3D_Classification.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "name": "3D Classification.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "machine_shape": "hm",
      "mount_file_id": "1Yt50EayGY38jEb-yfDor6aqJvnrBEEDl",
      "authorship_tag": "ABX9TyN6Dcr4JE9JdeQU1UKdSIPJ",
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Meddebma/Complete-Python-3-Bootcamp/blob/master/3D_Classification.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wiFnM05m6uL2"
      },
      "source": [
        "%pip install \"monai--weekly[nibabel, skimage, pillow, tensorboard, gdown, ignite, torchvision, itk, tqdm, lmdb, psutil]\""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WHt5hRV4MYoT",
        "outputId": "99f30d59-adf3-4be6-e776-e2d8282cb767"
      },
      "source": [
        "# Copyright 2020 MONAI Consortium\n",
        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
        "# you may not use this file except in compliance with the License.\n",
        "# You may obtain a copy of the License at\n",
        "#     http://www.apache.org/licenses/LICENSE-2.0\n",
        "# Unless required by applicable law or agreed to in writing, software\n",
        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
        "# See the License for the specific language governing permissions and\n",
        "# limitations under the License.\n",
        "\n",
        "import logging\n",
        "import os\n",
        "import sys\n",
        "import tempfile\n",
        "import shutil\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "import torch\n",
        "from torch.utils.tensorboard import SummaryWriter\n",
        "import numpy as np\n",
        "\n",
        "import monai\n",
        "from monai.networks.nets import DenseNet121, Classifier\n",
        "from monai.apps import download_and_extract\n",
        "from monai.config import print_config\n",
        "from monai.utils import first\n",
        "from monai.visualize import GradCAM\n",
        "from monai.metrics import ConfusionMatrixMetric, get_confusion_matrix\n",
        "from monai.data import CacheDataset, DataLoader, ImageDataset, ITKReader, PILReader, Dataset, partition_dataset_classes, partition_dataset\n",
        "from sklearn.metrics import (\n",
        "    classification_report, confusion_matrix,\n",
        "    ConfusionMatrixDisplay, \n",
        ")\n",
        "from monai.data import decollate_batch\n",
        "from monai.metrics import ROCAUCMetric\n",
        "from torch.utils.data import DataLoader \n",
        "from monai.transforms import Transform\n",
        "from monai.transforms import (\n",
        "    AddChannel,\n",
        "    Compose,\n",
        "    LoadImage,\n",
        "    RandRotate90,\n",
        "    Invertd,\n",
        "    Lambda,\n",
        "    Resize,\n",
        "    Orientation,\n",
        "    Orientationd,\n",
        "    EnsureChannelFirstd,\n",
        "    ScaleIntensity,\n",
        "    AsChannelFirst,\n",
        "    AsChannelLast,\n",
        "    Activations,\n",
        "    AddChannel,\n",
        "    AddChanneld,\n",
        "    ToTensord,\n",
        "    Resized,\n",
        "    AsDiscrete,\n",
        "    Spacingd,\n",
        "    EnsureChannelFirst,\n",
        "    RepeatChannel,\n",
        "    RemoveRepeatedChannel,\n",
        "    ScaleIntensityRanged,\n",
        "    SplitChannel,\n",
        "    SqueezeDim,\n",
        "    ToTensor,\n",
        "    Randomizable,\n",
        "    LoadImaged,\n",
        "    ToTensord,\n",
        "    MaskIntensityd,\n",
        "    Flip,\n",
        "    MaskIntensity,\n",
        "    ScaleIntensityd,\n",
        "    EnsureType,\n",
        "    EnsureTyped,\n",
        ")\n",
        "import numpy as np\n",
        "import statsmodels.formula.api as smf\n",
        "import statsmodels.api as sm\n",
        "import pandas as pd\n",
        "import seaborn as sns \n",
        "from glob import glob\n",
        "import logging\n",
        "from abc import ABC, abstractmethod\n",
        "from typing import Any, Callable, Dict, Generator, Hashable, Iterable, List, Optional, Tuple\n",
        "\n",
        "import numpy as np\n",
        "import torch\n",
        "\n",
        "import numpy as np\n",
        "from monai.transforms import Transform, AddChannel, AsChannelFirst\n",
        "\n",
        "\n",
        "logging.basicConfig(stream=sys.stdout, level=logging.INFO)\n",
        "print_config()"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "MONAI version: 0.7.dev2132\n",
            "Numpy version: 1.19.5\n",
            "Pytorch version: 1.9.0+cu102\n",
            "MONAI flags: HAS_EXT = False, USE_COMPILED = False\n",
            "MONAI rev id: bcbf21ec88dd95929b1d6439ebbf37b3cb266988\n",
            "\n",
            "Optional dependencies:\n",
            "Pytorch Ignite version: 0.4.5\n",
            "Nibabel version: 3.0.2\n",
            "scikit-image version: 0.16.2\n",
            "Pillow version: 7.1.2\n",
            "Tensorboard version: 2.5.0\n",
            "gdown version: 3.6.4\n",
            "TorchVision version: 0.10.0+cu102\n",
            "tqdm version: 4.62.0\n",
            "lmdb version: 0.99\n",
            "psutil version: 5.4.8\n",
            "pandas version: 1.1.5\n",
            "einops version: NOT INSTALLED or UNKNOWN VERSION.\n",
            "\n",
            "For details about installing the optional dependencies, please visit:\n",
            "    https://docs.monai.io/en/latest/installation.html#installing-the-recommended-dependencies\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BvUL9l6t_tDM"
      },
      "source": [
        "**Data Direction**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mHeS8Rt6_zWD"
      },
      "source": [
        "data_dir= \"/content/drive/My Drive/Spleen_AI\""
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "O1S2v-B-AITM",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "810d0dd7-4169-4b95-b3b9-38b2758fd083"
      },
      "source": [
        "labels_all = pd.read_excel (r'/content/drive/MyDrive/Spleen_AI/labels.xlsx')\n",
        "l=labels_all[\"label\"].to_numpy()\n",
        "print(l)\n",
        "#print(labels_all[\"label\"])"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[1 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 1 0 1 0 1\n",
            " 0 0 0 0 0 0 1 0 0 1 0 1 0 0 0 0 1 1 1 0 0 0 1 1 1 0 0 0 0 0 0 1 1 0 1 0 1\n",
            " 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 0 1 1 0 0 1 1 1 1 1 1 0 0 0 1 0 1 1 0 1 1\n",
            " 1 0 0 1 1 1 1 1 1 1 1 1 1]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GyjyR96rS0Bl",
        "outputId": "51372b0d-6f70-4919-c649-464faea519df"
      },
      "source": [
        "images = sorted(glob(os.path.join(data_dir, \"Projekt2f/\", \"*.nii.gz\")))\n",
        "segmentations = sorted(glob(os.path.join(data_dir, \"Projekt2f/segmentation/*\", \"*.nii.gz\")))\n",
        "\n",
        "#for i in images:\n",
        " # print(os.path.basename(i))\n",
        "\n",
        "labels= l\n",
        "#labels= np.array([0,0,0,0,0,0,1,0,0,0,0,0,0,1,1,1,1,1,1,1,1,1,1], dtype=np.int64)\n",
        "print(len(images))\n",
        "print(len(segmentations))\n",
        "print(len(labels))\n"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "124\n",
            "124\n",
            "124\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JgOMkqS2-i4q"
      },
      "source": [
        "**Check Data Shape**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "B6e1LbkoDgKM"
      },
      "source": [
        "device = torch.device(\"cuda\")\n",
        "#AddChannel(), Orientation(axcodes=\"PRI\"),Flip(spatial_axis=0),\n",
        "transforms=Compose([ Flip(spatial_axis=(0,-1)), ToTensor()])\n",
        "check = ImageDataset(image_files=images[:2], labels=labels[:2],\n",
        "                        transform=transforms)\n",
        "check_loader = DataLoader(check, batch_size=1,\n",
        "                          num_workers=1, pin_memory=torch.cuda.is_available())\n",
        "for data,i in check_loader:\n",
        "  inputs = data[0].to(device)\n",
        "  plt.imshow(data[0][:, :, 35], cmap=\"gray\")\n",
        "  print(inputs.shape)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AJJwFbO47yGV"
      },
      "source": [
        "**Loading Data**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "scT_G0HLMNjH",
        "outputId": "30d3c8da-e006-40dd-8a4b-51032db29677"
      },
      "source": [
        "train_files = [{\"image\": i, \"seg\": s, \"label\": l} for i, s, l in zip(images[:100], segmentations[:100], labels[:100])]\n",
        "val_files  = [{\"image\": i, \"seg\": s, \"label\": l} for i, s, l in zip(images[100:], segmentations[100:], labels[100:])]\n",
        "\n",
        "data = [{\"image\": i, \"seg\": s, \"label\": l} for i, s, l in zip(images, segmentations, labels)]\n",
        "\n",
        "transforms = Compose([\n",
        "    LoadImaged(keys=[\"image\", \"seg\"]),\n",
        "    AddChanneld(keys=[\"image\", \"seg\"]),\n",
        "    Spacingd(keys=[\"image\", \"seg\"], pixdim=(1.5, 1.5, 2.0), mode=(\"bilinear\", \"nearest\")),\n",
        "    Orientationd(keys=[\"image\", \"seg\"], axcodes=\"PLI\"),\n",
        "    Resized(keys=[\"image\", \"seg\"], spatial_size=(96, 96, 96)),\n",
        "    ScaleIntensityRanged(keys=['image'], a_min=-57, a_max=164, b_min=0.0, b_max=1.0, clip=True),\n",
        "    MaskIntensityd(keys=\"image\", mask_key=\"seg\"),\n",
        "    EnsureTyped(keys=[\"image\", \"seg\"]),\n",
        "])\n",
        "\n",
        "\n",
        "data_partition= partition_dataset_classes(data, labels.tolist(), shuffle=True, ratios=[0.8, 0.2])\n",
        "\n",
        "# create a training data loader\n",
        "train_ds = CacheDataset(data=data_partition, transform=transforms)\n",
        "train_loader = DataLoader(train_ds, batch_size=2, shuffle=True, num_workers=4, pin_memory=torch.cuda.is_available())\n",
        "\n",
        "# create a validation data loader\n",
        "val_ds = CacheDataset(data=data_partition, transform=transforms)\n",
        "val_loader = DataLoader(val_ds, batch_size=2, shuffle=True, num_workers=4, pin_memory=torch.cuda.is_available())\n",
        "\n",
        "post_pred = Compose([EnsureType(), Activations(softmax=True)])\n",
        "post_label = Compose([EnsureType(), AsDiscrete(to_onehot=True, n_classes=2)])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\rLoading dataset:   0%|          | 0/2 [00:00<?, ?it/s]"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true,
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "xxkzWn0Yco_B",
        "outputId": "488a81c9-c262-4367-9c23-9325a3b9ae60"
      },
      "source": [
        "device = torch.device(\"cuda\")\n",
        "for data in train_loader:\n",
        "  image, label = data[0].to(device), data[1].to(device)\n",
        "  plt.figure(\"check\", (18, 6))\n",
        "  plt.subplot(1, 2, 1)\n",
        "  plt.imshow(data[0][0, 0, :, :, 75], cmap=\"gray\")\n",
        "  plt.subplot(1, 2, 2)\n",
        "  plt.imshow(data[1][0, 0, :, :, 75], cmap=\"gray\")\n",
        "\n",
        "  #print(inputs.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Adding an axes using the same arguments as a previous axes currently reuses the earlier instance.  In a future version, a new instance will always be created and returned.  Meanwhile, this warning can be suppressed, and the future behavior ensured, by passing a unique label to each axes instance.\n",
            "Adding an axes using the same arguments as a previous axes currently reuses the earlier instance.  In a future version, a new instance will always be created and returned.  Meanwhile, this warning can be suppressed, and the future behavior ensured, by passing a unique label to each axes instance.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1296x432 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "o46nSnXueNbT"
      },
      "source": [
        "for idx, (images, segmentations, labels) in enumerate(train_loader):\n",
        "      \n",
        "      print(idx,  images.shape, segmentations.shape , labels)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RtjR-k758HU7"
      },
      "source": [
        "**Training**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CGT1akrQ-MKe"
      },
      "source": [
        "# Create DenseNet121, CrossEntropyLoss and Adam optimizer\n",
        "device = torch.device(\"cuda\")\n",
        "model = monai.networks.nets.DenseNet121(spatial_dims=3, in_channels=1, out_channels=2).to(device)\n",
        "loss_function = torch.nn.CrossEntropyLoss()\n",
        "optimizer = torch.optim.Adam(model.parameters(), 1e-5)\n",
        "auc_metric = ROCAUCMetric()"
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CqYvjh7BQyM2",
        "outputId": "d175a9c9-74ac-4124-e2e5-dbf00373dd94"
      },
      "source": [
        "# start a typical PyTorch training\n",
        "val_interval = 2\n",
        "best_metric = -1\n",
        "best_metric_epoch = -1\n",
        "writer = SummaryWriter()\n",
        "max_epochs = 400\n",
        "for epoch in range(max_epochs):\n",
        "        print(\"-\" * 10)\n",
        "        print(f\"epoch {epoch + 1}/{max_epochs}\")\n",
        "        model.train()\n",
        "        epoch_loss = 0\n",
        "        step = 0\n",
        "        for batch_data in train_loader:\n",
        "            step += 1\n",
        "            inputs, labels = batch_data[\"image\"].to(device), batch_data[\"label\"].to(device)\n",
        "            optimizer.zero_grad()\n",
        "            outputs = model(inputs)\n",
        "            loss = loss_function(outputs, labels)\n",
        "            loss.backward()\n",
        "            optimizer.step()\n",
        "            epoch_loss += loss.item()\n",
        "            epoch_len = len(train_ds) // train_loader.batch_size\n",
        "            print(f\"{step}/{epoch_len}, train_loss: {loss.item():.4f}\")\n",
        "            writer.add_scalar(\"train_loss\", loss.item(), epoch_len * epoch + step)\n",
        "        epoch_loss /= step\n",
        "        print(f\"epoch {epoch + 1} average loss: {epoch_loss:.4f}\")\n",
        "\n",
        "        if (epoch + 1) % val_interval == 0:\n",
        "            model.eval()\n",
        "            with torch.no_grad():\n",
        "                y_pred = torch.tensor([], dtype=torch.float32, device=device)\n",
        "                y = torch.tensor([], dtype=torch.long, device=device)\n",
        "                for val_data in val_loader:\n",
        "                    val_images, val_labels = val_data[\"image\"].to(device), val_data[\"label\"].to(device)\n",
        "                    y_pred = torch.cat([y_pred, model(val_images)], dim=0)\n",
        "                    y = torch.cat([y, val_labels], dim=0)\n",
        "\n",
        "                acc_value = torch.eq(y_pred.argmax(dim=1), y)\n",
        "                acc_metric = acc_value.sum().item() / len(acc_value)\n",
        "                y_onehot = [post_label(i) for i in decollate_batch(y)]\n",
        "                y_pred_act = [post_pred(i) for i in decollate_batch(y_pred)]\n",
        "                auc_metric(y_pred_act, y_onehot)\n",
        "                auc_result = auc_metric.aggregate()\n",
        "                auc_metric.reset()\n",
        "                del y_pred_act, y_onehot\n",
        "                if acc_metric > best_metric:\n",
        "                    best_metric = acc_metric\n",
        "                    best_metric_epoch = epoch + 1\n",
        "                    torch.save(model.state_dict(), \"/content/drive/My Drive/Spleen_AI/best_metric_model_classification3d.pth\")\n",
        "                    print(\"saved new best metric model\")\n",
        "                print(\n",
        "                    \"current epoch: {} current accuracy: {:.4f} current AUC: {:.4f} best accuracy: {:.4f} at epoch {}\".format(\n",
        "                        epoch + 1, acc_metric, auc_result, best_metric, best_metric_epoch\n",
        "                    )\n",
        "                )\n",
        "                writer.add_scalar(\"val_accuracy\", acc_metric, epoch + 1)\n",
        "print(f\"train completed, best_metric: {best_metric:.4f} at epoch: {best_metric_epoch}\")\n",
        "writer.close()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "----------\n",
            "epoch 1/400\n",
            "1/50, train_loss: 0.0531\n",
            "2/50, train_loss: 0.0691\n",
            "3/50, train_loss: 0.6785\n",
            "4/50, train_loss: 0.0697\n",
            "5/50, train_loss: 0.5392\n",
            "6/50, train_loss: 0.0970\n",
            "7/50, train_loss: 0.1158\n",
            "8/50, train_loss: 0.0865\n",
            "9/50, train_loss: 0.0586\n",
            "10/50, train_loss: 0.1770\n",
            "11/50, train_loss: 0.4609\n",
            "12/50, train_loss: 0.0605\n",
            "13/50, train_loss: 1.0035\n",
            "14/50, train_loss: 0.4252\n",
            "15/50, train_loss: 0.0569\n",
            "16/50, train_loss: 0.0918\n",
            "17/50, train_loss: 0.0768\n",
            "18/50, train_loss: 0.0703\n",
            "19/50, train_loss: 0.8685\n",
            "20/50, train_loss: 0.4686\n",
            "21/50, train_loss: 0.0828\n",
            "22/50, train_loss: 0.0576\n",
            "23/50, train_loss: 0.0718\n",
            "24/50, train_loss: 0.4636\n",
            "25/50, train_loss: 0.0495\n",
            "26/50, train_loss: 0.0658\n",
            "27/50, train_loss: 0.0918\n",
            "28/50, train_loss: 0.0700\n",
            "29/50, train_loss: 0.4214\n",
            "30/50, train_loss: 0.0572\n",
            "31/50, train_loss: 0.0676\n",
            "32/50, train_loss: 0.0608\n",
            "33/50, train_loss: 0.4824\n",
            "34/50, train_loss: 0.0952\n",
            "35/50, train_loss: 0.4427\n",
            "36/50, train_loss: 0.8345\n",
            "37/50, train_loss: 0.0764\n",
            "38/50, train_loss: 0.0721\n",
            "39/50, train_loss: 0.4675\n",
            "40/50, train_loss: 0.8863\n",
            "41/50, train_loss: 0.0597\n",
            "42/50, train_loss: 0.4759\n",
            "43/50, train_loss: 0.0501\n",
            "44/50, train_loss: 0.4217\n",
            "45/50, train_loss: 0.0568\n",
            "46/50, train_loss: 0.0514\n",
            "47/50, train_loss: 0.0741\n",
            "48/50, train_loss: 0.1208\n",
            "49/50, train_loss: 1.0262\n",
            "50/50, train_loss: 0.0677\n",
            "epoch 1 average loss: 0.2570\n",
            "----------\n",
            "epoch 2/400\n",
            "1/50, train_loss: 0.4650\n",
            "2/50, train_loss: 0.5119\n",
            "3/50, train_loss: 0.0876\n",
            "4/50, train_loss: 0.1174\n",
            "5/50, train_loss: 0.4234\n",
            "6/50, train_loss: 0.0586\n",
            "7/50, train_loss: 0.0723\n",
            "8/50, train_loss: 0.0892\n",
            "9/50, train_loss: 0.0617\n",
            "10/50, train_loss: 0.0649\n",
            "11/50, train_loss: 0.0603\n",
            "12/50, train_loss: 0.5445\n",
            "13/50, train_loss: 0.8676\n",
            "14/50, train_loss: 0.7937\n",
            "15/50, train_loss: 0.0533\n",
            "16/50, train_loss: 0.0479\n",
            "17/50, train_loss: 0.0653\n",
            "18/50, train_loss: 0.0694\n",
            "19/50, train_loss: 0.0649\n",
            "20/50, train_loss: 0.4635\n",
            "21/50, train_loss: 0.4687\n",
            "22/50, train_loss: 0.4491\n",
            "23/50, train_loss: 0.0523\n",
            "24/50, train_loss: 0.0517\n",
            "25/50, train_loss: 0.4051\n",
            "26/50, train_loss: 0.0600\n",
            "27/50, train_loss: 0.0445\n",
            "28/50, train_loss: 0.0440\n",
            "29/50, train_loss: 0.0642\n",
            "30/50, train_loss: 0.9583\n",
            "31/50, train_loss: 0.0823\n",
            "32/50, train_loss: 0.0431\n",
            "33/50, train_loss: 0.0863\n",
            "34/50, train_loss: 0.8107\n",
            "35/50, train_loss: 0.0439\n",
            "36/50, train_loss: 0.4519\n",
            "37/50, train_loss: 0.7604\n",
            "38/50, train_loss: 0.0502\n",
            "39/50, train_loss: 0.4485\n",
            "40/50, train_loss: 0.0669\n",
            "41/50, train_loss: 0.4343\n",
            "42/50, train_loss: 0.0485\n",
            "43/50, train_loss: 0.0548\n",
            "44/50, train_loss: 0.0460\n",
            "45/50, train_loss: 0.0463\n",
            "46/50, train_loss: 0.0770\n",
            "47/50, train_loss: 0.0782\n",
            "48/50, train_loss: 0.4722\n",
            "49/50, train_loss: 0.0444\n",
            "50/50, train_loss: 0.0688\n",
            "epoch 2 average loss: 0.2359\n",
            "saved new best metric model\n",
            "current epoch: 2 current accuracy: 0.6667 current AUC: 0.6639 best accuracy: 0.6667 at epoch 2\n",
            "----------\n",
            "epoch 3/400\n",
            "1/50, train_loss: 0.8480\n",
            "2/50, train_loss: 0.4340\n",
            "3/50, train_loss: 0.0441\n",
            "4/50, train_loss: 0.7692\n",
            "5/50, train_loss: 0.0546\n",
            "6/50, train_loss: 0.7331\n",
            "7/50, train_loss: 0.0945\n",
            "8/50, train_loss: 0.7927\n",
            "9/50, train_loss: 0.4921\n",
            "10/50, train_loss: 0.0419\n",
            "11/50, train_loss: 0.0558\n",
            "12/50, train_loss: 0.4674\n",
            "13/50, train_loss: 0.4650\n",
            "14/50, train_loss: 0.0483\n",
            "15/50, train_loss: 0.5574\n",
            "16/50, train_loss: 0.8152\n",
            "17/50, train_loss: 0.4368\n",
            "18/50, train_loss: 0.4741\n",
            "19/50, train_loss: 0.7840\n",
            "20/50, train_loss: 0.9279\n",
            "21/50, train_loss: 0.7470\n",
            "22/50, train_loss: 0.0449\n",
            "23/50, train_loss: 0.0433\n",
            "24/50, train_loss: 0.0674\n",
            "25/50, train_loss: 0.4803\n",
            "26/50, train_loss: 0.0585\n",
            "27/50, train_loss: 0.8905\n",
            "28/50, train_loss: 0.0399\n",
            "29/50, train_loss: 0.0611\n",
            "30/50, train_loss: 0.4998\n",
            "31/50, train_loss: 0.0698\n",
            "32/50, train_loss: 0.4219\n",
            "33/50, train_loss: 0.0755\n",
            "34/50, train_loss: 0.4972\n",
            "35/50, train_loss: 0.9022\n",
            "36/50, train_loss: 0.4434\n",
            "37/50, train_loss: 0.0726\n",
            "38/50, train_loss: 0.9221\n",
            "39/50, train_loss: 0.4522\n",
            "40/50, train_loss: 0.4397\n",
            "41/50, train_loss: 0.4460\n",
            "42/50, train_loss: 0.0463\n",
            "43/50, train_loss: 0.4681\n",
            "44/50, train_loss: 0.1030\n",
            "45/50, train_loss: 0.8462\n",
            "46/50, train_loss: 0.0661\n",
            "47/50, train_loss: 0.0948\n",
            "48/50, train_loss: 0.0614\n",
            "49/50, train_loss: 0.0751\n",
            "50/50, train_loss: 0.5318\n",
            "epoch 3 average loss: 0.3861\n",
            "----------\n",
            "epoch 4/400\n",
            "1/50, train_loss: 0.4439\n",
            "2/50, train_loss: 0.5304\n",
            "3/50, train_loss: 0.4440\n",
            "4/50, train_loss: 0.0767\n",
            "5/50, train_loss: 0.4425\n",
            "6/50, train_loss: 0.7646\n",
            "7/50, train_loss: 0.4394\n",
            "8/50, train_loss: 0.0955\n",
            "9/50, train_loss: 0.0566\n",
            "10/50, train_loss: 0.0708\n",
            "11/50, train_loss: 0.3718\n",
            "12/50, train_loss: 0.7823\n",
            "13/50, train_loss: 0.0548\n",
            "14/50, train_loss: 0.4357\n",
            "15/50, train_loss: 0.9505\n",
            "16/50, train_loss: 0.4021\n",
            "17/50, train_loss: 0.0832\n",
            "18/50, train_loss: 0.1056\n",
            "19/50, train_loss: 0.3972\n",
            "20/50, train_loss: 0.1077\n",
            "21/50, train_loss: 0.8543\n",
            "22/50, train_loss: 0.0429\n",
            "23/50, train_loss: 0.4326\n",
            "24/50, train_loss: 0.9608\n",
            "25/50, train_loss: 0.0630\n",
            "26/50, train_loss: 0.4851\n",
            "27/50, train_loss: 0.5682\n",
            "28/50, train_loss: 0.0489\n",
            "29/50, train_loss: 0.0531\n",
            "30/50, train_loss: 0.4507\n",
            "31/50, train_loss: 0.9774\n",
            "32/50, train_loss: 0.0760\n",
            "33/50, train_loss: 0.4689\n",
            "34/50, train_loss: 0.4766\n",
            "35/50, train_loss: 0.7725\n",
            "36/50, train_loss: 0.5989\n",
            "37/50, train_loss: 0.0808\n",
            "38/50, train_loss: 0.8131\n",
            "39/50, train_loss: 0.8522\n",
            "40/50, train_loss: 0.4117\n",
            "41/50, train_loss: 0.0653\n",
            "42/50, train_loss: 0.0695\n",
            "43/50, train_loss: 0.9830\n",
            "44/50, train_loss: 0.4652\n",
            "45/50, train_loss: 0.3972\n",
            "46/50, train_loss: 0.4164\n",
            "47/50, train_loss: 1.0461\n",
            "48/50, train_loss: 0.9755\n",
            "49/50, train_loss: 0.0775\n",
            "50/50, train_loss: 0.9392\n",
            "epoch 4 average loss: 0.4396\n",
            "current epoch: 4 current accuracy: 0.6667 current AUC: 0.6891 best accuracy: 0.6667 at epoch 2\n",
            "----------\n",
            "epoch 5/400\n",
            "1/50, train_loss: 0.8926\n",
            "2/50, train_loss: 0.7348\n",
            "3/50, train_loss: 0.5031\n",
            "4/50, train_loss: 0.0931\n",
            "5/50, train_loss: 0.4802\n",
            "6/50, train_loss: 0.4786\n",
            "7/50, train_loss: 0.1659\n",
            "8/50, train_loss: 0.4602\n",
            "9/50, train_loss: 0.0986\n",
            "10/50, train_loss: 0.7500\n",
            "11/50, train_loss: 0.5193\n",
            "12/50, train_loss: 0.0662\n",
            "13/50, train_loss: 0.1049\n",
            "14/50, train_loss: 0.0702\n",
            "15/50, train_loss: 0.1100\n",
            "16/50, train_loss: 0.4033\n",
            "17/50, train_loss: 0.4222\n",
            "18/50, train_loss: 0.4234\n",
            "19/50, train_loss: 0.4181\n",
            "20/50, train_loss: 0.9421\n",
            "21/50, train_loss: 0.1165\n",
            "22/50, train_loss: 0.4143\n",
            "23/50, train_loss: 0.0839\n",
            "24/50, train_loss: 0.0822\n",
            "25/50, train_loss: 0.0585\n",
            "26/50, train_loss: 0.0677\n",
            "27/50, train_loss: 0.1175\n",
            "28/50, train_loss: 0.0719\n",
            "29/50, train_loss: 0.8555\n",
            "30/50, train_loss: 0.4947\n",
            "31/50, train_loss: 0.8978\n",
            "32/50, train_loss: 0.0943\n",
            "33/50, train_loss: 0.0861\n",
            "34/50, train_loss: 0.0554\n",
            "35/50, train_loss: 0.4482\n",
            "36/50, train_loss: 0.4953\n",
            "37/50, train_loss: 0.0705\n",
            "38/50, train_loss: 0.0651\n",
            "39/50, train_loss: 0.8441\n",
            "40/50, train_loss: 0.7935\n",
            "41/50, train_loss: 0.6490\n",
            "42/50, train_loss: 0.0514\n",
            "43/50, train_loss: 0.0496\n",
            "44/50, train_loss: 0.1041\n",
            "45/50, train_loss: 0.0884\n",
            "46/50, train_loss: 0.8251\n",
            "47/50, train_loss: 0.0538\n",
            "48/50, train_loss: 0.5477\n",
            "49/50, train_loss: 0.0869\n",
            "50/50, train_loss: 0.4040\n",
            "epoch 5 average loss: 0.3442\n",
            "----------\n",
