | *************************************************************** 165 | Ground-truth Labels: 2,4,7,8,9 166 | Ground-truth Num of Instances: 2,1,2,2,1 167 | Our Recovered Labels: 2,7,8,9 | LeAcc: 0.900 168 | Our Recovered Num of Instances: 1,4,1,2 | LnAcc: 0.500 | IRec: 0.625 169 | Start Experiment 13 170 | start_id: 33651 171 | 7/8, Acc for this batch: 0.875 172 | *************************************************************** 173 | Ground-truth Labels: 2,3,5,6,7,8 174 | Ground-truth Num of Instances: 1,2,1,1,2,1 175 | Our Recovered Labels: 0,7,8 | LeAcc: 0.500 176 | Our Recovered Num of Instances: 4,1,1 | LnAcc: 0.400 | IRec: 0.250 177 | Start Experiment 14 178 | start_id: 17000 179 | 7/8, Acc for this batch: 0.875 180 | *************************************************************** 181 | Ground-truth Labels: 0,1,3,6,8,9 182 | Ground-truth Num of Instances: 2,2,1,1,1,1 183 | Our Recovered Labels: 1,8,9 | LeAcc: 0.700 184 | Our Recovered Num of Instances: 7,1,1 | LnAcc: 0.600 | IRec: 0.500 185 | Start Experiment 15 186 | start_id: 43072 187 | 8/8, Acc for this batch: 1.000 188 | *************************************************************** 189 | Ground-truth Labels: 0,1,3,4,5,8 190 | Ground-truth Num of Instances: 3,1,1,1,1,1 191 | Our Recovered Labels: 5 | LeAcc: 0.500 192 | Our Recovered Num of Instances: 2 | LnAcc: 0.400 | IRec: 0.125 193 | Start Experiment 16 194 | start_id: 32079 195 | 7/8, Acc for this batch: 0.875 196 | *************************************************************** 197 | Ground-truth Labels: 1,3,5,6,7,8,9 198 | Ground-truth Num of Instances: 1,1,1,2,1,1,1 199 | Our Recovered Labels: 6,9 | LeAcc: 0.500 200 | Our Recovered Num of Instances: 8,1 | LnAcc: 0.400 | IRec: 0.375 201 | Start Experiment 17 202 | start_id: 59661 203 | 6/8, Acc for this batch: 0.750 204 | *************************************************************** 205 | Ground-truth Labels: 3,4,5,6,8,9 206 | Ground-truth Num of Instances: 1,2,1,1,1,2 207 | Our Recovered Labels: 3,5,6,7,9 | LeAcc: 0.700 208 | Our Recovered Num of Instances: 1,2,2,1,2 | LnAcc: 0.500 | IRec: 0.625 209 | Start Experiment 18 210 | start_id: 44664 211 | 8/8, Acc for this batch: 1.000 212 | *************************************************************** 213 | Ground-truth Labels: 1,5,6,7,8,9 214 | Ground-truth Num of Instances: 1,1,2,1,1,2 215 | Our Recovered Labels: 6,7,8,9 | LeAcc: 0.800 216 | Our Recovered Num of Instances: 2,1,7,2 | LnAcc: 0.700 | IRec: 0.750 217 | Start Experiment 19 218 | start_id: 57949 219 | 7/8, Acc for this batch: 0.875 220 | *************************************************************** 221 | Ground-truth Labels: 0,2,3,5,8,9 222 | Ground-truth Num of Instances: 1,2,2,1,1,1 223 | Our Recovered Labels: 2,9 | LeAcc: 0.600 224 | Our Recovered Num of Instances: 1,2 | LnAcc: 0.400 | IRec: 0.250 225 | Start Experiment 20 226 | start_id: 27278 227 | 8/8, Acc for this batch: 1.000 228 | *************************************************************** 229 | Ground-truth Labels: 0,1,3,5,6,7,8 230 | Ground-truth Num of Instances: 1,1,1,1,1,2,1 231 | Our Recovered Labels: 0 | LeAcc: 0.400 232 | Our Recovered Num of Instances: 2 | LnAcc: 0.300 | IRec: 0.125 233 | --------------------------------------------------------------- 234 | Mean Ours LeAcc: 0.660 | Mean Ours LnAcc: 0.490 | Mean Ours IRec: 0.463 235 | -------------------------------------------------------------------------------- /new_logs/ExtremeDistribution/exp3.log: -------------------------------------------------------------------------------- 1 | running experiments at 2022-11-13-15:35:47 2 | exp_name: extreme_distribution 3 | cpu: False 4 | num_tries: 20 5 | num_images: 216 6 | seed: 12 7 | alpha: 1 8 | simplified: False 9 | compare: False 10 | estimate: False 11 | analysis: False 12 | ratio: 0.0 13 | data_path: data 14 | dataset: CIFAR100 15 | split: train 16 | distribution: custom_imbalanced 17 | start_id: 0 18 | num_classes: 100 19 | num_uniform_cls: 32 20 | num_target_cls: 5 21 | max_size: 32 22 | model: vgg16 23 | trained_model: False 24 | iter_train: False 25 | iters: 1000 26 | epochs: 10 27 | batch_size: 128 28 | lr: 0.1 29 | optimizer: SGD 30 | scheduler: linear 31 | weight_decay: 0.0005 32 | warmup: False 33 | epoch_interval: 10 34 | iter_interval: 100 35 | mid_save: False 36 | model_path: models 37 | dryrun: False 38 | batchnorm: False 39 | dropout: False 40 | silu: False 41 | leaky_relu: False 42 | n_dim: 300 43 | n_hidden: 1 44 | defense: False 45 | defense_method: dp 46 | noise_std: 0.001 47 | clip_bound: 4 48 | sparse_ratio: 10 49 | prune_ratio: 10 50 | rec_img: False 51 | fix_labels: False 52 | gt_labels: False 53 | optim: ig 54 | restarts: 1 55 | cost_fn: sim 56 | indices: def 57 | weights: equal 58 | rec_lr: None 59 | rec_optimizer: adam 60 | signed: False 61 | boxed: False 62 | scoring_choice: loss 63 | init: randn 64 | tv: 1e-06 65 | l2: 1e-06 66 | max_iterations: 8000 67 | loss_thresh: 0.0001 68 | save_image: False 69 | log_dir: logs/extreme_distribution 70 | image_dir: images/extreme_distribution 71 | 72 | Start Experiment 1 73 | start_id: 0 74 | 0/216, Acc for this batch: 0.000 75 | *************************************************************** 76 | Ground-truth Labels: 0,18,92 77 | Ground-truth Num of Instances: 1,214,1 78 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 79 | Our Recovered Num of Instances: 1,214,1 | LnAcc: 1.000 | IRec: 1.000 80 | Start Experiment 2 81 | start_id: 21372 82 | 0/216, Acc for this batch: 0.000 83 | *************************************************************** 84 | Ground-truth Labels: 0,18,92 85 | Ground-truth Num of Instances: 1,214,1 86 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 87 | Our Recovered Num of Instances: 1,214,1 | LnAcc: 1.000 | IRec: 1.000 88 | Start Experiment 3 89 | start_id: 41345 90 | 0/216, Acc for this batch: 0.000 91 | *************************************************************** 92 | Ground-truth Labels: 0,18,92 93 | Ground-truth Num of Instances: 1,214,1 94 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 95 | Our Recovered Num of Instances: 1,214,1 | LnAcc: 1.000 | IRec: 1.000 96 | Start Experiment 4 97 | start_id: 