tutorials/adv_robust_gkp_demo.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "machine_shape": "hm",
      "gpuType": "V100",
      "collapsed_sections": [
        "cTOGA-N9VbUE",
        "u-Tbwum8ViOy"
      ]
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive', force_remount=True)\n",
        "\n",
        "!nvidia-smi"
      ],
      "metadata": {
        "id": "gylQve5WVZH8"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install dill\n",
        "!pip install adversarial-robustness-toolbox"
      ],
      "metadata": {
        "id": "YXLXlHmnVWu1"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "%cd /content/drive/MyDrive/adv_robust_gkp/"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LEBxnQVyBIgy",
        "outputId": "021270df-dec4-46b9-ced0-76c8acfe1c9d"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/content/drive/MyDrive/adv_robust_gkp\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Prune & Group according to SR-GKP (one-shot, post-train, data-agnostic)"
      ],
      "metadata": {
        "id": "cTOGA-N9VbUE"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!python gkp_main.py \\\n",
        "--exp_desc resnet20_srgkp_prune_demo \\\n",
        "--setting_dir /content/drive/MyDrive/adv_robust_gkp/settings/resnet20_srgkp_prune.json \\\n",
        "--dataset cifar10 \\\n",
        "--model_dir /content/drive/MyDrive/adv_robust_gkp/ckpts/resnet20_cifar10_ba.pt \\\n",
        "--output_folder_dir /content/drive/MyDrive/adv_robust_gkp/output/resnet20_pruned/ \\\n",
        "--task gkp_prune \\\n",
        "--adv_attack no_attack"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hi02oZBL3KTr",
        "outputId": "7652b674-70aa-4e0d-cd72-6ae8700c1564"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "2023-12-10 04:11:06,243 | INFO : Input setting file /content/drive/MyDrive/adv_robust_gkp/settings/resnet20_srgkp_prune.json loaded.\n",
            "2023-12-10 04:11:06,260 | INFO : Input setting file /content/drive/MyDrive/adv_robust_gkp/settings/resnet20_srgkp_prune.json saved to /content/drive/MyDrive/adv_robust_gkp/output/resnet20_pruned/input_setting.json.\n",
            "2023-12-10 04:11:06,263 | INFO : Using normalized dataset.\n",
            "2023-12-10 04:11:06,263 | INFO : Experiment (resnet20_srgkp_prune_demo) input registered to setting.\n",
            "2023-12-10 04:11:06,285 | INFO : Experiment resnet20_srgkp_prune_demo (task: gkp_prune) started at 2023-12-10T04:11:06.236574.\n",
            "2023-12-10 04:11:06,289 | INFO : All modules imported\n",
            "/content/drive/MyDrive/adv_robust_gkp/ckpts/resnet20_cifar10_ba.pt\n",
            "2023-12-10 04:11:08,082 | INFO : Model /content/drive/MyDrive/adv_robust_gkp/ckpts/resnet20_cifar10_ba.pt (type: <class 'models.resnet.CifarResNet'> loaded and saved to /content/drive/MyDrive/adv_robust_gkp/output/resnet20_pruned/input_model.pt.\n",
            "2023-12-10 04:11:08,082 | INFO : Model /content/drive/MyDrive/adv_robust_gkp/ckpts/resnet20_cifar10_ba.pt loaded (type: <class 'models.resnet.CifarResNet'>).\n",
            "2023-12-10 04:11:08,082 | INFO : Performing one-shot GKP pruning...\n",
            "2023-12-10 04:11:08,092 | INFO : clustering_method: Smoothness Snaking; n_clusters: 8; pruning_target_layers: [2, 3, 4]; grouping_target_layers: [2, 3, 4].\n",
            "2023-12-10 04:11:08,093 | INFO : pruning_strategy: smooth_cost_beam; pruning_rate: 0.4375.\n",
            "2023-12-10 04:11:08,093 | INFO : \tmetric: euclidean; inner_outer_balancer: auto; cost_smooth_balancer: 0.9; eval_kept_kernel_number: 1; beam_width: 2; smoothness_check_step: 3.\n",
            "2023-12-10 04:11:08,103 | INFO : Layer 2-0-0; Shape (16, 16, 3, 3) -> (16, 144); Method: Smoothness Snaking.\n",
            "2023-12-10 04:11:08,406 | INFO : Layer 2-0-2; Shape (16, 16, 3, 3) -> (16, 144); Method: Smoothness Snaking.\n",
            "2023-12-10 04:11:08,708 | INFO : Layer 2-1-0; Shape (16, 16, 3, 3) -> (16, 144); Method: Smoothness Snaking.\n",
            "2023-12-10 04:11:09,020 | INFO : Layer 2-1-2; Shape (16, 16, 3, 3) -> (16, 144); Method: Smoothness Snaking.\n",
            "2023-12-10 04:11:09,322 | INFO : Layer 2-2-0; Shape (16, 16, 3, 3) -> (16, 144); Method: Smoothness Snaking.\n",
            "2023-12-10 04:11:09,624 | INFO : Layer 2-2-2; Shape (16, 16, 3, 3) -> (16, 144); Method: Smoothness Snaking.\n",
            "2023-12-10 04:11:09,934 | INFO : Layer 3-0-0; Shape (32, 16, 3, 3) -> (32, 144); Method: Smoothness Snaking.\n",
            "2023-12-10 04:11:10,268 | INFO : Layer 3-0-2; Shape (32, 32, 3, 3) -> (32, 288); Method: Smoothness Snaking.\n",
            "2023-12-10 04:11:14,200 | INFO : Layer 3-1-0; Shape (32, 32, 3, 3) -> (32, 288); Method: Smoothness Snaking.\n",
            "2023-12-10 04:11:18,079 | INFO : Layer 3-1-2; Shape (32, 32, 3, 3) -> (32, 288); Method: Smoothness Snaking.\n",
            "2023-12-10 04:11:21,981 | INFO : Layer 3-2-0; Shape (32, 32, 3, 3) -> (32, 288); Method: Smoothness Snaking.\n",
            "2023-12-10 04:11:25,874 | INFO : Layer 3-2-2; Shape (32, 32, 3, 3) -> (32, 288); Method: Smoothness Snaking.\n",
            "2023-12-10 04:11:29,801 | INFO : Layer 4-0-0; Shape (64, 32, 3, 3) -> (64, 288); Method: Smoothness Snaking.\n",
            "2023-12-10 04:11:33,888 | INFO : Layer 4-0-2; Shape (64, 64, 3, 3) -> (64, 576); Method: Smoothness Snaking.\n",
            "2023-12-10 04:12:27,297 | INFO : Layer 4-1-0; Shape (64, 64, 3, 3) -> (64, 576); Method: Smoothness Snaking.\n",
            "2023-12-10 04:13:20,696 | INFO : Layer 4-1-2; Shape (64, 64, 3, 3) -> (64, 576); Method: Smoothness Snaking.\n",
            "2023-12-10 04:14:13,933 | INFO : Layer 4-2-0; Shape (64, 64, 3, 3) -> (64, 576); Method: Smoothness Snaking.\n",
            "2023-12-10 04:15:07,560 | INFO : Layer 4-2-2; Shape (64, 64, 3, 3) -> (64, 576); Method: Smoothness Snaking.\n",
            "2023-12-10 04:16:00,831 | INFO : Model /content/drive/MyDrive/adv_robust_gkp/ckpts/resnet20_cifar10_ba.pt trained and saved to /content/drive/MyDrive/adv_robust_gkp/output/resnet20_pruned/pruned.pt.\n",
            "2023-12-10 04:16:00,832 | INFO : Model /content/drive/MyDrive/adv_robust_gkp/ckpts/resnet20_cifar10_ba.pt <class 'prune_cifar.ModifiedResNet'> now pruned.\n",
            "GroupResNet pruning rate: 0.4375; kernel_gcd: 16\n",
            "2023-12-10 04:16:00,868 | INFO : Model /content/drive/MyDrive/adv_robust_gkp/ckpts/resnet20_cifar10_ba.pt trained and saved to /content/drive/MyDrive/adv_robust_gkp/output/resnet20_pruned/grouped.pt.\n",
            "2023-12-10 04:16:00,868 | INFO : Model /content/drive/MyDrive/adv_robust_gkp/ckpts/resnet20_cifar10_ba.pt <class 'group_cifar.GroupResNet'> now grouped.\n",
            "2023-12-10 04:16:00,870 | INFO : Saving output setting file to /content/drive/MyDrive/adv_robust_gkp/output/resnet20_pruned/output_setting.json...\n",
            "2023-12-10 04:16:00,872 | INFO : Experiment resnet20_srgkp_prune_demo (task: gkp_prune) ended at 2023-12-10T04:16:00.869104 duration: 0:04:54.632527.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Finetune (in a vanilla manner) the pruned & grouped model."