            "epoch 6/400\n",
            "1/50, train_loss: 0.0879\n",
            "2/50, train_loss: 0.6032\n",
            "3/50, train_loss: 0.0503\n",
            "4/50, train_loss: 0.8920\n",
            "5/50, train_loss: 0.0968\n",
            "6/50, train_loss: 0.4822\n",
            "7/50, train_loss: 0.1023\n",
            "8/50, train_loss: 0.0531\n",
            "9/50, train_loss: 0.5465\n",
            "10/50, train_loss: 0.4623\n",
            "11/50, train_loss: 0.4478\n",
            "12/50, train_loss: 0.8117\n",
            "13/50, train_loss: 0.5508\n",
            "14/50, train_loss: 0.4036\n",
            "15/50, train_loss: 0.0683\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "Traceback (most recent call last):\n",
            "  File \"/usr/lib/python3.7/multiprocessing/queues.py\", line 232, in _feed\n",
            "    close()\n",
            "  File \"/usr/lib/python3.7/multiprocessing/connection.py\", line 177, in close\n",
            "    self._close()\n",
            "  File \"/usr/lib/python3.7/multiprocessing/connection.py\", line 361, in _close\n",
            "    _close(self._handle)\n",
            "OSError: [Errno 9] Bad file descriptor\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "\u001b[1;30;43mDie letzten 5000 Zeilen der Streamingausgabe wurden abgeschnitten.\u001b[0m\n",
            "11/50, train_loss: 0.0065\n",
            "12/50, train_loss: 0.0755\n",
            "13/50, train_loss: 0.0060\n",
            "14/50, train_loss: 0.0066\n",
            "15/50, train_loss: 0.0848\n",
            "16/50, train_loss: 0.0081\n",
            "17/50, train_loss: 0.1196\n",
            "18/50, train_loss: 0.0787\n",
            "19/50, train_loss: 0.0899\n",
            "20/50, train_loss: 0.1038\n",
            "21/50, train_loss: 0.0062\n",
            "22/50, train_loss: 0.1522\n",
            "23/50, train_loss: 0.0059\n",
            "24/50, train_loss: 0.0635\n",
            "25/50, train_loss: 0.1124\n",
            "26/50, train_loss: 0.0792\n",
            "27/50, train_loss: 0.0059\n",
            "28/50, train_loss: 0.1131\n",
            "29/50, train_loss: 0.1214\n",
            "30/50, train_loss: 0.1285\n",
            "31/50, train_loss: 0.1009\n",
            "32/50, train_loss: 0.0064\n",
            "33/50, train_loss: 0.2027\n",
            "34/50, train_loss: 0.1356\n",
            "35/50, train_loss: 0.0748\n",
            "36/50, train_loss: 0.1845\n",
            "37/50, train_loss: 0.0079\n",
            "38/50, train_loss: 0.0053\n",
            "39/50, train_loss: 0.0060\n",
            "40/50, train_loss: 0.0094\n",
            "41/50, train_loss: 0.0075\n",
            "42/50, train_loss: 0.0815\n",
            "43/50, train_loss: 0.0748\n",
            "44/50, train_loss: 0.1405\n",
            "45/50, train_loss: 0.1960\n",
            "46/50, train_loss: 0.0065\n",
            "47/50, train_loss: 0.1530\n",
            "48/50, train_loss: 0.0663\n",
            "49/50, train_loss: 0.0796\n",
            "50/50, train_loss: 0.0070\n",
            "epoch 84 average loss: 0.0738\n",
            "current epoch: 84 current accuracy: 0.7500 current AUC: 0.6891 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 85/400\n",
            "1/50, train_loss: 0.1018\n",
            "2/50, train_loss: 0.0074\n",
            "3/50, train_loss: 0.0069\n",
            "4/50, train_loss: 0.0068\n",
            "5/50, train_loss: 0.0714\n",
            "6/50, train_loss: 0.1180\n",
            "7/50, train_loss: 0.0079\n",
            "8/50, train_loss: 0.1867\n",
            "9/50, train_loss: 0.0839\n",
            "10/50, train_loss: 0.0062\n",
            "11/50, train_loss: 0.0058\n",
            "12/50, train_loss: 0.0057\n",
            "13/50, train_loss: 0.0067\n",
            "14/50, train_loss: 0.0841\n",
            "15/50, train_loss: 0.0073\n",
            "16/50, train_loss: 0.0070\n",
            "17/50, train_loss: 0.0072\n",
            "18/50, train_loss: 0.0086\n",
            "19/50, train_loss: 0.0781\n",
            "20/50, train_loss: 0.0071\n",
            "21/50, train_loss: 0.0059\n",
            "22/50, train_loss: 0.1654\n",
            "23/50, train_loss: 0.0060\n",
            "24/50, train_loss: 0.1234\n",
            "25/50, train_loss: 0.1413\n",
            "26/50, train_loss: 0.0057\n",
            "27/50, train_loss: 0.0628\n",
            "28/50, train_loss: 0.0061\n",
            "29/50, train_loss: 0.0790\n",
            "30/50, train_loss: 0.0847\n",
            "31/50, train_loss: 0.1268\n",
            "32/50, train_loss: 0.2014\n",
            "33/50, train_loss: 0.0822\n",
            "34/50, train_loss: 0.0089\n",
            "35/50, train_loss: 0.0060\n",
            "36/50, train_loss: 0.0058\n",
            "37/50, train_loss: 0.0823\n",
            "38/50, train_loss: 0.1418\n",
            "39/50, train_loss: 0.1338\n",
            "40/50, train_loss: 0.0927\n",
            "41/50, train_loss: 0.0751\n",
            "42/50, train_loss: 0.0051\n",
            "43/50, train_loss: 0.0070\n",
            "44/50, train_loss: 0.0845\n",
            "45/50, train_loss: 0.0062\n",
            "46/50, train_loss: 0.0872\n",
            "47/50, train_loss: 0.1495\n",
            "48/50, train_loss: 0.0823\n",
            "49/50, train_loss: 0.0711\n",
            "50/50, train_loss: 0.1456\n",
            "epoch 85 average loss: 0.0618\n",
            "----------\n",
            "epoch 86/400\n",
            "1/50, train_loss: 0.0056\n",
            "2/50, train_loss: 0.0054\n",
            "3/50, train_loss: 0.0060\n",
            "4/50, train_loss: 0.0058\n",
            "5/50, train_loss: 0.1103\n",
            "6/50, train_loss: 0.0097\n",
            "7/50, train_loss: 0.0053\n",
            "8/50, train_loss: 0.0056\n",
            "9/50, train_loss: 0.1404\n",
            "10/50, train_loss: 0.0061\n",
            "11/50, train_loss: 0.1457\n",
            "12/50, train_loss: 0.0914\n",
            "13/50, train_loss: 0.1467\n",
            "14/50, train_loss: 0.1672\n",
            "15/50, train_loss: 0.0837\n",
            "16/50, train_loss: 0.0054\n",
            "17/50, train_loss: 0.0649\n",
            "18/50, train_loss: 0.0062\n",
            "19/50, train_loss: 0.0870\n",
            "20/50, train_loss: 0.0086\n",
            "21/50, train_loss: 0.0063\n",
            "22/50, train_loss: 0.0056\n",
            "23/50, train_loss: 0.0074\n",
            "24/50, train_loss: 0.1216\n",
            "25/50, train_loss: 0.0075\n",
            "26/50, train_loss: 0.0061\n",
            "27/50, train_loss: 0.0807\n",
            "28/50, train_loss: 0.1459\n",
            "29/50, train_loss: 0.0838\n",
            "30/50, train_loss: 0.1109\n",
            "31/50, train_loss: 0.0678\n",
            "32/50, train_loss: 0.0069\n",
            "33/50, train_loss: 0.0076\n",
            "34/50, train_loss: 0.0862\n",
            "35/50, train_loss: 0.0788\n",
            "36/50, train_loss: 0.0989\n",
            "37/50, train_loss: 0.0782\n",
            "38/50, train_loss: 0.1304\n",
            "39/50, train_loss: 0.0043\n",
            "40/50, train_loss: 0.0060\n",
            "41/50, train_loss: 0.0775\n",
            "42/50, train_loss: 0.0091\n",
            "43/50, train_loss: 0.0844\n",
            "44/50, train_loss: 0.0066\n",
            "45/50, train_loss: 0.0707\n",
            "46/50, train_loss: 0.0061\n",
            "47/50, train_loss: 0.1344\n",
            "48/50, train_loss: 0.0085\n",
            "49/50, train_loss: 0.0050\n",
            "50/50, train_loss: 0.1399\n",
            "epoch 86 average loss: 0.0558\n",
            "current epoch: 86 current accuracy: 0.6667 current AUC: 0.7059 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 87/400\n",
            "1/50, train_loss: 0.0053\n",
            "2/50, train_loss: 0.0743\n",
            "3/50, train_loss: 0.1200\n",
            "4/50, train_loss: 0.0642\n",
            "5/50, train_loss: 0.0053\n",
            "6/50, train_loss: 0.0806\n",
            "7/50, train_loss: 0.1467\n",
            "8/50, train_loss: 0.1531\n",
            "9/50, train_loss: 0.1396\n",
            "10/50, train_loss: 0.0074\n",
            "11/50, train_loss: 0.1128\n",
            "12/50, train_loss: 0.1922\n",
            "13/50, train_loss: 0.0060\n",
            "14/50, train_loss: 0.0055\n",
            "15/50, train_loss: 0.1234\n",
            "16/50, train_loss: 0.0068\n",
            "17/50, train_loss: 0.0808\n",
            "18/50, train_loss: 0.0056\n",
            "19/50, train_loss: 0.1433\n",
            "20/50, train_loss: 0.0801\n",
            "21/50, train_loss: 0.0745\n",
            "22/50, train_loss: 0.0054\n",
            "23/50, train_loss: 0.0053\n",
            "24/50, train_loss: 0.0840\n",
            "25/50, train_loss: 0.1833\n",
            "26/50, train_loss: 0.1425\n",
            "27/50, train_loss: 0.0073\n",
            "28/50, train_loss: 0.0770\n",
            "29/50, train_loss: 0.0061\n",
            "30/50, train_loss: 0.0058\n",
            "31/50, train_loss: 0.0637\n",
            "32/50, train_loss: 0.0059\n",
            "33/50, train_loss: 0.0792\n",
            "34/50, train_loss: 0.1221\n",
            "35/50, train_loss: 0.0961\n",
            "36/50, train_loss: 0.0065\n",
            "37/50, train_loss: 0.0060\n",
            "38/50, train_loss: 0.0931\n",
            "39/50, train_loss: 0.1161\n",
            "40/50, train_loss: 0.0068\n",
            "41/50, train_loss: 0.1391\n",
            "42/50, train_loss: 0.0064\n",
            "43/50, train_loss: 0.0050\n",
            "44/50, train_loss: 0.0067\n",
            "45/50, train_loss: 0.0070\n",
            "46/50, train_loss: 0.0671\n",
            "47/50, train_loss: 0.0058\n",
            "48/50, train_loss: 0.0058\n",
            "49/50, train_loss: 0.0060\n",
            "50/50, train_loss: 0.0592\n",
            "epoch 87 average loss: 0.0610\n",
            "----------\n",
            "epoch 88/400\n",
            "1/50, train_loss: 0.0883\n",
            "2/50, train_loss: 0.0065\n",
            "3/50, train_loss: 0.0066\n",
            "4/50, train_loss: 0.0755\n",
            "5/50, train_loss: 0.1202\n",
            "6/50, train_loss: 0.1432\n",
            "7/50, train_loss: 0.0064\n",
            "8/50, train_loss: 0.0064\n",
            "9/50, train_loss: 0.0075\n",
            "10/50, train_loss: 0.0053\n",
            "11/50, train_loss: 0.1794\n",
            "12/50, train_loss: 0.1101\n",
            "13/50, train_loss: 0.0672\n",
            "14/50, train_loss: 0.1240\n",
            "15/50, train_loss: 0.1001\n",
            "16/50, train_loss: 0.0062\n",
            "17/50, train_loss: 0.0805\n",
            "18/50, train_loss: 0.1283\n",
            "19/50, train_loss: 0.0085\n",
            "20/50, train_loss: 0.0074\n",
            "21/50, train_loss: 0.0987\n",
            "22/50, train_loss: 0.0771\n",
            "23/50, train_loss: 0.0069\n",
            "24/50, train_loss: 0.1232\n",
            "25/50, train_loss: 0.0055\n",
            "26/50, train_loss: 0.0047\n",
            "27/50, train_loss: 0.0056\n",
            "28/50, train_loss: 0.0038\n",
            "29/50, train_loss: 0.0868\n",
            "30/50, train_loss: 0.0048\n",
            "31/50, train_loss: 0.0830\n",
            "32/50, train_loss: 0.0075\n",
            "33/50, train_loss: 0.1088\n",
            "34/50, train_loss: 0.0645\n",
            "35/50, train_loss: 0.1258\n",
            "36/50, train_loss: 0.1837\n",
            "37/50, train_loss: 0.0718\n",
            "38/50, train_loss: 0.0803\n",
            "39/50, train_loss: 0.0064\n",
            "40/50, train_loss: 0.0060\n",
            "41/50, train_loss: 0.0755\n",
            "42/50, train_loss: 0.0914\n",
            "43/50, train_loss: 0.0052\n",
            "44/50, train_loss: 0.0794\n",
            "45/50, train_loss: 0.0054\n",
            "46/50, train_loss: 0.0656\n",
            "47/50, train_loss: 0.0048\n",
            "48/50, train_loss: 0.0061\n",
            "49/50, train_loss: 0.0057\n",
            "50/50, train_loss: 0.2120\n",
            "epoch 88 average loss: 0.0597\n",
            "current epoch: 88 current accuracy: 0.7500 current AUC: 0.6891 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 89/400\n",
            "1/50, train_loss: 0.1035\n",
            "2/50, train_loss: 0.0729\n",
            "3/50, train_loss: 0.0608\n",
            "4/50, train_loss: 0.0795\n",
            "5/50, train_loss: 0.1370\n",
            "6/50, train_loss: 0.0791\n",
            "7/50, train_loss: 0.0602\n",
            "8/50, train_loss: 0.1177\n",
            "9/50, train_loss: 0.0077\n",
            "10/50, train_loss: 0.0084\n",
            "11/50, train_loss: 0.0062\n",
            "12/50, train_loss: 0.0060\n",
            "13/50, train_loss: 0.0058\n",
            "14/50, train_loss: 0.1133\n",
            "15/50, train_loss: 0.0084\n",
            "16/50, train_loss: 0.1102\n",
            "17/50, train_loss: 0.0047\n",
            "18/50, train_loss: 0.1578\n",
            "19/50, train_loss: 0.0056\n",
            "20/50, train_loss: 0.1232\n",
            "21/50, train_loss: 0.0053\n",
            "22/50, train_loss: 0.0057\n",
            "23/50, train_loss: 0.0042\n",
            "24/50, train_loss: 0.0769\n",
            "25/50, train_loss: 0.1378\n",
            "26/50, train_loss: 0.0084\n",
            "27/50, train_loss: 0.0049\n",
            "28/50, train_loss: 0.0597\n",
            "29/50, train_loss: 0.0068\n",
            "30/50, train_loss: 0.0904\n",
            "31/50, train_loss: 0.0063\n",
            "32/50, train_loss: 0.0052\n",
            "33/50, train_loss: 0.0048\n",
            "34/50, train_loss: 0.0056\n",
            "35/50, train_loss: 0.0682\n",
            "36/50, train_loss: 0.1105\n",
            "37/50, train_loss: 0.0045\n",
            "38/50, train_loss: 0.1720\n",
            "39/50, train_loss: 0.0052\n",
            "40/50, train_loss: 0.0061\n",
            "41/50, train_loss: 0.2515\n",
            "42/50, train_loss: 0.0069\n",
            "43/50, train_loss: 0.0717\n",
            "44/50, train_loss: 0.0963\n",
            "45/50, train_loss: 0.1889\n",
            "46/50, train_loss: 0.0043\n",
            "47/50, train_loss: 0.0688\n",
            "48/50, train_loss: 0.0768\n",
            "49/50, train_loss: 0.0834\n",
            "50/50, train_loss: 0.1310\n",
            "epoch 89 average loss: 0.0607\n",
            "----------\n",
            "epoch 90/400\n",
            "1/50, train_loss: 0.0047\n",
            "2/50, train_loss: 0.0077\n",
            "3/50, train_loss: 0.0618\n",
            "4/50, train_loss: 0.0761\n",
            "5/50, train_loss: 0.0615\n",
            "6/50, train_loss: 0.0062\n",
            "7/50, train_loss: 0.0716\n",
            "8/50, train_loss: 0.1948\n",
            "9/50, train_loss: 0.0054\n",
            "10/50, train_loss: 0.0091\n",
            "11/50, train_loss: 0.0056\n",
            "12/50, train_loss: 0.0066\n",
            "13/50, train_loss: 0.1913\n",
            "14/50, train_loss: 0.1551\n",
            "15/50, train_loss: 0.0064\n",
            "16/50, train_loss: 0.0046\n",
            "17/50, train_loss: 0.0051\n",
            "18/50, train_loss: 0.0078\n",
            "19/50, train_loss: 0.0865\n",
            "20/50, train_loss: 0.1123\n",
            "21/50, train_loss: 0.0049\n",
            "22/50, train_loss: 0.0815\n",
            "23/50, train_loss: 0.0059\n",
            "24/50, train_loss: 0.0775\n",
            "25/50, train_loss: 0.0747\n",
            "26/50, train_loss: 0.1110\n",
            "27/50, train_loss: 0.1280\n",
            "28/50, train_loss: 0.0062\n",
            "29/50, train_loss: 0.2062\n",
            "30/50, train_loss: 0.1115\n",
            "31/50, train_loss: 0.0681\n",
            "32/50, train_loss: 0.0065\n",
            "33/50, train_loss: 0.0659\n",
            "34/50, train_loss: 0.0791\n",
            "35/50, train_loss: 0.0043\n",
            "36/50, train_loss: 0.1594\n",
            "37/50, train_loss: 0.0054\n",
            "38/50, train_loss: 0.0067\n",
            "39/50, train_loss: 0.0057\n",
            "40/50, train_loss: 0.0925\n",
            "41/50, train_loss: 0.0785\n",
            "42/50, train_loss: 0.0062\n",
            "43/50, train_loss: 0.0092\n",
            "44/50, train_loss: 0.0845\n",
            "45/50, train_loss: 0.1000\n",
            "46/50, train_loss: 0.0057\n",
            "47/50, train_loss: 0.0070\n",
            "48/50, train_loss: 0.0041\n",
            "49/50, train_loss: 0.0058\n",
            "50/50, train_loss: 0.1287\n",
            "epoch 90 average loss: 0.0562\n",
            "current epoch: 90 current accuracy: 0.6667 current AUC: 0.6807 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 91/400\n",
            "1/50, train_loss: 0.1613\n",
            "2/50, train_loss: 0.0062\n",
            "3/50, train_loss: 0.0868\n",
            "4/50, train_loss: 0.0649\n",
            "5/50, train_loss: 0.0061\n",
            "6/50, train_loss: 0.0061\n",
            "7/50, train_loss: 0.1130\n",
            "8/50, train_loss: 0.0831\n",
            "9/50, train_loss: 0.1091\n",
            "10/50, train_loss: 0.0057\n",
            "11/50, train_loss: 0.1003\n",
            "12/50, train_loss: 0.0972\n",
            "13/50, train_loss: 0.0062\n",
            "14/50, train_loss: 0.1646\n",
            "15/50, train_loss: 0.1162\n",
            "16/50, train_loss: 0.0059\n",
            "17/50, train_loss: 0.0669\n",
            "18/50, train_loss: 0.2010\n",
            "19/50, train_loss: 0.0057\n",
            "20/50, train_loss: 0.0748\n",
            "21/50, train_loss: 0.0062\n",
            "22/50, train_loss: 0.0851\n",
            "23/50, train_loss: 0.0060\n",
            "24/50, train_loss: 0.1038\n",
            "25/50, train_loss: 0.0063\n",
            "26/50, train_loss: 0.0048\n",
            "27/50, train_loss: 0.0060\n",
            "28/50, train_loss: 0.0828\n",
            "29/50, train_loss: 0.1468\n",
            "30/50, train_loss: 0.0777\n",
            "31/50, train_loss: 0.0994\n",
            "32/50, train_loss: 0.0924\n",
            "33/50, train_loss: 0.0074\n",
            "34/50, train_loss: 0.0060\n",
            "35/50, train_loss: 0.0069\n",
            "36/50, train_loss: 0.0068\n",
            "37/50, train_loss: 0.0718\n",
            "38/50, train_loss: 0.0813\n",
            "39/50, train_loss: 0.0641\n",
            "40/50, train_loss: 0.1236\n",
            "41/50, train_loss: 0.0069\n",
            "42/50, train_loss: 0.1846\n",
            "43/50, train_loss: 0.0053\n",
            "44/50, train_loss: 0.0053\n",
            "45/50, train_loss: 0.0620\n",
            "46/50, train_loss: 0.0071\n",
            "47/50, train_loss: 0.0057\n",
            "48/50, train_loss: 0.0590\n",
            "49/50, train_loss: 0.1743\n",
            "50/50, train_loss: 0.0787\n",
            "epoch 91 average loss: 0.0631\n",
            "----------\n",
            "epoch 92/400\n",
            "1/50, train_loss: 0.0061\n",
            "2/50, train_loss: 0.0576\n",
            "3/50, train_loss: 0.1375\n",
            "4/50, train_loss: 0.0735\n",
            "5/50, train_loss: 0.1613\n",
            "6/50, train_loss: 0.1575\n",
            "7/50, train_loss: 0.0957\n",
            "8/50, train_loss: 0.0634\n",
            "9/50, train_loss: 0.0063\n",
            "10/50, train_loss: 0.0066\n",
            "11/50, train_loss: 0.0841\n",
            "12/50, train_loss: 0.0081\n",
            "13/50, train_loss: 0.0112\n",
            "14/50, train_loss: 0.1060\n",
            "15/50, train_loss: 0.1206\n",
            "16/50, train_loss: 0.1541\n",
            "17/50, train_loss: 0.0047\n",
            "18/50, train_loss: 0.1102\n",
            "19/50, train_loss: 0.0066\n",
            "20/50, train_loss: 0.0641\n",
            "21/50, train_loss: 0.0075\n",
            "22/50, train_loss: 0.0682\n",
            "23/50, train_loss: 0.0994\n",
            "24/50, train_loss: 0.0773\n",
            "25/50, train_loss: 0.0056\n",
            "26/50, train_loss: 0.0588\n",
            "27/50, train_loss: 0.1160\n",
            "28/50, train_loss: 0.0084\n",
            "29/50, train_loss: 0.1120\n",
            "30/50, train_loss: 0.0079\n",
            "31/50, train_loss: 0.0071\n",
            "32/50, train_loss: 0.0083\n",
            "33/50, train_loss: 0.0815\n",
            "34/50, train_loss: 0.2332\n",
            "35/50, train_loss: 0.0995\n",
            "36/50, train_loss: 0.0071\n",
            "37/50, train_loss: 0.0080\n",
            "38/50, train_loss: 0.0072\n",
            "39/50, train_loss: 0.0061\n",
            "40/50, train_loss: 0.0066\n",
            "41/50, train_loss: 0.0058\n",
            "42/50, train_loss: 0.1071\n",
            "43/50, train_loss: 0.0785\n",
            "44/50, train_loss: 0.0733\n",
            "45/50, train_loss: 0.0065\n",
            "46/50, train_loss: 0.0076\n",
            "47/50, train_loss: 0.0688\n",
            "48/50, train_loss: 0.0787\n",
            "49/50, train_loss: 0.0065\n",
            "50/50, train_loss: 0.0055\n",
            "epoch 92 average loss: 0.0580\n",
            "current epoch: 92 current accuracy: 0.7500 current AUC: 0.7059 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 93/400\n",
            "1/50, train_loss: 0.0076\n",
            "2/50, train_loss: 0.0080\n",
            "3/50, train_loss: 0.1462\n",
            "4/50, train_loss: 0.0065\n",
            "5/50, train_loss: 0.0074\n",
            "6/50, train_loss: 0.2018\n",
            "7/50, train_loss: 0.0069\n",
            "8/50, train_loss: 0.0063\n",
            "9/50, train_loss: 0.2003\n",
            "10/50, train_loss: 0.0888\n",
            "11/50, train_loss: 0.0069\n",
            "12/50, train_loss: 0.0661\n",
            "13/50, train_loss: 0.0057\n",
            "14/50, train_loss: 0.1969\n",
            "15/50, train_loss: 0.0664\n",
            "16/50, train_loss: 0.0062\n",
            "17/50, train_loss: 0.0873\n",
            "18/50, train_loss: 0.0060\n",
            "19/50, train_loss: 0.0064\n",
            "20/50, train_loss: 0.1132\n",
            "21/50, train_loss: 0.0727\n",
            "22/50, train_loss: 0.0784\n",
            "23/50, train_loss: 0.0731\n",
            "24/50, train_loss: 0.0061\n",
            "25/50, train_loss: 0.0931\n",
            "26/50, train_loss: 0.1234\n",
            "27/50, train_loss: 0.0603\n",
            "28/50, train_loss: 0.0055\n",
            "29/50, train_loss: 0.1285\n",
            "30/50, train_loss: 0.0793\n",
            "31/50, train_loss: 0.0673\n",
            "32/50, train_loss: 0.0059\n",
            "33/50, train_loss: 0.0072\n",
            "34/50, train_loss: 0.1190\n",
            "35/50, train_loss: 0.0072\n",
            "36/50, train_loss: 0.0077\n",
            "37/50, train_loss: 0.0629\n",
            "38/50, train_loss: 0.0957\n",
            "39/50, train_loss: 0.0861\n",
            "40/50, train_loss: 0.0058\n",
            "41/50, train_loss: 0.0597\n",
            "42/50, train_loss: 0.0657\n",
            "43/50, train_loss: 0.0054\n",
            "44/50, train_loss: 0.0081\n",
            "45/50, train_loss: 0.0740\n",
            "46/50, train_loss: 0.0073\n",
            "47/50, train_loss: 0.1385\n",
            "48/50, train_loss: 0.0044\n",
            "49/50, train_loss: 0.0832\n",
            "50/50, train_loss: 0.0084\n",
            "epoch 93 average loss: 0.0576\n",
            "----------\n",
            "epoch 94/400\n",
            "1/50, train_loss: 0.0059\n",
            "2/50, train_loss: 0.1140\n",
            "3/50, train_loss: 0.0065\n",
            "4/50, train_loss: 0.1211\n",
            "5/50, train_loss: 0.0571\n",
            "6/50, train_loss: 0.0079\n",
            "7/50, train_loss: 0.0072\n",
            "8/50, train_loss: 0.1379\n",
            "9/50, train_loss: 0.1193\n",
            "10/50, train_loss: 0.0076\n",
            "11/50, train_loss: 0.0651\n",
            "12/50, train_loss: 0.0053\n",
            "13/50, train_loss: 0.0061\n",
            "14/50, train_loss: 0.0898\n",
            "15/50, train_loss: 0.0054\n",
            "16/50, train_loss: 0.0062\n",
            "17/50, train_loss: 0.0058\n",
            "18/50, train_loss: 0.0614\n",
            "19/50, train_loss: 0.0563\n",
            "20/50, train_loss: 0.0061\n",
            "21/50, train_loss: 0.0793\n",
            "22/50, train_loss: 0.1331\n",
            "23/50, train_loss: 0.1221\n",
            "24/50, train_loss: 0.0054\n",
            "25/50, train_loss: 0.0041\n",
            "26/50, train_loss: 0.0885\n",
            "27/50, train_loss: 0.0605\n",
            "28/50, train_loss: 0.0075\n",
            "29/50, train_loss: 0.0772\n",
            "30/50, train_loss: 0.1055\n",
            "31/50, train_loss: 0.0054\n",
            "32/50, train_loss: 0.0065\n",
            "33/50, train_loss: 0.0071\n",
            "34/50, train_loss: 0.0967\n",
            "35/50, train_loss: 0.0058\n",
            "36/50, train_loss: 0.0054\n",
            "37/50, train_loss: 0.0044\n",
            "38/50, train_loss: 0.0551\n",
            "39/50, train_loss: 0.0785\n",
            "40/50, train_loss: 0.0798\n",
            "41/50, train_loss: 0.0600\n",
            "42/50, train_loss: 0.0052\n",
            "43/50, train_loss: 0.1262\n",
            "44/50, train_loss: 0.0855\n",
            "45/50, train_loss: 0.0709\n",
            "46/50, train_loss: 0.0742\n",
            "47/50, train_loss: 0.1836\n",
            "48/50, train_loss: 0.0058\n",
            "49/50, train_loss: 0.1363\n",
            "50/50, train_loss: 0.0047\n",
            "epoch 94 average loss: 0.0535\n",
            "current epoch: 94 current accuracy: 0.7500 current AUC: 0.6639 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 95/400\n",
            "1/50, train_loss: 0.1241\n",
            "2/50, train_loss: 0.0051\n",
            "3/50, train_loss: 0.0067\n",
            "4/50, train_loss: 0.0680\n",
            "5/50, train_loss: 0.0696\n",
            "6/50, train_loss: 0.0928\n",
            "7/50, train_loss: 0.0062\n",
            "8/50, train_loss: 0.0048\n",
            "9/50, train_loss: 0.0639\n",
            "10/50, train_loss: 0.0057\n",
            "11/50, train_loss: 0.0061\n",
            "12/50, train_loss: 0.0831\n",
            "13/50, train_loss: 0.0963\n",
            "14/50, train_loss: 0.0057\n",
            "15/50, train_loss: 0.0675\n",
            "16/50, train_loss: 0.0744\n",
            "17/50, train_loss: 0.1040\n",
            "18/50, train_loss: 0.0575\n",
            "19/50, train_loss: 0.1334\n",
            "20/50, train_loss: 0.0063\n",
            "21/50, train_loss: 0.1182\n",
            "22/50, train_loss: 0.0060\n",
            "23/50, train_loss: 0.0054\n",
            "24/50, train_loss: 0.0745\n",
            "25/50, train_loss: 0.0059\n",
            "26/50, train_loss: 0.0617\n",
            "27/50, train_loss: 0.0064\n",
            "28/50, train_loss: 0.0068\n",
            "29/50, train_loss: 0.0052\n",
            "30/50, train_loss: 0.0051\n",
            "31/50, train_loss: 0.0642\n",
            "32/50, train_loss: 0.0813\n",
            "33/50, train_loss: 0.1061\n",
            "34/50, train_loss: 0.0043\n",
            "35/50, train_loss: 0.1398\n",
            "36/50, train_loss: 0.2335\n",
            "37/50, train_loss: 0.0953\n",
            "38/50, train_loss: 0.0546\n",
            "39/50, train_loss: 0.0053\n",
            "40/50, train_loss: 0.1025\n",
            "41/50, train_loss: 0.0045\n",
            "42/50, train_loss: 0.0777\n",
            "43/50, train_loss: 0.0051\n",
            "44/50, train_loss: 0.0520\n",
            "45/50, train_loss: 0.0791\n",