14129 98 | 0/216, Acc for this batch: 0.000 99 | *************************************************************** 100 | Ground-truth Labels: 0,18,92 101 | Ground-truth Num of Instances: 1,214,1 102 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 103 | Our Recovered Num of Instances: 1,214,1 | LnAcc: 1.000 | IRec: 1.000 104 | Start Experiment 5 105 | start_id: 35526 106 | 0/216, Acc for this batch: 0.000 107 | *************************************************************** 108 | Ground-truth Labels: 0,18,92 109 | Ground-truth Num of Instances: 1,214,1 110 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 111 | Our Recovered Num of Instances: 1,214,1 | LnAcc: 1.000 | IRec: 1.000 112 | Start Experiment 6 113 | start_id: 8219 114 | 0/216, Acc for this batch: 0.000 115 | *************************************************************** 116 | Ground-truth Labels: 0,18,92 117 | Ground-truth Num of Instances: 1,214,1 118 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 119 | Our Recovered Num of Instances: 1,214,1 | LnAcc: 1.000 | IRec: 1.000 120 | Start Experiment 7 121 | start_id: 28529 122 | 0/216, Acc for this batch: 0.000 123 | *************************************************************** 124 | Ground-truth Labels: 0,18,92 125 | Ground-truth Num of Instances: 1,214,1 126 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 127 | Our Recovered Num of Instances: 1,214,1 | LnAcc: 1.000 | IRec: 1.000 128 | Start Experiment 8 129 | start_id: 49343 130 | 0/216, Acc for this batch: 0.000 131 | *************************************************************** 132 | Ground-truth Labels: 0,18,92 133 | Ground-truth Num of Instances: 1,214,1 134 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 135 | Our Recovered Num of Instances: 1,214,1 | LnAcc: 1.000 | IRec: 1.000 136 | Start Experiment 9 137 | start_id: 21336 138 | 0/216, Acc for this batch: 0.000 139 | *************************************************************** 140 | Ground-truth Labels: 0,18,92 141 | Ground-truth Num of Instances: 1,214,1 142 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 143 | Our Recovered Num of Instances: 1,214,1 | LnAcc: 1.000 | IRec: 1.000 144 | Start Experiment 10 145 | start_id: 41220 146 | 0/216, Acc for this batch: 0.000 147 | *************************************************************** 148 | Ground-truth Labels: 0,18,92 149 | Ground-truth Num of Instances: 1,214,1 150 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 151 | Our Recovered Num of Instances: 1,214,1 | LnAcc: 1.000 | IRec: 1.000 152 | Start Experiment 11 153 | start_id: 14103 154 | 0/216, Acc for this batch: 0.000 155 | *************************************************************** 156 | Ground-truth Labels: 0,18,92 157 | Ground-truth Num of Instances: 1,214,1 158 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 159 | Our Recovered Num of Instances: 1,214,1 | LnAcc: 1.000 | IRec: 1.000 160 | Start Experiment 12 161 | start_id: 35396 162 | 0/216, Acc for this batch: 0.000 163 | *************************************************************** 164 | Ground-truth Labels: 0,18,92 165 | Ground-truth Num of Instances: 1,214,1 166 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 167 | Our Recovered Num of Instances: 1,214,1 | LnAcc: 1.000 | IRec: 1.000 168 | Start Experiment 13 169 | start_id: 7956 170 | 0/216, Acc for this batch: 0.000 171 | *************************************************************** 172 | Ground-truth Labels: 0,18,92 173 | Ground-truth Num of Instances: 1,214,1 174 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 175 | Our Recovered Num of Instances: 1,214,1 | LnAcc: 1.000 | IRec: 1.000 176 | Start Experiment 14 177 | start_id: 28457 178 | 0/216, Acc for this batch: 0.000 179 | *************************************************************** 180 | Ground-truth Labels: 0,18,92 181 | Ground-truth Num of Instances: 1,214,1 182 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 183 | Our Recovered Num of Instances: 1,214,1 | LnAcc: 1.000 | IRec: 1.000 184 | Start Experiment 15 185 | start_id: 49146 186 | 0/216, Acc for this batch: 0.000 187 | *************************************************************** 188 | Ground-truth Labels: 0,18,92 189 | Ground-truth Num of Instances: 1,214,1 190 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 191 | Our Recovered Num of Instances: 1,214,1 | LnAcc: 1.000 | IRec: 1.000 192 | Start Experiment 16 193 | start_id: 21107 194 | 0/216, Acc for this batch: 0.000 195 | *************************************************************** 196 | Ground-truth Labels: 0,18,92 197 | Ground-truth Num of Instances: 1,214,1 198 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 199 | Our Recovered Num of Instances: 1,214,1 | LnAcc: 1.000 | IRec: 1.000 200 | Start Experiment 17 201 | start_id: 41095 202 | 0/216, Acc for this batch: 0.000 203 | *************************************************************** 204 | Ground-truth Labels: 0,18,92 205 | Ground-truth Num of Instances: 1,214,1 206 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 207 | Our Recovered Num of Instances: 1,214,1 | LnAcc: 1.000 | IRec: 1.000 208 | Start Experiment 18 209 | start_id: 13914 210 | 0/216, Acc for this batch: 0.000 211 | *************************************************************** 212 | Ground-truth Labels: 0,18,92 213 | Ground-truth Num of Instances: 1,214,1 214 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 215 | Our Recovered Num of Instances: 1,214,1 | LnAcc: 1.000 | IRec: 1.000 216 | Start Experiment 19 217 | start_id: 35248 218 | 0/216, Acc for this batch: 0.000 219 | *************************************************************** 220 | Ground-truth Labels: 0,18,92 221 | Ground-truth Num of Instances: 1,214,1 222 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 223 | Our Recovered Num of Instances: 1,214,1 | LnAcc: 1.000 | IRec: 1.000 224 | Start Experiment 20 225 | start_id: 7650 226 | 0/216, Acc for this batch: 0.000 227 | *************************************************************** 228 | Ground-truth Labels: 0,18,92 229 | Ground-truth Num of Instances: 1,214,1 230 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 231 | Our