      ],
      "metadata": {
        "id": "u-Tbwum8ViOy"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!python main.py \\\n",
        "--exp_desc resnet20_srgkp_finetune \\\n",
        "--setting_dir /content/drive/MyDrive/adv_robust_gkp/settings/gkp_cifar10_finetune.json \\\n",
        "--dataset cifar10 \\\n",
        "--model_dir /content/drive/MyDrive/adv_robust_gkp/output/resnet20_pruned/grouped.pt \\\n",
        "--output_folder_dir /content/drive/MyDrive/adv_robust_gkp/output/resnet20_finetuned/ \\\n",
        "--task finetune \\\n",
        "--adv_attack attack"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KqrSY7lo4IWo",
        "outputId": "89c57567-53a6-4794-c4e4-f518ab63aab1"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "2023-12-10 04:18:12,929 | INFO : Input setting file /content/drive/MyDrive/adv_robust_gkp/settings/gkp_cifar10_finetune.json loaded.\n",
            "2023-12-10 04:18:12,932 | INFO : Input setting file /content/drive/MyDrive/adv_robust_gkp/settings/gkp_cifar10_finetune.json saved to /content/drive/MyDrive/adv_robust_gkp/output/resnet20_finetuned/input_setting.json.\n",
            "2023-12-10 04:18:12,933 | INFO : Using normalized dataset.\n",
            "2023-12-10 04:18:12,933 | INFO : Experiment (resnet20_srgkp_finetune) input registered to setting.\n",
            "2023-12-10 04:18:12,955 | INFO : Experiment resnet20_srgkp_finetune (task: finetune) started at 2023-12-10T04:18:12.923248.\n",
            "Files already downloaded and verified\n",
            "Files already downloaded and verified\n",
            "2023-12-10 04:18:15,377 | INFO : Dataset cifar10 loaded.\n",
            "/content/drive/MyDrive/adv_robust_gkp/output/resnet20_pruned/grouped.pt\n",
            "2023-12-10 04:18:15,385 | INFO : All modules imported\n",
            "2023-12-10 04:18:17,194 | INFO : Model /content/drive/MyDrive/adv_robust_gkp/output/resnet20_pruned/grouped.pt (type: <class 'group_cifar.GroupResNet'> loaded and saved to /content/drive/MyDrive/adv_robust_gkp/output/resnet20_finetuned/input_model.pt.\n",
            "2023-12-10 04:18:17,198 | INFO : Model /content/drive/MyDrive/adv_robust_gkp/output/resnet20_pruned/grouped.pt successfully wrapped as an ART classifier.\n",
            "2023-12-10 04:18:17,198 | INFO : Testing baseline performance of model /content/drive/MyDrive/adv_robust_gkp/output/resnet20_pruned/grouped.pt...\n",
            "2023-12-10 04:18:20,755 | INFO : \tBenign test acc: 25.40% \t(Best: 25.40% from Epoch #0) \tNew best model ckpt best_benign_test_acc.pt saved.\n",
            "2023-12-10 04:18:29,760 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 24.39 \t(Best: 24.39 from Epoch #0) \tNew best model ckpt best_<FGSM with eps=0.01>.pt saved.\n",
            "2023-12-10 04:18:37,424 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.84 \t(Best: 14.84 from Epoch #0) \tNew best model ckpt best_<FGSM with eps=0.1>.pt saved.\n",
            "2023-12-10 04:18:37,425 | INFO : Training classifier for 300 epochs...\n",
            "2023-12-10 04:18:50,225 | INFO : Epoch #1 trained (lr = [0.01]).\n",
            "2023-12-10 04:18:51,703 | INFO : \tBenign test acc: 85.03% \t(Best: 85.03% from Epoch #1) \tNew best model ckpt best_benign_test_acc.pt saved.\n",
            "2023-12-10 04:18:59,796 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.55 \t(Best: 64.55 from Epoch #1) \tNew best model ckpt best_<FGSM with eps=0.01>.pt saved.\n",
            "2023-12-10 04:19:07,370 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.08 \t(Best: 14.84 from Epoch #0) \n",
            "2023-12-10 04:19:19,845 | INFO : Epoch #2 trained (lr = [0.01]).\n",
            "2023-12-10 04:19:21,311 | INFO : \tBenign test acc: 80.90% \t(Best: 85.03% from Epoch #1) \n",
            "2023-12-10 04:19:29,356 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 60.13 \t(Best: 64.55 from Epoch #1) \n",
            "2023-12-10 04:19:36,998 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.11 \t(Best: 15.11 from Epoch #2) \tNew best model ckpt best_<FGSM with eps=0.1>.pt saved.\n",
            "2023-12-10 04:19:49,455 | INFO : Epoch #3 trained (lr = [0.01]).\n",
            "2023-12-10 04:19:50,904 | INFO : \tBenign test acc: 83.03% \t(Best: 85.03% from Epoch #1) \n",
            "2023-12-10 04:19:58,630 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 62.20 \t(Best: 64.55 from Epoch #1) \n",
            "2023-12-10 04:20:06,137 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 13.22 \t(Best: 15.11 from Epoch #2) \n",
            "2023-12-10 04:20:18,673 | INFO : Epoch #4 trained (lr = [0.01]).\n",
            "2023-12-10 04:20:20,129 | INFO : \tBenign test acc: 85.02% \t(Best: 85.03% from Epoch #1) \n",
            "2023-12-10 04:20:27,978 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.39 \t(Best: 64.55 from Epoch #1) \n",
            "2023-12-10 04:20:35,544 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.90 \t(Best: 16.90 from Epoch #4) \tNew best model ckpt best_<FGSM with eps=0.1>.pt saved.\n",
            "2023-12-10 04:20:48,009 | INFO : Epoch #5 trained (lr = [0.01]).\n",
            "2023-12-10 04:20:49,535 | INFO : \tBenign test acc: 85.41% \t(Best: 85.41% from Epoch #5) \tNew best model ckpt best_benign_test_acc.pt saved.\n",
            "2023-12-10 04:20:57,410 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.63 \t(Best: 65.63 from Epoch #5) \tNew best model ckpt best_<FGSM with eps=0.01>.pt saved.\n",
            "2023-12-10 04:21:04,920 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.97 \t(Best: 16.90 from Epoch #4) \n",
            "2023-12-10 04:21:17,503 | INFO : Epoch #6 trained (lr = [0.01]).\n",
            "2023-12-10 04:21:19,067 | INFO : \tBenign test acc: 85.55% \t(Best: 85.55% from Epoch #6) \tNew best model ckpt best_benign_test_acc.pt saved.\n",
            "2023-12-10 04:21:26,822 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.56 \t(Best: 65.63 from Epoch #5) \n",
            "2023-12-10 04:21:34,337 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 13.63 \t(Best: 16.90 from Epoch #4) \n",
            "2023-12-10 04:21:46,838 | INFO : Epoch #7 trained (lr = [0.01]).\n",
            "2023-12-10 04:21:48,316 | INFO : \tBenign test acc: 86.39% \t(Best: 86.39% from Epoch #7) \tNew best model ckpt best_benign_test_acc.pt saved.\n",
            "2023-12-10 04:21:56,198 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.20 \t(Best: 65.63 from Epoch #5) \n",
            "2023-12-10 04:22:03,703 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 12.85 \t(Best: 16.90 from Epoch #4) \n",
            "2023-12-10 04:22:16,409 | INFO : Epoch #8 trained (lr = [0.01]).\n",
            "2023-12-10 04:22:17,880 | INFO : \tBenign test acc: 85.78% \t(Best: 86.39% from Epoch #7) \n",
            "2023-12-10 04:22:25,717 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.02 \t(Best: 65.63 from Epoch #5) \n",
            "2023-12-10 04:22:33,742 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.38 \t(Best: 16.90 from Epoch #4) \n",
            "2023-12-10 04:22:46,149 | INFO : Epoch #9 trained (lr = [0.01]).\n",
            "2023-12-10 04:22:47,624 | INFO : \tBenign test acc: 86.20% \t(Best: 86.39% from Epoch #7) \n",
            "2023-12-10 04:22:55,529 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.31 \t(Best: 65.63 from Epoch #5) \n",
            "2023-12-10 04:23:03,081 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.39 \t(Best: 17.39 from Epoch #9) \tNew best model ckpt best_<FGSM with eps=0.1>.pt saved.\n",
            "2023-12-10 04:23:15,479 | INFO : Epoch #10 trained (lr = [0.01]).\n",
            "2023-12-10 04:23:16,963 | INFO : \tBenign test acc: 85.93% \t(Best: 86.39% from Epoch #7) \n",
            "2023-12-10 04:23:24,699 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.58 \t(Best: 65.63 from Epoch #5) \n",
            "2023-12-10 04:23:32,419 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.54 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:23:44,991 | INFO : Epoch #11 trained (lr = [0.01]).\n",
            "2023-12-10 04:23:46,448 | INFO : \tBenign test acc: 85.80% \t(Best: 86.39% from Epoch #7) \n",
            "2023-12-10 04:23:54,209 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.56 \t(Best: 65.63 from Epoch #5) \n",
            "2023-12-10 04:24:01,758 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.83 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:24:14,262 | INFO : Epoch #12 trained (lr = [0.01]).\n",
            "2023-12-10 04:24:15,730 | INFO : \tBenign test acc: 87.10% \t(Best: 87.10% from Epoch #12) \tNew best model ckpt best_benign_test_acc.pt saved.\n",
            "2023-12-10 04:24:23,500 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 67.56 \t(Best: 67.56 from Epoch #12) \tNew best model ckpt best_<FGSM with eps=0.01>.pt saved.\n",
            "2023-12-10 04:24:30,959 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.53 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:24:43,438 | INFO : Epoch #13 trained (lr = [0.01]).\n",
            "2023-12-10 04:24:44,861 | INFO : \tBenign test acc: 86.51% \t(Best: 87.10% from Epoch #12) \n",
            "2023-12-10 04:24:52,644 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 61.48 \t(Best: 67.56 from Epoch #12) \n",
            "2023-12-10 04:25:00,233 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.88 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:25:12,706 | INFO : Epoch #14 trained (lr = [0.01]).\n",
            "2023-12-10 04:25:14,209 | INFO : \tBenign test acc: 86.37% \t(Best: 87.10% from Epoch #12) \n",
            "2023-12-10 04:25:21,960 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.37 \t(Best: 67.56 from Epoch #12) \n",
            "2023-12-10 04:25:29,530 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 13.94 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:25:41,995 | INFO : Epoch #15 trained (lr = [0.01]).\n",
            "2023-12-10 04:25:43,506 | INFO : \tBenign test acc: 87.76% \t(Best: 87.76% from Epoch #15) \tNew best model ckpt best_benign_test_acc.pt saved.\n",
            "2023-12-10 04:25:51,394 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.26 \t(Best: 67.56 from Epoch #12) \n",
            "2023-12-10 04:25:59,028 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.44 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:26:11,594 | INFO : Epoch #16 trained (lr = [0.01]).\n",
            "2023-12-10 04:26:13,046 | INFO : \tBenign test acc: 86.01% \t(Best: 87.76% from Epoch #15) \n",
            "2023-12-10 04:26:20,961 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.74 \t(Best: 67.56 from Epoch #12) \n",
            "2023-12-10 04:26:28,448 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.41 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:26:40,985 | INFO : Epoch #17 trained (lr = [0.01]).\n",
            "2023-12-10 04:26:42,467 | INFO : \tBenign test acc: 88.47% \t(Best: 88.47% from Epoch #17) \tNew best model ckpt best_benign_test_acc.pt saved.\n",
            "2023-12-10 04:26:50,307 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.70 \t(Best: 67.56 from Epoch #12) \n",
            "2023-12-10 04:26:57,815 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.17 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:27:10,269 | INFO : Epoch #18 trained (lr = [0.01]).\n",
            "2023-12-10 04:27:11,737 | INFO : \tBenign test acc: 87.96% \t(Best: 88.47% from Epoch #17) \n",
            "2023-12-10 04:27:19,648 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.84 \t(Best: 67.56 from Epoch #12) \n",
            "2023-12-10 04:27:27,067 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.08 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:27:39,539 | INFO : Epoch #19 trained (lr = [0.01]).\n",
            "2023-12-10 04:27:41,015 | INFO : \tBenign test acc: 87.57% \t(Best: 88.47% from Epoch #17) \n",
            "2023-12-10 04:27:48,768 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 66.39 \t(Best: 67.56 from Epoch #12) \n",
            "2023-12-10 04:27:56,267 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 13.80 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:28:08,702 | INFO : Epoch #20 trained (lr = [0.01]).\n",
            "2023-12-10 04:28:10,185 | INFO : \tBenign test acc: 86.53% \t(Best: 88.47% from Epoch #17) \n",
            "2023-12-10 04:28:18,134 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 62.68 \t(Best: 67.56 from Epoch #12) \n",
            "2023-12-10 04:28:26,713 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.27 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:28:39,201 | INFO : Epoch #21 trained (lr = [0.01]).\n",
            "2023-12-10 04:28:40,719 | INFO : \tBenign test acc: 88.11% \t(Best: 88.47% from Epoch #17) \n",
            "2023-12-10 04:28:48,475 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.83 \t(Best: 67.56 from Epoch #12) \n",
            "2023-12-10 04:28:56,088 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.90 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:29:08,578 | INFO : Epoch #22 trained (lr = [0.01]).\n",
            "2023-12-10 04:29:10,102 | INFO : \tBenign test acc: 88.65% \t(Best: 88.65% from Epoch #22) \tNew best model ckpt best_benign_test_acc.pt saved.\n",
            "2023-12-10 04:29:17,927 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.90 \t(Best: 67.56 from Epoch #12) \n",
            "2023-12-10 04:29:25,360 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 12.83 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:29:38,022 | INFO : Epoch #23 trained (lr = [0.01]).\n",
            "2023-12-10 04:29:39,541 | INFO : \tBenign test acc: 86.94% \t(Best: 88.65% from Epoch #22) \n",
            "2023-12-10 04:29:47,256 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 59.90 \t(Best: 67.56 from Epoch #12) \n",
            "2023-12-10 04:29:54,795 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 13.15 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:30:07,619 | INFO : Epoch #24 trained (lr = [0.01]).\n",
            "2023-12-10 04:30:09,069 | INFO : \tBenign test acc: 88.53% \t(Best: 88.65% from Epoch #22) \n",
            "2023-12-10 04:30:17,061 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 67.56 \t(Best: 67.56 from Epoch #24) \tNew best model ckpt best_<FGSM with eps=0.01>.pt saved.\n",
            "2023-12-10 04:30:24,559 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.74 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:30:37,123 | INFO : Epoch #25 trained (lr = [0.01]).\n",
            "2023-12-10 04:30:38,611 | INFO : \tBenign test acc: 86.57% \t(Best: 88.65% from Epoch #22) \n",
            "2023-12-10 04:30:46,505 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 60.77 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:30:53,949 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 12.63 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:31:06,526 | INFO : Epoch #26 trained (lr = [0.01]).\n",
            "2023-12-10 04:31:08,023 | INFO : \tBenign test acc: 87.17% \t(Best: 88.65% from Epoch #22) \n",
            "2023-12-10 04:31:15,892 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 66.25 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:31:23,410 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.16 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:31:35,994 | INFO : Epoch #27 trained (lr = [0.01]).\n",
            "2023-12-10 04:31:37,417 | INFO : \tBenign test acc: 88.06% \t(Best: 88.65% from Epoch #22) \n",
            "2023-12-10 04:31:45,215 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.98 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:31:52,786 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.01 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:32:05,433 | INFO : Epoch #28 trained (lr = [0.01]).\n",
            "2023-12-10 04:32:06,960 | INFO : \tBenign test acc: 88.09% \t(Best: 88.65% from Epoch #22) \n",
            "2023-12-10 04:32:14,759 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 66.36 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:32:22,358 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.35 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:32:34,766 | INFO : Epoch #29 trained (lr = [0.01]).\n",
            "2023-12-10 04:32:36,309 | INFO : \tBenign test acc: 87.02% \t(Best: 88.65% from Epoch #22) \n",
            "2023-12-10 04:32:44,095 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 66.06 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:32:51,686 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.31 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:33:04,581 | INFO : Epoch #30 trained (lr = [0.01]).\n",
            "2023-12-10 04:33:06,065 | INFO : \tBenign test acc: 86.96% \t(Best: 88.65% from Epoch #22) \n",
            "2023-12-10 04:33:13,741 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 66.70 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:33:21,230 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.71 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:33:33,798 | INFO : Epoch #31 trained (lr = [0.01]).\n",
            "2023-12-10 04:33:35,266 | INFO : \tBenign test acc: 86.88% \t(Best: 88.65% from Epoch #22) \n",
            "2023-12-10 04:33:43,057 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.02 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:33:50,456 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.65 \t(Best: 17.39 from Epoch #9) \n",
            "2023-12-10 04:34:02,681 | INFO : Epoch #32 trained (lr = [0.01]).\n",
            "2023-12-10 04:34:04,151 | INFO : \tBenign test acc: 85.61% \t(Best: 88.65% from Epoch #22) \n",
            "2023-12-10 04:34:11,705 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 61.46 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:34:19,173 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.44 \t(Best: 17.44 from Epoch #32) \tNew best model ckpt best_<FGSM with eps=0.1>.pt saved.\n",
            "2023-12-10 04:34:31,478 | INFO : Epoch #33 trained (lr = [0.01]).\n",
            "2023-12-10 04:34:32,892 | INFO : \tBenign test acc: 87.73% \t(Best: 88.65% from Epoch #22) \n",
            "2023-12-10 04:34:40,533 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.30 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:34:48,066 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.35 \t(Best: 17.44 from Epoch #32) \n",
            "2023-12-10 04:35:00,495 | INFO : Epoch #34 trained (lr = [0.01]).\n",
            "2023-12-10 04:35:01,954 | INFO : \tBenign test acc: 89.18% \t(Best: 89.18% from Epoch #34) \tNew best model ckpt best_benign_test_acc.pt saved.\n",
            "2023-12-10 04:35:09,542 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.49 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:35:16,933 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 13.16 \t(Best: 17.44 from Epoch #32) \n",
            "2023-12-10 04:35:29,261 | INFO : Epoch #35 trained (lr = [0.01]).\n",
            "2023-12-10 04:35:30,690 | INFO : \tBenign test acc: 88.45% \t(Best: 89.18% from Epoch #34) \n",