            "46/50, train_loss: 0.0047\n",
            "47/50, train_loss: 0.0065\n",
            "48/50, train_loss: 0.0963\n",
            "49/50, train_loss: 0.0048\n",
            "50/50, train_loss: 0.0055\n",
            "epoch 95 average loss: 0.0520\n",
            "----------\n",
            "epoch 96/400\n",
            "1/50, train_loss: 0.0715\n",
            "2/50, train_loss: 0.1150\n",
            "3/50, train_loss: 0.1219\n",
            "4/50, train_loss: 0.1008\n",
            "5/50, train_loss: 0.0053\n",
            "6/50, train_loss: 0.0057\n",
            "7/50, train_loss: 0.0718\n",
            "8/50, train_loss: 0.0744\n",
            "9/50, train_loss: 0.0863\n",
            "10/50, train_loss: 0.0094\n",
            "11/50, train_loss: 0.0049\n",
            "12/50, train_loss: 0.0052\n",
            "13/50, train_loss: 0.0736\n",
            "14/50, train_loss: 0.0047\n",
            "15/50, train_loss: 0.0757\n",
            "16/50, train_loss: 0.0050\n",
            "17/50, train_loss: 0.1015\n",
            "18/50, train_loss: 0.0058\n",
            "19/50, train_loss: 0.0045\n",
            "20/50, train_loss: 0.2580\n",
            "21/50, train_loss: 0.1080\n",
            "22/50, train_loss: 0.0071\n",
            "23/50, train_loss: 0.1032\n",
            "24/50, train_loss: 0.0052\n",
            "25/50, train_loss: 0.0061\n",
            "26/50, train_loss: 0.0634\n",
            "27/50, train_loss: 0.0040\n",
            "28/50, train_loss: 0.0047\n",
            "29/50, train_loss: 0.0654\n",
            "30/50, train_loss: 0.0623\n",
            "31/50, train_loss: 0.0067\n",
            "32/50, train_loss: 0.0054\n",
            "33/50, train_loss: 0.1019\n",
            "34/50, train_loss: 0.0549\n",
            "35/50, train_loss: 0.0924\n",
            "36/50, train_loss: 0.0608\n",
            "37/50, train_loss: 0.0050\n",
            "38/50, train_loss: 0.0062\n",
            "39/50, train_loss: 0.0070\n",
            "40/50, train_loss: 0.0581\n",
            "41/50, train_loss: 0.0065\n",
            "42/50, train_loss: 0.0591\n",
            "43/50, train_loss: 0.0676\n",
            "44/50, train_loss: 0.1206\n",
            "45/50, train_loss: 0.0747\n",
            "46/50, train_loss: 0.0059\n",
            "47/50, train_loss: 0.0755\n",
            "48/50, train_loss: 0.0059\n",
            "49/50, train_loss: 0.0052\n",
            "50/50, train_loss: 0.0544\n",
            "epoch 96 average loss: 0.0501\n",
            "current epoch: 96 current accuracy: 0.7500 current AUC: 0.6975 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 97/400\n",
            "1/50, train_loss: 0.0037\n",
            "2/50, train_loss: 0.0073\n",
            "3/50, train_loss: 0.0787\n",
            "4/50, train_loss: 0.0073\n",
            "5/50, train_loss: 0.0040\n",
            "6/50, train_loss: 0.0624\n",
            "7/50, train_loss: 0.0057\n",
            "8/50, train_loss: 0.0061\n",
            "9/50, train_loss: 0.0079\n",
            "10/50, train_loss: 0.0857\n",
            "11/50, train_loss: 0.0942\n",
            "12/50, train_loss: 0.0902\n",
            "13/50, train_loss: 0.0071\n",
            "14/50, train_loss: 0.0054\n",
            "15/50, train_loss: 0.0049\n",
            "16/50, train_loss: 0.1707\n",
            "17/50, train_loss: 0.1165\n",
            "18/50, train_loss: 0.0047\n",
            "19/50, train_loss: 0.0072\n",
            "20/50, train_loss: 0.0660\n",
            "21/50, train_loss: 0.0987\n",
            "22/50, train_loss: 0.0542\n",
            "23/50, train_loss: 0.0651\n",
            "24/50, train_loss: 0.1179\n",
            "25/50, train_loss: 0.0067\n",
            "26/50, train_loss: 0.0045\n",
            "27/50, train_loss: 0.0565\n",
            "28/50, train_loss: 0.0958\n",
            "29/50, train_loss: 0.0755\n",
            "30/50, train_loss: 0.0601\n",
            "31/50, train_loss: 0.0607\n",
            "32/50, train_loss: 0.0060\n",
            "33/50, train_loss: 0.0665\n",
            "34/50, train_loss: 0.0053\n",
            "35/50, train_loss: 0.0050\n",
            "36/50, train_loss: 0.0512\n",
            "37/50, train_loss: 0.0047\n",
            "38/50, train_loss: 0.0067\n",
            "39/50, train_loss: 0.0062\n",
            "40/50, train_loss: 0.1167\n",
            "41/50, train_loss: 0.0961\n",
            "42/50, train_loss: 0.0057\n",
            "43/50, train_loss: 0.0059\n",
            "44/50, train_loss: 0.0561\n",
            "45/50, train_loss: 0.0049\n",
            "46/50, train_loss: 0.0052\n",
            "47/50, train_loss: 0.0064\n",
            "48/50, train_loss: 0.0071\n",
            "49/50, train_loss: 0.0649\n",
            "50/50, train_loss: 0.0056\n",
            "epoch 97 average loss: 0.0412\n",
            "----------\n",
            "epoch 98/400\n",
            "1/50, train_loss: 0.1163\n",
            "2/50, train_loss: 0.0646\n",
            "3/50, train_loss: 0.0574\n",
            "4/50, train_loss: 0.0062\n",
            "5/50, train_loss: 0.0051\n",
            "6/50, train_loss: 0.0906\n",
            "7/50, train_loss: 0.1918\n",
            "8/50, train_loss: 0.1602\n",
            "9/50, train_loss: 0.0056\n",
            "10/50, train_loss: 0.0054\n",
            "11/50, train_loss: 0.0567\n",
            "12/50, train_loss: 0.0058\n",
            "13/50, train_loss: 0.0047\n",
            "14/50, train_loss: 0.0525\n",
            "15/50, train_loss: 0.0047\n",
            "16/50, train_loss: 0.0054\n",
            "17/50, train_loss: 0.0072\n",
            "18/50, train_loss: 0.0044\n",
            "19/50, train_loss: 0.0063\n",
            "20/50, train_loss: 0.0054\n",
            "21/50, train_loss: 0.0058\n",
            "22/50, train_loss: 0.0048\n",
            "23/50, train_loss: 0.0938\n",
            "24/50, train_loss: 0.0718\n",
            "25/50, train_loss: 0.0482\n",
            "26/50, train_loss: 0.0049\n",
            "27/50, train_loss: 0.0041\n",
            "28/50, train_loss: 0.0529\n",
            "29/50, train_loss: 0.0808\n",
            "30/50, train_loss: 0.0045\n",
            "31/50, train_loss: 0.0683\n",
            "32/50, train_loss: 0.1005\n",
            "33/50, train_loss: 0.0051\n",
            "34/50, train_loss: 0.0078\n",
            "35/50, train_loss: 0.0558\n",
            "36/50, train_loss: 0.0047\n",
            "37/50, train_loss: 0.0054\n",
            "38/50, train_loss: 0.0714\n",
            "39/50, train_loss: 0.0055\n",
            "40/50, train_loss: 0.0480\n",
            "41/50, train_loss: 0.1089\n",
            "42/50, train_loss: 0.0043\n",
            "43/50, train_loss: 0.0054\n",
            "44/50, train_loss: 0.0035\n",
            "45/50, train_loss: 0.0045\n",
            "46/50, train_loss: 0.0565\n",
            "47/50, train_loss: 0.0041\n",
            "48/50, train_loss: 0.1051\n",
            "49/50, train_loss: 0.0040\n",
            "50/50, train_loss: 0.0053\n",
            "epoch 98 average loss: 0.0380\n",
            "current epoch: 98 current accuracy: 0.7083 current AUC: 0.6639 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 99/400\n",
            "1/50, train_loss: 0.0051\n",
            "2/50, train_loss: 0.2416\n",
            "3/50, train_loss: 0.0988\n",
            "4/50, train_loss: 0.0052\n",
            "5/50, train_loss: 0.0514\n",
            "6/50, train_loss: 0.2308\n",
            "7/50, train_loss: 0.0048\n",
            "8/50, train_loss: 0.0724\n",
            "9/50, train_loss: 0.0706\n",
            "10/50, train_loss: 0.1009\n",
            "11/50, train_loss: 0.0987\n",
            "12/50, train_loss: 0.0944\n",
            "13/50, train_loss: 0.0647\n",
            "14/50, train_loss: 0.0602\n",
            "15/50, train_loss: 0.0934\n",
            "16/50, train_loss: 0.0052\n",
            "17/50, train_loss: 0.0939\n",
            "18/50, train_loss: 0.1028\n",
            "19/50, train_loss: 0.0532\n",
            "20/50, train_loss: 0.0049\n",
            "21/50, train_loss: 0.0664\n",
            "22/50, train_loss: 0.0070\n",
            "23/50, train_loss: 0.0634\n",
            "24/50, train_loss: 0.0903\n",
            "25/50, train_loss: 0.0057\n",
            "26/50, train_loss: 0.1015\n",
            "27/50, train_loss: 0.1064\n",
            "28/50, train_loss: 0.1029\n",
            "29/50, train_loss: 0.0060\n",
            "30/50, train_loss: 0.0056\n",
            "31/50, train_loss: 0.0066\n",
            "32/50, train_loss: 0.0116\n",
            "33/50, train_loss: 0.0053\n",
            "34/50, train_loss: 0.0062\n",
            "35/50, train_loss: 0.0053\n",
            "36/50, train_loss: 0.0046\n",
            "37/50, train_loss: 0.0772\n",
            "38/50, train_loss: 0.0587\n",
            "39/50, train_loss: 0.0647\n",
            "40/50, train_loss: 0.0662\n",
            "41/50, train_loss: 0.0922\n",
            "42/50, train_loss: 0.0059\n",
            "43/50, train_loss: 0.0687\n",
            "44/50, train_loss: 0.1517\n",
            "45/50, train_loss: 0.0061\n",
            "46/50, train_loss: 0.0652\n",
            "47/50, train_loss: 0.0811\n",
            "48/50, train_loss: 0.0540\n",
            "49/50, train_loss: 0.0050\n",
            "50/50, train_loss: 0.0062\n",
            "epoch 99 average loss: 0.0590\n",
            "----------\n",
            "epoch 100/400\n",
            "1/50, train_loss: 0.0053\n",
            "2/50, train_loss: 0.0466\n",
            "3/50, train_loss: 0.0057\n",
            "4/50, train_loss: 0.0052\n",
            "5/50, train_loss: 0.0067\n",
            "6/50, train_loss: 0.0756\n",
            "7/50, train_loss: 0.1447\n",
            "8/50, train_loss: 0.0742\n",
            "9/50, train_loss: 0.0057\n",
            "10/50, train_loss: 0.0538\n",
            "11/50, train_loss: 0.0570\n",
            "12/50, train_loss: 0.0582\n",
            "13/50, train_loss: 0.0050\n",
            "14/50, train_loss: 0.0571\n",
            "15/50, train_loss: 0.0058\n",
            "16/50, train_loss: 0.0056\n",
            "17/50, train_loss: 0.0691\n",
            "18/50, train_loss: 0.0049\n",
            "19/50, train_loss: 0.0045\n",
            "20/50, train_loss: 0.0055\n",
            "21/50, train_loss: 0.0052\n",
            "22/50, train_loss: 0.0052\n",
            "23/50, train_loss: 0.0610\n",
            "24/50, train_loss: 0.0660\n",
            "25/50, train_loss: 0.0048\n",
            "26/50, train_loss: 0.0851\n",
            "27/50, train_loss: 0.0048\n",
            "28/50, train_loss: 0.0547\n",
            "29/50, train_loss: 0.0055\n",
            "30/50, train_loss: 0.0055\n",
            "31/50, train_loss: 0.0537\n",
            "32/50, train_loss: 0.1376\n",
            "33/50, train_loss: 0.0057\n",
            "34/50, train_loss: 0.0888\n",
            "35/50, train_loss: 0.0624\n",
            "36/50, train_loss: 0.0048\n",
            "37/50, train_loss: 0.1018\n",
            "38/50, train_loss: 0.1005\n",
            "39/50, train_loss: 0.0055\n",
            "40/50, train_loss: 0.0960\n",
            "41/50, train_loss: 0.0055\n",
            "42/50, train_loss: 0.0598\n",
            "43/50, train_loss: 0.0044\n",
            "44/50, train_loss: 0.0048\n",
            "45/50, train_loss: 0.0038\n",
            "46/50, train_loss: 0.0045\n",
            "47/50, train_loss: 0.0037\n",
            "48/50, train_loss: 0.0041\n",
            "49/50, train_loss: 0.0051\n",
            "50/50, train_loss: 0.0077\n",
            "epoch 100 average loss: 0.0351\n",
            "current epoch: 100 current accuracy: 0.7500 current AUC: 0.7311 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 101/400\n",
            "1/50, train_loss: 0.0046\n",
            "2/50, train_loss: 0.0049\n",
            "3/50, train_loss: 0.0615\n",
            "4/50, train_loss: 0.0043\n",
            "5/50, train_loss: 0.0691\n",
            "6/50, train_loss: 0.0522\n",
            "7/50, train_loss: 0.0046\n",
            "8/50, train_loss: 0.0042\n",
            "9/50, train_loss: 0.0040\n",
            "10/50, train_loss: 0.0484\n",
            "11/50, train_loss: 0.0713\n",
            "12/50, train_loss: 0.0982\n",
            "13/50, train_loss: 0.0491\n",
            "14/50, train_loss: 0.0046\n",
            "15/50, train_loss: 0.0050\n",
            "16/50, train_loss: 0.1159\n",
            "17/50, train_loss: 0.0648\n",
            "18/50, train_loss: 0.1428\n",
            "19/50, train_loss: 0.0046\n",
            "20/50, train_loss: 0.0054\n",
            "21/50, train_loss: 0.1016\n",
            "22/50, train_loss: 0.0047\n",
            "23/50, train_loss: 0.0050\n",
            "24/50, train_loss: 0.1194\n",
            "25/50, train_loss: 0.0050\n",
            "26/50, train_loss: 0.1324\n",
            "27/50, train_loss: 0.0044\n",
            "28/50, train_loss: 0.0552\n",
            "29/50, train_loss: 0.0994\n",
            "30/50, train_loss: 0.0032\n",
            "31/50, train_loss: 0.0052\n",
            "32/50, train_loss: 0.0043\n",
            "33/50, train_loss: 0.0614\n",
            "34/50, train_loss: 0.0046\n",
            "35/50, train_loss: 0.0533\n",
            "36/50, train_loss: 0.0042\n",
            "37/50, train_loss: 0.0972\n",
            "38/50, train_loss: 0.0529\n",
            "39/50, train_loss: 0.0053\n",
            "40/50, train_loss: 0.0039\n",
            "41/50, train_loss: 0.0044\n",
            "42/50, train_loss: 0.0044\n",
            "43/50, train_loss: 0.0552\n",
            "44/50, train_loss: 0.0045\n",
            "45/50, train_loss: 0.0539\n",
            "46/50, train_loss: 0.1144\n",
            "47/50, train_loss: 0.0613\n",
            "48/50, train_loss: 0.0062\n",
            "49/50, train_loss: 0.1121\n",
            "50/50, train_loss: 0.0582\n",
            "epoch 101 average loss: 0.0423\n",
            "----------\n",
            "epoch 102/400\n",
            "1/50, train_loss: 0.0559\n",
            "2/50, train_loss: 0.0053\n",
            "3/50, train_loss: 0.0069\n",
            "4/50, train_loss: 0.0529\n",
            "5/50, train_loss: 0.1210\n",
            "6/50, train_loss: 0.0501\n",
            "7/50, train_loss: 0.0037\n",
            "8/50, train_loss: 0.0856\n",
            "9/50, train_loss: 0.0512\n",
            "10/50, train_loss: 0.0059\n",
            "11/50, train_loss: 0.0040\n",
            "12/50, train_loss: 0.0052\n",
            "13/50, train_loss: 0.0546\n",
            "14/50, train_loss: 0.0040\n",
            "15/50, train_loss: 0.0824\n",
            "16/50, train_loss: 0.0045\n",
            "17/50, train_loss: 0.0047\n",
            "18/50, train_loss: 0.0045\n",
            "19/50, train_loss: 0.0047\n",
            "20/50, train_loss: 0.1379\n",
            "21/50, train_loss: 0.0582\n",
            "22/50, train_loss: 0.0046\n",
            "23/50, train_loss: 0.0479\n",
            "24/50, train_loss: 0.0630\n",
            "25/50, train_loss: 0.0838\n",
            "26/50, train_loss: 0.0458\n",
            "27/50, train_loss: 0.0518\n",
            "28/50, train_loss: 0.0046\n",
            "29/50, train_loss: 0.0055\n",
            "30/50, train_loss: 0.0040\n",
            "31/50, train_loss: 0.0034\n",
            "32/50, train_loss: 0.0583\n",
            "33/50, train_loss: 0.0533\n",
            "34/50, train_loss: 0.0495\n",
            "35/50, train_loss: 0.0049\n",
            "36/50, train_loss: 0.0042\n",
            "37/50, train_loss: 0.0046\n",
            "38/50, train_loss: 0.1088\n",
            "39/50, train_loss: 0.0550\n",
            "40/50, train_loss: 0.0052\n",
            "41/50, train_loss: 0.0042\n",
            "42/50, train_loss: 0.0993\n",
            "43/50, train_loss: 0.0979\n",
            "44/50, train_loss: 0.0038\n",
            "45/50, train_loss: 0.0570\n",
            "46/50, train_loss: 0.0876\n",
            "47/50, train_loss: 0.1660\n",
            "48/50, train_loss: 0.0510\n",
            "49/50, train_loss: 0.1779\n",
            "50/50, train_loss: 0.0044\n",
            "epoch 102 average loss: 0.0442\n",
            "current epoch: 102 current accuracy: 0.7083 current AUC: 0.6807 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 103/400\n",
            "1/50, train_loss: 0.0039\n",
            "2/50, train_loss: 0.0033\n",
            "3/50, train_loss: 0.0606\n",
            "4/50, train_loss: 0.0048\n",
            "5/50, train_loss: 0.0048\n",
            "6/50, train_loss: 0.0058\n",
            "7/50, train_loss: 0.0538\n",
            "8/50, train_loss: 0.0043\n",
            "9/50, train_loss: 0.0046\n",
            "10/50, train_loss: 0.0043\n",
            "11/50, train_loss: 0.0830\n",
            "12/50, train_loss: 0.0051\n",
            "13/50, train_loss: 0.0052\n",
            "14/50, train_loss: 0.0044\n",
            "15/50, train_loss: 0.0656\n",
            "16/50, train_loss: 0.0056\n",
            "17/50, train_loss: 0.0056\n",
            "18/50, train_loss: 0.0047\n",
            "19/50, train_loss: 0.0974\n",
            "20/50, train_loss: 0.0047\n",
            "21/50, train_loss: 0.0054\n",
            "22/50, train_loss: 0.0040\n",
            "23/50, train_loss: 0.0044\n",
            "24/50, train_loss: 0.0050\n",
            "25/50, train_loss: 0.1498\n",
            "26/50, train_loss: 0.0046\n",
            "27/50, train_loss: 0.0490\n",
            "28/50, train_loss: 0.0962\n",
            "29/50, train_loss: 0.0044\n",
            "30/50, train_loss: 0.0063\n",
            "31/50, train_loss: 0.0044\n",
            "32/50, train_loss: 0.0594\n",
            "33/50, train_loss: 0.0593\n",
            "34/50, train_loss: 0.0046\n",
            "35/50, train_loss: 0.0594\n",
            "36/50, train_loss: 0.0458\n",
            "37/50, train_loss: 0.1228\n",
            "38/50, train_loss: 0.0449\n",
            "39/50, train_loss: 0.0052\n",
            "40/50, train_loss: 0.0048\n",
            "41/50, train_loss: 0.0595\n",
            "42/50, train_loss: 0.0049\n",
            "43/50, train_loss: 0.0043\n",
            "44/50, train_loss: 0.0068\n",
            "45/50, train_loss: 0.0993\n",
            "46/50, train_loss: 0.0059\n",
            "47/50, train_loss: 0.0452\n",
            "48/50, train_loss: 0.0045\n",
            "49/50, train_loss: 0.0037\n",
            "50/50, train_loss: 0.0032\n",
            "epoch 103 average loss: 0.0282\n",
            "----------\n",
            "epoch 104/400\n",
            "1/50, train_loss: 0.1383\n",
            "2/50, train_loss: 0.1153\n",
            "3/50, train_loss: 0.0041\n",
            "4/50, train_loss: 0.0974\n",
            "5/50, train_loss: 0.0762\n",
            "6/50, train_loss: 0.0938\n",
            "7/50, train_loss: 0.0773\n",
            "8/50, train_loss: 0.0542\n",
            "9/50, train_loss: 0.0664\n",
            "10/50, train_loss: 0.0598\n",
            "11/50, train_loss: 0.0036\n",
            "12/50, train_loss: 0.0577\n",
            "13/50, train_loss: 0.0043\n",
            "14/50, train_loss: 0.0051\n",
            "15/50, train_loss: 0.0038\n",
            "16/50, train_loss: 0.1121\n",
            "17/50, train_loss: 0.0050\n",
            "18/50, train_loss: 0.0664\n",
            "19/50, train_loss: 0.0039\n",
            "20/50, train_loss: 0.0572\n",
            "21/50, train_loss: 0.0046\n",
            "22/50, train_loss: 0.0050\n",
            "23/50, train_loss: 0.0445\n",
            "24/50, train_loss: 0.0773\n",
            "25/50, train_loss: 0.0627\n",
            "26/50, train_loss: 0.0040\n",
            "27/50, train_loss: 0.0045\n",
            "28/50, train_loss: 0.0999\n",
            "29/50, train_loss: 0.0700\n",
            "30/50, train_loss: 0.0488\n",
            "31/50, train_loss: 0.0749\n",
            "32/50, train_loss: 0.0040\n",
            "33/50, train_loss: 0.0047\n",
            "34/50, train_loss: 0.0569\n",
            "35/50, train_loss: 0.0967\n",
            "36/50, train_loss: 0.0424\n",
            "37/50, train_loss: 0.0672\n",
            "38/50, train_loss: 0.0964\n",
            "39/50, train_loss: 0.0051\n",
            "40/50, train_loss: 0.0043\n",
            "41/50, train_loss: 0.0049\n",
            "42/50, train_loss: 0.0033\n",
            "43/50, train_loss: 0.0488\n",
            "44/50, train_loss: 0.1069\n",
            "45/50, train_loss: 0.0691\n",
            "46/50, train_loss: 0.0040\n",
            "47/50, train_loss: 0.0658\n",
            "48/50, train_loss: 0.0871\n",
            "49/50, train_loss: 0.0040\n",
            "50/50, train_loss: 0.0585\n",
            "epoch 104 average loss: 0.0486\n",
            "current epoch: 104 current accuracy: 0.7500 current AUC: 0.7059 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 105/400\n",
            "1/50, train_loss: 0.0463\n",
            "2/50, train_loss: 0.0779\n",
            "3/50, train_loss: 0.0047\n",
            "4/50, train_loss: 0.0713\n",
            "5/50, train_loss: 0.0716\n",
            "6/50, train_loss: 0.0463\n",
            "7/50, train_loss: 0.0586\n",
            "8/50, train_loss: 0.0041\n",
            "9/50, train_loss: 0.0543\n",
            "10/50, train_loss: 0.0641\n",
            "11/50, train_loss: 0.0617\n",
            "12/50, train_loss: 0.0461\n",
            "13/50, train_loss: 0.1080\n",
            "14/50, train_loss: 0.0040\n",
            "15/50, train_loss: 0.0037\n",
            "16/50, train_loss: 0.0040\n",
            "17/50, train_loss: 0.0048\n",
            "18/50, train_loss: 0.0052\n",
            "19/50, train_loss: 0.0499\n",
            "20/50, train_loss: 0.0454\n",
            "21/50, train_loss: 0.0045\n",
            "22/50, train_loss: 0.0040\n",
            "23/50, train_loss: 0.0031\n",
            "24/50, train_loss: 0.0039\n",
            "25/50, train_loss: 0.0038\n",
            "26/50, train_loss: 0.0943\n",
            "27/50, train_loss: 0.0047\n",
            "28/50, train_loss: 0.0035\n",
            "29/50, train_loss: 0.0563\n",
            "30/50, train_loss: 0.0041\n",
            "31/50, train_loss: 0.0046\n",
            "32/50, train_loss: 0.0032\n",
            "33/50, train_loss: 0.0886\n",
            "34/50, train_loss: 0.0041\n",
            "35/50, train_loss: 0.1011\n",
            "36/50, train_loss: 0.0491\n",
            "37/50, train_loss: 0.0990\n",
            "38/50, train_loss: 0.0042\n",
            "39/50, train_loss: 0.0041\n",
            "40/50, train_loss: 0.0488\n",
            "41/50, train_loss: 0.0047\n",
            "42/50, train_loss: 0.0797\n",
            "43/50, train_loss: 0.0038\n",
            "44/50, train_loss: 0.0042\n",
            "45/50, train_loss: 0.0429\n",
            "46/50, train_loss: 0.0039\n",
            "47/50, train_loss: 0.0813\n",
            "48/50, train_loss: 0.0050\n",
            "49/50, train_loss: 0.0586\n",
            "50/50, train_loss: 0.1034\n",
            "epoch 105 average loss: 0.0362\n",
            "----------\n",
            "epoch 106/400\n",
            "1/50, train_loss: 0.0034\n",
            "2/50, train_loss: 0.1146\n",
            "3/50, train_loss: 0.0837\n",
            "4/50, train_loss: 0.0625\n",
            "5/50, train_loss: 0.0515\n",
            "6/50, train_loss: 0.0425\n",
            "7/50, train_loss: 0.0723\n",
            "8/50, train_loss: 0.0037\n",
            "9/50, train_loss: 0.0033\n",
            "10/50, train_loss: 0.0042\n",
            "11/50, train_loss: 0.0928\n",
            "12/50, train_loss: 0.0052\n",
            "13/50, train_loss: 0.0033\n",
            "14/50, train_loss: 0.0531\n",
            "15/50, train_loss: 0.0471\n",
            "16/50, train_loss: 0.0040\n",
            "17/50, train_loss: 0.0046\n",
            "18/50, train_loss: 0.0755\n",
            "19/50, train_loss: 0.0041\n",
            "20/50, train_loss: 0.0035\n",
            "21/50, train_loss: 0.0038\n",
            "22/50, train_loss: 0.0039\n",
            "23/50, train_loss: 0.0050\n",
            "24/50, train_loss: 0.0713\n",
            "25/50, train_loss: 0.0914\n",
            "26/50, train_loss: 0.1082\n",
            "27/50, train_loss: 0.0738\n",
            "28/50, train_loss: 0.0567\n",
            "29/50, train_loss: 0.0037\n",
            "30/50, train_loss: 0.0042\n",
            "31/50, train_loss: 0.0039\n",
            "32/50, train_loss: 0.0042\n",
            "33/50, train_loss: 0.0681\n",
            "34/50, train_loss: 0.0033\n",
            "35/50, train_loss: 0.0462\n",
            "36/50, train_loss: 0.0606\n",
            "37/50, train_loss: 0.0446\n",
            "38/50, train_loss: 0.0509\n",
            "39/50, train_loss: 0.0045\n",
            "40/50, train_loss: 0.0041\n",
            "41/50, train_loss: 0.0534\n",
            "42/50, train_loss: 0.1293\n",
            "43/50, train_loss: 0.0027\n",
            "44/50, train_loss: 0.0734\n",
            "45/50, train_loss: 0.0433\n",
            "46/50, train_loss: 0.0405\n",
            "47/50, train_loss: 0.0411\n",
            "48/50, train_loss: 0.0031\n",
            "49/50, train_loss: 0.0725\n",
            "50/50, train_loss: 0.0044\n",
            "epoch 106 average loss: 0.0382\n",
            "current epoch: 106 current accuracy: 0.7500 current AUC: 0.6723 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 107/400\n",
            "1/50, train_loss: 0.0025\n",
            "2/50, train_loss: 0.0501\n",
            "3/50, train_loss: 0.0922\n",
            "4/50, train_loss: 0.0791\n",
            "5/50, train_loss: 0.0033\n",
            "6/50, train_loss: 0.0495\n",
            "7/50, train_loss: 0.0033\n",
            "8/50, train_loss: 0.0417\n",
            "9/50, train_loss: 0.0049\n",
            "10/50, train_loss: 0.1081\n",
            "11/50, train_loss: 0.0443\n",
            "12/50, train_loss: 0.0616\n",
            "13/50, train_loss: 0.0039\n",
            "14/50, train_loss: 0.0488\n",
            "15/50, train_loss: 0.0502\n",
            "16/50, train_loss: 0.0036\n",
            "17/50, train_loss: 0.0723\n",
            "18/50, train_loss: 0.0448\n",
            "19/50, train_loss: 0.0738\n",
            "20/50, train_loss: 0.0567\n",
            "21/50, train_loss: 0.0039\n",
            "22/50, train_loss: 0.0802\n",
            "23/50, train_loss: 0.0396\n",
            "24/50, train_loss: 0.0035\n",
            "25/50, train_loss: 0.0639\n",
            "26/50, train_loss: 0.0425\n",
            "27/50, train_loss: 0.0964\n",
            "28/50, train_loss: 0.0027\n",
            "29/50, train_loss: 0.0047\n",
            "30/50, train_loss: 0.0035\n",
            "31/50, train_loss: 0.0033\n",
            "32/50, train_loss: 0.0449\n",
            "33/50, train_loss: 0.0870\n",
            "34/50, train_loss: 0.1127\n",
            "35/50, train_loss: 0.0038\n",
            "36/50, train_loss: 0.0459\n",
            "37/50, train_loss: 0.0421\n",
            "38/50, train_loss: 0.0511\n",
            "39/50, train_loss: 0.0913\n",
            "40/50, train_loss: 0.0447\n",
            "41/50, train_loss: 0.0864\n",
            "42/50, train_loss: 0.0799\n",
            "43/50, train_loss: 0.0046\n",
            "44/50, train_loss: 0.0034\n",
            "45/50, train_loss: 0.0458\n",
            "46/50, train_loss: 0.0034\n",
            "47/50, train_loss: 0.0040\n",
            "48/50, train_loss: 0.0036\n",
            "49/50, train_loss: 0.0036\n",
            "50/50, train_loss: 0.0831\n",
            "epoch 107 average loss: 0.0416\n",
            "----------\n",
            "epoch 108/400\n",
            "1/50, train_loss: 0.0047\n",
            "2/50, train_loss: 0.0036\n",
            "3/50, train_loss: 0.0761\n",
            "4/50, train_loss: 0.0630\n",
            "5/50, train_loss: 0.0034\n",
            "6/50, train_loss: 0.0032\n",