Recovered Num of Instances: 1,214,1 | LnAcc: 1.000 | IRec: 1.000 232 | --------------------------------------------------------------- 233 | Mean Ours LeAcc: 1.000 | Mean Ours LnAcc: 1.000 | Mean Ours IRec: 1.000 234 | Mean Rec Instances: Class 0-1, Class 18-214, Class 92-1 235 | -------------------------------------------------------------------------------- /new_logs/ExtremeDistribution/exp4.log: -------------------------------------------------------------------------------- 1 | running experiments at 2022-11-13-15:38:47 2 | exp_name: extreme_distribution 3 | cpu: False 4 | num_tries: 20 5 | num_images: 648 6 | seed: 12 7 | alpha: 1 8 | simplified: False 9 | compare: False 10 | estimate: False 11 | analysis: False 12 | ratio: 0.0 13 | data_path: data 14 | dataset: CIFAR100 15 | split: train 16 | distribution: custom_imbalanced 17 | start_id: 0 18 | num_classes: 100 19 | num_uniform_cls: 32 20 | num_target_cls: 5 21 | max_size: 32 22 | model: vgg16 23 | trained_model: False 24 | iter_train: False 25 | iters: 1000 26 | epochs: 10 27 | batch_size: 128 28 | lr: 0.1 29 | optimizer: SGD 30 | scheduler: linear 31 | weight_decay: 0.0005 32 | warmup: False 33 | epoch_interval: 10 34 | iter_interval: 100 35 | mid_save: False 36 | model_path: models 37 | dryrun: False 38 | batchnorm: False 39 | dropout: False 40 | silu: False 41 | leaky_relu: False 42 | n_dim: 300 43 | n_hidden: 1 44 | defense: False 45 | defense_method: dp 46 | noise_std: 0.001 47 | clip_bound: 4 48 | sparse_ratio: 10 49 | prune_ratio: 10 50 | rec_img: False 51 | fix_labels: False 52 | gt_labels: False 53 | optim: ig 54 | restarts: 1 55 | cost_fn: sim 56 | indices: def 57 | weights: equal 58 | rec_lr: None 59 | rec_optimizer: adam 60 | signed: False 61 | boxed: False 62 | scoring_choice: loss 63 | init: randn 64 | tv: 1e-06 65 | l2: 1e-06 66 | max_iterations: 8000 67 | loss_thresh: 0.0001 68 | save_image: False 69 | log_dir: logs/extreme_distribution 70 | image_dir: images/extreme_distribution 71 | 72 | Start Experiment 1 73 | start_id: 0 74 | 0/648, Acc for this batch: 0.000 75 | *************************************************************** 76 | Ground-truth Labels: 0,18,92 77 | Ground-truth Num of Instances: 1,646,1 78 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 79 | Our Recovered Num of Instances: 1,646,1 | LnAcc: 1.000 | IRec: 1.000 80 | Start Experiment 2 81 | start_id: 14513 82 | 0/648, Acc for this batch: 0.000 83 | *************************************************************** 84 | Ground-truth Labels: 0,18,92 85 | Ground-truth Num of Instances: 1,646,1 86 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 87 | Our Recovered Num of Instances: 1,646,1 | LnAcc: 1.000 | IRec: 1.000 88 | Start Experiment 3 89 | start_id: 30093 90 | 0/648, Acc for this batch: 0.000 91 | *************************************************************** 92 | Ground-truth Labels: 0,18,92 93 | Ground-truth Num of Instances: 1,646,1 94 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 95 | Our Recovered Num of Instances: 1,646,1 | LnAcc: 1.000 | IRec: 1.000 96 | Start Experiment 4 97 | start_id: 42725 98 | 0/648, Acc for this batch: 0.000 99 | *************************************************************** 100 | Ground-truth Labels: 0,18,92 101 | Ground-truth Num of Instances: 1,646,1 102 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 103 | Our Recovered Num of Instances: 1,646,1 | LnAcc: 1.000 | IRec: 1.000 104 | Start Experiment 5 105 | start_id: 9646 106 | 0/648, Acc for this batch: 0.000 107 | *************************************************************** 108 | Ground-truth Labels: 0,18,92 109 | Ground-truth Num of Instances: 1,646,1 110 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 111 | Our Recovered Num of Instances: 1,646,1 | LnAcc: 1.000 | IRec: 1.000 112 | Start Experiment 6 113 | start_id: 22468 114 | 0/648, Acc for this batch: 0.000 115 | *************************************************************** 116 | Ground-truth Labels: 0,18,92 117 | Ground-truth Num of Instances: 1,646,1 118 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 119 | Our Recovered Num of Instances: 1,646,1 | LnAcc: 1.000 | IRec: 1.000 120 | Start Experiment 7 121 | start_id: 36803 122 | 0/648, Acc for this batch: 0.000 123 | *************************************************************** 124 | Ground-truth Labels: 0,18,92 125 | Ground-truth Num of Instances: 1,646,1 126 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 127 | Our Recovered Num of Instances: 1,646,1 | LnAcc: 1.000 | IRec: 1.000 128 | Start Experiment 8 129 | start_id: 2558 130 | 0/648, Acc for this batch: 0.000 131 | *************************************************************** 132 | Ground-truth Labels: 0,18,92 133 | Ground-truth Num of Instances: 1,646,1 134 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 135 | Our Recovered Num of Instances: 1,646,1 | LnAcc: 1.000 | IRec: 1.000 136 | Start Experiment 9 137 | start_id: 17229 138 | 0/648, Acc for this batch: 0.000 139 | *************************************************************** 140 | Ground-truth Labels: 0,18,92 141 | Ground-truth Num of Instances: 1,646,1 142 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 143 | Our Recovered Num of Instances: 1,646,1 | LnAcc: 1.000 | IRec: 1.000 144 | Start Experiment 10 145 | start_id: 31960 146 | 0/648, Acc for this batch: 0.000 147 | *************************************************************** 148 | Ground-truth Labels: 0,18,92 149 | Ground-truth Num of Instances: 1,646,1 150 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 151 | Our Recovered Num of Instances: 1,646,1 | LnAcc: 1.000 | IRec: 1.000 152 | Start Experiment 11 153 | start_id: 44939 154 | 0/648, Acc for this batch: 0.000 155 | *************************************************************** 156 | Ground-truth Labels: 0,18,92 157 | Ground-truth Num of Instances: 1,646,1 158 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 159 | Our Recovered Num of Instances: 1,646,1 | LnAcc: 1.000 | IRec: 1.000 160 | Start Experiment 12 161 | start_id: 11133 162 | 0/648, Acc for this batch: 0.000 163 | *************************************************************** 