            "2023-12-10 04:35:38,307 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.63 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:35:45,697 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 13.71 \t(Best: 17.44 from Epoch #32) \n",
            "2023-12-10 04:35:58,206 | INFO : Epoch #36 trained (lr = [0.01]).\n",
            "2023-12-10 04:35:59,613 | INFO : \tBenign test acc: 87.51% \t(Best: 89.18% from Epoch #34) \n",
            "2023-12-10 04:36:07,141 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 61.33 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:36:14,405 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 13.71 \t(Best: 17.44 from Epoch #32) \n",
            "2023-12-10 04:36:26,750 | INFO : Epoch #37 trained (lr = [0.01]).\n",
            "2023-12-10 04:36:28,203 | INFO : \tBenign test acc: 86.95% \t(Best: 89.18% from Epoch #34) \n",
            "2023-12-10 04:36:35,862 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.73 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:36:43,257 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.82 \t(Best: 17.82 from Epoch #37) \tNew best model ckpt best_<FGSM with eps=0.1>.pt saved.\n",
            "2023-12-10 04:36:55,481 | INFO : Epoch #38 trained (lr = [0.01]).\n",
            "2023-12-10 04:36:56,888 | INFO : \tBenign test acc: 86.33% \t(Best: 89.18% from Epoch #34) \n",
            "2023-12-10 04:37:04,584 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.50 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:37:11,869 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.81 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:37:24,243 | INFO : Epoch #39 trained (lr = [0.01]).\n",
            "2023-12-10 04:37:25,686 | INFO : \tBenign test acc: 88.28% \t(Best: 89.18% from Epoch #34) \n",
            "2023-12-10 04:37:33,588 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.30 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:37:40,903 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.38 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:37:53,125 | INFO : Epoch #40 trained (lr = [0.01]).\n",
            "2023-12-10 04:37:54,557 | INFO : \tBenign test acc: 87.85% \t(Best: 89.18% from Epoch #34) \n",
            "2023-12-10 04:38:02,184 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.28 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:38:09,605 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.51 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:38:22,058 | INFO : Epoch #41 trained (lr = [0.01]).\n",
            "2023-12-10 04:38:23,466 | INFO : \tBenign test acc: 87.97% \t(Best: 89.18% from Epoch #34) \n",
            "2023-12-10 04:38:31,127 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.35 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:38:38,426 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.45 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:38:50,688 | INFO : Epoch #42 trained (lr = [0.01]).\n",
            "2023-12-10 04:38:52,102 | INFO : \tBenign test acc: 87.06% \t(Best: 89.18% from Epoch #34) \n",
            "2023-12-10 04:38:59,858 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.47 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:39:07,149 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.00 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:39:19,408 | INFO : Epoch #43 trained (lr = [0.01]).\n",
            "2023-12-10 04:39:20,937 | INFO : \tBenign test acc: 89.25% \t(Best: 89.25% from Epoch #43) \tNew best model ckpt best_benign_test_acc.pt saved.\n",
            "2023-12-10 04:39:28,499 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 67.16 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:39:35,773 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.71 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:39:47,974 | INFO : Epoch #44 trained (lr = [0.01]).\n",
            "2023-12-10 04:39:49,467 | INFO : \tBenign test acc: 85.87% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:39:57,282 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 62.41 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:40:04,639 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.24 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:40:17,348 | INFO : Epoch #45 trained (lr = [0.01]).\n",
            "2023-12-10 04:40:18,824 | INFO : \tBenign test acc: 88.52% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:40:26,310 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.59 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:40:33,742 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.32 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:40:45,984 | INFO : Epoch #46 trained (lr = [0.01]).\n",
            "2023-12-10 04:40:47,518 | INFO : \tBenign test acc: 87.36% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:40:55,036 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.20 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:41:02,336 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 13.24 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:41:14,639 | INFO : Epoch #47 trained (lr = [0.01]).\n",
            "2023-12-10 04:41:16,084 | INFO : \tBenign test acc: 88.90% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:41:23,597 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.40 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:41:31,180 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.21 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:41:43,432 | INFO : Epoch #48 trained (lr = [0.01]).\n",
            "2023-12-10 04:41:44,954 | INFO : \tBenign test acc: 88.26% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:41:52,457 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 62.97 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:41:59,828 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.24 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:42:12,387 | INFO : Epoch #49 trained (lr = [0.01]).\n",
            "2023-12-10 04:42:13,811 | INFO : \tBenign test acc: 87.94% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:42:21,369 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.55 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:42:28,758 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.93 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:42:41,442 | INFO : Epoch #50 trained (lr = [0.01]).\n",
            "2023-12-10 04:42:42,873 | INFO : \tBenign test acc: 86.32% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:42:50,437 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 62.87 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:42:57,782 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.93 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:43:10,120 | INFO : Epoch #51 trained (lr = [0.01]).\n",
            "2023-12-10 04:43:11,630 | INFO : \tBenign test acc: 87.01% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:43:19,141 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 60.64 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:43:26,492 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.43 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:43:38,954 | INFO : Epoch #52 trained (lr = [0.01]).\n",
            "2023-12-10 04:43:40,402 | INFO : \tBenign test acc: 88.84% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:43:48,037 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.92 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:43:55,485 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 13.15 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:44:07,854 | INFO : Epoch #53 trained (lr = [0.01]).\n",
            "2023-12-10 04:44:09,305 | INFO : \tBenign test acc: 88.22% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:44:17,112 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.74 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:44:24,470 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.81 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:44:36,765 | INFO : Epoch #54 trained (lr = [0.01]).\n",
            "2023-12-10 04:44:38,225 | INFO : \tBenign test acc: 87.55% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:44:45,985 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.95 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:44:53,385 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.76 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:45:05,778 | INFO : Epoch #55 trained (lr = [0.01]).\n",
            "2023-12-10 04:45:07,200 | INFO : \tBenign test acc: 86.93% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:45:14,923 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.90 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:45:22,316 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.19 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:45:34,586 | INFO : Epoch #56 trained (lr = [0.01]).\n",
            "2023-12-10 04:45:36,032 | INFO : \tBenign test acc: 88.23% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:45:43,782 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.55 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:45:51,120 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 13.17 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:46:03,526 | INFO : Epoch #57 trained (lr = [0.01]).\n",
            "2023-12-10 04:46:04,977 | INFO : \tBenign test acc: 86.53% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:46:12,669 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 61.32 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:46:19,906 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.71 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:46:32,179 | INFO : Epoch #58 trained (lr = [0.01]).\n",
            "2023-12-10 04:46:33,586 | INFO : \tBenign test acc: 86.42% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:46:41,275 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 59.50 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:46:48,580 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 12.87 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:47:01,039 | INFO : Epoch #59 trained (lr = [0.01]).\n",
            "2023-12-10 04:47:02,487 | INFO : \tBenign test acc: 87.25% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:47:10,124 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.46 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:47:17,495 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.72 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:47:29,822 | INFO : Epoch #60 trained (lr = [0.01]).\n",
            "2023-12-10 04:47:31,272 | INFO : \tBenign test acc: 87.98% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:47:38,852 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 61.74 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:47:46,170 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 13.51 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:47:58,585 | INFO : Epoch #61 trained (lr = [0.01]).\n",
            "2023-12-10 04:48:00,083 | INFO : \tBenign test acc: 87.42% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:48:07,575 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.93 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:48:14,879 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 13.79 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:48:27,275 | INFO : Epoch #62 trained (lr = [0.01]).\n",
            "2023-12-10 04:48:28,716 | INFO : \tBenign test acc: 87.12% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:48:36,263 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 62.98 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:48:43,594 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 13.87 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:48:55,864 | INFO : Epoch #63 trained (lr = [0.01]).\n",
            "2023-12-10 04:48:57,312 | INFO : \tBenign test acc: 87.84% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:49:04,968 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.55 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:49:12,298 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.41 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:49:24,814 | INFO : Epoch #64 trained (lr = [0.01]).\n",
            "2023-12-10 04:49:26,284 | INFO : \tBenign test acc: 87.87% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:49:33,839 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.12 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:49:41,449 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 12.28 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:49:53,767 | INFO : Epoch #65 trained (lr = [0.01]).\n",
            "2023-12-10 04:49:55,280 | INFO : \tBenign test acc: 88.16% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:50:02,898 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.73 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:50:10,381 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.11 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:50:22,670 | INFO : Epoch #66 trained (lr = [0.01]).\n",
            "2023-12-10 04:50:24,118 | INFO : \tBenign test acc: 88.33% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:50:31,663 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.83 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:50:39,051 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.50 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:50:51,475 | INFO : Epoch #67 trained (lr = [0.01]).\n",
            "2023-12-10 04:50:52,927 | INFO : \tBenign test acc: 87.90% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:51:00,460 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.96 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:51:07,805 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 13.85 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:51:20,127 | INFO : Epoch #68 trained (lr = [0.01]).\n",
            "2023-12-10 04:51:21,543 | INFO : \tBenign test acc: 88.65% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:51:29,069 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.59 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:51:36,659 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.90 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:51:48,959 | INFO : Epoch #69 trained (lr = [0.01]).\n",
            "2023-12-10 04:51:50,403 | INFO : \tBenign test acc: 88.50% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:51:57,927 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.77 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:52:05,279 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.70 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:52:17,574 | INFO : Epoch #70 trained (lr = [0.01]).\n",
            "2023-12-10 04:52:18,988 | INFO : \tBenign test acc: 89.02% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:52:26,787 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 66.68 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:52:34,265 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.62 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:52:46,473 | INFO : Epoch #71 trained (lr = [0.01]).\n",
            "2023-12-10 04:52:47,889 | INFO : \tBenign test acc: 87.08% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:52:55,824 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.17 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:53:03,157 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.18 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:53:15,405 | INFO : Epoch #72 trained (lr = [0.01]).\n",
            "2023-12-10 04:53:16,805 | INFO : \tBenign test acc: 87.79% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:53:24,432 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 62.92 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:53:31,761 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.12 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:53:44,030 | INFO : Epoch #73 trained (lr = [0.01]).\n",
            "2023-12-10 04:53:45,435 | INFO : \tBenign test acc: 86.65% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:53:53,090 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 62.68 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:54:00,369 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.23 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:54:13,073 | INFO : Epoch #74 trained (lr = [0.01]).\n",
            "2023-12-10 04:54:14,567 | INFO : \tBenign test acc: 88.44% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:54:22,247 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 62.82 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:54:29,492 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 13.45 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:54:41,799 | INFO : Epoch #75 trained (lr = [0.01]).\n",
            "2023-12-10 04:54:43,205 | INFO : \tBenign test acc: 87.60% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:54:50,964 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.33 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:54:58,319 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.95 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:55:10,721 | INFO : Epoch #76 trained (lr = [0.01]).\n",
            "2023-12-10 04:55:12,229 | INFO : \tBenign test acc: 86.61% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:55:19,847 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.11 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:55:27,124 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.38 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:55:39,736 | INFO : Epoch #77 trained (lr = [0.01]).\n",
            "2023-12-10 04:55:41,240 | INFO : \tBenign test acc: 88.27% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:55:48,843 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.48 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:55:56,116 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 12.10 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:56:08,486 | INFO : Epoch #78 trained (lr = [0.01]).\n",
            "2023-12-10 04:56:09,981 | INFO : \tBenign test acc: 87.13% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:56:17,569 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 62.89 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:56:24,901 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.13 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:56:37,106 | INFO : Epoch #79 trained (lr = [0.01]).\n",
            "2023-12-10 04:56:38,608 | INFO : \tBenign test acc: 88.75% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:56:46,156 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.96 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:56:53,478 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 12.12 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:57:05,845 | INFO : Epoch #80 trained (lr = [0.01]).\n",