            "7/50, train_loss: 0.0037\n",
            "8/50, train_loss: 0.1082\n",
            "9/50, train_loss: 0.0042\n",
            "10/50, train_loss: 0.0037\n",
            "11/50, train_loss: 0.0779\n",
            "12/50, train_loss: 0.0522\n",
            "13/50, train_loss: 0.0669\n",
            "14/50, train_loss: 0.0030\n",
            "15/50, train_loss: 0.0438\n",
            "16/50, train_loss: 0.1069\n",
            "17/50, train_loss: 0.0041\n",
            "18/50, train_loss: 0.0034\n",
            "19/50, train_loss: 0.0496\n",
            "20/50, train_loss: 0.0037\n",
            "21/50, train_loss: 0.0449\n",
            "22/50, train_loss: 0.0454\n",
            "23/50, train_loss: 0.0524\n",
            "24/50, train_loss: 0.0037\n",
            "25/50, train_loss: 0.0037\n",
            "26/50, train_loss: 0.0764\n",
            "27/50, train_loss: 0.0520\n",
            "28/50, train_loss: 0.0043\n",
            "29/50, train_loss: 0.0037\n",
            "30/50, train_loss: 0.0031\n",
            "31/50, train_loss: 0.0038\n",
            "32/50, train_loss: 0.0027\n",
            "33/50, train_loss: 0.0869\n",
            "34/50, train_loss: 0.0042\n",
            "35/50, train_loss: 0.0439\n",
            "36/50, train_loss: 0.0475\n",
            "37/50, train_loss: 0.1373\n",
            "38/50, train_loss: 0.0038\n",
            "39/50, train_loss: 0.0434\n",
            "40/50, train_loss: 0.0734\n",
            "41/50, train_loss: 0.0035\n",
            "42/50, train_loss: 0.0041\n",
            "43/50, train_loss: 0.0035\n",
            "44/50, train_loss: 0.0030\n",
            "45/50, train_loss: 0.0032\n",
            "46/50, train_loss: 0.0030\n",
            "47/50, train_loss: 0.0033\n",
            "48/50, train_loss: 0.0036\n",
            "49/50, train_loss: 0.0037\n",
            "50/50, train_loss: 0.0411\n",
            "epoch 108 average loss: 0.0299\n",
            "current epoch: 108 current accuracy: 0.7500 current AUC: 0.6471 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 109/400\n",
            "1/50, train_loss: 0.0551\n",
            "2/50, train_loss: 0.0037\n",
            "3/50, train_loss: 0.0031\n",
            "4/50, train_loss: 0.0033\n",
            "5/50, train_loss: 0.0426\n",
            "6/50, train_loss: 0.0385\n",
            "7/50, train_loss: 0.0485\n",
            "8/50, train_loss: 0.0627\n",
            "9/50, train_loss: 0.0037\n",
            "10/50, train_loss: 0.0032\n",
            "11/50, train_loss: 0.1350\n",
            "12/50, train_loss: 0.0467\n",
            "13/50, train_loss: 0.0027\n",
            "14/50, train_loss: 0.0466\n",
            "15/50, train_loss: 0.0034\n",
            "16/50, train_loss: 0.0041\n",
            "17/50, train_loss: 0.1118\n",
            "18/50, train_loss: 0.0674\n",
            "19/50, train_loss: 0.0032\n",
            "20/50, train_loss: 0.0946\n",
            "21/50, train_loss: 0.0037\n",
            "22/50, train_loss: 0.0756\n",
            "23/50, train_loss: 0.0511\n",
            "24/50, train_loss: 0.0375\n",
            "25/50, train_loss: 0.0034\n",
            "26/50, train_loss: 0.0034\n",
            "27/50, train_loss: 0.0476\n",
            "28/50, train_loss: 0.0036\n",
            "29/50, train_loss: 0.0510\n",
            "30/50, train_loss: 0.0603\n",
            "31/50, train_loss: 0.0370\n",
            "32/50, train_loss: 0.0028\n",
            "33/50, train_loss: 0.0490\n",
            "34/50, train_loss: 0.0618\n",
            "35/50, train_loss: 0.0767\n",
            "36/50, train_loss: 0.0038\n",
            "37/50, train_loss: 0.1226\n",
            "38/50, train_loss: 0.0526\n",
            "39/50, train_loss: 0.0765\n",
            "40/50, train_loss: 0.0842\n",
            "41/50, train_loss: 0.0783\n",
            "42/50, train_loss: 0.0811\n",
            "43/50, train_loss: 0.0393\n",
            "44/50, train_loss: 0.0036\n",
            "45/50, train_loss: 0.0033\n",
            "46/50, train_loss: 0.1059\n",
            "47/50, train_loss: 0.0883\n",
            "48/50, train_loss: 0.0710\n",
            "49/50, train_loss: 0.0027\n",
            "50/50, train_loss: 0.0031\n",
            "epoch 109 average loss: 0.0432\n",
            "----------\n",
            "epoch 110/400\n",
            "1/50, train_loss: 0.0034\n",
            "2/50, train_loss: 0.0829\n",
            "3/50, train_loss: 0.0030\n",
            "4/50, train_loss: 0.0981\n",
            "5/50, train_loss: 0.0782\n",
            "6/50, train_loss: 0.0029\n",
            "7/50, train_loss: 0.0045\n",
            "8/50, train_loss: 0.0027\n",
            "9/50, train_loss: 0.0586\n",
            "10/50, train_loss: 0.0032\n",
            "11/50, train_loss: 0.0032\n",
            "12/50, train_loss: 0.0513\n",
            "13/50, train_loss: 0.0031\n",
            "14/50, train_loss: 0.0045\n",
            "15/50, train_loss: 0.0033\n",
            "16/50, train_loss: 0.0025\n",
            "17/50, train_loss: 0.0038\n",
            "18/50, train_loss: 0.0639\n",
            "19/50, train_loss: 0.0494\n",
            "20/50, train_loss: 0.0033\n",
            "21/50, train_loss: 0.0044\n",
            "22/50, train_loss: 0.0033\n",
            "23/50, train_loss: 0.0031\n",
            "24/50, train_loss: 0.0031\n",
            "25/50, train_loss: 0.0544\n",
            "26/50, train_loss: 0.0398\n",
            "27/50, train_loss: 0.0031\n",
            "28/50, train_loss: 0.0410\n",
            "29/50, train_loss: 0.0651\n",
            "30/50, train_loss: 0.0025\n",
            "31/50, train_loss: 0.0748\n",
            "32/50, train_loss: 0.0535\n",
            "33/50, train_loss: 0.0028\n",
            "34/50, train_loss: 0.0040\n",
            "35/50, train_loss: 0.0036\n",
            "36/50, train_loss: 0.0032\n",
            "37/50, train_loss: 0.0032\n",
            "38/50, train_loss: 0.0525\n",
            "39/50, train_loss: 0.0033\n",
            "40/50, train_loss: 0.0503\n",
            "41/50, train_loss: 0.0893\n",
            "42/50, train_loss: 0.0031\n",
            "43/50, train_loss: 0.0580\n",
            "44/50, train_loss: 0.0034\n",
            "45/50, train_loss: 0.0550\n",
            "46/50, train_loss: 0.0027\n",
            "47/50, train_loss: 0.0406\n",
            "48/50, train_loss: 0.0450\n",
            "49/50, train_loss: 0.0025\n",
            "50/50, train_loss: 0.0925\n",
            "epoch 110 average loss: 0.0278\n",
            "current epoch: 110 current accuracy: 0.7500 current AUC: 0.6723 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 111/400\n",
            "1/50, train_loss: 0.0454\n",
            "2/50, train_loss: 0.0031\n",
            "3/50, train_loss: 0.0035\n",
            "4/50, train_loss: 0.0039\n",
            "5/50, train_loss: 0.0695\n",
            "6/50, train_loss: 0.0720\n",
            "7/50, train_loss: 0.0527\n",
            "8/50, train_loss: 0.0040\n",
            "9/50, train_loss: 0.0024\n",
            "10/50, train_loss: 0.0032\n",
            "11/50, train_loss: 0.0566\n",
            "12/50, train_loss: 0.0028\n",
            "13/50, train_loss: 0.0032\n",
            "14/50, train_loss: 0.0038\n",
            "15/50, train_loss: 0.0481\n",
            "16/50, train_loss: 0.0031\n",
            "17/50, train_loss: 0.0420\n",
            "18/50, train_loss: 0.0027\n",
            "19/50, train_loss: 0.0511\n",
            "20/50, train_loss: 0.0550\n",
            "21/50, train_loss: 0.0028\n",
            "22/50, train_loss: 0.0029\n",
            "23/50, train_loss: 0.0032\n",
            "24/50, train_loss: 0.0029\n",
            "25/50, train_loss: 0.0492\n",
            "26/50, train_loss: 0.1004\n",
            "27/50, train_loss: 0.0029\n",
            "28/50, train_loss: 0.0028\n",
            "29/50, train_loss: 0.0712\n",
            "30/50, train_loss: 0.0033\n",
            "31/50, train_loss: 0.0433\n",
            "32/50, train_loss: 0.0709\n",
            "33/50, train_loss: 0.0033\n",
            "34/50, train_loss: 0.0386\n",
            "35/50, train_loss: 0.0924\n",
            "36/50, train_loss: 0.0034\n",
            "37/50, train_loss: 0.0449\n",
            "38/50, train_loss: 0.0029\n",
            "39/50, train_loss: 0.0032\n",
            "40/50, train_loss: 0.0033\n",
            "41/50, train_loss: 0.0802\n",
            "42/50, train_loss: 0.0030\n",
            "43/50, train_loss: 0.0649\n",
            "44/50, train_loss: 0.0775\n",
            "45/50, train_loss: 0.0380\n",
            "46/50, train_loss: 0.0031\n",
            "47/50, train_loss: 0.0023\n",
            "48/50, train_loss: 0.0030\n",
            "49/50, train_loss: 0.0449\n",
            "50/50, train_loss: 0.0585\n",
            "epoch 111 average loss: 0.0290\n",
            "----------\n",
            "epoch 112/400\n",
            "1/50, train_loss: 0.0464\n",
            "2/50, train_loss: 0.0028\n",
            "3/50, train_loss: 0.1008\n",
            "4/50, train_loss: 0.0031\n",
            "5/50, train_loss: 0.0030\n",
            "6/50, train_loss: 0.0024\n",
            "7/50, train_loss: 0.0384\n",
            "8/50, train_loss: 0.0026\n",
            "9/50, train_loss: 0.0024\n",
            "10/50, train_loss: 0.0428\n",
            "11/50, train_loss: 0.0587\n",
            "12/50, train_loss: 0.0498\n",
            "13/50, train_loss: 0.0684\n",
            "14/50, train_loss: 0.0024\n",
            "15/50, train_loss: 0.0361\n",
            "16/50, train_loss: 0.0689\n",
            "17/50, train_loss: 0.0036\n",
            "18/50, train_loss: 0.0736\n",
            "19/50, train_loss: 0.0038\n",
            "20/50, train_loss: 0.1353\n",
            "21/50, train_loss: 0.0411\n",
            "22/50, train_loss: 0.0590\n",
            "23/50, train_loss: 0.0027\n",
            "24/50, train_loss: 0.0035\n",
            "25/50, train_loss: 0.0415\n",
            "26/50, train_loss: 0.0837\n",
            "27/50, train_loss: 0.0033\n",
            "28/50, train_loss: 0.1015\n",
            "29/50, train_loss: 0.0399\n",
            "30/50, train_loss: 0.0444\n",
            "31/50, train_loss: 0.0028\n",
            "32/50, train_loss: 0.0589\n",
            "33/50, train_loss: 0.0508\n",
            "34/50, train_loss: 0.0033\n",
            "35/50, train_loss: 0.0028\n",
            "36/50, train_loss: 0.0775\n",
            "37/50, train_loss: 0.0025\n",
            "38/50, train_loss: 0.0026\n",
            "39/50, train_loss: 0.0032\n",
            "40/50, train_loss: 0.0028\n",
            "41/50, train_loss: 0.0729\n",
            "42/50, train_loss: 0.0717\n",
            "43/50, train_loss: 0.0880\n",
            "44/50, train_loss: 0.0392\n",
            "45/50, train_loss: 0.0544\n",
            "46/50, train_loss: 0.0024\n",
            "47/50, train_loss: 0.0032\n",
            "48/50, train_loss: 0.0029\n",
            "49/50, train_loss: 0.0426\n",
            "50/50, train_loss: 0.0032\n",
            "epoch 112 average loss: 0.0351\n",
            "current epoch: 112 current accuracy: 0.7500 current AUC: 0.6639 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 113/400\n",
            "1/50, train_loss: 0.0892\n",
            "2/50, train_loss: 0.0846\n",
            "3/50, train_loss: 0.0441\n",
            "4/50, train_loss: 0.0693\n",
            "5/50, train_loss: 0.0371\n",
            "6/50, train_loss: 0.0758\n",
            "7/50, train_loss: 0.0460\n",
            "8/50, train_loss: 0.0579\n",
            "9/50, train_loss: 0.0510\n",
            "10/50, train_loss: 0.0031\n",
            "11/50, train_loss: 0.0035\n",
            "12/50, train_loss: 0.0391\n",
            "13/50, train_loss: 0.0024\n",
            "14/50, train_loss: 0.0742\n",
            "15/50, train_loss: 0.0033\n",
            "16/50, train_loss: 0.0438\n",
            "17/50, train_loss: 0.0031\n",
            "18/50, train_loss: 0.0380\n",
            "19/50, train_loss: 0.0791\n",
            "20/50, train_loss: 0.0034\n",
            "21/50, train_loss: 0.0438\n",
            "22/50, train_loss: 0.0571\n",
            "23/50, train_loss: 0.0031\n",
            "24/50, train_loss: 0.0501\n",
            "25/50, train_loss: 0.0025\n",
            "26/50, train_loss: 0.0419\n",
            "27/50, train_loss: 0.0379\n",
            "28/50, train_loss: 0.0745\n",
            "29/50, train_loss: 0.0032\n",
            "30/50, train_loss: 0.0036\n",
            "31/50, train_loss: 0.0396\n",
            "32/50, train_loss: 0.0421\n",
            "33/50, train_loss: 0.0735\n",
            "34/50, train_loss: 0.0481\n",
            "35/50, train_loss: 0.0499\n",
            "36/50, train_loss: 0.0044\n",
            "37/50, train_loss: 0.0028\n",
            "38/50, train_loss: 0.0776\n",
            "39/50, train_loss: 0.0034\n",
            "40/50, train_loss: 0.1069\n",
            "41/50, train_loss: 0.0025\n",
            "42/50, train_loss: 0.0342\n",
            "43/50, train_loss: 0.1248\n",
            "44/50, train_loss: 0.0037\n",
            "45/50, train_loss: 0.0029\n",
            "46/50, train_loss: 0.0024\n",
            "47/50, train_loss: 0.0674\n",
            "48/50, train_loss: 0.0030\n",
            "49/50, train_loss: 0.0910\n",
            "50/50, train_loss: 0.0029\n",
            "epoch 113 average loss: 0.0390\n",
            "----------\n",
            "epoch 114/400\n",
            "1/50, train_loss: 0.0398\n",
            "2/50, train_loss: 0.0363\n",
            "3/50, train_loss: 0.0033\n",
            "4/50, train_loss: 0.0028\n",
            "5/50, train_loss: 0.0481\n",
            "6/50, train_loss: 0.0906\n",
            "7/50, train_loss: 0.0031\n",
            "8/50, train_loss: 0.0041\n",
            "9/50, train_loss: 0.0030\n",
            "10/50, train_loss: 0.0878\n",
            "11/50, train_loss: 0.0470\n",
            "12/50, train_loss: 0.1428\n",
            "13/50, train_loss: 0.0735\n",
            "14/50, train_loss: 0.0416\n",
            "15/50, train_loss: 0.0029\n",
            "16/50, train_loss: 0.0732\n",
            "17/50, train_loss: 0.0596\n",
            "18/50, train_loss: 0.0775\n",
            "19/50, train_loss: 0.0022\n",
            "20/50, train_loss: 0.0029\n",
            "21/50, train_loss: 0.0031\n",
            "22/50, train_loss: 0.0030\n",
            "23/50, train_loss: 0.0031\n",
            "24/50, train_loss: 0.0436\n",
            "25/50, train_loss: 0.0034\n",
            "26/50, train_loss: 0.0578\n",
            "27/50, train_loss: 0.0505\n",
            "28/50, train_loss: 0.0036\n",
            "29/50, train_loss: 0.0028\n",
            "30/50, train_loss: 0.0031\n",
            "31/50, train_loss: 0.0676\n",
            "32/50, train_loss: 0.0028\n",
            "33/50, train_loss: 0.0531\n",
            "34/50, train_loss: 0.0439\n",
            "35/50, train_loss: 0.0028\n",
            "36/50, train_loss: 0.0415\n",
            "37/50, train_loss: 0.0031\n",
            "38/50, train_loss: 0.0357\n",
            "39/50, train_loss: 0.0036\n",
            "40/50, train_loss: 0.0019\n",
            "41/50, train_loss: 0.0512\n",
            "42/50, train_loss: 0.0028\n",
            "43/50, train_loss: 0.0024\n",
            "44/50, train_loss: 0.0031\n",
            "45/50, train_loss: 0.0542\n",
            "46/50, train_loss: 0.0728\n",
            "47/50, train_loss: 0.0349\n",
            "48/50, train_loss: 0.0029\n",
            "49/50, train_loss: 0.0774\n",
            "50/50, train_loss: 0.0029\n",
            "epoch 114 average loss: 0.0315\n",
            "current epoch: 114 current accuracy: 0.7500 current AUC: 0.6723 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 115/400\n",
            "1/50, train_loss: 0.0863\n",
            "2/50, train_loss: 0.0802\n",
            "3/50, train_loss: 0.0031\n",
            "4/50, train_loss: 0.0031\n",
            "5/50, train_loss: 0.0024\n",
            "6/50, train_loss: 0.1928\n",
            "7/50, train_loss: 0.0027\n",
            "8/50, train_loss: 0.0030\n",
            "9/50, train_loss: 0.0421\n",
            "10/50, train_loss: 0.0029\n",
            "11/50, train_loss: 0.0029\n",
            "12/50, train_loss: 0.0034\n",
            "13/50, train_loss: 0.0567\n",
            "14/50, train_loss: 0.0034\n",
            "15/50, train_loss: 0.0552\n",
            "16/50, train_loss: 0.0496\n",
            "17/50, train_loss: 0.0853\n",
            "18/50, train_loss: 0.0608\n",
            "19/50, train_loss: 0.0026\n",
            "20/50, train_loss: 0.0727\n",
            "21/50, train_loss: 0.0501\n",
            "22/50, train_loss: 0.0026\n",
            "23/50, train_loss: 0.0752\n",
            "24/50, train_loss: 0.0591\n",
            "25/50, train_loss: 0.0033\n",
            "26/50, train_loss: 0.0809\n",
            "27/50, train_loss: 0.0036\n",
            "28/50, train_loss: 0.0036\n",
            "29/50, train_loss: 0.0373\n",
            "30/50, train_loss: 0.0730\n",
            "31/50, train_loss: 0.1290\n",
            "32/50, train_loss: 0.0033\n",
            "33/50, train_loss: 0.0033\n",
            "34/50, train_loss: 0.0560\n",
            "35/50, train_loss: 0.0361\n",
            "36/50, train_loss: 0.0442\n",
            "37/50, train_loss: 0.0033\n",
            "38/50, train_loss: 0.0582\n",
            "39/50, train_loss: 0.0030\n",
            "40/50, train_loss: 0.0028\n",
            "41/50, train_loss: 0.0407\n",
            "42/50, train_loss: 0.0037\n",
            "43/50, train_loss: 0.0346\n",
            "44/50, train_loss: 0.0038\n",
            "45/50, train_loss: 0.0858\n",
            "46/50, train_loss: 0.0026\n",
            "47/50, train_loss: 0.0050\n",
            "48/50, train_loss: 0.0031\n",
            "49/50, train_loss: 0.0046\n",
            "50/50, train_loss: 0.0528\n",
            "epoch 115 average loss: 0.0355\n",
            "----------\n",
            "epoch 116/400\n",
            "1/50, train_loss: 0.0020\n",
            "2/50, train_loss: 0.0038\n",
            "3/50, train_loss: 0.0031\n",
            "4/50, train_loss: 0.0026\n",
            "5/50, train_loss: 0.0047\n",
            "6/50, train_loss: 0.0040\n",
            "7/50, train_loss: 0.0531\n",
            "8/50, train_loss: 0.0640\n",
            "9/50, train_loss: 0.0031\n",
            "10/50, train_loss: 0.0025\n",
            "11/50, train_loss: 0.0031\n",
            "12/50, train_loss: 0.1001\n",
            "13/50, train_loss: 0.0032\n",
            "14/50, train_loss: 0.0403\n",
            "15/50, train_loss: 0.0383\n",
            "16/50, train_loss: 0.0326\n",
            "17/50, train_loss: 0.0382\n",
            "18/50, train_loss: 0.0038\n",
            "19/50, train_loss: 0.0026\n",
            "20/50, train_loss: 0.0032\n",
            "21/50, train_loss: 0.0025\n",
            "22/50, train_loss: 0.0026\n",
            "23/50, train_loss: 0.0028\n",
            "24/50, train_loss: 0.0637\n",
            "25/50, train_loss: 0.0036\n",
            "26/50, train_loss: 0.0594\n",
            "27/50, train_loss: 0.0034\n",
            "28/50, train_loss: 0.0045\n",
            "29/50, train_loss: 0.0031\n",
            "30/50, train_loss: 0.0034\n",
            "31/50, train_loss: 0.0533\n",
            "32/50, train_loss: 0.0454\n",
            "33/50, train_loss: 0.0032\n",
            "34/50, train_loss: 0.0492\n",
            "35/50, train_loss: 0.0030\n",
            "36/50, train_loss: 0.0029\n",
            "37/50, train_loss: 0.0026\n",
            "38/50, train_loss: 0.0032\n",
            "39/50, train_loss: 0.0032\n",
            "40/50, train_loss: 0.0031\n",
            "41/50, train_loss: 0.0034\n",
            "42/50, train_loss: 0.1224\n",
            "43/50, train_loss: 0.0034\n",
            "44/50, train_loss: 0.0737\n",
            "45/50, train_loss: 0.0513\n",
            "46/50, train_loss: 0.0437\n",
            "47/50, train_loss: 0.0460\n",
            "48/50, train_loss: 0.0029\n",
            "49/50, train_loss: 0.0032\n",
            "50/50, train_loss: 0.0029\n",
            "epoch 116 average loss: 0.0216\n",
            "current epoch: 116 current accuracy: 0.7500 current AUC: 0.7395 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 117/400\n",
            "1/50, train_loss: 0.0721\n",
            "2/50, train_loss: 0.0612\n",
            "3/50, train_loss: 0.0042\n",
            "4/50, train_loss: 0.0458\n",
            "5/50, train_loss: 0.0020\n",
            "6/50, train_loss: 0.0025\n",
            "7/50, train_loss: 0.0484\n",
            "8/50, train_loss: 0.1426\n",
            "9/50, train_loss: 0.0677\n",
            "10/50, train_loss: 0.0037\n",
            "11/50, train_loss: 0.0373\n",
            "12/50, train_loss: 0.0027\n",
            "13/50, train_loss: 0.0036\n",
            "14/50, train_loss: 0.0034\n",
            "15/50, train_loss: 0.0032\n",
            "16/50, train_loss: 0.0030\n",
            "17/50, train_loss: 0.0034\n",
            "18/50, train_loss: 0.0028\n",
            "19/50, train_loss: 0.0028\n",
            "20/50, train_loss: 0.0683\n",
            "21/50, train_loss: 0.0027\n",
            "22/50, train_loss: 0.0428\n",
            "23/50, train_loss: 0.0027\n",
            "24/50, train_loss: 0.0731\n",
            "25/50, train_loss: 0.0468\n",
            "26/50, train_loss: 0.0022\n",
            "27/50, train_loss: 0.0033\n",
            "28/50, train_loss: 0.0646\n",
            "29/50, train_loss: 0.0032\n",
            "30/50, train_loss: 0.0872\n",
            "31/50, train_loss: 0.0033\n",
            "32/50, train_loss: 0.0034\n",
            "33/50, train_loss: 0.0433\n",
            "34/50, train_loss: 0.0026\n",
            "35/50, train_loss: 0.0386\n",
            "36/50, train_loss: 0.0940\n",
            "37/50, train_loss: 0.0417\n",
            "38/50, train_loss: 0.0029\n",
            "39/50, train_loss: 0.0033\n",
            "40/50, train_loss: 0.0415\n",
            "41/50, train_loss: 0.0546\n",
            "42/50, train_loss: 0.0033\n",
            "43/50, train_loss: 0.0437\n",
            "44/50, train_loss: 0.0391\n",
            "45/50, train_loss: 0.0028\n",
            "46/50, train_loss: 0.0463\n",
            "47/50, train_loss: 0.0027\n",
            "48/50, train_loss: 0.0031\n",
            "49/50, train_loss: 0.0041\n",
            "50/50, train_loss: 0.0363\n",
            "epoch 117 average loss: 0.0284\n",
            "----------\n",
            "epoch 118/400\n",
            "1/50, train_loss: 0.0029\n",
            "2/50, train_loss: 0.0025\n",
            "3/50, train_loss: 0.0026\n",
            "4/50, train_loss: 0.0367\n",
            "5/50, train_loss: 0.0018\n",
            "6/50, train_loss: 0.0026\n",
            "7/50, train_loss: 0.0032\n",
            "8/50, train_loss: 0.0028\n",
            "9/50, train_loss: 0.0017\n",
            "10/50, train_loss: 0.0456\n",
            "11/50, train_loss: 0.0029\n",
            "12/50, train_loss: 0.0033\n",
            "13/50, train_loss: 0.0309\n",
            "14/50, train_loss: 0.0026\n",
            "15/50, train_loss: 0.1138\n",
            "16/50, train_loss: 0.0029\n",
            "17/50, train_loss: 0.0035\n",
            "18/50, train_loss: 0.0378\n",
            "19/50, train_loss: 0.0324\n",
            "20/50, train_loss: 0.0377\n",
            "21/50, train_loss: 0.0907\n",
            "22/50, train_loss: 0.0847\n",
            "23/50, train_loss: 0.0028\n",
            "24/50, train_loss: 0.0343\n",
            "25/50, train_loss: 0.0028\n",
            "26/50, train_loss: 0.0333\n",
            "27/50, train_loss: 0.0037\n",
            "28/50, train_loss: 0.0609\n",
            "29/50, train_loss: 0.0020\n",
            "30/50, train_loss: 0.0482\n",
            "31/50, train_loss: 0.0025\n",
            "32/50, train_loss: 0.0493\n",
            "33/50, train_loss: 0.0020\n",
            "34/50, train_loss: 0.0027\n",
            "35/50, train_loss: 0.0038\n",
            "36/50, train_loss: 0.0640\n",
            "37/50, train_loss: 0.0392\n",
            "38/50, train_loss: 0.0031\n",
            "39/50, train_loss: 0.0030\n",
            "40/50, train_loss: 0.0027\n",
            "41/50, train_loss: 0.0023\n",
            "42/50, train_loss: 0.0032\n",
            "43/50, train_loss: 0.0030\n",
            "44/50, train_loss: 0.0780\n",
            "45/50, train_loss: 0.0028\n",
            "46/50, train_loss: 0.0310\n",
            "47/50, train_loss: 0.0025\n",
            "48/50, train_loss: 0.0027\n",
            "49/50, train_loss: 0.0036\n",
            "50/50, train_loss: 0.0378\n",
            "epoch 118 average loss: 0.0215\n",
            "current epoch: 118 current accuracy: 0.7500 current AUC: 0.7311 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 119/400\n",
            "1/50, train_loss: 0.0033\n",
            "2/50, train_loss: 0.0033\n",
            "3/50, train_loss: 0.0029\n",
            "4/50, train_loss: 0.0543\n",
            "5/50, train_loss: 0.0704\n",
            "6/50, train_loss: 0.0028\n",
            "7/50, train_loss: 0.0027\n",
            "8/50, train_loss: 0.0023\n",
            "9/50, train_loss: 0.0023\n",
            "10/50, train_loss: 0.0026\n",
            "11/50, train_loss: 0.0027\n",
            "12/50, train_loss: 0.0021\n",
            "13/50, train_loss: 0.0402\n",
            "14/50, train_loss: 0.0531\n",
            "15/50, train_loss: 0.0317\n",
            "16/50, train_loss: 0.0407\n",
            "17/50, train_loss: 0.0026\n",
            "18/50, train_loss: 0.1666\n",
            "19/50, train_loss: 0.0593\n",
            "20/50, train_loss: 0.0023\n",
            "21/50, train_loss: 0.0027\n",
            "22/50, train_loss: 0.0030\n",
            "23/50, train_loss: 0.0674\n",
            "24/50, train_loss: 0.0027\n",
            "25/50, train_loss: 0.0343\n",
            "26/50, train_loss: 0.0373\n",
            "27/50, train_loss: 0.0642\n",
            "28/50, train_loss: 0.0023\n",
            "29/50, train_loss: 0.0385\n",
            "30/50, train_loss: 0.0356\n",
            "31/50, train_loss: 0.0032\n",
            "32/50, train_loss: 0.0020\n",
            "33/50, train_loss: 0.0025\n",
            "34/50, train_loss: 0.0382\n",
            "35/50, train_loss: 0.0037\n",
            "36/50, train_loss: 0.0026\n",
            "37/50, train_loss: 0.0023\n",
            "38/50, train_loss: 0.0346\n",
            "39/50, train_loss: 0.0452\n",
            "40/50, train_loss: 0.0681\n",
            "41/50, train_loss: 0.0019\n",
            "42/50, train_loss: 0.0387\n",
            "43/50, train_loss: 0.0455\n",
            "44/50, train_loss: 0.0392\n",
            "45/50, train_loss: 0.0558\n",
            "46/50, train_loss: 0.0604\n",
            "47/50, train_loss: 0.0322\n",
            "48/50, train_loss: 0.0655\n",
            "49/50, train_loss: 0.0028\n",
            "50/50, train_loss: 0.0028\n",
            "epoch 119 average loss: 0.0277\n",
            "----------\n",
            "epoch 120/400\n",
            "1/50, train_loss: 0.0024\n",
            "2/50, train_loss: 0.0702\n",
            "3/50, train_loss: 0.0021\n",
            "4/50, train_loss: 0.0706\n",
            "5/50, train_loss: 0.0025\n",
            "6/50, train_loss: 0.0044\n",
            "7/50, train_loss: 0.0420\n",
            "8/50, train_loss: 0.0391\n",
            "9/50, train_loss: 0.0609\n",
            "10/50, train_loss: 0.0410\n",
            "11/50, train_loss: 0.0315\n",
            "12/50, train_loss: 0.0023\n",
            "13/50, train_loss: 0.0019\n",
            "14/50, train_loss: 0.0694\n",
            "15/50, train_loss: 0.0027\n",
            "16/50, train_loss: 0.0503\n",
            "17/50, train_loss: 0.0366\n",
            "18/50, train_loss: 0.0972\n",
            "19/50, train_loss: 0.0699\n",
            "20/50, train_loss: 0.0501\n",