164 | Ground-truth Labels: 0,18,92 165 | Ground-truth Num of Instances: 1,646,1 166 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 167 | Our Recovered Num of Instances: 1,646,1 | LnAcc: 1.000 | IRec: 1.000 168 | Start Experiment 13 169 | start_id: 25396 170 | 0/648, Acc for this batch: 0.000 171 | *************************************************************** 172 | Ground-truth Labels: 0,18,92 173 | Ground-truth Num of Instances: 1,646,1 174 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 175 | Our Recovered Num of Instances: 1,646,1 | LnAcc: 1.000 | IRec: 1.000 176 | Start Experiment 14 177 | start_id: 38677 178 | 0/648, Acc for this batch: 0.000 179 | *************************************************************** 180 | Ground-truth Labels: 0,18,92 181 | Ground-truth Num of Instances: 1,646,1 182 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 183 | Our Recovered Num of Instances: 1,646,1 | LnAcc: 1.000 | IRec: 1.000 184 | Start Experiment 15 185 | start_id: 4787 186 | 0/648, Acc for this batch: 0.000 187 | *************************************************************** 188 | Ground-truth Labels: 0,18,92 189 | Ground-truth Num of Instances: 1,646,1 190 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 191 | Our Recovered Num of Instances: 1,646,1 | LnAcc: 1.000 | IRec: 1.000 192 | Start Experiment 16 193 | start_id: 19493 194 | 0/648, Acc for this batch: 0.000 195 | *************************************************************** 196 | Ground-truth Labels: 0,18,92 197 | Ground-truth Num of Instances: 1,646,1 198 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 199 | Our Recovered Num of Instances: 1,646,1 | LnAcc: 1.000 | IRec: 1.000 200 | Start Experiment 17 201 | start_id: 33289 202 | 0/648, Acc for this batch: 0.000 203 | *************************************************************** 204 | Ground-truth Labels: 0,18,92 205 | Ground-truth Num of Instances: 1,646,1 206 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 207 | Our Recovered Num of Instances: 1,646,1 | LnAcc: 1.000 | IRec: 1.000 208 | Start Experiment 18 209 | start_id: 47091 210 | 0/648, Acc for this batch: 0.000 211 | *************************************************************** 212 | Ground-truth Labels: 0,18,92 213 | Ground-truth Num of Instances: 1,646,1 214 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 215 | Our Recovered Num of Instances: 1,646,1 | LnAcc: 1.000 | IRec: 1.000 216 | Start Experiment 19 217 | start_id: 13173 218 | 0/648, Acc for this batch: 0.000 219 | *************************************************************** 220 | Ground-truth Labels: 0,18,92 221 | Ground-truth Num of Instances: 1,646,1 222 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 223 | Our Recovered Num of Instances: 1,646,1 | LnAcc: 1.000 | IRec: 1.000 224 | Start Experiment 20 225 | start_id: 27898 226 | 0/648, Acc for this batch: 0.000 227 | *************************************************************** 228 | Ground-truth Labels: 0,18,92 229 | Ground-truth Num of Instances: 1,646,1 230 | Our Recovered Labels: 0,18,92 | LeAcc: 1.000 231 | Our Recovered Num of Instances: 1,646,1 | LnAcc: 1.000 | IRec: 1.000 232 | --------------------------------------------------------------- 233 | Mean Ours LeAcc: 1.000 | Mean Ours LnAcc: 1.000 | Mean Ours IRec: 1.000 234 | Mean Rec Instances: Class 0-1, Class 18-646, Class 92-1 235 | -------------------------------------------------------------------------------- /new_logs/TrainingStage/MNIST/exp1.log: -------------------------------------------------------------------------------- 1 | running experiments at 2022-11-17-09:11:47 2 | exp_name: train_stage2 3 | cpu: False 4 | num_tries: 20 5 | num_images: 8 6 | seed: 12 7 | alpha: 1 8 | simplified: False 9 | compare: False 10 | estimate: False 11 | analysis: False 12 | ratio: 0.0 13 | data_path: data 14 | dataset: MNIST_GRAY 15 | split: train 16 | distribution: random 17 | start_id: 0 18 | num_classes: 100 19 | num_uniform_cls: 32 20 | num_target_cls: 5 21 | max_size: 32 22 | model: dnn 23 | trained_model: True 24 | iter_train: True 25 | iters: 1000 26 | epochs: 10 27 | batch_size: 128 28 | lr: 0.1 29 | optimizer: SGD 30 | scheduler: linear 31 | weight_decay: 0.0005 32 | warmup: False 33 | epoch_interval: 10 34 | iter_interval: 100 35 | mid_save: False 36 | model_path: models 37 | dryrun: False 38 | batchnorm: False 39 | dropout: False 40 | silu: False 41 | leaky_relu: False 42 | n_dim: 300 43 | n_hidden: 1 44 | defense: False 45 | defense_method: dp 46 | noise_std: 0.001 47 | clip_bound: 4 48 | sparse_ratio: 10 49 | prune_ratio: 10 50 | rec_img: False 51 | fix_labels: False 52 | gt_labels: False 53 | optim: ig 54 | restarts: 1 55 | cost_fn: sim 56 | indices: def 57 | weights: equal 58 | rec_lr: None 59 | rec_optimizer: adam 60 | signed: False 61 | boxed: False 62 | scoring_choice: loss 63 | init: randn 64 | tv: 1e-06 65 | l2: 1e-06 66 | max_iterations: 8000 67 | loss_thresh: 0.0001 68 | save_image: False 69 | log_dir: logs/train_stage2 70 | image_dir: images/train_stage2 71 | 72 | Model loaded from file dnn_MNIST_GRAY_Iter1000.pth. 73 | Start Experiment 1 74 | start_id: 0 75 | 8/8, Acc for this batch: 1.000 76 | *************************************************************** 77 | Ground-truth Labels: 0,3,4,5,8 78 | Ground-truth Num of Instances: 3,1,2,1,1 79 | Our Recovered Labels: 0,4,5 | LeAcc: 0.800 80 | Our Recovered Num of Instances: 1,4,2 | LnAcc: 0.500 | IRec: 0.500 81 | Start Experiment 2 82 | start_id: 11879 83 | 8/8, Acc for this batch: 1.000 84 | *************************************************************** 85 | Ground-truth Labels: 0,1,2,5,7,9 86 | Ground-truth Num of Instances: 1,1,1,2,1,2 87 | Our Recovered Labels: 0,1,9 | LeAcc: 0.700 88 | Our Recovered Num of Instances: 1,2,8 | LnAcc: 0.500 | IRec: 0.500 89 | Start Experiment 3 90 | start_id: 41168 91 | 7/8, Acc for this batch: 0.875 92 | *************************************************************** 93 | Ground-truth Labels: 0,4,5,7,9 94 | Ground-truth Num of Instances: 1,2,1,1,3 95 | Our Recovered Labels: 0,4,7 | LeAcc: 0.800 96 | Our Recovered Num of Instances: 1,2,4 | LnAcc: 0.700 | IRec: 0.500 97 | Start