            "2023-12-10 04:57:07,314 | INFO : \tBenign test acc: 88.57% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:57:14,980 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.87 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:57:22,347 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.48 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:57:34,584 | INFO : Epoch #81 trained (lr = [0.01]).\n",
            "2023-12-10 04:57:36,044 | INFO : \tBenign test acc: 88.22% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:57:43,534 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.95 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:57:51,032 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.66 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:58:03,261 | INFO : Epoch #82 trained (lr = [0.01]).\n",
            "2023-12-10 04:58:04,706 | INFO : \tBenign test acc: 87.59% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:58:12,302 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.49 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:58:19,644 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.67 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:58:31,890 | INFO : Epoch #83 trained (lr = [0.01]).\n",
            "2023-12-10 04:58:33,345 | INFO : \tBenign test acc: 88.64% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:58:40,866 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.57 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:58:48,238 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.39 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:59:00,533 | INFO : Epoch #84 trained (lr = [0.01]).\n",
            "2023-12-10 04:59:01,988 | INFO : \tBenign test acc: 86.76% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:59:09,782 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 60.83 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:59:17,406 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 13.49 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:59:29,672 | INFO : Epoch #85 trained (lr = [0.01]).\n",
            "2023-12-10 04:59:31,094 | INFO : \tBenign test acc: 87.35% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 04:59:38,687 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 61.04 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 04:59:46,130 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 13.52 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 04:59:58,292 | INFO : Epoch #86 trained (lr = [0.01]).\n",
            "2023-12-10 04:59:59,728 | INFO : \tBenign test acc: 87.13% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 05:00:07,373 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.90 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 05:00:14,752 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.72 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:00:27,406 | INFO : Epoch #87 trained (lr = [0.01]).\n",
            "2023-12-10 05:00:28,814 | INFO : \tBenign test acc: 88.19% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 05:00:36,386 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.39 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 05:00:43,723 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 12.76 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:00:56,351 | INFO : Epoch #88 trained (lr = [0.01]).\n",
            "2023-12-10 05:00:57,818 | INFO : \tBenign test acc: 86.88% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 05:01:05,456 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 61.92 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 05:01:12,708 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.18 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:01:25,551 | INFO : Epoch #89 trained (lr = [0.01]).\n",
            "2023-12-10 05:01:27,016 | INFO : \tBenign test acc: 88.62% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 05:01:34,655 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.42 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 05:01:41,938 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.14 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:01:54,232 | INFO : Epoch #90 trained (lr = [0.01]).\n",
            "2023-12-10 05:01:55,663 | INFO : \tBenign test acc: 87.68% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 05:02:03,464 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.83 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 05:02:10,876 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 13.09 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:02:23,231 | INFO : Epoch #91 trained (lr = [0.01]).\n",
            "2023-12-10 05:02:24,660 | INFO : \tBenign test acc: 88.72% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 05:02:32,476 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 66.09 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 05:02:39,733 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.97 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:02:51,969 | INFO : Epoch #92 trained (lr = [0.01]).\n",
            "2023-12-10 05:02:53,395 | INFO : \tBenign test acc: 87.76% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 05:03:01,046 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 60.07 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 05:03:08,301 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.19 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:03:20,742 | INFO : Epoch #93 trained (lr = [0.01]).\n",
            "2023-12-10 05:03:22,170 | INFO : \tBenign test acc: 88.17% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 05:03:29,774 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.43 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 05:03:37,100 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 13.90 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:03:49,446 | INFO : Epoch #94 trained (lr = [0.01]).\n",
            "2023-12-10 05:03:50,894 | INFO : \tBenign test acc: 88.26% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 05:03:58,684 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.79 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 05:04:06,018 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 12.44 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:04:18,298 | INFO : Epoch #95 trained (lr = [0.01]).\n",
            "2023-12-10 05:04:19,722 | INFO : \tBenign test acc: 88.00% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 05:04:27,501 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.69 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 05:04:34,883 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.04 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:04:47,175 | INFO : Epoch #96 trained (lr = [0.01]).\n",
            "2023-12-10 05:04:48,656 | INFO : \tBenign test acc: 86.94% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 05:04:56,248 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.60 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 05:05:03,615 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.07 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:05:15,917 | INFO : Epoch #97 trained (lr = [0.01]).\n",
            "2023-12-10 05:05:17,373 | INFO : \tBenign test acc: 88.19% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 05:05:25,063 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.12 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 05:05:32,494 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.71 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:05:44,716 | INFO : Epoch #98 trained (lr = [0.01]).\n",
            "2023-12-10 05:05:46,243 | INFO : \tBenign test acc: 87.81% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 05:05:53,848 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 62.12 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 05:06:01,298 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.11 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:06:13,826 | INFO : Epoch #99 trained (lr = [0.01]).\n",
            "2023-12-10 05:06:15,287 | INFO : \tBenign test acc: 87.81% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 05:06:22,820 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 62.30 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 05:06:30,236 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.21 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:06:42,698 | INFO : Epoch #100 trained (lr = [0.01]).\n",
            "2023-12-10 05:06:44,114 | INFO : \tBenign test acc: 88.35% \t(Best: 89.25% from Epoch #43) \n",
            "2023-12-10 05:06:51,756 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.90 \t(Best: 67.56 from Epoch #24) \n",
            "2023-12-10 05:06:59,204 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.10 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:07:11,691 | INFO : Epoch #101 trained (lr = [0.001]).\n",
            "2023-12-10 05:07:13,168 | INFO : \tBenign test acc: 91.40% \t(Best: 91.40% from Epoch #101) \tNew best model ckpt best_benign_test_acc.pt saved.\n",
            "2023-12-10 05:07:20,866 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 68.72 \t(Best: 68.72 from Epoch #101) \tNew best model ckpt best_<FGSM with eps=0.01>.pt saved.\n",
            "2023-12-10 05:07:28,301 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.08 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:07:40,835 | INFO : Epoch #102 trained (lr = [0.001]).\n",
            "2023-12-10 05:07:42,285 | INFO : \tBenign test acc: 91.30% \t(Best: 91.40% from Epoch #101) \n",
            "2023-12-10 05:07:50,108 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 68.62 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:07:57,810 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.36 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:08:10,567 | INFO : Epoch #103 trained (lr = [0.001]).\n",
            "2023-12-10 05:08:12,053 | INFO : \tBenign test acc: 91.66% \t(Best: 91.66% from Epoch #103) \tNew best model ckpt best_benign_test_acc.pt saved.\n",
            "2023-12-10 05:08:19,990 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 68.07 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:08:27,554 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.21 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:08:40,129 | INFO : Epoch #104 trained (lr = [0.001]).\n",
            "2023-12-10 05:08:41,603 | INFO : \tBenign test acc: 91.70% \t(Best: 91.70% from Epoch #104) \tNew best model ckpt best_benign_test_acc.pt saved.\n",
            "2023-12-10 05:08:49,871 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 67.40 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:08:57,998 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.32 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:09:10,552 | INFO : Epoch #105 trained (lr = [0.001]).\n",
            "2023-12-10 05:09:12,070 | INFO : \tBenign test acc: 91.82% \t(Best: 91.82% from Epoch #105) \tNew best model ckpt best_benign_test_acc.pt saved.\n",
            "2023-12-10 05:09:19,916 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 68.12 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:09:27,475 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.58 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:09:40,007 | INFO : Epoch #106 trained (lr = [0.001]).\n",
            "2023-12-10 05:09:41,500 | INFO : \tBenign test acc: 91.63% \t(Best: 91.82% from Epoch #105) \n",
            "2023-12-10 05:09:49,563 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 68.42 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:09:57,087 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.08 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:10:09,651 | INFO : Epoch #107 trained (lr = [0.001]).\n",
            "2023-12-10 05:10:11,136 | INFO : \tBenign test acc: 91.69% \t(Best: 91.82% from Epoch #105) \n",
            "2023-12-10 05:10:18,969 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 67.85 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:10:26,638 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.68 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:10:39,311 | INFO : Epoch #108 trained (lr = [0.001]).\n",
            "2023-12-10 05:10:40,835 | INFO : \tBenign test acc: 91.91% \t(Best: 91.91% from Epoch #108) \tNew best model ckpt best_benign_test_acc.pt saved.\n",
            "2023-12-10 05:10:48,597 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 67.35 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:10:56,302 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.27 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:11:08,998 | INFO : Epoch #109 trained (lr = [0.001]).\n",
            "2023-12-10 05:11:10,505 | INFO : \tBenign test acc: 91.97% \t(Best: 91.97% from Epoch #109) \tNew best model ckpt best_benign_test_acc.pt saved.\n",
            "2023-12-10 05:11:18,415 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 67.95 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:11:25,967 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.11 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:11:38,478 | INFO : Epoch #110 trained (lr = [0.001]).\n",
            "2023-12-10 05:11:39,927 | INFO : \tBenign test acc: 91.72% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:11:47,837 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 67.19 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:11:55,357 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.70 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:12:07,897 | INFO : Epoch #111 trained (lr = [0.001]).\n",
            "2023-12-10 05:12:09,314 | INFO : \tBenign test acc: 91.73% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:12:17,257 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 67.93 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:12:24,769 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.80 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:12:37,234 | INFO : Epoch #112 trained (lr = [0.001]).\n",
            "2023-12-10 05:12:38,656 | INFO : \tBenign test acc: 91.74% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:12:46,554 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 67.25 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:12:54,089 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.36 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:13:06,717 | INFO : Epoch #113 trained (lr = [0.001]).\n",
            "2023-12-10 05:13:08,151 | INFO : \tBenign test acc: 91.90% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:13:16,095 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 67.02 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:13:23,564 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.41 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:13:36,395 | INFO : Epoch #114 trained (lr = [0.001]).\n",
            "2023-12-10 05:13:37,823 | INFO : \tBenign test acc: 91.80% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:13:45,643 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 67.34 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:13:53,128 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.19 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:14:05,901 | INFO : Epoch #115 trained (lr = [0.001]).\n",
            "2023-12-10 05:14:07,338 | INFO : \tBenign test acc: 91.71% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:14:15,239 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 66.50 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:14:22,800 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.29 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:14:35,292 | INFO : Epoch #116 trained (lr = [0.001]).\n",
            "2023-12-10 05:14:36,807 | INFO : \tBenign test acc: 91.89% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:14:44,671 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 67.34 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:14:52,366 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.76 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:15:04,833 | INFO : Epoch #117 trained (lr = [0.001]).\n",
            "2023-12-10 05:15:06,359 | INFO : \tBenign test acc: 91.61% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:15:14,184 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 66.77 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:15:21,833 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.75 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:15:34,332 | INFO : Epoch #118 trained (lr = [0.001]).\n",
            "2023-12-10 05:15:35,769 | INFO : \tBenign test acc: 91.84% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:15:43,618 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 67.59 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:15:51,191 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.10 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:16:03,682 | INFO : Epoch #119 trained (lr = [0.001]).\n",
            "2023-12-10 05:16:05,133 | INFO : \tBenign test acc: 91.80% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:16:12,930 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 67.23 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:16:20,481 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.84 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:16:32,942 | INFO : Epoch #120 trained (lr = [0.001]).\n",
            "2023-12-10 05:16:34,439 | INFO : \tBenign test acc: 91.93% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:16:42,200 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 66.98 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:16:49,770 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.38 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:17:02,334 | INFO : Epoch #121 trained (lr = [0.001]).\n",
            "2023-12-10 05:17:03,787 | INFO : \tBenign test acc: 91.87% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:17:11,814 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 67.00 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:17:19,487 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.39 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:17:31,906 | INFO : Epoch #122 trained (lr = [0.001]).\n",
            "2023-12-10 05:17:33,352 | INFO : \tBenign test acc: 91.91% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:17:41,195 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.58 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:17:48,782 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.34 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:18:01,369 | INFO : Epoch #123 trained (lr = [0.001]).\n",
            "2023-12-10 05:18:02,817 | INFO : \tBenign test acc: 91.55% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:18:10,679 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 66.30 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:18:18,249 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.17 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:18:30,623 | INFO : Epoch #124 trained (lr = [0.001]).\n",