            "21/50, train_loss: 0.0325\n",
            "22/50, train_loss: 0.0026\n",
            "23/50, train_loss: 0.0486\n",
            "24/50, train_loss: 0.0742\n",
            "25/50, train_loss: 0.0025\n",
            "26/50, train_loss: 0.0328\n",
            "27/50, train_loss: 0.0022\n",
            "28/50, train_loss: 0.0025\n",
            "29/50, train_loss: 0.0022\n",
            "30/50, train_loss: 0.0375\n",
            "31/50, train_loss: 0.0450\n",
            "32/50, train_loss: 0.0404\n",
            "33/50, train_loss: 0.0029\n",
            "34/50, train_loss: 0.0021\n",
            "35/50, train_loss: 0.0427\n",
            "36/50, train_loss: 0.0401\n",
            "37/50, train_loss: 0.0901\n",
            "38/50, train_loss: 0.0601\n",
            "39/50, train_loss: 0.0025\n",
            "40/50, train_loss: 0.0031\n",
            "41/50, train_loss: 0.0359\n",
            "42/50, train_loss: 0.0027\n",
            "43/50, train_loss: 0.0023\n",
            "44/50, train_loss: 0.0374\n",
            "45/50, train_loss: 0.0028\n",
            "46/50, train_loss: 0.0351\n",
            "47/50, train_loss: 0.1313\n",
            "48/50, train_loss: 0.0623\n",
            "49/50, train_loss: 0.0613\n",
            "50/50, train_loss: 0.0337\n",
            "epoch 120 average loss: 0.0344\n",
            "current epoch: 120 current accuracy: 0.7500 current AUC: 0.7227 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 121/400\n",
            "1/50, train_loss: 0.0499\n",
            "2/50, train_loss: 0.0341\n",
            "3/50, train_loss: 0.0349\n",
            "4/50, train_loss: 0.0358\n",
            "5/50, train_loss: 0.0025\n",
            "6/50, train_loss: 0.0808\n",
            "7/50, train_loss: 0.0027\n",
            "8/50, train_loss: 0.0558\n",
            "9/50, train_loss: 0.0656\n",
            "10/50, train_loss: 0.0363\n",
            "11/50, train_loss: 0.0405\n",
            "12/50, train_loss: 0.0032\n",
            "13/50, train_loss: 0.0596\n",
            "14/50, train_loss: 0.0022\n",
            "15/50, train_loss: 0.0023\n",
            "16/50, train_loss: 0.0408\n",
            "17/50, train_loss: 0.0312\n",
            "18/50, train_loss: 0.0456\n",
            "19/50, train_loss: 0.0023\n",
            "20/50, train_loss: 0.0023\n",
            "21/50, train_loss: 0.0033\n",
            "22/50, train_loss: 0.0333\n",
            "23/50, train_loss: 0.0645\n",
            "24/50, train_loss: 0.0292\n",
            "25/50, train_loss: 0.0337\n",
            "26/50, train_loss: 0.0483\n",
            "27/50, train_loss: 0.0501\n",
            "28/50, train_loss: 0.0034\n",
            "29/50, train_loss: 0.0027\n",
            "30/50, train_loss: 0.0604\n",
            "31/50, train_loss: 0.0028\n",
            "32/50, train_loss: 0.0027\n",
            "33/50, train_loss: 0.0613\n",
            "34/50, train_loss: 0.0028\n",
            "35/50, train_loss: 0.0366\n",
            "36/50, train_loss: 0.1109\n",
            "37/50, train_loss: 0.0024\n",
            "38/50, train_loss: 0.0336\n",
            "39/50, train_loss: 0.0032\n",
            "40/50, train_loss: 0.0373\n",
            "41/50, train_loss: 0.0692\n",
            "42/50, train_loss: 0.0355\n",
            "43/50, train_loss: 0.0021\n",
            "44/50, train_loss: 0.0294\n",
            "45/50, train_loss: 0.0028\n",
            "46/50, train_loss: 0.0638\n",
            "47/50, train_loss: 0.0029\n",
            "48/50, train_loss: 0.0679\n",
            "49/50, train_loss: 0.0030\n",
            "50/50, train_loss: 0.0887\n",
            "epoch 121 average loss: 0.0323\n",
            "----------\n",
            "epoch 122/400\n",
            "1/50, train_loss: 0.0379\n",
            "2/50, train_loss: 0.0027\n",
            "3/50, train_loss: 0.0300\n",
            "4/50, train_loss: 0.0024\n",
            "5/50, train_loss: 0.0341\n",
            "6/50, train_loss: 0.0676\n",
            "7/50, train_loss: 0.0359\n",
            "8/50, train_loss: 0.0028\n",
            "9/50, train_loss: 0.0544\n",
            "10/50, train_loss: 0.0023\n",
            "11/50, train_loss: 0.0420\n",
            "12/50, train_loss: 0.0749\n",
            "13/50, train_loss: 0.0358\n",
            "14/50, train_loss: 0.0388\n",
            "15/50, train_loss: 0.0418\n",
            "16/50, train_loss: 0.0530\n",
            "17/50, train_loss: 0.0326\n",
            "18/50, train_loss: 0.0433\n",
            "19/50, train_loss: 0.0025\n",
            "20/50, train_loss: 0.0356\n",
            "21/50, train_loss: 0.0024\n",
            "22/50, train_loss: 0.0027\n",
            "23/50, train_loss: 0.0024\n",
            "24/50, train_loss: 0.0699\n",
            "25/50, train_loss: 0.0030\n",
            "26/50, train_loss: 0.0509\n",
            "27/50, train_loss: 0.0581\n",
            "28/50, train_loss: 0.0319\n",
            "29/50, train_loss: 0.0018\n",
            "30/50, train_loss: 0.0024\n",
            "31/50, train_loss: 0.0685\n",
            "32/50, train_loss: 0.0590\n",
            "33/50, train_loss: 0.0030\n",
            "34/50, train_loss: 0.0025\n",
            "35/50, train_loss: 0.0372\n",
            "36/50, train_loss: 0.0716\n",
            "37/50, train_loss: 0.0325\n",
            "38/50, train_loss: 0.0016\n",
            "39/50, train_loss: 0.0340\n",
            "40/50, train_loss: 0.0311\n",
            "41/50, train_loss: 0.0027\n",
            "42/50, train_loss: 0.0888\n",
            "43/50, train_loss: 0.0392\n",
            "44/50, train_loss: 0.0830\n",
            "45/50, train_loss: 0.0019\n",
            "46/50, train_loss: 0.0024\n",
            "47/50, train_loss: 0.0025\n",
            "48/50, train_loss: 0.0535\n",
            "49/50, train_loss: 0.0346\n",
            "50/50, train_loss: 0.0019\n",
            "epoch 122 average loss: 0.0310\n",
            "current epoch: 122 current accuracy: 0.7500 current AUC: 0.7059 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 123/400\n",
            "1/50, train_loss: 0.0028\n",
            "2/50, train_loss: 0.0444\n",
            "3/50, train_loss: 0.0022\n",
            "4/50, train_loss: 0.0458\n",
            "5/50, train_loss: 0.0758\n",
            "6/50, train_loss: 0.0327\n",
            "7/50, train_loss: 0.0023\n",
            "8/50, train_loss: 0.0024\n",
            "9/50, train_loss: 0.0611\n",
            "10/50, train_loss: 0.0406\n",
            "11/50, train_loss: 0.0531\n",
            "12/50, train_loss: 0.0027\n",
            "13/50, train_loss: 0.0438\n",
            "14/50, train_loss: 0.0026\n",
            "15/50, train_loss: 0.0494\n",
            "16/50, train_loss: 0.0309\n",
            "17/50, train_loss: 0.0356\n",
            "18/50, train_loss: 0.0022\n",
            "19/50, train_loss: 0.0027\n",
            "20/50, train_loss: 0.0364\n",
            "21/50, train_loss: 0.0021\n",
            "22/50, train_loss: 0.0027\n",
            "23/50, train_loss: 0.0368\n",
            "24/50, train_loss: 0.0453\n",
            "25/50, train_loss: 0.0279\n",
            "26/50, train_loss: 0.0028\n",
            "27/50, train_loss: 0.0746\n",
            "28/50, train_loss: 0.0619\n",
            "29/50, train_loss: 0.0618\n",
            "30/50, train_loss: 0.0026\n",
            "31/50, train_loss: 0.0030\n",
            "32/50, train_loss: 0.0026\n",
            "33/50, train_loss: 0.0650\n",
            "34/50, train_loss: 0.0339\n",
            "35/50, train_loss: 0.0991\n",
            "36/50, train_loss: 0.0350\n",
            "37/50, train_loss: 0.0025\n",
            "38/50, train_loss: 0.0026\n",
            "39/50, train_loss: 0.0023\n",
            "40/50, train_loss: 0.0024\n",
            "41/50, train_loss: 0.0028\n",
            "42/50, train_loss: 0.0842\n",
            "43/50, train_loss: 0.0024\n",
            "44/50, train_loss: 0.0027\n",
            "45/50, train_loss: 0.0356\n",
            "46/50, train_loss: 0.0024\n",
            "47/50, train_loss: 0.0565\n",
            "48/50, train_loss: 0.0550\n",
            "49/50, train_loss: 0.0338\n",
            "50/50, train_loss: 0.0024\n",
            "epoch 123 average loss: 0.0283\n",
            "----------\n",
            "epoch 124/400\n",
            "1/50, train_loss: 0.0908\n",
            "2/50, train_loss: 0.0683\n",
            "3/50, train_loss: 0.0413\n",
            "4/50, train_loss: 0.0544\n",
            "5/50, train_loss: 0.0029\n",
            "6/50, train_loss: 0.0023\n",
            "7/50, train_loss: 0.0468\n",
            "8/50, train_loss: 0.0470\n",
            "9/50, train_loss: 0.0017\n",
            "10/50, train_loss: 0.0625\n",
            "11/50, train_loss: 0.0021\n",
            "12/50, train_loss: 0.0030\n",
            "13/50, train_loss: 0.0518\n",
            "14/50, train_loss: 0.0581\n",
            "15/50, train_loss: 0.0350\n",
            "16/50, train_loss: 0.0023\n",
            "17/50, train_loss: 0.0314\n",
            "18/50, train_loss: 0.0339\n",
            "19/50, train_loss: 0.0022\n",
            "20/50, train_loss: 0.0035\n",
            "21/50, train_loss: 0.0301\n",
            "22/50, train_loss: 0.0286\n",
            "23/50, train_loss: 0.0024\n",
            "24/50, train_loss: 0.0022\n",
            "25/50, train_loss: 0.0494\n",
            "26/50, train_loss: 0.0025\n",
            "27/50, train_loss: 0.0417\n",
            "28/50, train_loss: 0.0024\n",
            "29/50, train_loss: 0.0025\n",
            "30/50, train_loss: 0.0396\n",
            "31/50, train_loss: 0.0024\n",
            "32/50, train_loss: 0.0501\n",
            "33/50, train_loss: 0.0703\n",
            "34/50, train_loss: 0.0021\n",
            "35/50, train_loss: 0.0332\n",
            "36/50, train_loss: 0.0029\n",
            "37/50, train_loss: 0.0490\n",
            "38/50, train_loss: 0.0278\n",
            "39/50, train_loss: 0.0484\n",
            "40/50, train_loss: 0.0024\n",
            "41/50, train_loss: 0.0029\n",
            "42/50, train_loss: 0.0022\n",
            "43/50, train_loss: 0.0361\n",
            "44/50, train_loss: 0.0023\n",
            "45/50, train_loss: 0.0022\n",
            "46/50, train_loss: 0.0024\n",
            "47/50, train_loss: 0.0796\n",
            "48/50, train_loss: 0.0606\n",
            "49/50, train_loss: 0.0326\n",
            "50/50, train_loss: 0.0022\n",
            "epoch 124 average loss: 0.0271\n",
            "current epoch: 124 current accuracy: 0.7500 current AUC: 0.6639 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 125/400\n",
            "1/50, train_loss: 0.0025\n",
            "2/50, train_loss: 0.0030\n",
            "3/50, train_loss: 0.0696\n",
            "4/50, train_loss: 0.0389\n",
            "5/50, train_loss: 0.0024\n",
            "6/50, train_loss: 0.0024\n",
            "7/50, train_loss: 0.0024\n",
            "8/50, train_loss: 0.0560\n",
            "9/50, train_loss: 0.0398\n",
            "10/50, train_loss: 0.0815\n",
            "11/50, train_loss: 0.0293\n",
            "12/50, train_loss: 0.0039\n",
            "13/50, train_loss: 0.0025\n",
            "14/50, train_loss: 0.0022\n",
            "15/50, train_loss: 0.0022\n",
            "16/50, train_loss: 0.0022\n",
            "17/50, train_loss: 0.0402\n",
            "18/50, train_loss: 0.0536\n",
            "19/50, train_loss: 0.0476\n",
            "20/50, train_loss: 0.0023\n",
            "21/50, train_loss: 0.0623\n",
            "22/50, train_loss: 0.0293\n",
            "23/50, train_loss: 0.0567\n",
            "24/50, train_loss: 0.0275\n",
            "25/50, train_loss: 0.0523\n",
            "26/50, train_loss: 0.0328\n",
            "27/50, train_loss: 0.0026\n",
            "28/50, train_loss: 0.0483\n",
            "29/50, train_loss: 0.0027\n",
            "30/50, train_loss: 0.0378\n",
            "31/50, train_loss: 0.0020\n",
            "32/50, train_loss: 0.0646\n",
            "33/50, train_loss: 0.0369\n",
            "34/50, train_loss: 0.0325\n",
            "35/50, train_loss: 0.0681\n",
            "36/50, train_loss: 0.0336\n",
            "37/50, train_loss: 0.0317\n",
            "38/50, train_loss: 0.0020\n",
            "39/50, train_loss: 0.0578\n",
            "40/50, train_loss: 0.0515\n",
            "41/50, train_loss: 0.0354\n",
            "42/50, train_loss: 0.0328\n",
            "43/50, train_loss: 0.0021\n",
            "44/50, train_loss: 0.0020\n",
            "45/50, train_loss: 0.0383\n",
            "46/50, train_loss: 0.0356\n",
            "47/50, train_loss: 0.0018\n",
            "48/50, train_loss: 0.0023\n",
            "49/50, train_loss: 0.0020\n",
            "50/50, train_loss: 0.0021\n",
            "epoch 125 average loss: 0.0274\n",
            "----------\n",
            "epoch 126/400\n",
            "1/50, train_loss: 0.0031\n",
            "2/50, train_loss: 0.0324\n",
            "3/50, train_loss: 0.0019\n",
            "4/50, train_loss: 0.0021\n",
            "5/50, train_loss: 0.0770\n",
            "6/50, train_loss: 0.0599\n",
            "7/50, train_loss: 0.0565\n",
            "8/50, train_loss: 0.0648\n",
            "9/50, train_loss: 0.0280\n",
            "10/50, train_loss: 0.0482\n",
            "11/50, train_loss: 0.0299\n",
            "12/50, train_loss: 0.0020\n",
            "13/50, train_loss: 0.0021\n",
            "14/50, train_loss: 0.0017\n",
            "15/50, train_loss: 0.0028\n",
            "16/50, train_loss: 0.0021\n",
            "17/50, train_loss: 0.0026\n",
            "18/50, train_loss: 0.0027\n",
            "19/50, train_loss: 0.0025\n",
            "20/50, train_loss: 0.0020\n",
            "21/50, train_loss: 0.0017\n",
            "22/50, train_loss: 0.0331\n",
            "23/50, train_loss: 0.0313\n",
            "24/50, train_loss: 0.0348\n",
            "25/50, train_loss: 0.1009\n",
            "26/50, train_loss: 0.0348\n",
            "27/50, train_loss: 0.0021\n",
            "28/50, train_loss: 0.0471\n",
            "29/50, train_loss: 0.0272\n",
            "30/50, train_loss: 0.0021\n",
            "31/50, train_loss: 0.0024\n",
            "32/50, train_loss: 0.0294\n",
            "33/50, train_loss: 0.0387\n",
            "34/50, train_loss: 0.0025\n",
            "35/50, train_loss: 0.0518\n",
            "36/50, train_loss: 0.0458\n",
            "37/50, train_loss: 0.0303\n",
            "38/50, train_loss: 0.0026\n",
            "39/50, train_loss: 0.0024\n",
            "40/50, train_loss: 0.0030\n",
            "41/50, train_loss: 0.0566\n",
            "42/50, train_loss: 0.0622\n",
            "43/50, train_loss: 0.0429\n",
            "44/50, train_loss: 0.0019\n",
            "45/50, train_loss: 0.0020\n",
            "46/50, train_loss: 0.0019\n",
            "47/50, train_loss: 0.0959\n",
            "48/50, train_loss: 0.0362\n",
            "49/50, train_loss: 0.0363\n",
            "50/50, train_loss: 0.0376\n",
            "epoch 126 average loss: 0.0264\n",
            "current epoch: 126 current accuracy: 0.7500 current AUC: 0.6807 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 127/400\n",
            "1/50, train_loss: 0.0024\n",
            "2/50, train_loss: 0.0023\n",
            "3/50, train_loss: 0.0024\n",
            "4/50, train_loss: 0.0019\n",
            "5/50, train_loss: 0.0439\n",
            "6/50, train_loss: 0.0028\n",
            "7/50, train_loss: 0.0019\n",
            "8/50, train_loss: 0.0482\n",
            "9/50, train_loss: 0.0022\n",
            "10/50, train_loss: 0.0258\n",
            "11/50, train_loss: 0.0373\n",
            "12/50, train_loss: 0.0315\n",
            "13/50, train_loss: 0.0274\n",
            "14/50, train_loss: 0.0021\n",
            "15/50, train_loss: 0.0494\n",
            "16/50, train_loss: 0.0593\n",
            "17/50, train_loss: 0.0022\n",
            "18/50, train_loss: 0.0023\n",
            "19/50, train_loss: 0.0290\n",
            "20/50, train_loss: 0.0018\n",
            "21/50, train_loss: 0.0025\n",
            "22/50, train_loss: 0.0023\n",
            "23/50, train_loss: 0.0946\n",
            "24/50, train_loss: 0.0514\n",
            "25/50, train_loss: 0.0502\n",
            "26/50, train_loss: 0.0023\n",
            "27/50, train_loss: 0.0023\n",
            "28/50, train_loss: 0.0369\n",
            "29/50, train_loss: 0.0636\n",
            "30/50, train_loss: 0.0301\n",
            "31/50, train_loss: 0.0539\n",
            "32/50, train_loss: 0.0022\n",
            "33/50, train_loss: 0.0303\n",
            "34/50, train_loss: 0.0017\n",
            "35/50, train_loss: 0.0290\n",
            "36/50, train_loss: 0.0016\n",
            "37/50, train_loss: 0.0454\n",
            "38/50, train_loss: 0.0027\n",
            "39/50, train_loss: 0.0024\n",
            "40/50, train_loss: 0.0550\n",
            "41/50, train_loss: 0.0019\n",
            "42/50, train_loss: 0.0391\n",
            "43/50, train_loss: 0.0023\n",
            "44/50, train_loss: 0.0017\n",
            "45/50, train_loss: 0.0284\n",
            "46/50, train_loss: 0.0324\n",
            "47/50, train_loss: 0.0023\n",
            "48/50, train_loss: 0.0019\n",
            "49/50, train_loss: 0.0021\n",
            "50/50, train_loss: 0.0019\n",
            "epoch 127 average loss: 0.0210\n",
            "----------\n",
            "epoch 128/400\n",
            "1/50, train_loss: 0.0019\n",
            "2/50, train_loss: 0.0024\n",
            "3/50, train_loss: 0.0020\n",
            "4/50, train_loss: 0.0322\n",
            "5/50, train_loss: 0.0016\n",
            "6/50, train_loss: 0.0275\n",
            "7/50, train_loss: 0.0022\n",
            "8/50, train_loss: 0.0318\n",
            "9/50, train_loss: 0.0019\n",
            "10/50, train_loss: 0.0511\n",
            "11/50, train_loss: 0.0526\n",
            "12/50, train_loss: 0.0569\n",
            "13/50, train_loss: 0.0267\n",
            "14/50, train_loss: 0.0019\n",
            "15/50, train_loss: 0.0462\n",
            "16/50, train_loss: 0.0268\n",
            "17/50, train_loss: 0.0023\n",
            "18/50, train_loss: 0.0380\n",
            "19/50, train_loss: 0.0012\n",
            "20/50, train_loss: 0.0342\n",
            "21/50, train_loss: 0.0422\n",
            "22/50, train_loss: 0.0020\n",
            "23/50, train_loss: 0.0021\n",
            "24/50, train_loss: 0.0275\n",
            "25/50, train_loss: 0.0357\n",
            "26/50, train_loss: 0.0025\n",
            "27/50, train_loss: 0.0023\n",
            "28/50, train_loss: 0.0527\n",
            "29/50, train_loss: 0.0022\n",
            "30/50, train_loss: 0.0307\n",
            "31/50, train_loss: 0.0733\n",
            "32/50, train_loss: 0.0555\n",
            "33/50, train_loss: 0.0023\n",
            "34/50, train_loss: 0.0661\n",
            "35/50, train_loss: 0.0300\n",
            "36/50, train_loss: 0.0295\n",
            "37/50, train_loss: 0.0482\n",
            "38/50, train_loss: 0.0022\n",
            "39/50, train_loss: 0.0022\n",
            "40/50, train_loss: 0.0256\n",
            "41/50, train_loss: 0.0023\n",
            "42/50, train_loss: 0.0016\n",
            "43/50, train_loss: 0.0019\n",
            "44/50, train_loss: 0.0020\n",
            "45/50, train_loss: 0.0617\n",
            "46/50, train_loss: 0.0015\n",
            "47/50, train_loss: 0.0394\n",
            "48/50, train_loss: 0.0026\n",
            "49/50, train_loss: 0.0406\n",
            "50/50, train_loss: 0.0331\n",
            "epoch 128 average loss: 0.0233\n",
            "current epoch: 128 current accuracy: 0.7500 current AUC: 0.7311 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 129/400\n",
            "1/50, train_loss: 0.0360\n",
            "2/50, train_loss: 0.0015\n",
            "3/50, train_loss: 0.0492\n",
            "4/50, train_loss: 0.0020\n",
            "5/50, train_loss: 0.0018\n",
            "6/50, train_loss: 0.0314\n",
            "7/50, train_loss: 0.0025\n",
            "8/50, train_loss: 0.0021\n",
            "9/50, train_loss: 0.1045\n",
            "10/50, train_loss: 0.0018\n",
            "11/50, train_loss: 0.0400\n",
            "12/50, train_loss: 0.0023\n",
            "13/50, train_loss: 0.0017\n",
            "14/50, train_loss: 0.0513\n",
            "15/50, train_loss: 0.0018\n",
            "16/50, train_loss: 0.0386\n",
            "17/50, train_loss: 0.0018\n",
            "18/50, train_loss: 0.0028\n",
            "19/50, train_loss: 0.0017\n",
            "20/50, train_loss: 0.0292\n",
            "21/50, train_loss: 0.0452\n",
            "22/50, train_loss: 0.0290\n",
            "23/50, train_loss: 0.0636\n",
            "24/50, train_loss: 0.0283\n",
            "25/50, train_loss: 0.0272\n",
            "26/50, train_loss: 0.0483\n",
            "27/50, train_loss: 0.0021\n",
            "28/50, train_loss: 0.0020\n",
            "29/50, train_loss: 0.0018\n",
            "30/50, train_loss: 0.0019\n",
            "31/50, train_loss: 0.0418\n",
            "32/50, train_loss: 0.0501\n",
            "33/50, train_loss: 0.0017\n",
            "34/50, train_loss: 0.0327\n",
            "35/50, train_loss: 0.0281\n",
            "36/50, train_loss: 0.0015\n",
            "37/50, train_loss: 0.0278\n",
            "38/50, train_loss: 0.0379\n",
            "39/50, train_loss: 0.0324\n",
            "40/50, train_loss: 0.0018\n",
            "41/50, train_loss: 0.0261\n",
            "42/50, train_loss: 0.0022\n",
            "43/50, train_loss: 0.0021\n",
            "44/50, train_loss: 0.0597\n",
            "45/50, train_loss: 0.0019\n",
            "46/50, train_loss: 0.0311\n",
            "47/50, train_loss: 0.0020\n",
            "48/50, train_loss: 0.0013\n",
            "49/50, train_loss: 0.0309\n",
            "50/50, train_loss: 0.0020\n",
            "epoch 129 average loss: 0.0214\n",
            "----------\n",
            "epoch 130/400\n",
            "1/50, train_loss: 0.0015\n",
            "2/50, train_loss: 0.0267\n",
            "3/50, train_loss: 0.0327\n",
            "4/50, train_loss: 0.0020\n",
            "5/50, train_loss: 0.0017\n",
            "6/50, train_loss: 0.0539\n",
            "7/50, train_loss: 0.0016\n",
            "8/50, train_loss: 0.0471\n",
            "9/50, train_loss: 0.0025\n",
            "10/50, train_loss: 0.0500\n",
            "11/50, train_loss: 0.0490\n",
            "12/50, train_loss: 0.0272\n",
            "13/50, train_loss: 0.0017\n",
            "14/50, train_loss: 0.0697\n",
            "15/50, train_loss: 0.0019\n",
            "16/50, train_loss: 0.0302\n",
            "17/50, train_loss: 0.0016\n",
            "18/50, train_loss: 0.0888\n",
            "19/50, train_loss: 0.0307\n",
            "20/50, train_loss: 0.0017\n",
            "21/50, train_loss: 0.0023\n",
            "22/50, train_loss: 0.0017\n",
            "23/50, train_loss: 0.0018\n",
            "24/50, train_loss: 0.0021\n",
            "25/50, train_loss: 0.0305\n",
            "26/50, train_loss: 0.0296\n",
            "27/50, train_loss: 0.0309\n",
            "28/50, train_loss: 0.0017\n",
            "29/50, train_loss: 0.0275\n",
            "30/50, train_loss: 0.0020\n",
            "31/50, train_loss: 0.0457\n",
            "32/50, train_loss: 0.0267\n",
            "33/50, train_loss: 0.0019\n",
            "34/50, train_loss: 0.0261\n",
            "35/50, train_loss: 0.0017\n",
            "36/50, train_loss: 0.0013\n",
            "37/50, train_loss: 0.0020\n",
            "38/50, train_loss: 0.0017\n",
            "39/50, train_loss: 0.0433\n",
            "40/50, train_loss: 0.0018\n",
            "41/50, train_loss: 0.0320\n",
            "42/50, train_loss: 0.0022\n",
            "43/50, train_loss: 0.0029\n",
            "44/50, train_loss: 0.0019\n",
            "45/50, train_loss: 0.0019\n",
            "46/50, train_loss: 0.0476\n",
            "47/50, train_loss: 0.0022\n",
            "48/50, train_loss: 0.0015\n",
            "49/50, train_loss: 0.0018\n",
            "50/50, train_loss: 0.0014\n",
            "epoch 130 average loss: 0.0180\n",
            "current epoch: 130 current accuracy: 0.7083 current AUC: 0.6723 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 131/400\n",
            "1/50, train_loss: 0.0430\n",
            "2/50, train_loss: 0.0020\n",
            "3/50, train_loss: 0.0018\n",
            "4/50, train_loss: 0.0397\n",
            "5/50, train_loss: 0.0027\n",
            "6/50, train_loss: 0.0023\n",
            "7/50, train_loss: 0.0370\n",
            "8/50, train_loss: 0.0269\n",
            "9/50, train_loss: 0.0240\n",
            "10/50, train_loss: 0.0302\n",
            "11/50, train_loss: 0.0020\n",
            "12/50, train_loss: 0.0020\n",
            "13/50, train_loss: 0.0241\n",
            "14/50, train_loss: 0.0018\n",
            "15/50, train_loss: 0.0017\n",
            "16/50, train_loss: 0.0309\n",
            "17/50, train_loss: 0.0015\n",
            "18/50, train_loss: 0.0248\n",
            "19/50, train_loss: 0.0018\n",
            "20/50, train_loss: 0.0017\n",
            "21/50, train_loss: 0.0612\n",
            "22/50, train_loss: 0.0013\n",
            "23/50, train_loss: 0.0343\n",
            "24/50, train_loss: 0.0286\n",
            "25/50, train_loss: 0.0018\n",
            "26/50, train_loss: 0.0494\n",
            "27/50, train_loss: 0.0257\n",
            "28/50, train_loss: 0.0544\n",
            "29/50, train_loss: 0.0021\n",
            "30/50, train_loss: 0.0393\n",
            "31/50, train_loss: 0.0017\n",
            "32/50, train_loss: 0.0667\n",
            "33/50, train_loss: 0.0016\n",
            "34/50, train_loss: 0.0018\n",
            "35/50, train_loss: 0.0019\n",
            "36/50, train_loss: 0.0016\n",
            "37/50, train_loss: 0.0014\n",
            "38/50, train_loss: 0.0017\n",
            "39/50, train_loss: 0.0016\n",
            "40/50, train_loss: 0.0017\n",
            "41/50, train_loss: 0.0955\n",
            "42/50, train_loss: 0.0021\n",
            "43/50, train_loss: 0.0293\n",
            "44/50, train_loss: 0.0492\n",
            "45/50, train_loss: 0.0282\n",
            "46/50, train_loss: 0.0420\n",
            "47/50, train_loss: 0.0309\n",
            "48/50, train_loss: 0.0021\n",
            "49/50, train_loss: 0.0015\n",
            "50/50, train_loss: 0.0017\n",
            "epoch 131 average loss: 0.0193\n",
            "----------\n",
            "epoch 132/400\n",
            "1/50, train_loss: 0.0022\n",
            "2/50, train_loss: 0.0256\n",
            "3/50, train_loss: 0.0460\n",
            "4/50, train_loss: 0.0293\n",
            "5/50, train_loss: 0.0309\n",
            "6/50, train_loss: 0.0020\n",
            "7/50, train_loss: 0.0018\n",
            "8/50, train_loss: 0.0015\n",
            "9/50, train_loss: 0.0021\n",
            "10/50, train_loss: 0.0021\n",
            "11/50, train_loss: 0.0519\n",
            "12/50, train_loss: 0.0018\n",
            "13/50, train_loss: 0.0015\n",
            "14/50, train_loss: 0.0713\n",
            "15/50, train_loss: 0.0018\n",
            "16/50, train_loss: 0.0018\n",
            "17/50, train_loss: 0.0019\n",
            "18/50, train_loss: 0.0016\n",
            "19/50, train_loss: 0.0318\n",
            "20/50, train_loss: 0.0438\n",
            "21/50, train_loss: 0.0019\n",
            "22/50, train_loss: 0.0026\n",
            "23/50, train_loss: 0.0254\n",
            "24/50, train_loss: 0.0013\n",
            "25/50, train_loss: 0.0014\n",
            "26/50, train_loss: 0.0019\n",
            "27/50, train_loss: 0.0017\n",
            "28/50, train_loss: 0.0293\n",
            "29/50, train_loss: 0.0018\n",
            "30/50, train_loss: 0.0019\n",
            "31/50, train_loss: 0.0430\n",
            "32/50, train_loss: 0.0016\n",
            "33/50, train_loss: 0.0022\n",