Experiment 4 98 | start_id: 12852 99 | 8/8, Acc for this batch: 1.000 100 | *************************************************************** 101 | Ground-truth Labels: 0,1,4,5,8,9 102 | Ground-truth Num of Instances: 1,1,1,2,2,1 103 | Our Recovered Labels: 3,4,5,8 | LeAcc: 0.600 104 | Our Recovered Num of Instances: 1,1,4,2 | LnAcc: 0.500 | IRec: 0.625 105 | Start Experiment 5 106 | start_id: 49967 107 | 8/8, Acc for this batch: 1.000 108 | *************************************************************** 109 | Ground-truth Labels: 0,2,3,5,9 110 | Ground-truth Num of Instances: 1,1,2,1,3 111 | Our Recovered Labels: 3,5,9 | LeAcc: 0.800 112 | Our Recovered Num of Instances: 2,1,5 | LnAcc: 0.700 | IRec: 0.750 113 | Start Experiment 6 114 | start_id: 13121 115 | 8/8, Acc for this batch: 1.000 116 | *************************************************************** 117 | Ground-truth Labels: 4,5,6,7,8 118 | Ground-truth Num of Instances: 1,1,1,1,4 119 | Our Recovered Labels: 7,8 | LeAcc: 0.700 120 | Our Recovered Num of Instances: 2,8 | LnAcc: 0.500 | IRec: 0.625 121 | Start Experiment 7 122 | start_id: 58442 123 | 6/8, Acc for this batch: 0.750 124 | *************************************************************** 125 | Ground-truth Labels: 1,2,3,4,5,6,7 126 | Ground-truth Num of Instances: 1,1,1,1,1,1,2 127 | Our Recovered Labels: 2,3,7 | LeAcc: 0.600 128 | Our Recovered Num of Instances: 2,1,1 | LnAcc: 0.400 | IRec: 0.375 129 | Start Experiment 8 130 | start_id: 17366 131 | 7/8, Acc for this batch: 0.875 132 | *************************************************************** 133 | Ground-truth Labels: 0,1,6,8,9 134 | Ground-truth Num of Instances: 3,1,2,1,1 135 | Our Recovered Labels: 2,6,8,9 | LeAcc: 0.700 136 | Our Recovered Num of Instances: 4,1,1,1 | LnAcc: 0.600 | IRec: 0.375 137 | Start Experiment 9 138 | start_id: 35251 139 | 7/8, Acc for this batch: 0.875 140 | *************************************************************** 141 | Ground-truth Labels: 2,3,5,6,7,9 142 | Ground-truth Num of Instances: 2,1,1,1,2,1 143 | Our Recovered Labels: 2,3,5,7 | LeAcc: 0.800 144 | Our Recovered Num of Instances: 2,1,2,2 | LnAcc: 0.700 | IRec: 0.750 145 | Start Experiment 10 146 | start_id: 39317 147 | 8/8, Acc for this batch: 1.000 148 | *************************************************************** 149 | Ground-truth Labels: 1,2,3,5,6,7,9 150 | Ground-truth Num of Instances: 1,2,1,1,1,1,1 151 | Our Recovered Labels: 2,4,7,9 | LeAcc: 0.500 152 | Our Recovered Num of Instances: 2,2,2,2 | LnAcc: 0.300 | IRec: 0.500 153 | Start Experiment 11 154 | start_id: 32817 155 | 8/8, Acc for this batch: 1.000 156 | *************************************************************** 157 | Ground-truth Labels: 1,2,4,7,8,9 158 | Ground-truth Num of Instances: 1,1,1,3,1,1 159 | Our Recovered Labels: 7,8,9 | LeAcc: 0.700 160 | Our Recovered Num of Instances: 1,1,7 | LnAcc: 0.500 | IRec: 0.375 161 | Start Experiment 12 162 | start_id: 12441 163 | 7/8, Acc for this batch: 0.875 164 | *************************************************************** 165 | Ground-truth Labels: 2,4,7,8,9 166 | Ground-truth Num of Instances: 2,1,2,2,1 167 | Our Recovered Labels: 2,7,8,9 | LeAcc: 0.900 168 | Our Recovered Num of Instances: 1,3,1,2 | LnAcc: 0.500 | IRec: 0.625 169 | Start Experiment 13 170 | start_id: 33651 171 | 7/8, Acc for this batch: 0.875 172 | *************************************************************** 173 | Ground-truth Labels: 2,3,5,6,7,8 174 | Ground-truth Num of Instances: 1,2,1,1,2,1 175 | Our Recovered Labels: 0,7,8 | LeAcc: 0.500 176 | Our Recovered Num of Instances: 4,1,2 | LnAcc: 0.300 | IRec: 0.250 177 | Start Experiment 14 178 | start_id: 17000 179 | 7/8, Acc for this batch: 0.875 180 | *************************************************************** 181 | Ground-truth Labels: 0,1,3,6,8,9 182 | Ground-truth Num of Instances: 2,2,1,1,1,1 183 | Our Recovered Labels: 3,5,7,8,9 | LeAcc: 0.500 184 | Our Recovered Num of Instances: 2,1,1,3,1 | LnAcc: 0.300 | IRec: 0.375 185 | Start Experiment 15 186 | start_id: 43072 187 | 8/8, Acc for this batch: 1.000 188 | *************************************************************** 189 | Ground-truth Labels: 0,1,3,4,5,8 190 | Ground-truth Num of Instances: 3,1,1,1,1,1 191 | Our Recovered Labels: | LeAcc: 0.400 192 | Our Recovered Num of Instances: | LnAcc: 0.400 | IRec: 0.000 193 | Start Experiment 16 194 | start_id: 32079 195 | 7/8, Acc for this batch: 0.875 196 | *************************************************************** 197 | Ground-truth Labels: 1,3,5,6,7,8,9 198 | Ground-truth Num of Instances: 1,1,1,2,1,1,1 199 | Our Recovered Labels: 9 | LeAcc: 0.400 200 | Our Recovered Num of Instances: 1 | LnAcc: 0.400 | IRec: 0.125 201 | Start Experiment 17 202 | start_id: 59661 203 | 6/8, Acc for this batch: 0.750 204 | *************************************************************** 205 | Ground-truth Labels: 3,4,5,6,8,9 206 | Ground-truth Num of Instances: 1,2,1,1,1,2 207 | Our Recovered Labels: 3,5,6,7,9 | LeAcc: 0.700 208 | Our Recovered Num of Instances: 1,3,2,1,2 | LnAcc: 0.500 | IRec: 0.625 209 | Start Experiment 18 210 | start_id: 44664 211 | 8/8, Acc for this batch: 1.000 212 | *************************************************************** 213 | Ground-truth Labels: 1,5,6,7,8,9 214 | Ground-truth Num of Instances: 1,1,2,1,1,2 215 | Our Recovered Labels: 6,7,8,9 | LeAcc: 0.800 216 | Our Recovered Num of Instances: 4,1,2,2 | LnAcc: 0.600 | IRec: 0.750 217 | Start Experiment 19 218 | start_id: 57949 219 | 8/8, Acc for this batch: 1.000 220 | *************************************************************** 221 | Ground-truth Labels: 0,2,3,5,8,9 222 | Ground-truth Num of Instances: 1,2,2,1,1,1 223 | Our Recovered Labels: 2,9 | LeAcc: 0.600 224 | Our Recovered Num of Instances: 1,1 | LnAcc: 0.500 | IRec: 0.250 225 | Start Experiment 20 226 | start_id: 27278 227 | 8/8, Acc for this batch: 1.000 228 | *************************************************************** 229 | Ground-truth Labels: 0,1,3,5,6,7,8 230 | Ground-truth Num of Instances: 1,1,1,1,1,2,1 231 | Our Recovered Labels: 0 | LeAcc: 0.400 232 | Our Recovered Num of Instances: 2 | LnAcc: 0.300 | IRec: 0.125 233 | --------------------------------------------------------------- 234 | Mean Ours LeAcc: 0.645 | Mean Ours LnAcc: 0.485 | Mean Ours IRec: 0.450 235 | -------------------------------------------------------------------------------- /new_logs/TrainingStage/MNIST/exp2.log: -------------------------------------------------------------------------------- 1 | running experiments at 2022-11-17-09:15:35 2 | exp_name: train_stage2 3 | cpu: False 4 | num_tries: 20 5 | num_images: 8 6 | seed: 12 7 | alpha: 1 8 | simplified: False 9 | compare: False 10 | estimate: False 11 | analysis: False 12 | ratio: 0.0 13 | data_path: data 14 | dataset: MNIST_GRAY 15 | split: train 16 | distribution: random 17 | start_id: 0 18 | num_classes: 100 19 | num_uniform_cls: 32 20 | num_target_cls: 5 21 | max_size: 32 22 | model: dnn 23 | trained_model: True 24 | iter_train: True 25 | iters: 900 26 | epochs: 10 27 | batch_size: 128 28 | lr: 0.1 29 | optimizer: SGD 30 | scheduler: linear 31 | weight_decay: 0.0005 32 | warmup: False 33 | epoch_interval: 10 34 | iter_interval: 100 35 | mid_save: False 36 | model_path: models 37 | dryrun: False 38 | batchnorm: False 39 | dropout: False 40 | silu: False 41 | leaky_relu: False 42 | n_dim: 300 43 | n_hidden: 1 44 | defense: False 45 | defense_method: dp 46 | noise_std: 0.001 47 | clip_bound: 4 48 | sparse_ratio: 10 49 | prune_ratio: 10 50 | rec_img: False 51 | fix_labels: False 52 | gt_labels: False 53 | optim: ig 54 | restarts: 1 55 | cost_fn: sim 56 | indices: def 57 | weights: equal 58 | rec_lr: None 59 | rec_optimizer: adam 60 | signed: False 61 | boxed: False 62 | scoring_choice: loss 63 | init: randn 64 | tv: 1e-06 65 | l2: 1e-06 66 | max_iterations: 8000 67 | loss_thresh: 0.0001 68 | save_image: False 69 | log_dir: logs/train_stage2 70 | image_dir: images/train_stage2 71 | 72 | Model loaded from file dnn_MNIST_GRAY_Iter900.pth. 73 | Start Experiment 1 74 | start_id: 0 75 | 8/8, Acc for this batch: 1.000 76 | *************************************************************** 77 | Ground-truth Labels: 0,3,4,5,8 78 | Ground-truth Num of Instances: 3,1,2,1,1 79 | Our Recovered Labels: 0,4,5 | LeAcc: 0.800 80 | Our Recovered Num of Instances: 1,4,2 | LnAcc: 0.500 | IRec: 0.500 81 | Start Experiment 2 82 | start_id: 11879 83 | 8/8, Acc for this batch: 1.000 84 | *************************************************************** 85 | Ground-truth Labels: 0,1,2,5,7,9 86 | Ground-truth Num of Instances: 1,1,1,2,1,2 87 | Our Recovered Labels: 0,1,9 | LeAcc: 0.700 88 | Our Recovered Num of Instances: 1,2,7 | LnAcc: 0.500 | IRec: 0.500 89 | Start Experiment 3 90 | start_id: 41168 91 | 7/8, Acc for this batch: 0.875 92 | *************************************************************** 93 | Ground-truth Labels: 0,4,5,7,9 94 | Ground-truth Num of Instances: 1,2,1,1,3 95 | Our Recovered Labels: 0,4,7 | LeAcc: 0.800 96 | Our Recovered Num of Instances: 1,2,4 | LnAcc: 0.700 | IRec: 0.500 97 | Start Experiment 4 98 | start_id: 12852 99 | 8/8, Acc for this batch: 1.000 100 | *************************************************************** 101 | Ground-truth Labels: 0,1,4,5,8,9 102 | Ground-truth Num of Instances: 1,1,1,2,2,1 103 | Our Recovered Labels: 3,4,5,8 | LeAcc: 0.600 104 | Our Recovered Num of Instances: 1,1,4,2 | LnAcc: 0.500 | IRec: 0.625 105 | Start Experiment 5 106 | start_id: 49967 107 | 8/8, Acc for this batch: 1.000 108 | *************************************************************** 109 | Ground-truth Labels: 0,2,3,5,9 110 | Ground-truth Num of Instances: 1,1,2,1,3 111 | Our Recovered Labels: 3,5,9 | LeAcc: 0.800 112 | Our Recovered Num of Instances: 2,1,5 | LnAcc: 0.700 | IRec: 0.750 113 | Start Experiment 6 114 | start_id: 13121 115 | 8/8, Acc for this batch: 1.000 116 | *************************************************************** 117 | Ground-truth Labels: 4,5,6,7,8 118 | Ground-truth Num of Instances: 1,1,1,1,4 119 | Our Recovered Labels: 7,8 | LeAcc: 0.700 120 | Our Recovered Num of Instances: 2,8 | LnAcc: 0.500 | IRec: 0.625 121 | Start Experiment 7 122 | start_id: 58442 123 | 6/8, Acc for this batch: 0.750 124 | *************************************************************** 125 | Ground-truth Labels: 1,2,3,4,5,6,7 126 | Ground-truth Num of Instances: 1,1,1,1,1,1,2 127 | Our Recovered Labels: 2,3,7 | LeAcc: 0.600 128 | Our Recovered Num of Instances: 2,1,1 | LnAcc: 0.400 | IRec: 0.375 129 | Start Experiment 8 130 | start_id: 17366 131 | 7/8, Acc for this batch: 0.875 132 | *************************************************************** 133 | Ground-truth Labels: 0,1,6,8,9 134 | Ground-truth Num of Instances: 3,1,2,1,1 135 | Our Recovered Labels: 0,6,9 | LeAcc: 0.800 136 | Our Recovered Num of Instances: 6,1,1 | LnAcc: 0.600 | IRec: 0.625 137 | Start Experiment 9 138 | start_id: 35251 139 | 7/8, Acc for this batch: 0.875 140 | *************************************************************** 141 | Ground-truth Labels: 2,3,5,6,7,9 142 | Ground-truth Num of Instances: 2,1,1,1,2,1 143 | Our Recovered Labels: 2,3,5,7 | LeAcc: 0.800 144 | Our Recovered Num of Instances: 2,1,2,2 | LnAcc: 0.700 | IRec: 0.750 145 | Start Experiment 10 146 | start_id: 39317 147 | 8/8, Acc for this batch: 1.000 148 | *************************************************************** 149 | Ground-truth Labels: 1,2,3,5,6,7,9 150 | Ground-truth Num of Instances: 1,2,1,1,1,1,1 151 | Our Recovered Labels: 2,4,7,9 | LeAcc: 0.500 152 | Our Recovered Num of Instances: 2,2,2,2 | LnAcc: 0.300 | IRec: 0.500 153 | Start Experiment 11 154 | start_id: 32817 155 | 8/8, Acc for this batch: 1.000 156 | *************************************************************** 157 | Ground-truth Labels: 1,2,4,7,8,9 158 | Ground-truth Num of Instances: 1,1,1,3,1,1 159 | Our Recovered Labels: 7,8,9 | LeAcc: 0.700 160 | Our Recovered Num of Instances: 1,1,7 | LnAcc: 0.500 | IRec: 0.375 161 | Start Experiment 12 162 | start_id: 12441 163 | 7/8, Acc for this batch: 0.875 164 | *************************************************************** 165 | Ground-truth Labels: 2,4,7,8,9 166 | Ground-truth Num of Instances: 2,1,2,2,1 167 | Our Recovered Labels: 2,7,8,9 | LeAcc: 0.900 168 | Our Recovered