            "2023-12-10 05:18:32,075 | INFO : \tBenign test acc: 91.79% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:18:39,977 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 66.41 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:18:47,776 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.55 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:19:00,287 | INFO : Epoch #125 trained (lr = [0.001]).\n",
            "2023-12-10 05:19:01,747 | INFO : \tBenign test acc: 91.61% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:19:09,588 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 66.04 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:19:17,129 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.68 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:19:29,592 | INFO : Epoch #126 trained (lr = [0.001]).\n",
            "2023-12-10 05:19:31,081 | INFO : \tBenign test acc: 91.67% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:19:38,890 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 66.48 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:19:46,481 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.68 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:19:58,999 | INFO : Epoch #127 trained (lr = [0.001]).\n",
            "2023-12-10 05:20:00,484 | INFO : \tBenign test acc: 91.95% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:20:08,264 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 66.00 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:20:15,986 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.26 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:20:28,490 | INFO : Epoch #128 trained (lr = [0.001]).\n",
            "2023-12-10 05:20:29,993 | INFO : \tBenign test acc: 91.92% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:20:37,717 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 66.57 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:20:45,371 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.79 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:20:58,151 | INFO : Epoch #129 trained (lr = [0.001]).\n",
            "2023-12-10 05:20:59,572 | INFO : \tBenign test acc: 91.83% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:21:07,400 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.74 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:21:15,061 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.28 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:21:27,630 | INFO : Epoch #130 trained (lr = [0.001]).\n",
            "2023-12-10 05:21:29,086 | INFO : \tBenign test acc: 91.58% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:21:37,004 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.05 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:21:44,733 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.24 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:21:57,324 | INFO : Epoch #131 trained (lr = [0.001]).\n",
            "2023-12-10 05:21:58,760 | INFO : \tBenign test acc: 91.88% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:22:06,449 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.81 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:22:14,023 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.35 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:22:26,755 | INFO : Epoch #132 trained (lr = [0.001]).\n",
            "2023-12-10 05:22:28,184 | INFO : \tBenign test acc: 91.75% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:22:35,984 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.81 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:22:43,606 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.32 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:22:56,276 | INFO : Epoch #133 trained (lr = [0.001]).\n",
            "2023-12-10 05:22:57,765 | INFO : \tBenign test acc: 91.83% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:23:05,955 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 66.26 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:23:13,474 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.36 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:23:26,188 | INFO : Epoch #134 trained (lr = [0.001]).\n",
            "2023-12-10 05:23:27,632 | INFO : \tBenign test acc: 91.89% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:23:35,465 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.77 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:23:43,015 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.82 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:23:55,528 | INFO : Epoch #135 trained (lr = [0.001]).\n",
            "2023-12-10 05:23:57,033 | INFO : \tBenign test acc: 91.66% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:24:04,874 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 66.06 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:24:12,519 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.59 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:24:24,987 | INFO : Epoch #136 trained (lr = [0.001]).\n",
            "2023-12-10 05:24:26,448 | INFO : \tBenign test acc: 91.59% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:24:34,215 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.90 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:24:41,700 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.57 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:24:54,242 | INFO : Epoch #137 trained (lr = [0.001]).\n",
            "2023-12-10 05:24:55,711 | INFO : \tBenign test acc: 91.58% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:25:03,524 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.26 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:25:11,061 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.71 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:25:23,539 | INFO : Epoch #138 trained (lr = [0.001]).\n",
            "2023-12-10 05:25:24,977 | INFO : \tBenign test acc: 91.62% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:25:32,734 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 66.97 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:25:40,580 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.25 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:25:53,020 | INFO : Epoch #139 trained (lr = [0.001]).\n",
            "2023-12-10 05:25:54,449 | INFO : \tBenign test acc: 91.41% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:26:02,211 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 66.12 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:26:09,768 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.79 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:26:22,411 | INFO : Epoch #140 trained (lr = [0.001]).\n",
            "2023-12-10 05:26:23,835 | INFO : \tBenign test acc: 91.41% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:26:31,668 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.45 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:26:39,222 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.07 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:26:52,151 | INFO : Epoch #141 trained (lr = [0.001]).\n",
            "2023-12-10 05:26:53,623 | INFO : \tBenign test acc: 91.47% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:27:01,473 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.93 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:27:09,086 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.35 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:27:21,557 | INFO : Epoch #142 trained (lr = [0.001]).\n",
            "2023-12-10 05:27:22,986 | INFO : \tBenign test acc: 91.40% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:27:30,904 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.85 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:27:38,390 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.26 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:27:51,042 | INFO : Epoch #143 trained (lr = [0.001]).\n",
            "2023-12-10 05:27:52,486 | INFO : \tBenign test acc: 91.53% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:28:00,334 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.99 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:28:07,906 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.43 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:28:20,460 | INFO : Epoch #144 trained (lr = [0.001]).\n",
            "2023-12-10 05:28:21,879 | INFO : \tBenign test acc: 91.51% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:28:29,722 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.15 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:28:37,232 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.58 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:28:49,928 | INFO : Epoch #145 trained (lr = [0.001]).\n",
            "2023-12-10 05:28:51,436 | INFO : \tBenign test acc: 91.58% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:28:59,184 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.14 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:29:06,723 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.12 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:29:19,338 | INFO : Epoch #146 trained (lr = [0.001]).\n",
            "2023-12-10 05:29:20,847 | INFO : \tBenign test acc: 91.54% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:29:28,660 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.15 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:29:36,264 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.35 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:29:48,829 | INFO : Epoch #147 trained (lr = [0.001]).\n",
            "2023-12-10 05:29:50,325 | INFO : \tBenign test acc: 91.61% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:29:58,057 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.62 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:30:05,643 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.86 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:30:18,929 | INFO : Epoch #148 trained (lr = [0.001]).\n",
            "2023-12-10 05:30:20,402 | INFO : \tBenign test acc: 91.72% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:30:28,207 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.24 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:30:35,849 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.37 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:30:48,630 | INFO : Epoch #149 trained (lr = [0.001]).\n",
            "2023-12-10 05:30:50,066 | INFO : \tBenign test acc: 91.61% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:30:57,967 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.72 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:31:05,686 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.08 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:31:18,442 | INFO : Epoch #150 trained (lr = [0.001]).\n",
            "2023-12-10 05:31:19,878 | INFO : \tBenign test acc: 91.23% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:31:27,973 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.49 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:31:35,519 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.08 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:31:47,957 | INFO : Epoch #151 trained (lr = [0.001]).\n",
            "2023-12-10 05:31:49,400 | INFO : \tBenign test acc: 91.69% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:31:57,264 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.10 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:32:04,767 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.80 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:32:17,434 | INFO : Epoch #152 trained (lr = [0.001]).\n",
            "2023-12-10 05:32:18,893 | INFO : \tBenign test acc: 91.32% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:32:26,822 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.34 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:32:34,437 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.57 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:32:46,974 | INFO : Epoch #153 trained (lr = [0.001]).\n",
            "2023-12-10 05:32:48,430 | INFO : \tBenign test acc: 91.31% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:32:56,402 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.05 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:33:04,050 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.59 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:33:16,597 | INFO : Epoch #154 trained (lr = [0.001]).\n",
            "2023-12-10 05:33:18,101 | INFO : \tBenign test acc: 91.41% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:33:25,964 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.06 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:33:33,640 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.14 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:33:46,170 | INFO : Epoch #155 trained (lr = [0.001]).\n",
            "2023-12-10 05:33:47,649 | INFO : \tBenign test acc: 91.77% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:33:55,514 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.15 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:34:03,113 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.13 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:34:17,006 | INFO : Epoch #156 trained (lr = [0.001]).\n",
            "2023-12-10 05:34:18,500 | INFO : \tBenign test acc: 91.43% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:34:26,249 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.84 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:34:33,883 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.38 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:34:46,456 | INFO : Epoch #157 trained (lr = [0.001]).\n",
            "2023-12-10 05:34:47,917 | INFO : \tBenign test acc: 91.48% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:34:55,750 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.25 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:35:03,333 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.31 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:35:15,851 | INFO : Epoch #158 trained (lr = [0.001]).\n",
            "2023-12-10 05:35:17,290 | INFO : \tBenign test acc: 91.37% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:35:25,226 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.37 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:35:32,813 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.61 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:35:45,898 | INFO : Epoch #159 trained (lr = [0.001]).\n",
            "2023-12-10 05:35:47,359 | INFO : \tBenign test acc: 91.24% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:35:55,188 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.73 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:36:02,774 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.91 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:36:15,276 | INFO : Epoch #160 trained (lr = [0.001]).\n",
            "2023-12-10 05:36:16,700 | INFO : \tBenign test acc: 91.38% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:36:25,086 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.20 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:36:32,433 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.69 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:36:44,898 | INFO : Epoch #161 trained (lr = [0.001]).\n",
            "2023-12-10 05:36:46,306 | INFO : \tBenign test acc: 91.45% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:36:53,980 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.22 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:37:01,282 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.77 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:37:13,726 | INFO : Epoch #162 trained (lr = [0.001]).\n",
            "2023-12-10 05:37:15,213 | INFO : \tBenign test acc: 91.45% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:37:22,834 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.38 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:37:30,182 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.44 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:37:42,534 | INFO : Epoch #163 trained (lr = [0.001]).\n",
            "2023-12-10 05:37:44,010 | INFO : \tBenign test acc: 91.24% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:37:51,639 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.57 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:37:58,895 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 14.77 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:38:11,262 | INFO : Epoch #164 trained (lr = [0.001]).\n",
            "2023-12-10 05:38:12,710 | INFO : \tBenign test acc: 91.71% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:38:20,268 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.47 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:38:28,111 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.52 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:38:40,450 | INFO : Epoch #165 trained (lr = [0.001]).\n",
            "2023-12-10 05:38:41,945 | INFO : \tBenign test acc: 91.44% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:38:49,572 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.41 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:38:56,948 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.31 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:39:09,260 | INFO : Epoch #166 trained (lr = [0.001]).\n",
            "2023-12-10 05:39:10,736 | INFO : \tBenign test acc: 91.31% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:39:18,319 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.93 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:39:25,731 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.96 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:39:38,152 | INFO : Epoch #167 trained (lr = [0.001]).\n",
            "2023-12-10 05:39:39,562 | INFO : \tBenign test acc: 91.42% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:39:47,077 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.24 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:39:54,485 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.39 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:40:06,765 | INFO : Epoch #168 trained (lr = [0.001]).\n",
            "2023-12-10 05:40:08,274 | INFO : \tBenign test acc: 91.71% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:40:15,836 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.70 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:40:23,202 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.91 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:40:35,506 | INFO : Epoch #169 trained (lr = [0.001]).\n",
            "2023-12-10 05:40:36,925 | INFO : \tBenign test acc: 91.77% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:40:44,477 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.86 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:40:51,917 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.00 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:41:04,383 | INFO : Epoch #170 trained (lr = [0.001]).\n",