            "34/50, train_loss: 0.0344\n",
            "35/50, train_loss: 0.0317\n",
            "36/50, train_loss: 0.0018\n",
            "37/50, train_loss: 0.0017\n",
            "38/50, train_loss: 0.0016\n",
            "39/50, train_loss: 0.0340\n",
            "40/50, train_loss: 0.0020\n",
            "41/50, train_loss: 0.0317\n",
            "42/50, train_loss: 0.0275\n",
            "43/50, train_loss: 0.0279\n",
            "44/50, train_loss: 0.0530\n",
            "45/50, train_loss: 0.0696\n",
            "46/50, train_loss: 0.0019\n",
            "47/50, train_loss: 0.0253\n",
            "48/50, train_loss: 0.0018\n",
            "49/50, train_loss: 0.0592\n",
            "50/50, train_loss: 0.0022\n",
            "epoch 132 average loss: 0.0175\n",
            "current epoch: 132 current accuracy: 0.7083 current AUC: 0.6639 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 133/400\n",
            "1/50, train_loss: 0.0241\n",
            "2/50, train_loss: 0.0012\n",
            "3/50, train_loss: 0.0398\n",
            "4/50, train_loss: 0.0244\n",
            "5/50, train_loss: 0.0287\n",
            "6/50, train_loss: 0.0245\n",
            "7/50, train_loss: 0.0261\n",
            "8/50, train_loss: 0.0018\n",
            "9/50, train_loss: 0.0017\n",
            "10/50, train_loss: 0.0022\n",
            "11/50, train_loss: 0.0530\n",
            "12/50, train_loss: 0.0264\n",
            "13/50, train_loss: 0.0020\n",
            "14/50, train_loss: 0.0016\n",
            "15/50, train_loss: 0.0018\n",
            "16/50, train_loss: 0.0017\n",
            "17/50, train_loss: 0.0448\n",
            "18/50, train_loss: 0.0225\n",
            "19/50, train_loss: 0.0021\n",
            "20/50, train_loss: 0.0017\n",
            "21/50, train_loss: 0.0016\n",
            "22/50, train_loss: 0.0012\n",
            "23/50, train_loss: 0.0017\n",
            "24/50, train_loss: 0.0356\n",
            "25/50, train_loss: 0.0014\n",
            "26/50, train_loss: 0.0448\n",
            "27/50, train_loss: 0.0016\n",
            "28/50, train_loss: 0.0016\n",
            "29/50, train_loss: 0.0015\n",
            "30/50, train_loss: 0.0019\n",
            "31/50, train_loss: 0.0274\n",
            "32/50, train_loss: 0.0020\n",
            "33/50, train_loss: 0.0315\n",
            "34/50, train_loss: 0.0011\n",
            "35/50, train_loss: 0.0576\n",
            "36/50, train_loss: 0.0016\n",
            "37/50, train_loss: 0.0352\n",
            "38/50, train_loss: 0.0019\n",
            "39/50, train_loss: 0.0254\n",
            "40/50, train_loss: 0.0017\n",
            "41/50, train_loss: 0.0011\n",
            "42/50, train_loss: 0.0591\n",
            "43/50, train_loss: 0.0017\n",
            "44/50, train_loss: 0.0018\n",
            "45/50, train_loss: 0.0604\n",
            "46/50, train_loss: 0.0013\n",
            "47/50, train_loss: 0.0305\n",
            "48/50, train_loss: 0.0017\n",
            "49/50, train_loss: 0.0017\n",
            "50/50, train_loss: 0.0232\n",
            "epoch 133 average loss: 0.0159\n",
            "----------\n",
            "epoch 134/400\n",
            "1/50, train_loss: 0.0017\n",
            "2/50, train_loss: 0.0260\n",
            "3/50, train_loss: 0.0017\n",
            "4/50, train_loss: 0.0017\n",
            "5/50, train_loss: 0.0335\n",
            "6/50, train_loss: 0.0301\n",
            "7/50, train_loss: 0.0247\n",
            "8/50, train_loss: 0.0322\n",
            "9/50, train_loss: 0.0254\n",
            "10/50, train_loss: 0.0559\n",
            "11/50, train_loss: 0.0016\n",
            "12/50, train_loss: 0.0015\n",
            "13/50, train_loss: 0.0023\n",
            "14/50, train_loss: 0.0017\n",
            "15/50, train_loss: 0.0016\n",
            "16/50, train_loss: 0.0023\n",
            "17/50, train_loss: 0.0012\n",
            "18/50, train_loss: 0.0017\n",
            "19/50, train_loss: 0.0013\n",
            "20/50, train_loss: 0.0015\n",
            "21/50, train_loss: 0.0014\n",
            "22/50, train_loss: 0.0017\n",
            "23/50, train_loss: 0.0018\n",
            "24/50, train_loss: 0.0017\n",
            "25/50, train_loss: 0.0013\n",
            "26/50, train_loss: 0.0433\n",
            "27/50, train_loss: 0.0345\n",
            "28/50, train_loss: 0.0014\n",
            "29/50, train_loss: 0.0543\n",
            "30/50, train_loss: 0.0019\n",
            "31/50, train_loss: 0.0254\n",
            "32/50, train_loss: 0.0015\n",
            "33/50, train_loss: 0.0743\n",
            "34/50, train_loss: 0.0016\n",
            "35/50, train_loss: 0.0681\n",
            "36/50, train_loss: 0.0344\n",
            "37/50, train_loss: 0.0275\n",
            "38/50, train_loss: 0.0014\n",
            "39/50, train_loss: 0.0015\n",
            "40/50, train_loss: 0.0269\n",
            "41/50, train_loss: 0.0013\n",
            "42/50, train_loss: 0.0020\n",
            "43/50, train_loss: 0.0589\n",
            "44/50, train_loss: 0.0265\n",
            "45/50, train_loss: 0.0318\n",
            "46/50, train_loss: 0.0015\n",
            "47/50, train_loss: 0.0016\n",
            "48/50, train_loss: 0.0012\n",
            "49/50, train_loss: 0.0016\n",
            "50/50, train_loss: 0.0013\n",
            "epoch 134 average loss: 0.0157\n",
            "current epoch: 134 current accuracy: 0.7083 current AUC: 0.6891 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 135/400\n",
            "1/50, train_loss: 0.0392\n",
            "2/50, train_loss: 0.0014\n",
            "3/50, train_loss: 0.0014\n",
            "4/50, train_loss: 0.0271\n",
            "5/50, train_loss: 0.0014\n",
            "6/50, train_loss: 0.0325\n",
            "7/50, train_loss: 0.0017\n",
            "8/50, train_loss: 0.0015\n",
            "9/50, train_loss: 0.0014\n",
            "10/50, train_loss: 0.0017\n",
            "11/50, train_loss: 0.0256\n",
            "12/50, train_loss: 0.0325\n",
            "13/50, train_loss: 0.0014\n",
            "14/50, train_loss: 0.0016\n",
            "15/50, train_loss: 0.0608\n",
            "16/50, train_loss: 0.0015\n",
            "17/50, train_loss: 0.0015\n",
            "18/50, train_loss: 0.0013\n",
            "19/50, train_loss: 0.0017\n",
            "20/50, train_loss: 0.0017\n",
            "21/50, train_loss: 0.0590\n",
            "22/50, train_loss: 0.0015\n",
            "23/50, train_loss: 0.0012\n",
            "24/50, train_loss: 0.0290\n",
            "25/50, train_loss: 0.0016\n",
            "26/50, train_loss: 0.0015\n",
            "27/50, train_loss: 0.0013\n",
            "28/50, train_loss: 0.0266\n",
            "29/50, train_loss: 0.0299\n",
            "30/50, train_loss: 0.0344\n",
            "31/50, train_loss: 0.0015\n",
            "32/50, train_loss: 0.0337\n",
            "33/50, train_loss: 0.0622\n",
            "34/50, train_loss: 0.0229\n",
            "35/50, train_loss: 0.0014\n",
            "36/50, train_loss: 0.0012\n",
            "37/50, train_loss: 0.0504\n",
            "38/50, train_loss: 0.0262\n",
            "39/50, train_loss: 0.0015\n",
            "40/50, train_loss: 0.0012\n",
            "41/50, train_loss: 0.0714\n",
            "42/50, train_loss: 0.0588\n",
            "43/50, train_loss: 0.0504\n",
            "44/50, train_loss: 0.0013\n",
            "45/50, train_loss: 0.0277\n",
            "46/50, train_loss: 0.0264\n",
            "47/50, train_loss: 0.0213\n",
            "48/50, train_loss: 0.0010\n",
            "49/50, train_loss: 0.0013\n",
            "50/50, train_loss: 0.0307\n",
            "epoch 135 average loss: 0.0183\n",
            "----------\n",
            "epoch 136/400\n",
            "1/50, train_loss: 0.0017\n",
            "2/50, train_loss: 0.0016\n",
            "3/50, train_loss: 0.0317\n",
            "4/50, train_loss: 0.0017\n",
            "5/50, train_loss: 0.0012\n",
            "6/50, train_loss: 0.0269\n",
            "7/50, train_loss: 0.0383\n",
            "8/50, train_loss: 0.0017\n",
            "9/50, train_loss: 0.0459\n",
            "10/50, train_loss: 0.0012\n",
            "11/50, train_loss: 0.0015\n",
            "12/50, train_loss: 0.0637\n",
            "13/50, train_loss: 0.0544\n",
            "14/50, train_loss: 0.0289\n",
            "15/50, train_loss: 0.0016\n",
            "16/50, train_loss: 0.0268\n",
            "17/50, train_loss: 0.0016\n",
            "18/50, train_loss: 0.0416\n",
            "19/50, train_loss: 0.0014\n",
            "20/50, train_loss: 0.0322\n",
            "21/50, train_loss: 0.0013\n",
            "22/50, train_loss: 0.0016\n",
            "23/50, train_loss: 0.0249\n",
            "24/50, train_loss: 0.0018\n",
            "25/50, train_loss: 0.0015\n",
            "26/50, train_loss: 0.0009\n",
            "27/50, train_loss: 0.0259\n",
            "28/50, train_loss: 0.0018\n",
            "29/50, train_loss: 0.0018\n",
            "30/50, train_loss: 0.0240\n",
            "31/50, train_loss: 0.0026\n",
            "32/50, train_loss: 0.0010\n",
            "33/50, train_loss: 0.0551\n",
            "34/50, train_loss: 0.0016\n",
            "35/50, train_loss: 0.0239\n",
            "36/50, train_loss: 0.0472\n",
            "37/50, train_loss: 0.0299\n",
            "38/50, train_loss: 0.0014\n",
            "39/50, train_loss: 0.0017\n",
            "40/50, train_loss: 0.0014\n",
            "41/50, train_loss: 0.0255\n",
            "42/50, train_loss: 0.0451\n",
            "43/50, train_loss: 0.0262\n",
            "44/50, train_loss: 0.0013\n",
            "45/50, train_loss: 0.0863\n",
            "46/50, train_loss: 0.0250\n",
            "47/50, train_loss: 0.0242\n",
            "48/50, train_loss: 0.0017\n",
            "49/50, train_loss: 0.0013\n",
            "50/50, train_loss: 0.0013\n",
            "epoch 136 average loss: 0.0179\n",
            "current epoch: 136 current accuracy: 0.7500 current AUC: 0.7059 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 137/400\n",
            "1/50, train_loss: 0.0230\n",
            "2/50, train_loss: 0.0248\n",
            "3/50, train_loss: 0.0242\n",
            "4/50, train_loss: 0.0015\n",
            "5/50, train_loss: 0.0267\n",
            "6/50, train_loss: 0.0397\n",
            "7/50, train_loss: 0.0013\n",
            "8/50, train_loss: 0.0016\n",
            "9/50, train_loss: 0.0434\n",
            "10/50, train_loss: 0.0013\n",
            "11/50, train_loss: 0.0014\n",
            "12/50, train_loss: 0.0485\n",
            "13/50, train_loss: 0.0484\n",
            "14/50, train_loss: 0.0014\n",
            "15/50, train_loss: 0.0016\n",
            "16/50, train_loss: 0.0015\n",
            "17/50, train_loss: 0.0236\n",
            "18/50, train_loss: 0.0017\n",
            "19/50, train_loss: 0.0729\n",
            "20/50, train_loss: 0.0274\n",
            "21/50, train_loss: 0.0242\n",
            "22/50, train_loss: 0.0246\n",
            "23/50, train_loss: 0.0679\n",
            "24/50, train_loss: 0.0012\n",
            "25/50, train_loss: 0.0549\n",
            "26/50, train_loss: 0.0611\n",
            "27/50, train_loss: 0.0328\n",
            "28/50, train_loss: 0.0247\n",
            "29/50, train_loss: 0.0277\n",
            "30/50, train_loss: 0.0275\n",
            "31/50, train_loss: 0.0350\n",
            "32/50, train_loss: 0.0014\n",
            "33/50, train_loss: 0.0013\n",
            "34/50, train_loss: 0.0484\n",
            "35/50, train_loss: 0.0015\n",
            "36/50, train_loss: 0.0384\n",
            "37/50, train_loss: 0.0218\n",
            "38/50, train_loss: 0.0543\n",
            "39/50, train_loss: 0.0016\n",
            "40/50, train_loss: 0.0011\n",
            "41/50, train_loss: 0.0017\n",
            "42/50, train_loss: 0.0477\n",
            "43/50, train_loss: 0.0015\n",
            "44/50, train_loss: 0.0541\n",
            "45/50, train_loss: 0.0015\n",
            "46/50, train_loss: 0.0303\n",
            "47/50, train_loss: 0.0019\n",
            "48/50, train_loss: 0.0311\n",
            "49/50, train_loss: 0.0254\n",
            "50/50, train_loss: 0.0304\n",
            "epoch 137 average loss: 0.0239\n",
            "----------\n",
            "epoch 138/400\n",
            "1/50, train_loss: 0.0014\n",
            "2/50, train_loss: 0.0013\n",
            "3/50, train_loss: 0.0515\n",
            "4/50, train_loss: 0.0013\n",
            "5/50, train_loss: 0.0382\n",
            "6/50, train_loss: 0.0218\n",
            "7/50, train_loss: 0.0277\n",
            "8/50, train_loss: 0.0318\n",
            "9/50, train_loss: 0.0365\n",
            "10/50, train_loss: 0.0312\n",
            "11/50, train_loss: 0.0013\n",
            "12/50, train_loss: 0.0012\n",
            "13/50, train_loss: 0.0013\n",
            "14/50, train_loss: 0.0396\n",
            "15/50, train_loss: 0.0016\n",
            "16/50, train_loss: 0.0017\n",
            "17/50, train_loss: 0.0013\n",
            "18/50, train_loss: 0.0286\n",
            "19/50, train_loss: 0.0312\n",
            "20/50, train_loss: 0.0213\n",
            "21/50, train_loss: 0.0017\n",
            "22/50, train_loss: 0.0011\n",
            "23/50, train_loss: 0.0245\n",
            "24/50, train_loss: 0.0014\n",
            "25/50, train_loss: 0.0020\n",
            "26/50, train_loss: 0.0014\n",
            "27/50, train_loss: 0.0391\n",
            "28/50, train_loss: 0.0238\n",
            "29/50, train_loss: 0.0658\n",
            "30/50, train_loss: 0.0254\n",
            "31/50, train_loss: 0.0013\n",
            "32/50, train_loss: 0.0010\n",
            "33/50, train_loss: 0.0274\n",
            "34/50, train_loss: 0.0237\n",
            "35/50, train_loss: 0.0016\n",
            "36/50, train_loss: 0.0561\n",
            "37/50, train_loss: 0.0016\n",
            "38/50, train_loss: 0.0300\n",
            "39/50, train_loss: 0.0015\n",
            "40/50, train_loss: 0.0018\n",
            "41/50, train_loss: 0.0015\n",
            "42/50, train_loss: 0.0011\n",
            "43/50, train_loss: 0.0013\n",
            "44/50, train_loss: 0.0220\n",
            "45/50, train_loss: 0.0219\n",
            "46/50, train_loss: 0.0014\n",
            "47/50, train_loss: 0.0249\n",
            "48/50, train_loss: 0.0409\n",
            "49/50, train_loss: 0.0014\n",
            "50/50, train_loss: 0.1135\n",
            "epoch 138 average loss: 0.0187\n",
            "current epoch: 138 current accuracy: 0.7500 current AUC: 0.7311 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 139/400\n",
            "1/50, train_loss: 0.0447\n",
            "2/50, train_loss: 0.0015\n",
            "3/50, train_loss: 0.0431\n",
            "4/50, train_loss: 0.0635\n",
            "5/50, train_loss: 0.0569\n",
            "6/50, train_loss: 0.0417\n",
            "7/50, train_loss: 0.0222\n",
            "8/50, train_loss: 0.0016\n",
            "9/50, train_loss: 0.0309\n",
            "10/50, train_loss: 0.0267\n",
            "11/50, train_loss: 0.0010\n",
            "12/50, train_loss: 0.0364\n",
            "13/50, train_loss: 0.0275\n",
            "14/50, train_loss: 0.0209\n",
            "15/50, train_loss: 0.0014\n",
            "16/50, train_loss: 0.0323\n",
            "17/50, train_loss: 0.0299\n",
            "18/50, train_loss: 0.0014\n",
            "19/50, train_loss: 0.0015\n",
            "20/50, train_loss: 0.0218\n",
            "21/50, train_loss: 0.0014\n",
            "22/50, train_loss: 0.0230\n",
            "23/50, train_loss: 0.0362\n",
            "24/50, train_loss: 0.0394\n",
            "25/50, train_loss: 0.0014\n",
            "26/50, train_loss: 0.0270\n",
            "27/50, train_loss: 0.0222\n",
            "28/50, train_loss: 0.0458\n",
            "29/50, train_loss: 0.0224\n",
            "30/50, train_loss: 0.0469\n",
            "31/50, train_loss: 0.0010\n",
            "32/50, train_loss: 0.0514\n",
            "33/50, train_loss: 0.0017\n",
            "34/50, train_loss: 0.0424\n",
            "35/50, train_loss: 0.0016\n",
            "36/50, train_loss: 0.0305\n",
            "37/50, train_loss: 0.0020\n",
            "38/50, train_loss: 0.0706\n",
            "39/50, train_loss: 0.0015\n",
            "40/50, train_loss: 0.0014\n",
            "41/50, train_loss: 0.0222\n",
            "42/50, train_loss: 0.0017\n",
            "43/50, train_loss: 0.0286\n",
            "44/50, train_loss: 0.0269\n",
            "45/50, train_loss: 0.0258\n",
            "46/50, train_loss: 0.0015\n",
            "47/50, train_loss: 0.0327\n",
            "48/50, train_loss: 0.0015\n",
            "49/50, train_loss: 0.0015\n",
            "50/50, train_loss: 0.0018\n",
            "epoch 139 average loss: 0.0224\n",
            "----------\n",
            "epoch 140/400\n",
            "1/50, train_loss: 0.0022\n",
            "2/50, train_loss: 0.0008\n",
            "3/50, train_loss: 0.0016\n",
            "4/50, train_loss: 0.0016\n",
            "5/50, train_loss: 0.0015\n",
            "6/50, train_loss: 0.0308\n",
            "7/50, train_loss: 0.0022\n",
            "8/50, train_loss: 0.0012\n",
            "9/50, train_loss: 0.0286\n",
            "10/50, train_loss: 0.0436\n",
            "11/50, train_loss: 0.0494\n",
            "12/50, train_loss: 0.0421\n",
            "13/50, train_loss: 0.0228\n",
            "14/50, train_loss: 0.0017\n",
            "15/50, train_loss: 0.0397\n",
            "16/50, train_loss: 0.0016\n",
            "17/50, train_loss: 0.0019\n",
            "18/50, train_loss: 0.0015\n",
            "19/50, train_loss: 0.0015\n",
            "20/50, train_loss: 0.0015\n",
            "21/50, train_loss: 0.0261\n",
            "22/50, train_loss: 0.0012\n",
            "23/50, train_loss: 0.0235\n",
            "24/50, train_loss: 0.0015\n",
            "25/50, train_loss: 0.0505\n",
            "26/50, train_loss: 0.0834\n",
            "27/50, train_loss: 0.0213\n",
            "28/50, train_loss: 0.0015\n",
            "29/50, train_loss: 0.0013\n",
            "30/50, train_loss: 0.0234\n",
            "31/50, train_loss: 0.0016\n",
            "32/50, train_loss: 0.0278\n",
            "33/50, train_loss: 0.0268\n",
            "34/50, train_loss: 0.0018\n",
            "35/50, train_loss: 0.0356\n",
            "36/50, train_loss: 0.0654\n",
            "37/50, train_loss: 0.0011\n",
            "38/50, train_loss: 0.0246\n",
            "39/50, train_loss: 0.0013\n",
            "40/50, train_loss: 0.0017\n",
            "41/50, train_loss: 0.0297\n",
            "42/50, train_loss: 0.0017\n",
            "43/50, train_loss: 0.0020\n",
            "44/50, train_loss: 0.0018\n",
            "45/50, train_loss: 0.0020\n",
            "46/50, train_loss: 0.0239\n",
            "47/50, train_loss: 0.0347\n",
            "48/50, train_loss: 0.0029\n",
            "49/50, train_loss: 0.0471\n",
            "50/50, train_loss: 0.0345\n",
            "epoch 140 average loss: 0.0176\n",
            "current epoch: 140 current accuracy: 0.4583 current AUC: 0.6807 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 141/400\n",
            "1/50, train_loss: 0.0018\n",
            "2/50, train_loss: 0.0016\n",
            "3/50, train_loss: 0.0544\n",
            "4/50, train_loss: 0.0018\n",
            "5/50, train_loss: 0.0621\n",
            "6/50, train_loss: 0.0788\n",
            "7/50, train_loss: 0.0557\n",
            "8/50, train_loss: 0.0015\n",
            "9/50, train_loss: 0.0017\n",
            "10/50, train_loss: 0.0309\n",
            "11/50, train_loss: 0.0381\n",
            "12/50, train_loss: 0.0015\n",
            "13/50, train_loss: 0.0018\n",
            "14/50, train_loss: 0.0016\n",
            "15/50, train_loss: 0.0352\n",
            "16/50, train_loss: 0.0501\n",
            "17/50, train_loss: 0.0016\n",
            "18/50, train_loss: 0.0546\n",
            "19/50, train_loss: 0.0016\n",
            "20/50, train_loss: 0.0356\n",
            "21/50, train_loss: 0.0211\n",
            "22/50, train_loss: 0.0017\n",
            "23/50, train_loss: 0.0017\n",
            "24/50, train_loss: 0.0347\n",
            "25/50, train_loss: 0.0011\n",
            "26/50, train_loss: 0.0282\n",
            "27/50, train_loss: 0.0230\n",
            "28/50, train_loss: 0.0541\n",
            "29/50, train_loss: 0.0251\n",
            "30/50, train_loss: 0.0227\n",
            "31/50, train_loss: 0.0014\n",
            "32/50, train_loss: 0.0858\n",
            "33/50, train_loss: 0.0541\n",
            "34/50, train_loss: 0.0221\n",
            "35/50, train_loss: 0.0019\n",
            "36/50, train_loss: 0.0349\n",
            "37/50, train_loss: 0.0385\n",
            "38/50, train_loss: 0.0019\n",
            "39/50, train_loss: 0.0018\n",
            "40/50, train_loss: 0.0027\n",
            "41/50, train_loss: 0.0461\n",
            "42/50, train_loss: 0.0339\n",
            "43/50, train_loss: 0.0025\n",
            "44/50, train_loss: 0.1003\n",
            "45/50, train_loss: 0.0257\n",
            "46/50, train_loss: 0.0018\n",
            "47/50, train_loss: 0.0374\n",
            "48/50, train_loss: 0.0019\n",
            "49/50, train_loss: 0.0268\n",
            "50/50, train_loss: 0.0478\n",
            "epoch 141 average loss: 0.0259\n",
            "----------\n",
            "epoch 142/400\n",
            "1/50, train_loss: 0.0013\n",
            "2/50, train_loss: 0.0236\n",
            "3/50, train_loss: 0.0611\n",
            "4/50, train_loss: 0.0019\n",
            "5/50, train_loss: 0.0014\n",
            "6/50, train_loss: 0.0017\n",
            "7/50, train_loss: 0.0439\n",
            "8/50, train_loss: 0.0218\n",
            "9/50, train_loss: 0.0017\n",
            "10/50, train_loss: 0.0023\n",
            "11/50, train_loss: 0.0024\n",
            "12/50, train_loss: 0.0397\n",
            "13/50, train_loss: 0.0024\n",
            "14/50, train_loss: 0.0025\n",
            "15/50, train_loss: 0.0282\n",
            "16/50, train_loss: 0.0018\n",
            "17/50, train_loss: 0.0014\n",
            "18/50, train_loss: 0.0527\n",
            "19/50, train_loss: 0.0016\n",
            "20/50, train_loss: 0.0015\n",
            "21/50, train_loss: 0.0019\n",
            "22/50, train_loss: 0.0009\n",
            "23/50, train_loss: 0.0293\n",
            "24/50, train_loss: 0.0263\n",
            "25/50, train_loss: 0.0471\n",
            "26/50, train_loss: 0.0016\n",
            "27/50, train_loss: 0.0015\n",
            "28/50, train_loss: 0.0261\n",
            "29/50, train_loss: 0.0317\n",
            "30/50, train_loss: 0.0015\n",
            "31/50, train_loss: 0.0015\n",
            "32/50, train_loss: 0.0208\n",
            "33/50, train_loss: 0.0240\n",
            "34/50, train_loss: 0.0014\n",
            "35/50, train_loss: 0.0017\n",
            "36/50, train_loss: 0.0446\n",
            "37/50, train_loss: 0.0257\n",
            "38/50, train_loss: 0.0017\n",
            "39/50, train_loss: 0.0251\n",
            "40/50, train_loss: 0.0017\n",
            "41/50, train_loss: 0.0013\n",
            "42/50, train_loss: 0.0238\n",
            "43/50, train_loss: 0.0014\n",
            "44/50, train_loss: 0.0019\n",
            "45/50, train_loss: 0.0564\n",
            "46/50, train_loss: 0.0018\n",
            "47/50, train_loss: 0.0227\n",
            "48/50, train_loss: 0.0015\n",
            "49/50, train_loss: 0.0284\n",
            "50/50, train_loss: 0.0021\n",
            "epoch 142 average loss: 0.0150\n",
            "current epoch: 142 current accuracy: 0.7083 current AUC: 0.6807 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 143/400\n",
            "1/50, train_loss: 0.0018\n",
            "2/50, train_loss: 0.0016\n",
            "3/50, train_loss: 0.0017\n",
            "4/50, train_loss: 0.0017\n",
            "5/50, train_loss: 0.0016\n",
            "6/50, train_loss: 0.0015\n",
            "7/50, train_loss: 0.0784\n",
            "8/50, train_loss: 0.0018\n",
            "9/50, train_loss: 0.0019\n",
            "10/50, train_loss: 0.0014\n",
            "11/50, train_loss: 0.0619\n",
            "12/50, train_loss: 0.0016\n",
            "13/50, train_loss: 0.0015\n",
            "14/50, train_loss: 0.0018\n",
            "15/50, train_loss: 0.0273\n",
            "16/50, train_loss: 0.0013\n",
            "17/50, train_loss: 0.0016\n",
            "18/50, train_loss: 0.0014\n",
            "19/50, train_loss: 0.0219\n",
            "20/50, train_loss: 0.0192\n",
            "21/50, train_loss: 0.0015\n",
            "22/50, train_loss: 0.0196\n",
            "23/50, train_loss: 0.0015\n",
            "24/50, train_loss: 0.0018\n",
            "25/50, train_loss: 0.0015\n",
            "26/50, train_loss: 0.0309\n",
            "27/50, train_loss: 0.0304\n",
            "28/50, train_loss: 0.0017\n",
            "29/50, train_loss: 0.0415\n",
            "30/50, train_loss: 0.0023\n",
            "31/50, train_loss: 0.0216\n",
            "32/50, train_loss: 0.0307\n",
            "33/50, train_loss: 0.0225\n",
            "34/50, train_loss: 0.0019\n",
            "35/50, train_loss: 0.0293\n",
            "36/50, train_loss: 0.0018\n",
            "37/50, train_loss: 0.0963\n",
            "38/50, train_loss: 0.0787\n",
            "39/50, train_loss: 0.0465\n",
            "40/50, train_loss: 0.0270\n",
            "41/50, train_loss: 0.0216\n",
            "42/50, train_loss: 0.0015\n",
            "43/50, train_loss: 0.0016\n",
            "44/50, train_loss: 0.0362\n",
            "45/50, train_loss: 0.0018\n",
            "46/50, train_loss: 0.0018\n",
            "47/50, train_loss: 0.0018\n",
            "48/50, train_loss: 0.0016\n",
            "49/50, train_loss: 0.0253\n",
            "50/50, train_loss: 0.0442\n",
            "epoch 143 average loss: 0.0172\n",
            "----------\n",
            "epoch 144/400\n",
            "1/50, train_loss: 0.0019\n",
            "2/50, train_loss: 0.0014\n",
            "3/50, train_loss: 0.0289\n",
            "4/50, train_loss: 0.0286\n",
            "5/50, train_loss: 0.0015\n",
            "6/50, train_loss: 0.0412\n",
            "7/50, train_loss: 0.0016\n",
            "8/50, train_loss: 0.0019\n",
            "9/50, train_loss: 0.0021\n",
            "10/50, train_loss: 0.0021\n",
            "11/50, train_loss: 0.0017\n",
            "12/50, train_loss: 0.0225\n",
            "13/50, train_loss: 0.0014\n",
            "14/50, train_loss: 0.1030\n",
            "15/50, train_loss: 0.0610\n",
            "16/50, train_loss: 0.0273\n",
            "17/50, train_loss: 0.0014\n",
            "18/50, train_loss: 0.0016\n",
            "19/50, train_loss: 0.0020\n",
            "20/50, train_loss: 0.0013\n",
            "21/50, train_loss: 0.0499\n",
            "22/50, train_loss: 0.0382\n",
            "23/50, train_loss: 0.0017\n",
            "24/50, train_loss: 0.0017\n",
            "25/50, train_loss: 0.0019\n",
            "26/50, train_loss: 0.0016\n",
            "27/50, train_loss: 0.0022\n",
            "28/50, train_loss: 0.0305\n",
            "29/50, train_loss: 0.0342\n",
            "30/50, train_loss: 0.0017\n",
            "31/50, train_loss: 0.0338\n",
            "32/50, train_loss: 0.0014\n",
            "33/50, train_loss: 0.0241\n",
            "34/50, train_loss: 0.0211\n",
            "35/50, train_loss: 0.0362\n",
            "36/50, train_loss: 0.0572\n",
            "37/50, train_loss: 0.0017\n",
            "38/50, train_loss: 0.0255\n",
            "39/50, train_loss: 0.0205\n",
            "40/50, train_loss: 0.0215\n",
            "41/50, train_loss: 0.0017\n",
            "42/50, train_loss: 0.0269\n",
            "43/50, train_loss: 0.0320\n",
            "44/50, train_loss: 0.0277\n",
            "45/50, train_loss: 0.0017\n",
            "46/50, train_loss: 0.0260\n",