Num of Instances: 1,3,1,2 | LnAcc: 0.500 | IRec: 0.625 169 | Start Experiment 13 170 | start_id: 33651 171 | 7/8, Acc for this batch: 0.875 172 | *************************************************************** 173 | Ground-truth Labels: 2,3,5,6,7,8 174 | Ground-truth Num of Instances: 1,2,1,1,2,1 175 | Our Recovered Labels: 0,7,8 | LeAcc: 0.500 176 | Our Recovered Num of Instances: 4,1,2 | LnAcc: 0.300 | IRec: 0.250 177 | Start Experiment 14 178 | start_id: 17000 179 | 7/8, Acc for this batch: 0.875 180 | *************************************************************** 181 | Ground-truth Labels: 0,1,3,6,8,9 182 | Ground-truth Num of Instances: 2,2,1,1,1,1 183 | Our Recovered Labels: 3,5,7,8,9 | LeAcc: 0.500 184 | Our Recovered Num of Instances: 2,1,1,3,1 | LnAcc: 0.300 | IRec: 0.375 185 | Start Experiment 15 186 | start_id: 43072 187 | 8/8, Acc for this batch: 1.000 188 | *************************************************************** 189 | Ground-truth Labels: 0,1,3,4,5,8 190 | Ground-truth Num of Instances: 3,1,1,1,1,1 191 | Our Recovered Labels: | LeAcc: 0.400 192 | Our Recovered Num of Instances: | LnAcc: 0.400 | IRec: 0.000 193 | Start Experiment 16 194 | start_id: 32079 195 | 7/8, Acc for this batch: 0.875 196 | *************************************************************** 197 | Ground-truth Labels: 1,3,5,6,7,8,9 198 | Ground-truth Num of Instances: 1,1,1,2,1,1,1 199 | Our Recovered Labels: 6,9 | LeAcc: 0.500 200 | Our Recovered Num of Instances: 8,1 | LnAcc: 0.400 | IRec: 0.375 201 | Start Experiment 17 202 | start_id: 59661 203 | 6/8, Acc for this batch: 0.750 204 | *************************************************************** 205 | Ground-truth Labels: 3,4,5,6,8,9 206 | Ground-truth Num of Instances: 1,2,1,1,1,2 207 | Our Recovered Labels: 3,5,6,7,9 | LeAcc: 0.700 208 | Our Recovered Num of Instances: 1,3,2,1,2 | LnAcc: 0.500 | IRec: 0.625 209 | Start Experiment 18 210 | start_id: 44664 211 | 8/8, Acc for this batch: 1.000 212 | *************************************************************** 213 | Ground-truth Labels: 1,5,6,7,8,9 214 | Ground-truth Num of Instances: 1,1,2,1,1,2 215 | Our Recovered Labels: 6,7,8,9 | LeAcc: 0.800 216 | Our Recovered Num of Instances: 4,1,2,2 | LnAcc: 0.600 | IRec: 0.750 217 | Start Experiment 19 218 | start_id: 57949 219 | 8/8, Acc for this batch: 1.000 220 | *************************************************************** 221 | Ground-truth Labels: 0,2,3,5,8,9 222 | Ground-truth Num of Instances: 1,2,2,1,1,1 223 | Our Recovered Labels: 2,9 | LeAcc: 0.600 224 | Our Recovered Num of Instances: 1,1 | LnAcc: 0.500 | IRec: 0.250 225 | Start Experiment 20 226 | start_id: 27278 227 | 8/8, Acc for this batch: 1.000 228 | *************************************************************** 229 | Ground-truth Labels: 0,1,3,5,6,7,8 230 | Ground-truth Num of Instances: 1,1,1,1,1,2,1 231 | Our Recovered Labels: 0 | LeAcc: 0.400 232 | Our Recovered Num of Instances: 2 | LnAcc: 0.300 | IRec: 0.125 233 | --------------------------------------------------------------- 234 | Mean Ours LeAcc: 0.655 | Mean Ours LnAcc: 0.485 | Mean Ours IRec: 0.475 235 | -------------------------------------------------------------------------------- /new_logs/TrainingStage/MNIST/exp3.log: -------------------------------------------------------------------------------- 1 | running experiments at 2022-11-17-09:15:46 2 | exp_name: train_stage2 3 | cpu: False 4 | num_tries: 20 5 | num_images: 8 6 | seed: 12 7 | alpha: 1 8 | simplified: False 9 | compare: False 10 | estimate: False 11 | analysis: False 12 | ratio: 0.0 13 | data_path: data 14 | dataset: MNIST_GRAY 15 | split: train 16 | distribution: random 17 | start_id: 0 18 | num_classes: 100 19 | num_uniform_cls: 32 20 | num_target_cls: 5 21 | max_size: 32 22 | model: dnn 23 | trained_model: True 24 | iter_train: True 25 | iters: 800 26 | epochs: 10 27 | batch_size: 128 28 | lr: 0.1 29 | optimizer: SGD 30 | scheduler: linear 31 | weight_decay: 0.0005 32 | warmup: False 33 | epoch_interval: 10 34 | iter_interval: 100 35 | mid_save: False 36 | model_path: models 37 | dryrun: False 38 | batchnorm: False 39 | dropout: False 40 | silu: False 41 | leaky_relu: False 42 | n_dim: 300 43 | n_hidden: 1 44 | defense: False 45 | defense_method: dp 46 | noise_std: 0.001 47 | clip_bound: 4 48 | sparse_ratio: 10 49 | prune_ratio: 10 50 | rec_img: False 51 | fix_labels: False 52 | gt_labels: False 53 | optim: ig 54 | restarts: 1 55 | cost_fn: sim 56 | indices: def 57 | weights: equal 58 | rec_lr: None 59 | rec_optimizer: adam 60 | signed: False 61 | boxed: False 62 | scoring_choice: loss 63 | init: randn 64 | tv: 1e-06 65 | l2: 1e-06 66 | max_iterations: 8000 67 | loss_thresh: 0.0001 68 | save_image: False 69 | log_dir: logs/train_stage2 70 | image_dir: images/train_stage2 71 | 72 | Model loaded from file dnn_MNIST_GRAY_Iter800.pth. 73 | Start Experiment 1 74 | start_id: 0 75 | 8/8, Acc for this batch: 1.000 76 | *************************************************************** 77 | Ground-truth Labels: 0,3,4,5,8 78 | Ground-truth Num of Instances: 3,1,2,1,1 79 | Our Recovered Labels: 0,4,5 | LeAcc: 0.800 80 | Our Recovered Num of Instances: 1,4,2 | LnAcc: 0.500 | IRec: 0.500 81 | Start Experiment 2 82 | start_id: 11879 83 | 8/8, Acc for this batch: 1.000 84 | *************************************************************** 85 | Ground-truth Labels: 0,1,2,5,7,9 86 | Ground-truth Num of Instances: 1,1,1,2,1,2 87 | Our Recovered Labels: 0,1,9 | LeAcc: 0.700 88 | Our Recovered Num of Instances: 1,2,7 | LnAcc: 0.500 | IRec: 0.500 89 | Start Experiment 3 90 | start_id: 41168 91 | 7/8, Acc for this batch: 0.875 92 | *************************************************************** 93 | Ground-truth Labels: 0,4,5,7,9 94 | Ground-truth Num of Instances: 1,2,1,1,3 95 | Our Recovered Labels: 0,4,7 | LeAcc: 0.800 96 | Our Recovered Num of Instances: 1,2,4 | LnAcc: 0.700 | IRec: 0.500 97 | Start Experiment 4 98 | start_id: 12852 99 | 8/8, Acc for this batch: 1.000 100 | *************************************************************** 101 | Ground-truth Labels: 0,1,4,5,8,9 102 | Ground-truth Num of Instances: 1,1,1,2,2,1 103 | Our Recovered