            "2023-12-10 05:41:05,810 | INFO : \tBenign test acc: 91.45% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:41:13,360 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.09 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:41:20,630 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.97 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:41:33,102 | INFO : Epoch #171 trained (lr = [0.001]).\n",
            "2023-12-10 05:41:34,519 | INFO : \tBenign test acc: 91.35% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:41:42,172 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.65 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:41:49,505 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.76 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:42:01,983 | INFO : Epoch #172 trained (lr = [0.001]).\n",
            "2023-12-10 05:42:03,392 | INFO : \tBenign test acc: 91.55% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:42:11,064 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.35 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:42:18,424 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.58 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:42:30,769 | INFO : Epoch #173 trained (lr = [0.001]).\n",
            "2023-12-10 05:42:32,236 | INFO : \tBenign test acc: 91.26% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:42:39,870 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.80 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:42:47,156 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.74 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:42:59,475 | INFO : Epoch #174 trained (lr = [0.001]).\n",
            "2023-12-10 05:43:00,885 | INFO : \tBenign test acc: 91.30% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:43:08,470 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 62.87 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:43:15,819 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.06 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:43:28,100 | INFO : Epoch #175 trained (lr = [0.001]).\n",
            "2023-12-10 05:43:29,526 | INFO : \tBenign test acc: 91.58% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:43:37,203 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.68 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:43:44,465 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.27 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:43:56,744 | INFO : Epoch #176 trained (lr = [0.001]).\n",
            "2023-12-10 05:43:58,205 | INFO : \tBenign test acc: 91.30% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:44:05,828 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.95 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:44:13,109 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.16 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:44:25,422 | INFO : Epoch #177 trained (lr = [0.001]).\n",
            "2023-12-10 05:44:26,830 | INFO : \tBenign test acc: 91.26% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:44:34,494 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.70 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:44:41,815 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.45 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:44:54,232 | INFO : Epoch #178 trained (lr = [0.001]).\n",
            "2023-12-10 05:44:55,682 | INFO : \tBenign test acc: 91.35% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:45:03,267 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 62.76 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:45:10,516 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 15.75 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:45:22,836 | INFO : Epoch #179 trained (lr = [0.001]).\n",
            "2023-12-10 05:45:24,277 | INFO : \tBenign test acc: 91.47% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:45:31,885 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.03 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:45:39,180 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.08 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:45:51,581 | INFO : Epoch #180 trained (lr = [0.001]).\n",
            "2023-12-10 05:45:53,043 | INFO : \tBenign test acc: 91.26% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:46:01,157 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.52 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:46:08,506 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.20 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:46:20,749 | INFO : Epoch #181 trained (lr = [0.001]).\n",
            "2023-12-10 05:46:22,209 | INFO : \tBenign test acc: 91.19% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:46:29,888 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.54 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:46:37,198 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.57 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:46:49,529 | INFO : Epoch #182 trained (lr = [0.001]).\n",
            "2023-12-10 05:46:50,985 | INFO : \tBenign test acc: 91.26% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:46:58,623 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.54 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:47:05,982 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.33 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:47:18,266 | INFO : Epoch #183 trained (lr = [0.001]).\n",
            "2023-12-10 05:47:19,788 | INFO : \tBenign test acc: 91.37% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:47:27,444 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 65.05 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:47:34,834 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.06 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:47:47,106 | INFO : Epoch #184 trained (lr = [0.001]).\n",
            "2023-12-10 05:47:48,566 | INFO : \tBenign test acc: 91.18% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:47:56,230 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.88 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:48:03,559 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.26 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:48:15,859 | INFO : Epoch #185 trained (lr = [0.001]).\n",
            "2023-12-10 05:48:17,335 | INFO : \tBenign test acc: 91.45% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:48:24,886 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 62.79 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:48:32,203 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.50 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:48:44,469 | INFO : Epoch #186 trained (lr = [0.001]).\n",
            "2023-12-10 05:48:45,872 | INFO : \tBenign test acc: 91.33% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:48:53,417 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.75 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:49:00,820 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.62 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:49:13,431 | INFO : Epoch #187 trained (lr = [0.001]).\n",
            "2023-12-10 05:49:14,850 | INFO : \tBenign test acc: 90.86% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:49:22,312 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.73 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:49:29,640 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.61 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:49:41,894 | INFO : Epoch #188 trained (lr = [0.001]).\n",
            "2023-12-10 05:49:43,368 | INFO : \tBenign test acc: 91.38% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:49:50,940 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.63 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:49:58,704 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.14 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:50:11,213 | INFO : Epoch #189 trained (lr = [0.001]).\n",
            "2023-12-10 05:50:12,620 | INFO : \tBenign test acc: 91.46% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:50:20,202 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.87 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:50:27,488 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.00 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:50:39,754 | INFO : Epoch #190 trained (lr = [0.001]).\n",
            "2023-12-10 05:50:41,230 | INFO : \tBenign test acc: 91.27% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:50:48,894 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.66 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:50:56,322 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.32 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:51:08,810 | INFO : Epoch #191 trained (lr = [0.001]).\n",
            "2023-12-10 05:51:10,246 | INFO : \tBenign test acc: 91.69% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:51:17,868 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.04 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:51:25,241 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.53 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:51:37,607 | INFO : Epoch #192 trained (lr = [0.001]).\n",
            "2023-12-10 05:51:39,115 | INFO : \tBenign test acc: 91.29% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:51:46,740 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.46 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:51:54,108 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.36 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:52:06,479 | INFO : Epoch #193 trained (lr = [0.001]).\n",
            "2023-12-10 05:52:07,894 | INFO : \tBenign test acc: 90.65% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:52:15,699 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.88 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:52:22,883 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.76 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:52:35,145 | INFO : Epoch #194 trained (lr = [0.001]).\n",
            "2023-12-10 05:52:36,549 | INFO : \tBenign test acc: 91.59% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:52:44,163 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.89 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:52:51,366 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.01 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:53:03,681 | INFO : Epoch #195 trained (lr = [0.001]).\n",
            "2023-12-10 05:53:05,102 | INFO : \tBenign test acc: 91.21% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:53:12,682 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.36 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:53:19,945 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.37 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:53:32,664 | INFO : Epoch #196 trained (lr = [0.001]).\n",
            "2023-12-10 05:53:34,084 | INFO : \tBenign test acc: 91.43% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:53:41,758 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.34 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:53:49,084 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.09 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:54:01,512 | INFO : Epoch #197 trained (lr = [0.001]).\n",
            "2023-12-10 05:54:02,963 | INFO : \tBenign test acc: 91.50% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:54:10,494 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 62.53 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:54:17,729 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.73 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:54:30,524 | INFO : Epoch #198 trained (lr = [0.001]).\n",
            "2023-12-10 05:54:31,963 | INFO : \tBenign test acc: 90.92% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:54:39,622 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.39 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:54:47,066 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.60 \t(Best: 17.82 from Epoch #37) \n",
            "2023-12-10 05:54:59,312 | INFO : Epoch #199 trained (lr = [0.001]).\n",
            "2023-12-10 05:55:00,809 | INFO : \tBenign test acc: 91.27% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:55:08,475 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.17 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:55:15,859 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.88 \t(Best: 17.88 from Epoch #199) \tNew best model ckpt best_<FGSM with eps=0.1>.pt saved.\n",
            "2023-12-10 05:55:28,066 | INFO : Epoch #200 trained (lr = [0.001]).\n",
            "2023-12-10 05:55:29,603 | INFO : \tBenign test acc: 90.94% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:55:37,379 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.86 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:55:44,794 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.58 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 05:55:57,194 | INFO : Epoch #201 trained (lr = [0.0001]).\n",
            "2023-12-10 05:55:58,637 | INFO : \tBenign test acc: 91.62% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:56:06,183 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.89 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:56:13,600 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.95 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 05:56:25,968 | INFO : Epoch #202 trained (lr = [0.0001]).\n",
            "2023-12-10 05:56:27,451 | INFO : \tBenign test acc: 91.59% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:56:34,958 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.10 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:56:42,261 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.20 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 05:56:54,618 | INFO : Epoch #203 trained (lr = [0.0001]).\n",
            "2023-12-10 05:56:56,076 | INFO : \tBenign test acc: 91.57% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:57:03,709 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.53 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:57:11,053 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.81 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 05:57:24,018 | INFO : Epoch #204 trained (lr = [0.0001]).\n",
            "2023-12-10 05:57:25,451 | INFO : \tBenign test acc: 91.64% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:57:33,064 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.11 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:57:40,472 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.33 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 05:57:52,792 | INFO : Epoch #205 trained (lr = [0.0001]).\n",
            "2023-12-10 05:57:54,262 | INFO : \tBenign test acc: 91.80% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:58:01,908 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.35 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:58:09,159 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.39 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 05:58:21,408 | INFO : Epoch #206 trained (lr = [0.0001]).\n",
            "2023-12-10 05:58:22,844 | INFO : \tBenign test acc: 91.55% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:58:30,490 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.16 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:58:37,919 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.59 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 05:58:50,439 | INFO : Epoch #207 trained (lr = [0.0001]).\n",
            "2023-12-10 05:58:51,908 | INFO : \tBenign test acc: 91.78% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:58:59,581 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.12 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:59:06,901 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.38 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 05:59:19,247 | INFO : Epoch #208 trained (lr = [0.0001]).\n",
            "2023-12-10 05:59:20,652 | INFO : \tBenign test acc: 91.69% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:59:28,309 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.23 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 05:59:35,623 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.51 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 05:59:47,887 | INFO : Epoch #209 trained (lr = [0.0001]).\n",
            "2023-12-10 05:59:49,325 | INFO : \tBenign test acc: 91.79% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 05:59:57,002 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.97 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:00:04,432 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.20 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:00:16,722 | INFO : Epoch #210 trained (lr = [0.0001]).\n",
            "2023-12-10 06:00:18,124 | INFO : \tBenign test acc: 91.82% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:00:25,798 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.02 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:00:33,321 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.87 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:00:45,705 | INFO : Epoch #211 trained (lr = [0.0001]).\n",
            "2023-12-10 06:00:47,131 | INFO : \tBenign test acc: 91.79% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:00:54,796 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.87 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:01:02,077 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.60 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:01:14,356 | INFO : Epoch #212 trained (lr = [0.0001]).\n",
            "2023-12-10 06:01:15,796 | INFO : \tBenign test acc: 91.76% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:01:23,783 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.41 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:01:31,083 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.07 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:01:43,357 | INFO : Epoch #213 trained (lr = [0.0001]).\n",
            "2023-12-10 06:01:44,821 | INFO : \tBenign test acc: 91.79% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:01:52,450 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.12 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:01:59,877 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.31 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:02:12,214 | INFO : Epoch #214 trained (lr = [0.0001]).\n",
            "2023-12-10 06:02:13,666 | INFO : \tBenign test acc: 91.70% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:02:21,260 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.95 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:02:28,682 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.59 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:02:41,419 | INFO : Epoch #215 trained (lr = [0.0001]).\n",