            "47/50, train_loss: 0.0019\n",
            "48/50, train_loss: 0.0220\n",
            "49/50, train_loss: 0.0018\n",
            "50/50, train_loss: 0.0377\n",
            "epoch 144 average loss: 0.0184\n",
            "current epoch: 144 current accuracy: 0.7083 current AUC: 0.7059 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 145/400\n",
            "1/50, train_loss: 0.0019\n",
            "2/50, train_loss: 0.0021\n",
            "3/50, train_loss: 0.0017\n",
            "4/50, train_loss: 0.0017\n",
            "5/50, train_loss: 0.0233\n",
            "6/50, train_loss: 0.0016\n",
            "7/50, train_loss: 0.0486\n",
            "8/50, train_loss: 0.0219\n",
            "9/50, train_loss: 0.0540\n",
            "10/50, train_loss: 0.0207\n",
            "11/50, train_loss: 0.0300\n",
            "12/50, train_loss: 0.0190\n",
            "13/50, train_loss: 0.0017\n",
            "14/50, train_loss: 0.0016\n",
            "15/50, train_loss: 0.0338\n",
            "16/50, train_loss: 0.0020\n",
            "17/50, train_loss: 0.0229\n",
            "18/50, train_loss: 0.0233\n",
            "19/50, train_loss: 0.0017\n",
            "20/50, train_loss: 0.0205\n",
            "21/50, train_loss: 0.0229\n",
            "22/50, train_loss: 0.0216\n",
            "23/50, train_loss: 0.0316\n",
            "24/50, train_loss: 0.0386\n",
            "25/50, train_loss: 0.0016\n",
            "26/50, train_loss: 0.0197\n",
            "27/50, train_loss: 0.0018\n",
            "28/50, train_loss: 0.0018\n",
            "29/50, train_loss: 0.0017\n",
            "30/50, train_loss: 0.0019\n",
            "31/50, train_loss: 0.0271\n",
            "32/50, train_loss: 0.0431\n",
            "33/50, train_loss: 0.0257\n",
            "34/50, train_loss: 0.0017\n",
            "35/50, train_loss: 0.0208\n",
            "36/50, train_loss: 0.0481\n",
            "37/50, train_loss: 0.0337\n",
            "38/50, train_loss: 0.0013\n",
            "39/50, train_loss: 0.0016\n",
            "40/50, train_loss: 0.0016\n",
            "41/50, train_loss: 0.0393\n",
            "42/50, train_loss: 0.0288\n",
            "43/50, train_loss: 0.0021\n",
            "44/50, train_loss: 0.0340\n",
            "45/50, train_loss: 0.0012\n",
            "46/50, train_loss: 0.0012\n",
            "47/50, train_loss: 0.0014\n",
            "48/50, train_loss: 0.0434\n",
            "49/50, train_loss: 0.0278\n",
            "50/50, train_loss: 0.0016\n",
            "epoch 145 average loss: 0.0173\n",
            "----------\n",
            "epoch 146/400\n",
            "1/50, train_loss: 0.0227\n",
            "2/50, train_loss: 0.0226\n",
            "3/50, train_loss: 0.0272\n",
            "4/50, train_loss: 0.0019\n",
            "5/50, train_loss: 0.0013\n",
            "6/50, train_loss: 0.0263\n",
            "7/50, train_loss: 0.0314\n",
            "8/50, train_loss: 0.0230\n",
            "9/50, train_loss: 0.0429\n",
            "10/50, train_loss: 0.0020\n",
            "11/50, train_loss: 0.0018\n",
            "12/50, train_loss: 0.0179\n",
            "13/50, train_loss: 0.0394\n",
            "14/50, train_loss: 0.0290\n",
            "15/50, train_loss: 0.0015\n",
            "16/50, train_loss: 0.0014\n",
            "17/50, train_loss: 0.0217\n",
            "18/50, train_loss: 0.0177\n",
            "19/50, train_loss: 0.0174\n",
            "20/50, train_loss: 0.0015\n",
            "21/50, train_loss: 0.0214\n",
            "22/50, train_loss: 0.0218\n",
            "23/50, train_loss: 0.0536\n",
            "24/50, train_loss: 0.0398\n",
            "25/50, train_loss: 0.0018\n",
            "26/50, train_loss: 0.0218\n",
            "27/50, train_loss: 0.0012\n",
            "28/50, train_loss: 0.0359\n",
            "29/50, train_loss: 0.0009\n",
            "30/50, train_loss: 0.0331\n",
            "31/50, train_loss: 0.0335\n",
            "32/50, train_loss: 0.0201\n",
            "33/50, train_loss: 0.0015\n",
            "34/50, train_loss: 0.0192\n",
            "35/50, train_loss: 0.0015\n",
            "36/50, train_loss: 0.0451\n",
            "37/50, train_loss: 0.0216\n",
            "38/50, train_loss: 0.0019\n",
            "39/50, train_loss: 0.0013\n",
            "40/50, train_loss: 0.0484\n",
            "41/50, train_loss: 0.0014\n",
            "42/50, train_loss: 0.0014\n",
            "43/50, train_loss: 0.0014\n",
            "44/50, train_loss: 0.0014\n",
            "45/50, train_loss: 0.0013\n",
            "46/50, train_loss: 0.0016\n",
            "47/50, train_loss: 0.0015\n",
            "48/50, train_loss: 0.0016\n",
            "49/50, train_loss: 0.0343\n",
            "50/50, train_loss: 0.0015\n",
            "epoch 146 average loss: 0.0165\n",
            "current epoch: 146 current accuracy: 0.7500 current AUC: 0.7059 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 147/400\n",
            "1/50, train_loss: 0.0194\n",
            "2/50, train_loss: 0.0019\n",
            "3/50, train_loss: 0.0425\n",
            "4/50, train_loss: 0.0018\n",
            "5/50, train_loss: 0.0202\n",
            "6/50, train_loss: 0.0016\n",
            "7/50, train_loss: 0.0014\n",
            "8/50, train_loss: 0.0016\n",
            "9/50, train_loss: 0.0230\n",
            "10/50, train_loss: 0.0296\n",
            "11/50, train_loss: 0.0018\n",
            "12/50, train_loss: 0.0202\n",
            "13/50, train_loss: 0.0017\n",
            "14/50, train_loss: 0.0014\n",
            "15/50, train_loss: 0.0287\n",
            "16/50, train_loss: 0.0215\n",
            "17/50, train_loss: 0.0017\n",
            "18/50, train_loss: 0.0015\n",
            "19/50, train_loss: 0.0203\n",
            "20/50, train_loss: 0.0011\n",
            "21/50, train_loss: 0.0017\n",
            "22/50, train_loss: 0.0289\n",
            "23/50, train_loss: 0.0015\n",
            "24/50, train_loss: 0.0194\n",
            "25/50, train_loss: 0.0252\n",
            "26/50, train_loss: 0.0478\n",
            "27/50, train_loss: 0.0241\n",
            "28/50, train_loss: 0.0428\n",
            "29/50, train_loss: 0.0369\n",
            "30/50, train_loss: 0.0011\n",
            "31/50, train_loss: 0.0014\n",
            "32/50, train_loss: 0.0016\n",
            "33/50, train_loss: 0.0185\n",
            "34/50, train_loss: 0.0201\n",
            "35/50, train_loss: 0.0014\n",
            "36/50, train_loss: 0.0342\n",
            "37/50, train_loss: 0.0015\n",
            "38/50, train_loss: 0.0011\n",
            "39/50, train_loss: 0.0014\n",
            "40/50, train_loss: 0.0013\n",
            "41/50, train_loss: 0.0552\n",
            "42/50, train_loss: 0.0013\n",
            "43/50, train_loss: 0.0014\n",
            "44/50, train_loss: 0.0015\n",
            "45/50, train_loss: 0.0019\n",
            "46/50, train_loss: 0.0013\n",
            "47/50, train_loss: 0.0334\n",
            "48/50, train_loss: 0.0014\n",
            "49/50, train_loss: 0.0014\n",
            "50/50, train_loss: 0.0014\n",
            "epoch 147 average loss: 0.0131\n",
            "----------\n",
            "epoch 148/400\n",
            "1/50, train_loss: 0.0326\n",
            "2/50, train_loss: 0.0226\n",
            "3/50, train_loss: 0.0013\n",
            "4/50, train_loss: 0.0173\n",
            "5/50, train_loss: 0.0015\n",
            "6/50, train_loss: 0.0016\n",
            "7/50, train_loss: 0.0017\n",
            "8/50, train_loss: 0.0197\n",
            "9/50, train_loss: 0.0013\n",
            "10/50, train_loss: 0.0014\n",
            "11/50, train_loss: 0.0192\n",
            "12/50, train_loss: 0.0015\n",
            "13/50, train_loss: 0.0014\n",
            "14/50, train_loss: 0.0210\n",
            "15/50, train_loss: 0.0014\n",
            "16/50, train_loss: 0.0015\n",
            "17/50, train_loss: 0.0016\n",
            "18/50, train_loss: 0.0184\n",
            "19/50, train_loss: 0.0298\n",
            "20/50, train_loss: 0.0018\n",
            "21/50, train_loss: 0.0306\n",
            "22/50, train_loss: 0.0388\n",
            "23/50, train_loss: 0.0399\n",
            "24/50, train_loss: 0.0382\n",
            "25/50, train_loss: 0.0274\n",
            "26/50, train_loss: 0.0013\n",
            "27/50, train_loss: 0.0177\n",
            "28/50, train_loss: 0.0014\n",
            "29/50, train_loss: 0.0017\n",
            "30/50, train_loss: 0.0373\n",
            "31/50, train_loss: 0.0555\n",
            "32/50, train_loss: 0.0417\n",
            "33/50, train_loss: 0.0014\n",
            "34/50, train_loss: 0.0197\n",
            "35/50, train_loss: 0.0247\n",
            "36/50, train_loss: 0.0016\n",
            "37/50, train_loss: 0.0239\n",
            "38/50, train_loss: 0.0747\n",
            "39/50, train_loss: 0.0013\n",
            "40/50, train_loss: 0.0017\n",
            "41/50, train_loss: 0.0015\n",
            "42/50, train_loss: 0.0351\n",
            "43/50, train_loss: 0.0014\n",
            "44/50, train_loss: 0.0240\n",
            "45/50, train_loss: 0.0250\n",
            "46/50, train_loss: 0.0232\n",
            "47/50, train_loss: 0.0013\n",
            "48/50, train_loss: 0.0012\n",
            "49/50, train_loss: 0.0014\n",
            "50/50, train_loss: 0.0015\n",
            "epoch 148 average loss: 0.0159\n",
            "current epoch: 148 current accuracy: 0.7083 current AUC: 0.7143 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 149/400\n",
            "1/50, train_loss: 0.0185\n",
            "2/50, train_loss: 0.0206\n",
            "3/50, train_loss: 0.0214\n",
            "4/50, train_loss: 0.0348\n",
            "5/50, train_loss: 0.0016\n",
            "6/50, train_loss: 0.0228\n",
            "7/50, train_loss: 0.0018\n",
            "8/50, train_loss: 0.0336\n",
            "9/50, train_loss: 0.0014\n",
            "10/50, train_loss: 0.0199\n",
            "11/50, train_loss: 0.0012\n",
            "12/50, train_loss: 0.0238\n",
            "13/50, train_loss: 0.0016\n",
            "14/50, train_loss: 0.0017\n",
            "15/50, train_loss: 0.0016\n",
            "16/50, train_loss: 0.0014\n",
            "17/50, train_loss: 0.0013\n",
            "18/50, train_loss: 0.0324\n",
            "19/50, train_loss: 0.0011\n",
            "20/50, train_loss: 0.0248\n",
            "21/50, train_loss: 0.0013\n",
            "22/50, train_loss: 0.0294\n",
            "23/50, train_loss: 0.0009\n",
            "24/50, train_loss: 0.0013\n",
            "25/50, train_loss: 0.0188\n",
            "26/50, train_loss: 0.0012\n",
            "27/50, train_loss: 0.0012\n",
            "28/50, train_loss: 0.0014\n",
            "29/50, train_loss: 0.0011\n",
            "30/50, train_loss: 0.0010\n",
            "31/50, train_loss: 0.0787\n",
            "32/50, train_loss: 0.0210\n",
            "33/50, train_loss: 0.0014\n",
            "34/50, train_loss: 0.0012\n",
            "35/50, train_loss: 0.0164\n",
            "36/50, train_loss: 0.0328\n",
            "37/50, train_loss: 0.0014\n",
            "38/50, train_loss: 0.0015\n",
            "39/50, train_loss: 0.0173\n",
            "40/50, train_loss: 0.0016\n",
            "41/50, train_loss: 0.0255\n",
            "42/50, train_loss: 0.0014\n",
            "43/50, train_loss: 0.0013\n",
            "44/50, train_loss: 0.0234\n",
            "45/50, train_loss: 0.0169\n",
            "46/50, train_loss: 0.0011\n",
            "47/50, train_loss: 0.0362\n",
            "48/50, train_loss: 0.0010\n",
            "49/50, train_loss: 0.0009\n",
            "50/50, train_loss: 0.0011\n",
            "epoch 149 average loss: 0.0121\n",
            "----------\n",
            "epoch 150/400\n",
            "1/50, train_loss: 0.0194\n",
            "2/50, train_loss: 0.0306\n",
            "3/50, train_loss: 0.0424\n",
            "4/50, train_loss: 0.0189\n",
            "5/50, train_loss: 0.0354\n",
            "6/50, train_loss: 0.0228\n",
            "7/50, train_loss: 0.0207\n",
            "8/50, train_loss: 0.0013\n",
            "9/50, train_loss: 0.0326\n",
            "10/50, train_loss: 0.0319\n",
            "11/50, train_loss: 0.0298\n",
            "12/50, train_loss: 0.0013\n",
            "13/50, train_loss: 0.0268\n",
            "14/50, train_loss: 0.0192\n",
            "15/50, train_loss: 0.0008\n",
            "16/50, train_loss: 0.0314\n",
            "17/50, train_loss: 0.0014\n",
            "18/50, train_loss: 0.0314\n",
            "19/50, train_loss: 0.0192\n",
            "20/50, train_loss: 0.0248\n",
            "21/50, train_loss: 0.0211\n",
            "22/50, train_loss: 0.0320\n",
            "23/50, train_loss: 0.0012\n",
            "24/50, train_loss: 0.0010\n",
            "25/50, train_loss: 0.1026\n",
            "26/50, train_loss: 0.0182\n",
            "27/50, train_loss: 0.0015\n",
            "28/50, train_loss: 0.0010\n",
            "29/50, train_loss: 0.0207\n",
            "30/50, train_loss: 0.0273\n",
            "31/50, train_loss: 0.0013\n",
            "32/50, train_loss: 0.0015\n",
            "33/50, train_loss: 0.0527\n",
            "34/50, train_loss: 0.0017\n",
            "35/50, train_loss: 0.0324\n",
            "36/50, train_loss: 0.0187\n",
            "37/50, train_loss: 0.0202\n",
            "38/50, train_loss: 0.0012\n",
            "39/50, train_loss: 0.0220\n",
            "40/50, train_loss: 0.0185\n",
            "41/50, train_loss: 0.0015\n",
            "42/50, train_loss: 0.0014\n",
            "43/50, train_loss: 0.0206\n",
            "44/50, train_loss: 0.0018\n",
            "45/50, train_loss: 0.0014\n",
            "46/50, train_loss: 0.0461\n",
            "47/50, train_loss: 0.0221\n",
            "48/50, train_loss: 0.0018\n",
            "49/50, train_loss: 0.0195\n",
            "50/50, train_loss: 0.0428\n",
            "epoch 150 average loss: 0.0200\n",
            "current epoch: 150 current accuracy: 0.7500 current AUC: 0.7143 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 151/400\n",
            "1/50, train_loss: 0.0248\n",
            "2/50, train_loss: 0.0011\n",
            "3/50, train_loss: 0.0178\n",
            "4/50, train_loss: 0.0015\n",
            "5/50, train_loss: 0.0382\n",
            "6/50, train_loss: 0.0221\n",
            "7/50, train_loss: 0.0240\n",
            "8/50, train_loss: 0.0175\n",
            "9/50, train_loss: 0.0014\n",
            "10/50, train_loss: 0.0581\n",
            "11/50, train_loss: 0.0017\n",
            "12/50, train_loss: 0.0015\n",
            "13/50, train_loss: 0.0254\n",
            "14/50, train_loss: 0.0210\n",
            "15/50, train_loss: 0.0016\n",
            "16/50, train_loss: 0.0013\n",
            "17/50, train_loss: 0.0330\n",
            "18/50, train_loss: 0.0011\n",
            "19/50, train_loss: 0.0170\n",
            "20/50, train_loss: 0.0183\n",
            "21/50, train_loss: 0.0419\n",
            "22/50, train_loss: 0.0015\n",
            "23/50, train_loss: 0.0313\n",
            "24/50, train_loss: 0.0313\n",
            "25/50, train_loss: 0.0013\n",
            "26/50, train_loss: 0.0011\n",
            "27/50, train_loss: 0.0014\n",
            "28/50, train_loss: 0.0015\n",
            "29/50, train_loss: 0.0015\n",
            "30/50, train_loss: 0.0011\n",
            "31/50, train_loss: 0.0015\n",
            "32/50, train_loss: 0.0013\n",
            "33/50, train_loss: 0.0298\n",
            "34/50, train_loss: 0.0011\n",
            "35/50, train_loss: 0.0012\n",
            "36/50, train_loss: 0.0171\n",
            "37/50, train_loss: 0.0015\n",
            "38/50, train_loss: 0.0015\n",
            "39/50, train_loss: 0.0016\n",
            "40/50, train_loss: 0.0014\n",
            "41/50, train_loss: 0.0012\n",
            "42/50, train_loss: 0.0190\n",
            "43/50, train_loss: 0.0256\n",
            "44/50, train_loss: 0.0012\n",
            "45/50, train_loss: 0.0327\n",
            "46/50, train_loss: 0.0014\n",
            "47/50, train_loss: 0.0012\n",
            "48/50, train_loss: 0.0232\n",
            "49/50, train_loss: 0.0012\n",
            "50/50, train_loss: 0.0011\n",
            "epoch 151 average loss: 0.0122\n",
            "----------\n",
            "epoch 152/400\n",
            "1/50, train_loss: 0.0312\n",
            "2/50, train_loss: 0.0008\n",
            "3/50, train_loss: 0.0013\n",
            "4/50, train_loss: 0.0012\n",
            "5/50, train_loss: 0.0013\n",
            "6/50, train_loss: 0.0015\n",
            "7/50, train_loss: 0.0011\n",
            "8/50, train_loss: 0.0184\n",
            "9/50, train_loss: 0.0303\n",
            "10/50, train_loss: 0.0011\n",
            "11/50, train_loss: 0.0011\n",
            "12/50, train_loss: 0.0014\n",
            "13/50, train_loss: 0.0182\n",
            "14/50, train_loss: 0.0203\n",
            "15/50, train_loss: 0.0340\n",
            "16/50, train_loss: 0.0013\n",
            "17/50, train_loss: 0.0244\n",
            "18/50, train_loss: 0.0012\n",
            "19/50, train_loss: 0.0172\n",
            "20/50, train_loss: 0.0011\n",
            "21/50, train_loss: 0.0184\n",
            "22/50, train_loss: 0.0350\n",
            "23/50, train_loss: 0.0493\n",
            "24/50, train_loss: 0.0015\n",
            "25/50, train_loss: 0.0014\n",
            "26/50, train_loss: 0.0188\n",
            "27/50, train_loss: 0.0334\n",
            "28/50, train_loss: 0.0251\n",
            "29/50, train_loss: 0.0012\n",
            "30/50, train_loss: 0.0181\n",
            "31/50, train_loss: 0.0012\n",
            "32/50, train_loss: 0.0014\n",
            "33/50, train_loss: 0.0344\n",
            "34/50, train_loss: 0.0229\n",
            "35/50, train_loss: 0.0193\n",
            "36/50, train_loss: 0.0012\n",
            "37/50, train_loss: 0.0261\n",
            "38/50, train_loss: 0.0225\n",
            "39/50, train_loss: 0.0015\n",
            "40/50, train_loss: 0.0015\n",
            "41/50, train_loss: 0.0011\n",
            "42/50, train_loss: 0.0164\n",
            "43/50, train_loss: 0.0302\n",
            "44/50, train_loss: 0.0012\n",
            "45/50, train_loss: 0.0013\n",
            "46/50, train_loss: 0.0297\n",
            "47/50, train_loss: 0.0011\n",
            "48/50, train_loss: 0.0010\n",
            "49/50, train_loss: 0.0012\n",
            "50/50, train_loss: 0.0013\n",
            "epoch 152 average loss: 0.0126\n",
            "current epoch: 152 current accuracy: 0.7500 current AUC: 0.7395 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 153/400\n",
            "1/50, train_loss: 0.0332\n",
            "2/50, train_loss: 0.0411\n",
            "3/50, train_loss: 0.0010\n",
            "4/50, train_loss: 0.0012\n",
            "5/50, train_loss: 0.0014\n",
            "6/50, train_loss: 0.0200\n",
            "7/50, train_loss: 0.0203\n",
            "8/50, train_loss: 0.0016\n",
            "9/50, train_loss: 0.0157\n",
            "10/50, train_loss: 0.0182\n",
            "11/50, train_loss: 0.0012\n",
            "12/50, train_loss: 0.0009\n",
            "13/50, train_loss: 0.0009\n",
            "14/50, train_loss: 0.0204\n",
            "15/50, train_loss: 0.0291\n",
            "16/50, train_loss: 0.0011\n",
            "17/50, train_loss: 0.0013\n",
            "18/50, train_loss: 0.0013\n",
            "19/50, train_loss: 0.0013\n",
            "20/50, train_loss: 0.0012\n",
            "21/50, train_loss: 0.0012\n",
            "22/50, train_loss: 0.0157\n",
            "23/50, train_loss: 0.0010\n",
            "24/50, train_loss: 0.0013\n",
            "25/50, train_loss: 0.0012\n",
            "26/50, train_loss: 0.0010\n",
            "27/50, train_loss: 0.0012\n",
            "28/50, train_loss: 0.0279\n",
            "29/50, train_loss: 0.0182\n",
            "30/50, train_loss: 0.0365\n",
            "31/50, train_loss: 0.0202\n",
            "32/50, train_loss: 0.0015\n",
            "33/50, train_loss: 0.0270\n",
            "34/50, train_loss: 0.0186\n",
            "35/50, train_loss: 0.0295\n",
            "36/50, train_loss: 0.0011\n",
            "37/50, train_loss: 0.0013\n",
            "38/50, train_loss: 0.0196\n",
            "39/50, train_loss: 0.0239\n",
            "40/50, train_loss: 0.0012\n",
            "41/50, train_loss: 0.0012\n",
            "42/50, train_loss: 0.0253\n",
            "43/50, train_loss: 0.0013\n",
            "44/50, train_loss: 0.0009\n",
            "45/50, train_loss: 0.0157\n",
            "46/50, train_loss: 0.0241\n",
            "47/50, train_loss: 0.0012\n",
            "48/50, train_loss: 0.0017\n",
            "49/50, train_loss: 0.0011\n",
            "50/50, train_loss: 0.0008\n",
            "epoch 153 average loss: 0.0107\n",
            "----------\n",
            "epoch 154/400\n",
            "1/50, train_loss: 0.0010\n",
            "2/50, train_loss: 0.0150\n",
            "3/50, train_loss: 0.0192\n",
            "4/50, train_loss: 0.0010\n",
            "5/50, train_loss: 0.0011\n",
            "6/50, train_loss: 0.0012\n",
            "7/50, train_loss: 0.0013\n",
            "8/50, train_loss: 0.0007\n",
            "9/50, train_loss: 0.0013\n",
            "10/50, train_loss: 0.0012\n",
            "11/50, train_loss: 0.0146\n",
            "12/50, train_loss: 0.0014\n",
            "13/50, train_loss: 0.0010\n",
            "14/50, train_loss: 0.0010\n",
            "15/50, train_loss: 0.0011\n",
            "16/50, train_loss: 0.0011\n",
            "17/50, train_loss: 0.0009\n",
            "18/50, train_loss: 0.0015\n",
            "19/50, train_loss: 0.0015\n",
            "20/50, train_loss: 0.0171\n",
            "21/50, train_loss: 0.0011\n",
            "22/50, train_loss: 0.0010\n",
            "23/50, train_loss: 0.0213\n",
            "24/50, train_loss: 0.0008\n",
            "25/50, train_loss: 0.0011\n",
            "26/50, train_loss: 0.0014\n",
            "27/50, train_loss: 0.0012\n",
            "28/50, train_loss: 0.0012\n",
            "29/50, train_loss: 0.0179\n",
            "30/50, train_loss: 0.0012\n",
            "31/50, train_loss: 0.0011\n",
            "32/50, train_loss: 0.0012\n",
            "33/50, train_loss: 0.0340\n",
            "34/50, train_loss: 0.0186\n",
            "35/50, train_loss: 0.0151\n",
            "36/50, train_loss: 0.0009\n",
            "37/50, train_loss: 0.0012\n",
            "38/50, train_loss: 0.0401\n",
            "39/50, train_loss: 0.0011\n",
            "40/50, train_loss: 0.0011\n",
            "41/50, train_loss: 0.0348\n",
            "42/50, train_loss: 0.0010\n",
            "43/50, train_loss: 0.0008\n",
            "44/50, train_loss: 0.0013\n",
            "45/50, train_loss: 0.0011\n",
            "46/50, train_loss: 0.0174\n",
            "47/50, train_loss: 0.0237\n",
            "48/50, train_loss: 0.0287\n",
            "49/50, train_loss: 0.0009\n",
            "50/50, train_loss: 0.0201\n",
            "epoch 154 average loss: 0.0075\n",
            "current epoch: 154 current accuracy: 0.7083 current AUC: 0.7395 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 155/400\n",
            "1/50, train_loss: 0.0158\n",
            "2/50, train_loss: 0.0010\n",
            "3/50, train_loss: 0.0010\n",
            "4/50, train_loss: 0.0301\n",
            "5/50, train_loss: 0.0010\n",
            "6/50, train_loss: 0.0157\n",
            "7/50, train_loss: 0.0010\n",
            "8/50, train_loss: 0.0010\n",
            "9/50, train_loss: 0.0009\n",
            "10/50, train_loss: 0.0279\n",
            "11/50, train_loss: 0.0012\n",
            "12/50, train_loss: 0.0322\n",
            "13/50, train_loss: 0.0212\n",
            "14/50, train_loss: 0.0162\n",
            "15/50, train_loss: 0.0012\n",
            "16/50, train_loss: 0.0169\n",
            "17/50, train_loss: 0.0011\n",
            "18/50, train_loss: 0.0152\n",
            "19/50, train_loss: 0.0269\n",
            "20/50, train_loss: 0.0385\n",
            "21/50, train_loss: 0.0011\n",
            "22/50, train_loss: 0.0008\n",
            "23/50, train_loss: 0.0011\n",
            "24/50, train_loss: 0.0185\n",
            "25/50, train_loss: 0.0011\n",
            "26/50, train_loss: 0.0011\n",
            "27/50, train_loss: 0.0009\n",
            "28/50, train_loss: 0.0011\n",
            "29/50, train_loss: 0.0011\n",
            "30/50, train_loss: 0.0304\n",
            "31/50, train_loss: 0.0232\n",
            "32/50, train_loss: 0.0176\n",
            "33/50, train_loss: 0.0209\n",
            "34/50, train_loss: 0.0152\n",
            "35/50, train_loss: 0.0012\n",
            "36/50, train_loss: 0.0012\n",
            "37/50, train_loss: 0.0309\n",
            "38/50, train_loss: 0.0008\n",
            "39/50, train_loss: 0.0163\n",
            "40/50, train_loss: 0.0012\n",
            "41/50, train_loss: 0.0010\n",
            "42/50, train_loss: 0.0009\n",
            "43/50, train_loss: 0.0009\n",
            "44/50, train_loss: 0.0007\n",
            "45/50, train_loss: 0.0011\n",
            "46/50, train_loss: 0.0010\n",
            "47/50, train_loss: 0.0349\n",
            "48/50, train_loss: 0.0206\n",
            "49/50, train_loss: 0.0258\n",
            "50/50, train_loss: 0.0150\n",
            "epoch 155 average loss: 0.0111\n",
            "----------\n",
            "epoch 156/400\n",
            "1/50, train_loss: 0.0351\n",
            "2/50, train_loss: 0.0228\n",
            "3/50, train_loss: 0.0009\n",
            "4/50, train_loss: 0.0158\n",
            "5/50, train_loss: 0.0169\n",
            "6/50, train_loss: 0.0199\n",
            "7/50, train_loss: 0.0009\n",
            "8/50, train_loss: 0.0008\n",
            "9/50, train_loss: 0.0408\n",
            "10/50, train_loss: 0.0009\n",
            "11/50, train_loss: 0.0011\n",
            "12/50, train_loss: 0.0197\n",
            "13/50, train_loss: 0.0171\n",
            "14/50, train_loss: 0.0010\n",
            "15/50, train_loss: 0.0011\n",
            "16/50, train_loss: 0.0190\n",
            "17/50, train_loss: 0.0314\n",
            "18/50, train_loss: 0.0010\n",
            "19/50, train_loss: 0.0156\n",
            "20/50, train_loss: 0.0189\n",
            "21/50, train_loss: 0.0007\n",
            "22/50, train_loss: 0.0009\n",
            "23/50, train_loss: 0.0010\n",
            "24/50, train_loss: 0.0010\n",
            "25/50, train_loss: 0.0010\n",
            "26/50, train_loss: 0.0011\n",
            "27/50, train_loss: 0.0183\n",
            "28/50, train_loss: 0.0010\n",
            "29/50, train_loss: 0.0014\n",
            "30/50, train_loss: 0.0009\n",
            "31/50, train_loss: 0.0155\n",
            "32/50, train_loss: 0.0417\n",
            "33/50, train_loss: 0.0008\n",
            "34/50, train_loss: 0.0270\n",
            "35/50, train_loss: 0.0007\n",
            "36/50, train_loss: 0.0007\n",
            "37/50, train_loss: 0.0010\n",
            "38/50, train_loss: 0.0010\n",
            "39/50, train_loss: 0.0146\n",
            "40/50, train_loss: 0.0010\n",
            "41/50, train_loss: 0.0161\n",
            "42/50, train_loss: 0.0164\n",
            "43/50, train_loss: 0.0154\n",
            "44/50, train_loss: 0.0008\n",
            "45/50, train_loss: 0.0011\n",
            "46/50, train_loss: 0.0262\n",
            "47/50, train_loss: 0.0361\n",
            "48/50, train_loss: 0.0147\n",
            "49/50, train_loss: 0.0270\n",
            "50/50, train_loss: 0.0352\n",
            "epoch 156 average loss: 0.0120\n",
            "current epoch: 156 current accuracy: 0.7500 current AUC: 0.6975 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 157/400\n",
            "1/50, train_loss: 0.0158\n",
            "2/50, train_loss: 0.0306\n",
            "3/50, train_loss: 0.0009\n",
            "4/50, train_loss: 0.0010\n",
            "5/50, train_loss: 0.0167\n",