Labels: 4,5,8 | LeAcc: 0.700 104 | Our Recovered Num of Instances: 2,3,2 | LnAcc: 0.500 | IRec: 0.625 105 | Start Experiment 5 106 | start_id: 49967 107 | 8/8, Acc for this batch: 1.000 108 | *************************************************************** 109 | Ground-truth Labels: 0,2,3,5,9 110 | Ground-truth Num of Instances: 1,1,2,1,3 111 | Our Recovered Labels: 3,5,9 | LeAcc: 0.800 112 | Our Recovered Num of Instances: 2,2,5 | LnAcc: 0.600 | IRec: 0.750 113 | Start Experiment 6 114 | start_id: 13121 115 | 8/8, Acc for this batch: 1.000 116 | *************************************************************** 117 | Ground-truth Labels: 4,5,6,7,8 118 | Ground-truth Num of Instances: 1,1,1,1,4 119 | Our Recovered Labels: 7,8 | LeAcc: 0.700 120 | Our Recovered Num of Instances: 2,8 | LnAcc: 0.500 | IRec: 0.625 121 | Start Experiment 7 122 | start_id: 58442 123 | 6/8, Acc for this batch: 0.750 124 | *************************************************************** 125 | Ground-truth Labels: 1,2,3,4,5,6,7 126 | Ground-truth Num of Instances: 1,1,1,1,1,1,2 127 | Our Recovered Labels: 2,3,7 | LeAcc: 0.600 128 | Our Recovered Num of Instances: 1,1,1 | LnAcc: 0.500 | IRec: 0.375 129 | Start Experiment 8 130 | start_id: 17366 131 | 7/8, Acc for this batch: 0.875 132 | *************************************************************** 133 | Ground-truth Labels: 0,1,6,8,9 134 | Ground-truth Num of Instances: 3,1,2,1,1 135 | Our Recovered Labels: 2,6,8,9 | LeAcc: 0.700 136 | Our Recovered Num of Instances: 4,1,1,1 | LnAcc: 0.600 | IRec: 0.375 137 | Start Experiment 9 138 | start_id: 35251 139 | 7/8, Acc for this batch: 0.875 140 | *************************************************************** 141 | Ground-truth Labels: 2,3,5,6,7,9 142 | Ground-truth Num of Instances: 2,1,1,1,2,1 143 | Our Recovered Labels: 2,3,5,7 | LeAcc: 0.800 144 | Our Recovered Num of Instances: 2,1,3,2 | LnAcc: 0.700 | IRec: 0.750 145 | Start Experiment 10 146 | start_id: 39317 147 | 8/8, Acc for this batch: 1.000 148 | *************************************************************** 149 | Ground-truth Labels: 1,2,3,5,6,7,9 150 | Ground-truth Num of Instances: 1,2,1,1,1,1,1 151 | Our Recovered Labels: 2,4,7,9 | LeAcc: 0.500 152 | Our Recovered Num of Instances: 3,2,1,2 | LnAcc: 0.300 | IRec: 0.500 153 | Start Experiment 11 154 | start_id: 32817 155 | 8/8, Acc for this batch: 1.000 156 | *************************************************************** 157 | Ground-truth Labels: 1,2,4,7,8,9 158 | Ground-truth Num of Instances: 1,1,1,3,1,1 159 | Our Recovered Labels: 7,8,9 | LeAcc: 0.700 160 | Our Recovered Num of Instances: 1,1,6 | LnAcc: 0.500 | IRec: 0.375 161 | Start Experiment 12 162 | start_id: 12441 163 | 7/8, Acc for this batch: 0.875 164 | *************************************************************** 165 | Ground-truth Labels: 2,4,7,8,9 166 | Ground-truth Num of Instances: 2,1,2,2,1 167 | Our Recovered Labels: 2,7,8,9 | LeAcc: 0.900 168 | Our Recovered Num of Instances: 1,3,1,2 | LnAcc: 0.500 | IRec: 0.625 169 | Start Experiment 13 170 | start_id: 33651 171 | 7/8, Acc for this batch: 0.875 172 | *************************************************************** 173 | Ground-truth Labels: 2,3,5,6,7,8 174 | Ground-truth Num of Instances: 1,2,1,1,2,1 175 | Our Recovered Labels: 0,3,7,8 | LeAcc: 0.600 176 | Our Recovered Num of Instances: 4,1,1,2 | LnAcc: 0.300 | IRec: 0.375 177 | Start Experiment 14 178 | start_id: 17000 179 | 7/8, Acc for this batch: 0.875 180 | *************************************************************** 181 | Ground-truth Labels: 0,1,3,6,8,9 182 | Ground-truth Num of Instances: 2,2,1,1,1,1 183 | Our Recovered Labels: 3,5,7,8,9 | LeAcc: 0.500 184 | Our Recovered Num of Instances: 1,1,1,4,1 | LnAcc: 0.400 | IRec: 0.375 185 | Start Experiment 15 186 | start_id: 43072 187 | 8/8, Acc for this batch: 1.000 188 | *************************************************************** 189 | Ground-truth Labels: 0,1,3,4,5,8 190 | Ground-truth Num of Instances: 3,1,1,1,1,1 191 | Our Recovered Labels: | LeAcc: 0.400 192 | Our Recovered Num of Instances: | LnAcc: 0.400 | IRec: 0.000 193 | Start Experiment 16 194 | start_id: 32079 195 | 7/8, Acc for this batch: 0.875 196 | *************************************************************** 197 | Ground-truth Labels: 1,3,5,6,7,8,9 198 | Ground-truth Num of Instances: 1,1,1,2,1,1,1 199 | Our Recovered Labels: 9 | LeAcc: 0.400 200 | Our Recovered Num of Instances: 1 | LnAcc: 0.400 | IRec: 0.125 201 | Start Experiment 17 202 | start_id: 59661 203 | 6/8, Acc for this batch: 0.750 204 | *************************************************************** 205 | Ground-truth Labels: 3,4,5,6,8,9 206 | Ground-truth Num of Instances: 1,2,1,1,1,2 207 | Our Recovered Labels: 3,5,6,7,9 | LeAcc: 0.700 208 | Our Recovered Num of Instances: 1,2,2,1,2 | LnAcc: 0.500 | IRec: 0.625 209 | Start Experiment 18 210 | start_id: 44664 211 | 8/8, Acc for this batch: 1.000 212 | *************************************************************** 213 | Ground-truth Labels: 1,5,6,7,8,9 214 | Ground-truth Num of Instances: 1,1,2,1,1,2 215 | Our Recovered Labels: 6,7,8,9 | LeAcc: 0.800 216 | Our Recovered Num of Instances: 3,1,4,2 | LnAcc: 0.600 | IRec: 0.750 217 | Start Experiment 19 218 | start_id: 57949 219 | 8/8, Acc for this batch: 1.000 220 | *************************************************************** 221 | Ground-truth Labels: 0,2,3,5,8,9 222 | Ground-truth Num of Instances: 1,2,2,1,1,1 223 | Our Recovered Labels: 2,9 | LeAcc: 0.600 224 | Our Recovered Num of Instances: 1,2 | LnAcc: 0.400 | IRec: 0.250 225 | Start Experiment 20 226 | start_id: 27278 227 | 8/8, Acc for this batch: 1.000 228 | *************************************************************** 229 | Ground-truth Labels: 0,1,3,5,6,7,8 230 | Ground-truth Num of Instances: 1,1,1,1,1,2,1 231 | Our Recovered Labels: 0 | LeAcc: 0.400 232 | Our Recovered Num of Instances: 1 | LnAcc: 0.400 | IRec: 0.125 233 | --------------------------------------------------------------- 234 | Mean Ours LeAcc: 0.655 | Mean Ours LnAcc: 0.490 | Mean Ours IRec: 0.456 235 | --------------------------------------------------------------------------------