            "2023-12-10 06:02:42,899 | INFO : \tBenign test acc: 91.75% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:02:50,524 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.26 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:02:57,867 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.03 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:03:10,131 | INFO : Epoch #216 trained (lr = [0.0001]).\n",
            "2023-12-10 06:03:11,587 | INFO : \tBenign test acc: 91.72% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:03:19,118 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.59 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:03:26,444 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.26 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:03:38,780 | INFO : Epoch #217 trained (lr = [0.0001]).\n",
            "2023-12-10 06:03:40,259 | INFO : \tBenign test acc: 91.79% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:03:47,808 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.89 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:03:55,158 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.21 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:04:07,495 | INFO : Epoch #218 trained (lr = [0.0001]).\n",
            "2023-12-10 06:04:08,986 | INFO : \tBenign test acc: 91.79% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:04:16,571 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.73 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:04:23,963 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.39 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:04:36,524 | INFO : Epoch #219 trained (lr = [0.0001]).\n",
            "2023-12-10 06:04:37,952 | INFO : \tBenign test acc: 91.86% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:04:45,455 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.09 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:04:52,764 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.30 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:05:05,242 | INFO : Epoch #220 trained (lr = [0.0001]).\n",
            "2023-12-10 06:05:06,712 | INFO : \tBenign test acc: 91.77% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:05:14,285 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.06 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:05:21,595 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.22 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:05:33,844 | INFO : Epoch #221 trained (lr = [0.0001]).\n",
            "2023-12-10 06:05:35,277 | INFO : \tBenign test acc: 91.86% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:05:42,895 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.15 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:05:50,273 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.80 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:06:02,787 | INFO : Epoch #222 trained (lr = [0.0001]).\n",
            "2023-12-10 06:06:04,186 | INFO : \tBenign test acc: 91.80% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:06:11,822 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.87 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:06:19,204 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.96 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:06:31,597 | INFO : Epoch #223 trained (lr = [0.0001]).\n",
            "2023-12-10 06:06:33,017 | INFO : \tBenign test acc: 91.64% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:06:40,639 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.74 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:06:48,001 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.65 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:07:00,275 | INFO : Epoch #224 trained (lr = [0.0001]).\n",
            "2023-12-10 06:07:01,743 | INFO : \tBenign test acc: 91.88% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:07:09,365 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.76 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:07:16,735 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.14 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:07:29,432 | INFO : Epoch #225 trained (lr = [0.0001]).\n",
            "2023-12-10 06:07:30,842 | INFO : \tBenign test acc: 91.88% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:07:38,473 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.78 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:07:45,844 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.83 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:07:58,238 | INFO : Epoch #226 trained (lr = [0.0001]).\n",
            "2023-12-10 06:07:59,663 | INFO : \tBenign test acc: 91.73% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:08:07,277 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.47 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:08:14,690 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.50 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:08:26,988 | INFO : Epoch #227 trained (lr = [0.0001]).\n",
            "2023-12-10 06:08:28,420 | INFO : \tBenign test acc: 91.87% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:08:36,108 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.34 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:08:43,311 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.29 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:08:55,550 | INFO : Epoch #228 trained (lr = [0.0001]).\n",
            "2023-12-10 06:08:56,957 | INFO : \tBenign test acc: 91.95% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:09:04,610 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.11 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:09:11,912 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.30 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:09:24,224 | INFO : Epoch #229 trained (lr = [0.0001]).\n",
            "2023-12-10 06:09:25,650 | INFO : \tBenign test acc: 91.92% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:09:33,784 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.02 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:09:41,158 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.49 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:09:53,539 | INFO : Epoch #230 trained (lr = [0.0001]).\n",
            "2023-12-10 06:09:54,986 | INFO : \tBenign test acc: 91.63% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:10:02,599 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.63 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:10:09,935 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.40 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:10:22,187 | INFO : Epoch #231 trained (lr = [0.0001]).\n",
            "2023-12-10 06:10:23,716 | INFO : \tBenign test acc: 91.70% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:10:31,502 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.92 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:10:38,899 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.60 \t(Best: 17.88 from Epoch #199) \n",
            "2023-12-10 06:10:51,118 | INFO : Epoch #232 trained (lr = [0.0001]).\n",
            "2023-12-10 06:10:52,610 | INFO : \tBenign test acc: 91.86% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:11:00,125 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.13 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:11:07,687 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 18.23 \t(Best: 18.23 from Epoch #232) \tNew best model ckpt best_<FGSM with eps=0.1>.pt saved.\n",
            "2023-12-10 06:11:19,945 | INFO : Epoch #233 trained (lr = [0.0001]).\n",
            "2023-12-10 06:11:21,442 | INFO : \tBenign test acc: 91.80% \t(Best: 91.97% from Epoch #109) \n",
            "2023-12-10 06:11:29,241 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.71 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:11:36,831 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.26 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:11:49,176 | INFO : Epoch #234 trained (lr = [0.0001]).\n",
            "2023-12-10 06:11:50,689 | INFO : \tBenign test acc: 91.97% \t(Best: 91.97% from Epoch #234) \tNew best model ckpt best_benign_test_acc.pt saved.\n",
            "2023-12-10 06:11:58,167 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.44 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:12:05,569 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.19 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:12:18,232 | INFO : Epoch #235 trained (lr = [0.0001]).\n",
            "2023-12-10 06:12:19,659 | INFO : \tBenign test acc: 91.93% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:12:27,313 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.76 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:12:34,757 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.26 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:12:47,035 | INFO : Epoch #236 trained (lr = [0.0001]).\n",
            "2023-12-10 06:12:48,485 | INFO : \tBenign test acc: 91.78% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:12:56,130 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.15 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:13:03,470 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.58 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:13:15,936 | INFO : Epoch #237 trained (lr = [0.0001]).\n",
            "2023-12-10 06:13:17,442 | INFO : \tBenign test acc: 91.86% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:13:25,028 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.49 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:13:32,370 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.70 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:13:44,751 | INFO : Epoch #238 trained (lr = [0.0001]).\n",
            "2023-12-10 06:13:46,194 | INFO : \tBenign test acc: 91.92% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:13:53,805 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.69 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:14:01,184 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.12 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:14:13,750 | INFO : Epoch #239 trained (lr = [0.0001]).\n",
            "2023-12-10 06:14:15,159 | INFO : \tBenign test acc: 91.78% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:14:22,783 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.07 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:14:30,174 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.44 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:14:42,720 | INFO : Epoch #240 trained (lr = [0.0001]).\n",
            "2023-12-10 06:14:44,161 | INFO : \tBenign test acc: 91.76% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:14:52,001 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.36 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:14:59,361 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.33 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:15:11,680 | INFO : Epoch #241 trained (lr = [0.0001]).\n",
            "2023-12-10 06:15:13,133 | INFO : \tBenign test acc: 91.87% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:15:20,966 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.59 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:15:28,412 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.28 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:15:40,815 | INFO : Epoch #242 trained (lr = [0.0001]).\n",
            "2023-12-10 06:15:42,278 | INFO : \tBenign test acc: 91.88% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:15:50,064 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.85 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:15:57,595 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.46 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:16:10,055 | INFO : Epoch #243 trained (lr = [0.0001]).\n",
            "2023-12-10 06:16:11,470 | INFO : \tBenign test acc: 91.75% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:16:19,389 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.65 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:16:26,707 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.56 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:16:38,998 | INFO : Epoch #244 trained (lr = [0.0001]).\n",
            "2023-12-10 06:16:40,511 | INFO : \tBenign test acc: 91.75% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:16:48,259 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.92 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:16:55,639 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.82 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:17:08,056 | INFO : Epoch #245 trained (lr = [0.0001]).\n",
            "2023-12-10 06:17:09,494 | INFO : \tBenign test acc: 91.75% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:17:17,269 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.64 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:17:24,541 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.35 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:17:36,825 | INFO : Epoch #246 trained (lr = [0.0001]).\n",
            "2023-12-10 06:17:38,320 | INFO : \tBenign test acc: 91.82% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:17:46,026 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.02 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:17:53,412 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.45 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:18:05,710 | INFO : Epoch #247 trained (lr = [0.0001]).\n",
            "2023-12-10 06:18:07,210 | INFO : \tBenign test acc: 91.75% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:18:14,892 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.48 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:18:22,250 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.05 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:18:34,598 | INFO : Epoch #248 trained (lr = [0.0001]).\n",
            "2023-12-10 06:18:36,080 | INFO : \tBenign test acc: 91.86% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:18:43,608 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.71 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:18:50,963 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.47 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:19:03,284 | INFO : Epoch #249 trained (lr = [0.0001]).\n",
            "2023-12-10 06:19:04,772 | INFO : \tBenign test acc: 91.80% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:19:12,403 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.25 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:19:19,960 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.11 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:19:32,597 | INFO : Epoch #250 trained (lr = [0.0001]).\n",
            "2023-12-10 06:19:34,102 | INFO : \tBenign test acc: 91.66% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:19:41,732 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.63 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:19:49,143 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.26 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:20:01,511 | INFO : Epoch #251 trained (lr = [0.0001]).\n",
            "2023-12-10 06:20:03,011 | INFO : \tBenign test acc: 91.83% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:20:10,604 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.67 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:20:18,080 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.91 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:20:30,607 | INFO : Epoch #252 trained (lr = [0.0001]).\n",
            "2023-12-10 06:20:32,148 | INFO : \tBenign test acc: 91.60% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:20:39,822 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.94 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:20:47,465 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.30 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:20:59,881 | INFO : Epoch #253 trained (lr = [0.0001]).\n",
            "2023-12-10 06:21:01,323 | INFO : \tBenign test acc: 91.80% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:21:08,961 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.32 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:21:16,305 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.77 \t(Best: 18.23 from Epoch #232) \n",
            "2023-12-10 06:21:28,691 | INFO : Epoch #254 trained (lr = [0.0001]).\n",
            "2023-12-10 06:21:30,166 | INFO : \tBenign test acc: 91.79% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:21:37,752 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.37 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:21:45,194 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 18.24 \t(Best: 18.24 from Epoch #254) \tNew best model ckpt best_<FGSM with eps=0.1>.pt saved.\n",
            "2023-12-10 06:21:57,715 | INFO : Epoch #255 trained (lr = [0.0001]).\n",
            "2023-12-10 06:21:59,130 | INFO : \tBenign test acc: 91.45% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:22:06,779 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.73 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:22:14,171 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.96 \t(Best: 18.24 from Epoch #254) \n",
            "2023-12-10 06:22:26,680 | INFO : Epoch #256 trained (lr = [0.0001]).\n",
            "2023-12-10 06:22:28,133 | INFO : \tBenign test acc: 91.86% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:22:35,795 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.77 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:22:43,144 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.61 \t(Best: 18.24 from Epoch #254) \n",
            "2023-12-10 06:22:55,559 | INFO : Epoch #257 trained (lr = [0.0001]).\n",
            "2023-12-10 06:22:57,051 | INFO : \tBenign test acc: 91.78% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:23:04,754 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.88 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:23:12,089 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.51 \t(Best: 18.24 from Epoch #254) \n",
            "2023-12-10 06:23:24,710 | INFO : Epoch #258 trained (lr = [0.0001]).\n",
            "2023-12-10 06:23:26,114 | INFO : \tBenign test acc: 91.76% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:23:33,799 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.46 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:23:41,129 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.21 \t(Best: 18.24 from Epoch #254) \n",
            "2023-12-10 06:23:53,441 | INFO : Epoch #259 trained (lr = [0.0001]).\n",
            "2023-12-10 06:23:54,891 | INFO : \tBenign test acc: 91.83% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:24:02,819 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.16 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:24:10,137 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.83 \t(Best: 18.24 from Epoch #254) \n",
            "2023-12-10 06:24:22,844 | INFO : Epoch #260 trained (lr = [0.0001]).\n",
            "2023-12-10 06:24:24,332 | INFO : \tBenign test acc: 91.74% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:24:32,052 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.55 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:24:39,425 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.21 \t(Best: 18.24 from Epoch #254) \n",
            "2023-12-10 06:24:51,644 | INFO : Epoch #261 trained (lr = [0.0001]).\n",
            "2023-12-10 06:24:53,113 | INFO : \tBenign test acc: 91.75% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:25:00,910 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.74 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:25:08,328 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 18.72 \t(Best: 18.72 from Epoch #261) \tNew best model ckpt best_<FGSM with eps=0.1>.pt saved.\n",
            "2023-12-10 06:25:20,548 | INFO : Epoch #262 trained (lr = [0.0001]).\n",
            "2023-12-10 06:25:22,021 | INFO : \tBenign test acc: 91.89% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:25:29,704 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.77 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:25:37,088 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.47 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:25:49,794 | INFO : Epoch #263 trained (lr = [0.0001]).\n",
            "2023-12-10 06:25:51,264 | INFO : \tBenign test acc: 91.88% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:25:58,772 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.23 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:26:06,132 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.90 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:26:18,386 | INFO : Epoch #264 trained (lr = [0.0001]).\n",
            "2023-12-10 06:26:19,900 | INFO : \tBenign test acc: 91.75% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:26:27,459 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.99 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:26:34,842 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.96 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:26:47,305 | INFO : Epoch #265 trained (lr = [0.0001]).\n",
            "2023-12-10 06:26:48,774 | INFO : \tBenign test acc: 91.73% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:26:56,365 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.30 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:27:03,721 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.35 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:27:15,952 | INFO : Epoch #266 trained (lr = [0.0001]).\n",
            "2023-12-10 06:27:17,393 | INFO : \tBenign test acc: 91.77% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:27:24,973 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.31 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:27:32,407 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.52 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:27:44,663 | INFO : Epoch #267 trained (lr = [0.0001]).\n",
            "2023-12-10 06:27:46,064 | INFO : \tBenign test acc: 91.77% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:27:53,594 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.12 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:28:01,253 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.39 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:28:13,766 | INFO : Epoch #268 trained (lr = [0.0001]).\n",
            "2023-12-10 06:28:15,189 | INFO : \tBenign test acc: 91.79% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:28:22,724 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.82 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:28:30,213 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.64 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:28:42,564 | INFO : Epoch #269 trained (lr = [0.0001]).\n",
            "2023-12-10 06:28:44,007 | INFO : \tBenign test acc: 91.84% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:28:51,681 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.57 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:28:59,065 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.72 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:29:11,539 | INFO : Epoch #270 trained (lr = [0.0001]).\n",
            "2023-12-10 06:29:12,972 | INFO : \tBenign test acc: 91.83% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:29:20,573 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.37 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:29:27,914 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.15 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:29:40,162 | INFO : Epoch #271 trained (lr = [0.0001]).\n",
            "2023-12-10 06:29:41,598 | INFO : \tBenign test acc: 91.75% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:29:49,463 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.82 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:29:56,762 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.61 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:30:09,143 | INFO : Epoch #272 trained (lr = [0.0001]).\n",
            "2023-12-10 06:30:10,574 | INFO : \tBenign test acc: 91.79% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:30:18,170 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.15 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:30:25,433 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.06 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:30:37,816 | INFO : Epoch #273 trained (lr = [0.0001]).\n",
            "2023-12-10 06:30:39,238 | INFO : \tBenign test acc: 91.68% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:30:46,896 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 62.90 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:30:54,248 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.24 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:31:06,506 | INFO : Epoch #274 trained (lr = [0.0001]).\n",
            "2023-12-10 06:31:07,925 | INFO : \tBenign test acc: 91.73% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:31:15,580 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.45 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:31:22,850 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.12 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:31:35,084 | INFO : Epoch #275 trained (lr = [0.0001]).\n",
            "2023-12-10 06:31:36,486 | INFO : \tBenign test acc: 91.85% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:31:44,161 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.53 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:31:51,518 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.57 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:32:03,793 | INFO : Epoch #276 trained (lr = [0.0001]).\n",
            "2023-12-10 06:32:05,195 | INFO : \tBenign test acc: 91.88% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:32:13,145 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.11 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:32:20,493 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.10 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:32:32,789 | INFO : Epoch #277 trained (lr = [0.0001]).\n",
            "2023-12-10 06:32:34,269 | INFO : \tBenign test acc: 91.72% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:32:41,899 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.49 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:32:49,229 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.02 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:33:01,658 | INFO : Epoch #278 trained (lr = [0.0001]).\n",
            "2023-12-10 06:33:03,090 | INFO : \tBenign test acc: 91.81% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:33:10,712 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.22 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:33:18,018 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.05 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:33:30,286 | INFO : Epoch #279 trained (lr = [0.0001]).\n",
            "2023-12-10 06:33:31,806 | INFO : \tBenign test acc: 91.81% \t(Best: 91.97% from Epoch #234) \n",
            "2023-12-10 06:33:39,874 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.58 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:33:47,270 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.27 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:33:59,594 | INFO : Epoch #280 trained (lr = [0.0001]).\n",
            "2023-12-10 06:34:01,097 | INFO : \tBenign test acc: 91.98% \t(Best: 91.98% from Epoch #280) \tNew best model ckpt best_benign_test_acc.pt saved.\n",
            "2023-12-10 06:34:08,801 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 62.97 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:34:16,292 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.13 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:34:28,619 | INFO : Epoch #281 trained (lr = [0.0001]).\n",
            "2023-12-10 06:34:30,129 | INFO : \tBenign test acc: 91.84% \t(Best: 91.98% from Epoch #280) \n",
            "2023-12-10 06:34:37,648 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.65 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:34:44,992 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 18.02 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:34:57,406 | INFO : Epoch #282 trained (lr = [0.0001]).\n",
            "2023-12-10 06:34:58,848 | INFO : \tBenign test acc: 91.75% \t(Best: 91.98% from Epoch #280) \n",
            "2023-12-10 06:35:06,369 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.52 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:35:13,772 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.22 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:35:26,327 | INFO : Epoch #283 trained (lr = [0.0001]).\n",
            "2023-12-10 06:35:27,806 | INFO : \tBenign test acc: 91.86% \t(Best: 91.98% from Epoch #280) \n",
            "2023-12-10 06:35:35,411 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.50 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:35:43,065 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.84 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:35:55,333 | INFO : Epoch #284 trained (lr = [0.0001]).\n",
            "2023-12-10 06:35:56,747 | INFO : \tBenign test acc: 91.88% \t(Best: 91.98% from Epoch #280) \n",
            "2023-12-10 06:36:04,317 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.88 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:36:11,640 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.97 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:36:23,972 | INFO : Epoch #285 trained (lr = [0.0001]).\n",
            "2023-12-10 06:36:25,409 | INFO : \tBenign test acc: 91.72% \t(Best: 91.98% from Epoch #280) \n",
            "2023-12-10 06:36:33,090 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.45 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:36:40,519 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.03 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:36:53,391 | INFO : Epoch #286 trained (lr = [0.0001]).\n",
            "2023-12-10 06:36:54,796 | INFO : \tBenign test acc: 91.81% \t(Best: 91.98% from Epoch #280) \n",
            "2023-12-10 06:37:02,427 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.30 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:37:09,866 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.24 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:37:22,068 | INFO : Epoch #287 trained (lr = [0.0001]).\n",
            "2023-12-10 06:37:23,563 | INFO : \tBenign test acc: 91.72% \t(Best: 91.98% from Epoch #280) \n",
            "2023-12-10 06:37:31,647 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.17 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:37:38,909 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.45 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:37:51,220 | INFO : Epoch #288 trained (lr = [0.0001]).\n",
            "2023-12-10 06:37:52,720 | INFO : \tBenign test acc: 91.74% \t(Best: 91.98% from Epoch #280) \n",
            "2023-12-10 06:38:00,470 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.26 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:38:07,735 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.84 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:38:20,201 | INFO : Epoch #289 trained (lr = [0.0001]).\n",
            "2023-12-10 06:38:21,603 | INFO : \tBenign test acc: 91.58% \t(Best: 91.98% from Epoch #280) \n",
            "2023-12-10 06:38:29,655 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.88 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:38:36,900 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.86 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:38:49,187 | INFO : Epoch #290 trained (lr = [0.0001]).\n",
            "2023-12-10 06:38:50,618 | INFO : \tBenign test acc: 91.79% \t(Best: 91.98% from Epoch #280) \n",
            "2023-12-10 06:38:58,240 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.00 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:39:05,547 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.39 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:39:17,774 | INFO : Epoch #291 trained (lr = [0.0001]).\n",
            "2023-12-10 06:39:19,204 | INFO : \tBenign test acc: 91.72% \t(Best: 91.98% from Epoch #280) \n",
            "2023-12-10 06:39:27,124 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.97 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:39:34,485 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.07 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:39:46,943 | INFO : Epoch #292 trained (lr = [0.0001]).\n",
            "2023-12-10 06:39:48,387 | INFO : \tBenign test acc: 91.66% \t(Best: 91.98% from Epoch #280) \n",
            "2023-12-10 06:39:56,082 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.34 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:40:03,455 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.46 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:40:15,729 | INFO : Epoch #293 trained (lr = [0.0001]).\n",
            "2023-12-10 06:40:17,237 | INFO : \tBenign test acc: 91.50% \t(Best: 91.98% from Epoch #280) \n",
            "2023-12-10 06:40:24,909 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 64.23 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:40:32,258 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.47 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:40:44,524 | INFO : Epoch #294 trained (lr = [0.0001]).\n",
            "2023-12-10 06:40:46,025 | INFO : \tBenign test acc: 91.66% \t(Best: 91.98% from Epoch #280) \n",
            "2023-12-10 06:40:53,677 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.20 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:41:01,027 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.78 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:41:13,352 | INFO : Epoch #295 trained (lr = [0.0001]).\n",
            "2023-12-10 06:41:14,815 | INFO : \tBenign test acc: 91.63% \t(Best: 91.98% from Epoch #280) \n",
            "2023-12-10 06:41:22,362 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.04 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:41:29,704 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.17 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:41:41,940 | INFO : Epoch #296 trained (lr = [0.0001]).\n",
            "2023-12-10 06:41:43,399 | INFO : \tBenign test acc: 91.75% \t(Best: 91.98% from Epoch #280) \n",
            "2023-12-10 06:41:50,969 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.57 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:41:58,412 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.22 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:42:10,805 | INFO : Epoch #297 trained (lr = [0.0001]).\n",
            "2023-12-10 06:42:12,250 | INFO : \tBenign test acc: 91.70% \t(Best: 91.98% from Epoch #280) \n",
            "2023-12-10 06:42:19,768 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.45 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:42:27,131 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.07 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:42:39,515 | INFO : Epoch #298 trained (lr = [0.0001]).\n",
            "2023-12-10 06:42:40,973 | INFO : \tBenign test acc: 91.78% \t(Best: 91.98% from Epoch #280) \n",
            "2023-12-10 06:42:48,630 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.45 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:42:56,126 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 16.36 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:43:08,389 | INFO : Epoch #299 trained (lr = [0.0001]).\n",
            "2023-12-10 06:43:09,785 | INFO : \tBenign test acc: 91.77% \t(Best: 91.98% from Epoch #280) \n",
            "2023-12-10 06:43:17,372 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.86 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:43:24,758 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.28 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:43:37,030 | INFO : Epoch #300 trained (lr = [0.0001]).\n",
            "2023-12-10 06:43:38,441 | INFO : \tBenign test acc: 91.67% \t(Best: 91.98% from Epoch #280) \n",
            "2023-12-10 06:43:46,081 | INFO : \tAdv test acc (<FGSM with eps=0.01>): 63.56 \t(Best: 68.72 from Epoch #101) \n",
            "2023-12-10 06:43:53,486 | INFO : \tAdv test acc (<FGSM with eps=0.1>): 17.27 \t(Best: 18.72 from Epoch #261) \n",
            "2023-12-10 06:43:53,487 | INFO : Model /content/drive/MyDrive/adv_robust_gkp/output/resnet20_pruned/grouped.pt finetuned, please check exp_ckpts/best_ckpts/ folder for best models.\n",
            "2023-12-10 06:43:53,488 | INFO : Reporting best performance...\n",
            "2023-12-10 06:43:53,488 | INFO : \tBest benign_test_acc: 91.98 from Epoch #280.\n",
            "2023-12-10 06:43:53,488 | INFO : \t\t<FGSM with eps=0.01>: 62.97.\n",
            "2023-12-10 06:43:53,488 | INFO : \t\t<FGSM with eps=0.1>: 16.13.\n",
            "2023-12-10 06:43:53,488 | INFO : \tBest <FGSM with eps=0.01>: 68.72 from Epoch #101.\n",
            "2023-12-10 06:43:53,488 | INFO : \t\tbenign_test_acc: 91.40.\n",
            "2023-12-10 06:43:53,488 | INFO : \t\t<FGSM with eps=0.1>: 15.08.\n",
            "2023-12-10 06:43:53,488 | INFO : \tBest <FGSM with eps=0.1>: 18.72 from Epoch #261.\n",
            "2023-12-10 06:43:53,488 | INFO : \t\tbenign_test_acc: 91.75.\n",
            "2023-12-10 06:43:53,488 | INFO : \t\t<FGSM with eps=0.01>: 63.74.\n",
            "2023-12-10 06:43:53,490 | INFO : Saving setting file to /content/drive/MyDrive/adv_robust_gkp/output/resnet20_finetuned/output_setting.json...\n",
            "2023-12-10 06:43:53,497 | INFO : Experiment resnet20_srgkp_finetune (task: finetune) ended at 2023-12-10T06:43:53.488988 duration: 2:25:40.565735.\n"
          ]
        }
      ]
    }
  ]
}