            "6/50, train_loss: 0.0009\n",
            "7/50, train_loss: 0.0009\n",
            "8/50, train_loss: 0.0158\n",
            "9/50, train_loss: 0.0136\n",
            "10/50, train_loss: 0.0011\n",
            "11/50, train_loss: 0.0279\n",
            "12/50, train_loss: 0.0009\n",
            "13/50, train_loss: 0.0009\n",
            "14/50, train_loss: 0.0009\n",
            "15/50, train_loss: 0.0305\n",
            "16/50, train_loss: 0.0249\n",
            "17/50, train_loss: 0.0012\n",
            "18/50, train_loss: 0.0382\n",
            "19/50, train_loss: 0.0466\n",
            "20/50, train_loss: 0.0171\n",
            "21/50, train_loss: 0.0007\n",
            "22/50, train_loss: 0.0008\n",
            "23/50, train_loss: 0.0219\n",
            "24/50, train_loss: 0.0329\n",
            "25/50, train_loss: 0.0339\n",
            "26/50, train_loss: 0.0010\n",
            "27/50, train_loss: 0.0011\n",
            "28/50, train_loss: 0.0140\n",
            "29/50, train_loss: 0.0200\n",
            "30/50, train_loss: 0.0009\n",
            "31/50, train_loss: 0.0213\n",
            "32/50, train_loss: 0.0154\n",
            "33/50, train_loss: 0.0164\n",
            "34/50, train_loss: 0.0350\n",
            "35/50, train_loss: 0.0323\n",
            "36/50, train_loss: 0.0242\n",
            "37/50, train_loss: 0.0243\n",
            "38/50, train_loss: 0.0010\n",
            "39/50, train_loss: 0.0010\n",
            "40/50, train_loss: 0.0189\n",
            "41/50, train_loss: 0.0290\n",
            "42/50, train_loss: 0.0186\n",
            "43/50, train_loss: 0.0009\n",
            "44/50, train_loss: 0.0154\n",
            "45/50, train_loss: 0.0011\n",
            "46/50, train_loss: 0.0012\n",
            "47/50, train_loss: 0.0273\n",
            "48/50, train_loss: 0.0175\n",
            "49/50, train_loss: 0.0012\n",
            "50/50, train_loss: 0.0010\n",
            "epoch 157 average loss: 0.0143\n",
            "----------\n",
            "epoch 158/400\n",
            "1/50, train_loss: 0.0365\n",
            "2/50, train_loss: 0.0165\n",
            "3/50, train_loss: 0.0346\n",
            "4/50, train_loss: 0.0016\n",
            "5/50, train_loss: 0.0168\n",
            "6/50, train_loss: 0.0008\n",
            "7/50, train_loss: 0.0011\n",
            "8/50, train_loss: 0.0278\n",
            "9/50, train_loss: 0.0140\n",
            "10/50, train_loss: 0.0180\n",
            "11/50, train_loss: 0.0165\n",
            "12/50, train_loss: 0.0010\n",
            "13/50, train_loss: 0.0008\n",
            "14/50, train_loss: 0.0016\n",
            "15/50, train_loss: 0.0010\n",
            "16/50, train_loss: 0.0146\n",
            "17/50, train_loss: 0.0136\n",
            "18/50, train_loss: 0.0011\n",
            "19/50, train_loss: 0.0009\n",
            "20/50, train_loss: 0.0186\n",
            "21/50, train_loss: 0.0363\n",
            "22/50, train_loss: 0.0291\n",
            "23/50, train_loss: 0.0318\n",
            "24/50, train_loss: 0.0009\n",
            "25/50, train_loss: 0.0007\n",
            "26/50, train_loss: 0.0193\n",
            "27/50, train_loss: 0.0010\n",
            "28/50, train_loss: 0.0009\n",
            "29/50, train_loss: 0.0152\n",
            "30/50, train_loss: 0.0009\n",
            "31/50, train_loss: 0.0010\n",
            "32/50, train_loss: 0.0269\n",
            "33/50, train_loss: 0.0156\n",
            "34/50, train_loss: 0.0212\n",
            "35/50, train_loss: 0.0008\n",
            "36/50, train_loss: 0.0008\n",
            "37/50, train_loss: 0.0007\n",
            "38/50, train_loss: 0.0012\n",
            "39/50, train_loss: 0.0141\n",
            "40/50, train_loss: 0.0163\n",
            "41/50, train_loss: 0.0007\n",
            "42/50, train_loss: 0.0270\n",
            "43/50, train_loss: 0.0150\n",
            "44/50, train_loss: 0.0013\n",
            "45/50, train_loss: 0.0009\n",
            "46/50, train_loss: 0.0276\n",
            "47/50, train_loss: 0.0011\n",
            "48/50, train_loss: 0.0010\n",
            "49/50, train_loss: 0.0283\n",
            "50/50, train_loss: 0.0009\n",
            "epoch 158 average loss: 0.0115\n",
            "current epoch: 158 current accuracy: 0.7083 current AUC: 0.6975 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 159/400\n",
            "1/50, train_loss: 0.0132\n",
            "2/50, train_loss: 0.0009\n",
            "3/50, train_loss: 0.0010\n",
            "4/50, train_loss: 0.0287\n",
            "5/50, train_loss: 0.0009\n",
            "6/50, train_loss: 0.0153\n",
            "7/50, train_loss: 0.0012\n",
            "8/50, train_loss: 0.0281\n",
            "9/50, train_loss: 0.0009\n",
            "10/50, train_loss: 0.0009\n",
            "11/50, train_loss: 0.0188\n",
            "12/50, train_loss: 0.0009\n",
            "13/50, train_loss: 0.0013\n",
            "14/50, train_loss: 0.0010\n",
            "15/50, train_loss: 0.0010\n",
            "16/50, train_loss: 0.0177\n",
            "17/50, train_loss: 0.0012\n",
            "18/50, train_loss: 0.0292\n",
            "19/50, train_loss: 0.0009\n",
            "20/50, train_loss: 0.0009\n",
            "21/50, train_loss: 0.0271\n",
            "22/50, train_loss: 0.0142\n",
            "23/50, train_loss: 0.0007\n",
            "24/50, train_loss: 0.0006\n",
            "25/50, train_loss: 0.0015\n",
            "26/50, train_loss: 0.0009\n",
            "27/50, train_loss: 0.0009\n",
            "28/50, train_loss: 0.0208\n",
            "29/50, train_loss: 0.0009\n",
            "30/50, train_loss: 0.0153\n",
            "31/50, train_loss: 0.0194\n",
            "32/50, train_loss: 0.0137\n",
            "33/50, train_loss: 0.0009\n",
            "34/50, train_loss: 0.0011\n",
            "35/50, train_loss: 0.0148\n",
            "36/50, train_loss: 0.0007\n",
            "37/50, train_loss: 0.0007\n",
            "38/50, train_loss: 0.0009\n",
            "39/50, train_loss: 0.0171\n",
            "40/50, train_loss: 0.0007\n",
            "41/50, train_loss: 0.0381\n",
            "42/50, train_loss: 0.0006\n",
            "43/50, train_loss: 0.0008\n",
            "44/50, train_loss: 0.0141\n",
            "45/50, train_loss: 0.0008\n",
            "46/50, train_loss: 0.0238\n",
            "47/50, train_loss: 0.0007\n",
            "48/50, train_loss: 0.0008\n",
            "49/50, train_loss: 0.0183\n",
            "50/50, train_loss: 0.0006\n",
            "epoch 159 average loss: 0.0083\n",
            "----------\n",
            "epoch 160/400\n",
            "1/50, train_loss: 0.0008\n",
            "2/50, train_loss: 0.0183\n",
            "3/50, train_loss: 0.0011\n",
            "4/50, train_loss: 0.0009\n",
            "5/50, train_loss: 0.0187\n",
            "6/50, train_loss: 0.0165\n",
            "7/50, train_loss: 0.0009\n",
            "8/50, train_loss: 0.0007\n",
            "9/50, train_loss: 0.0257\n",
            "10/50, train_loss: 0.0010\n",
            "11/50, train_loss: 0.0174\n",
            "12/50, train_loss: 0.0195\n",
            "13/50, train_loss: 0.0268\n",
            "14/50, train_loss: 0.0154\n",
            "15/50, train_loss: 0.0258\n",
            "16/50, train_loss: 0.0007\n",
            "17/50, train_loss: 0.0009\n",
            "18/50, train_loss: 0.0273\n",
            "19/50, train_loss: 0.0234\n",
            "20/50, train_loss: 0.0009\n",
            "21/50, train_loss: 0.0274\n",
            "22/50, train_loss: 0.0244\n",
            "23/50, train_loss: 0.0148\n",
            "24/50, train_loss: 0.0008\n",
            "25/50, train_loss: 0.0009\n",
            "26/50, train_loss: 0.0009\n",
            "27/50, train_loss: 0.0286\n",
            "28/50, train_loss: 0.0009\n",
            "29/50, train_loss: 0.0837\n",
            "30/50, train_loss: 0.0293\n",
            "31/50, train_loss: 0.0137\n",
            "32/50, train_loss: 0.0008\n",
            "33/50, train_loss: 0.0406\n",
            "34/50, train_loss: 0.0135\n",
            "35/50, train_loss: 0.0008\n",
            "36/50, train_loss: 0.0147\n",
            "37/50, train_loss: 0.0008\n",
            "38/50, train_loss: 0.0007\n",
            "39/50, train_loss: 0.0007\n",
            "40/50, train_loss: 0.0136\n",
            "41/50, train_loss: 0.0008\n",
            "42/50, train_loss: 0.0011\n",
            "43/50, train_loss: 0.0143\n",
            "44/50, train_loss: 0.0136\n",
            "45/50, train_loss: 0.0165\n",
            "46/50, train_loss: 0.0007\n",
            "47/50, train_loss: 0.0176\n",
            "48/50, train_loss: 0.0008\n",
            "49/50, train_loss: 0.0255\n",
            "50/50, train_loss: 0.0006\n",
            "epoch 160 average loss: 0.0129\n",
            "current epoch: 160 current accuracy: 0.7083 current AUC: 0.6891 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 161/400\n",
            "1/50, train_loss: 0.0452\n",
            "2/50, train_loss: 0.0009\n",
            "3/50, train_loss: 0.0008\n",
            "4/50, train_loss: 0.0008\n",
            "5/50, train_loss: 0.0163\n",
            "6/50, train_loss: 0.0152\n",
            "7/50, train_loss: 0.0258\n",
            "8/50, train_loss: 0.0569\n",
            "9/50, train_loss: 0.0280\n",
            "10/50, train_loss: 0.0165\n",
            "11/50, train_loss: 0.0204\n",
            "12/50, train_loss: 0.0169\n",
            "13/50, train_loss: 0.0010\n",
            "14/50, train_loss: 0.0007\n",
            "15/50, train_loss: 0.0195\n",
            "16/50, train_loss: 0.0009\n",
            "17/50, train_loss: 0.0138\n",
            "18/50, train_loss: 0.0242\n",
            "19/50, train_loss: 0.0008\n",
            "20/50, train_loss: 0.0009\n",
            "21/50, train_loss: 0.0202\n",
            "22/50, train_loss: 0.0008\n",
            "23/50, train_loss: 0.0007\n",
            "24/50, train_loss: 0.0008\n",
            "25/50, train_loss: 0.0127\n",
            "26/50, train_loss: 0.0310\n",
            "27/50, train_loss: 0.0144\n",
            "28/50, train_loss: 0.0318\n",
            "29/50, train_loss: 0.0009\n",
            "30/50, train_loss: 0.0009\n",
            "31/50, train_loss: 0.0007\n",
            "32/50, train_loss: 0.0125\n",
            "33/50, train_loss: 0.0009\n",
            "34/50, train_loss: 0.0009\n",
            "35/50, train_loss: 0.0009\n",
            "36/50, train_loss: 0.0007\n",
            "37/50, train_loss: 0.0133\n",
            "38/50, train_loss: 0.0009\n",
            "39/50, train_loss: 0.0007\n",
            "40/50, train_loss: 0.0390\n",
            "41/50, train_loss: 0.0260\n",
            "42/50, train_loss: 0.0008\n",
            "43/50, train_loss: 0.0010\n",
            "44/50, train_loss: 0.0157\n",
            "45/50, train_loss: 0.0148\n",
            "46/50, train_loss: 0.0149\n",
            "47/50, train_loss: 0.0006\n",
            "48/50, train_loss: 0.0160\n",
            "49/50, train_loss: 0.0009\n",
            "50/50, train_loss: 0.0013\n",
            "epoch 161 average loss: 0.0116\n",
            "----------\n",
            "epoch 162/400\n",
            "1/50, train_loss: 0.0007\n",
            "2/50, train_loss: 0.0141\n",
            "3/50, train_loss: 0.0137\n",
            "4/50, train_loss: 0.0008\n",
            "5/50, train_loss: 0.0174\n",
            "6/50, train_loss: 0.0130\n",
            "7/50, train_loss: 0.0008\n",
            "8/50, train_loss: 0.0007\n",
            "9/50, train_loss: 0.0246\n",
            "10/50, train_loss: 0.0006\n",
            "11/50, train_loss: 0.0163\n",
            "12/50, train_loss: 0.0158\n",
            "13/50, train_loss: 0.0159\n",
            "14/50, train_loss: 0.0009\n",
            "15/50, train_loss: 0.0007\n",
            "16/50, train_loss: 0.0008\n",
            "17/50, train_loss: 0.0385\n",
            "18/50, train_loss: 0.0121\n",
            "19/50, train_loss: 0.0323\n",
            "20/50, train_loss: 0.0008\n",
            "21/50, train_loss: 0.0251\n",
            "22/50, train_loss: 0.0230\n",
            "23/50, train_loss: 0.0185\n",
            "24/50, train_loss: 0.0009\n",
            "25/50, train_loss: 0.0008\n",
            "26/50, train_loss: 0.0008\n",
            "27/50, train_loss: 0.0007\n",
            "28/50, train_loss: 0.0181\n",
            "29/50, train_loss: 0.0008\n",
            "30/50, train_loss: 0.0274\n",
            "31/50, train_loss: 0.0291\n",
            "32/50, train_loss: 0.0008\n",
            "33/50, train_loss: 0.0189\n",
            "34/50, train_loss: 0.0156\n",
            "35/50, train_loss: 0.0246\n",
            "36/50, train_loss: 0.0007\n",
            "37/50, train_loss: 0.0164\n",
            "38/50, train_loss: 0.0008\n",
            "39/50, train_loss: 0.0006\n",
            "40/50, train_loss: 0.0008\n",
            "41/50, train_loss: 0.0240\n",
            "42/50, train_loss: 0.0007\n",
            "43/50, train_loss: 0.0008\n",
            "44/50, train_loss: 0.0274\n",
            "45/50, train_loss: 0.0143\n",
            "46/50, train_loss: 0.0007\n",
            "47/50, train_loss: 0.0138\n",
            "48/50, train_loss: 0.0009\n",
            "49/50, train_loss: 0.0008\n",
            "50/50, train_loss: 0.0005\n",
            "epoch 162 average loss: 0.0106\n",
            "current epoch: 162 current accuracy: 0.7500 current AUC: 0.7227 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 163/400\n",
            "1/50, train_loss: 0.0231\n",
            "2/50, train_loss: 0.0008\n",
            "3/50, train_loss: 0.0294\n",
            "4/50, train_loss: 0.0007\n",
            "5/50, train_loss: 0.0008\n",
            "6/50, train_loss: 0.0241\n",
            "7/50, train_loss: 0.0008\n",
            "8/50, train_loss: 0.0008\n",
            "9/50, train_loss: 0.0006\n",
            "10/50, train_loss: 0.0144\n",
            "11/50, train_loss: 0.0007\n",
            "12/50, train_loss: 0.0008\n",
            "13/50, train_loss: 0.0008\n",
            "14/50, train_loss: 0.0011\n",
            "15/50, train_loss: 0.0006\n",
            "16/50, train_loss: 0.0176\n",
            "17/50, train_loss: 0.0241\n",
            "18/50, train_loss: 0.0009\n",
            "19/50, train_loss: 0.0012\n",
            "20/50, train_loss: 0.0008\n",
            "21/50, train_loss: 0.0006\n",
            "22/50, train_loss: 0.0156\n",
            "23/50, train_loss: 0.0006\n",
            "24/50, train_loss: 0.0008\n",
            "25/50, train_loss: 0.0197\n",
            "26/50, train_loss: 0.0181\n",
            "27/50, train_loss: 0.0137\n",
            "28/50, train_loss: 0.0008\n",
            "29/50, train_loss: 0.0008\n",
            "30/50, train_loss: 0.0163\n",
            "31/50, train_loss: 0.0007\n",
            "32/50, train_loss: 0.0187\n",
            "33/50, train_loss: 0.0228\n",
            "34/50, train_loss: 0.0147\n",
            "35/50, train_loss: 0.0007\n",
            "36/50, train_loss: 0.0008\n",
            "37/50, train_loss: 0.0008\n",
            "38/50, train_loss: 0.0006\n",
            "39/50, train_loss: 0.0007\n",
            "40/50, train_loss: 0.0176\n",
            "41/50, train_loss: 0.0145\n",
            "42/50, train_loss: 0.0007\n",
            "43/50, train_loss: 0.0121\n",
            "44/50, train_loss: 0.0006\n",
            "45/50, train_loss: 0.0009\n",
            "46/50, train_loss: 0.0008\n",
            "47/50, train_loss: 0.0008\n",
            "48/50, train_loss: 0.0171\n",
            "49/50, train_loss: 0.0007\n",
            "50/50, train_loss: 0.0256\n",
            "epoch 163 average loss: 0.0077\n",
            "----------\n",
            "epoch 164/400\n",
            "1/50, train_loss: 0.0007\n",
            "2/50, train_loss: 0.0007\n",
            "3/50, train_loss: 0.0159\n",
            "4/50, train_loss: 0.0007\n",
            "5/50, train_loss: 0.0205\n",
            "6/50, train_loss: 0.0276\n",
            "7/50, train_loss: 0.0169\n",
            "8/50, train_loss: 0.0217\n",
            "9/50, train_loss: 0.0151\n",
            "10/50, train_loss: 0.0350\n",
            "11/50, train_loss: 0.0168\n",
            "12/50, train_loss: 0.0010\n",
            "13/50, train_loss: 0.0597\n",
            "14/50, train_loss: 0.0010\n",
            "15/50, train_loss: 0.0280\n",
            "16/50, train_loss: 0.0138\n",
            "17/50, train_loss: 0.0009\n",
            "18/50, train_loss: 0.0125\n",
            "19/50, train_loss: 0.0006\n",
            "20/50, train_loss: 0.0132\n",
            "21/50, train_loss: 0.0008\n",
            "22/50, train_loss: 0.0253\n",
            "23/50, train_loss: 0.0007\n",
            "24/50, train_loss: 0.0007\n",
            "25/50, train_loss: 0.0255\n",
            "26/50, train_loss: 0.0005\n",
            "27/50, train_loss: 0.0007\n",
            "28/50, train_loss: 0.0128\n",
            "29/50, train_loss: 0.0183\n",
            "30/50, train_loss: 0.0135\n",
            "31/50, train_loss: 0.0158\n",
            "32/50, train_loss: 0.0269\n",
            "33/50, train_loss: 0.0006\n",
            "34/50, train_loss: 0.0249\n",
            "35/50, train_loss: 0.0231\n",
            "36/50, train_loss: 0.0008\n",
            "37/50, train_loss: 0.0401\n",
            "38/50, train_loss: 0.0007\n",
            "39/50, train_loss: 0.0006\n",
            "40/50, train_loss: 0.0247\n",
            "41/50, train_loss: 0.0183\n",
            "42/50, train_loss: 0.0007\n",
            "43/50, train_loss: 0.0008\n",
            "44/50, train_loss: 0.0008\n",
            "45/50, train_loss: 0.0223\n",
            "46/50, train_loss: 0.0126\n",
            "47/50, train_loss: 0.0150\n",
            "48/50, train_loss: 0.0142\n",
            "49/50, train_loss: 0.0009\n",
            "50/50, train_loss: 0.0009\n",
            "epoch 164 average loss: 0.0129\n",
            "current epoch: 164 current accuracy: 0.7500 current AUC: 0.7059 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 165/400\n",
            "1/50, train_loss: 0.0009\n",
            "2/50, train_loss: 0.0168\n",
            "3/50, train_loss: 0.0006\n",
            "4/50, train_loss: 0.0156\n",
            "5/50, train_loss: 0.0227\n",
            "6/50, train_loss: 0.0007\n",
            "7/50, train_loss: 0.0141\n",
            "8/50, train_loss: 0.0160\n",
            "9/50, train_loss: 0.0007\n",
            "10/50, train_loss: 0.0010\n",
            "11/50, train_loss: 0.0007\n",
            "12/50, train_loss: 0.0288\n",
            "13/50, train_loss: 0.0008\n",
            "14/50, train_loss: 0.0008\n",
            "15/50, train_loss: 0.0279\n",
            "16/50, train_loss: 0.0007\n",
            "17/50, train_loss: 0.0152\n",
            "18/50, train_loss: 0.0006\n",
            "19/50, train_loss: 0.0342\n",
            "20/50, train_loss: 0.0008\n",
            "21/50, train_loss: 0.0006\n",
            "22/50, train_loss: 0.0143\n",
            "23/50, train_loss: 0.0007\n",
            "24/50, train_loss: 0.0009\n",
            "25/50, train_loss: 0.0238\n",
            "26/50, train_loss: 0.0125\n",
            "27/50, train_loss: 0.0006\n",
            "28/50, train_loss: 0.0008\n",
            "29/50, train_loss: 0.0008\n",
            "30/50, train_loss: 0.0007\n",
            "31/50, train_loss: 0.0007\n",
            "32/50, train_loss: 0.0007\n",
            "33/50, train_loss: 0.0010\n",
            "34/50, train_loss: 0.0119\n",
            "35/50, train_loss: 0.0006\n",
            "36/50, train_loss: 0.0307\n",
            "37/50, train_loss: 0.0007\n",
            "38/50, train_loss: 0.0255\n",
            "39/50, train_loss: 0.0144\n",
            "40/50, train_loss: 0.0237\n",
            "41/50, train_loss: 0.0007\n",
            "42/50, train_loss: 0.0164\n",
            "43/50, train_loss: 0.0152\n",
            "44/50, train_loss: 0.0175\n",
            "45/50, train_loss: 0.0132\n",
            "46/50, train_loss: 0.0153\n",
            "47/50, train_loss: 0.0149\n",
            "48/50, train_loss: 0.0008\n",
            "49/50, train_loss: 0.0008\n",
            "50/50, train_loss: 0.0006\n",
            "epoch 165 average loss: 0.0092\n",
            "----------\n",
            "epoch 166/400\n",
            "1/50, train_loss: 0.0150\n",
            "2/50, train_loss: 0.0007\n",
            "3/50, train_loss: 0.0139\n",
            "4/50, train_loss: 0.0007\n",
            "5/50, train_loss: 0.0133\n",
            "6/50, train_loss: 0.0007\n",
            "7/50, train_loss: 0.0007\n",
            "8/50, train_loss: 0.0446\n",
            "9/50, train_loss: 0.0008\n",
            "10/50, train_loss: 0.0141\n",
            "11/50, train_loss: 0.0008\n",
            "12/50, train_loss: 0.0007\n",
            "13/50, train_loss: 0.0006\n",
            "14/50, train_loss: 0.0304\n",
            "15/50, train_loss: 0.0007\n",
            "16/50, train_loss: 0.0007\n",
            "17/50, train_loss: 0.0147\n",
            "18/50, train_loss: 0.0252\n",
            "19/50, train_loss: 0.0264\n",
            "20/50, train_loss: 0.0119\n",
            "21/50, train_loss: 0.0286\n",
            "22/50, train_loss: 0.0006\n",
            "23/50, train_loss: 0.0008\n",
            "24/50, train_loss: 0.0159\n",
            "25/50, train_loss: 0.0005\n",
            "26/50, train_loss: 0.0179\n",
            "27/50, train_loss: 0.0292\n",
            "28/50, train_loss: 0.0181\n",
            "29/50, train_loss: 0.0189\n",
            "30/50, train_loss: 0.0138\n",
            "31/50, train_loss: 0.0005\n",
            "32/50, train_loss: 0.0123\n",
            "33/50, train_loss: 0.0127\n",
            "34/50, train_loss: 0.0237\n",
            "35/50, train_loss: 0.0150\n",
            "36/50, train_loss: 0.0133\n",
            "37/50, train_loss: 0.0008\n",
            "38/50, train_loss: 0.0008\n",
            "39/50, train_loss: 0.0009\n",
            "40/50, train_loss: 0.0149\n",
            "41/50, train_loss: 0.0189\n",
            "42/50, train_loss: 0.0130\n",
            "43/50, train_loss: 0.0005\n",
            "44/50, train_loss: 0.0115\n",
            "45/50, train_loss: 0.0222\n",
            "46/50, train_loss: 0.0284\n",
            "47/50, train_loss: 0.0215\n",
            "48/50, train_loss: 0.0263\n",
            "49/50, train_loss: 0.0009\n",
            "50/50, train_loss: 0.0308\n",
            "epoch 166 average loss: 0.0126\n",
            "current epoch: 166 current accuracy: 0.7083 current AUC: 0.7143 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 167/400\n",
            "1/50, train_loss: 0.0523\n",
            "2/50, train_loss: 0.0158\n",
            "3/50, train_loss: 0.0008\n",
            "4/50, train_loss: 0.0159\n",
            "5/50, train_loss: 0.0240\n",
            "6/50, train_loss: 0.0009\n",
            "7/50, train_loss: 0.0185\n",
            "8/50, train_loss: 0.0297\n",
            "9/50, train_loss: 0.0254\n",
            "10/50, train_loss: 0.0008\n",
            "11/50, train_loss: 0.0147\n",
            "12/50, train_loss: 0.0194\n",
            "13/50, train_loss: 0.0122\n",
            "14/50, train_loss: 0.0247\n",
            "15/50, train_loss: 0.0006\n",
            "16/50, train_loss: 0.0006\n",
            "17/50, train_loss: 0.0007\n",
            "18/50, train_loss: 0.0007\n",
            "19/50, train_loss: 0.0004\n",
            "20/50, train_loss: 0.0007\n",
            "21/50, train_loss: 0.0246\n",
            "22/50, train_loss: 0.0005\n",
            "23/50, train_loss: 0.0274\n",
            "24/50, train_loss: 0.0007\n",
            "25/50, train_loss: 0.0007\n",
            "26/50, train_loss: 0.0006\n",
            "27/50, train_loss: 0.0006\n",
            "28/50, train_loss: 0.0145\n",
            "29/50, train_loss: 0.0115\n",
            "30/50, train_loss: 0.0139\n",
            "31/50, train_loss: 0.0007\n",
            "32/50, train_loss: 0.0239\n",
            "33/50, train_loss: 0.0138\n",
            "34/50, train_loss: 0.0154\n",
            "35/50, train_loss: 0.0006\n",
            "36/50, train_loss: 0.0123\n",
            "37/50, train_loss: 0.0225\n",
            "38/50, train_loss: 0.0007\n",
            "39/50, train_loss: 0.0157\n",
            "40/50, train_loss: 0.0007\n",
            "41/50, train_loss: 0.0007\n",
            "42/50, train_loss: 0.0127\n",
            "43/50, train_loss: 0.0007\n",
            "44/50, train_loss: 0.0283\n",
            "45/50, train_loss: 0.0129\n",
            "46/50, train_loss: 0.0175\n",
            "47/50, train_loss: 0.0006\n",
            "48/50, train_loss: 0.0122\n",
            "49/50, train_loss: 0.0227\n",
            "50/50, train_loss: 0.0134\n",
            "epoch 167 average loss: 0.0116\n",
            "----------\n",
            "epoch 168/400\n",
            "1/50, train_loss: 0.0007\n",
            "2/50, train_loss: 0.0278\n",
            "3/50, train_loss: 0.0387\n",
            "4/50, train_loss: 0.0007\n",
            "5/50, train_loss: 0.0144\n",
            "6/50, train_loss: 0.0446\n",
            "7/50, train_loss: 0.0006\n",
            "8/50, train_loss: 0.0006\n",
            "9/50, train_loss: 0.0225\n",
            "10/50, train_loss: 0.0007\n",
            "11/50, train_loss: 0.0006\n",
            "12/50, train_loss: 0.0006\n",
            "13/50, train_loss: 0.0134\n",
            "14/50, train_loss: 0.0007\n",
            "15/50, train_loss: 0.0006\n",
            "16/50, train_loss: 0.0007\n",
            "17/50, train_loss: 0.0009\n",
            "18/50, train_loss: 0.0005\n",
            "19/50, train_loss: 0.0379\n",
            "20/50, train_loss: 0.0200\n",
            "21/50, train_loss: 0.0139\n",
            "22/50, train_loss: 0.0007\n",
            "23/50, train_loss: 0.0008\n",
            "24/50, train_loss: 0.0226\n",
            "25/50, train_loss: 0.0005\n",
            "26/50, train_loss: 0.0208\n",
            "27/50, train_loss: 0.0127\n",
            "28/50, train_loss: 0.0006\n",
            "29/50, train_loss: 0.0006\n",
            "30/50, train_loss: 0.0152\n",
            "31/50, train_loss: 0.0173\n",
            "32/50, train_loss: 0.0120\n",
            "33/50, train_loss: 0.0121\n",
            "34/50, train_loss: 0.0138\n",
            "35/50, train_loss: 0.0009\n",
            "36/50, train_loss: 0.0118\n",
            "37/50, train_loss: 0.0156\n",
            "38/50, train_loss: 0.0007\n",
            "39/50, train_loss: 0.0159\n",
            "40/50, train_loss: 0.0122\n",
            "41/50, train_loss: 0.0007\n",
            "42/50, train_loss: 0.0007\n",
            "43/50, train_loss: 0.0012\n",
            "44/50, train_loss: 0.0325\n",
            "45/50, train_loss: 0.0007\n",
            "46/50, train_loss: 0.0336\n",
            "47/50, train_loss: 0.0007\n",
            "48/50, train_loss: 0.0311\n",
            "49/50, train_loss: 0.0123\n",
            "50/50, train_loss: 0.0006\n",
            "epoch 168 average loss: 0.0108\n",
            "current epoch: 168 current accuracy: 0.7083 current AUC: 0.6891 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 169/400\n",
            "1/50, train_loss: 0.0008\n",
            "2/50, train_loss: 0.0006\n",
            "3/50, train_loss: 0.0008\n",
            "4/50, train_loss: 0.0007\n",
            "5/50, train_loss: 0.0007\n",
            "6/50, train_loss: 0.0328\n",
            "7/50, train_loss: 0.0007\n",
            "8/50, train_loss: 0.0006\n",
            "9/50, train_loss: 0.0008\n",
            "10/50, train_loss: 0.0209\n",
            "11/50, train_loss: 0.0320\n",
            "12/50, train_loss: 0.0007\n",
            "13/50, train_loss: 0.0262\n",
            "14/50, train_loss: 0.0183\n",
            "15/50, train_loss: 0.0008\n",
            "16/50, train_loss: 0.0250\n",
            "17/50, train_loss: 0.0008\n",
            "18/50, train_loss: 0.0013\n",
            "19/50, train_loss: 0.0222\n",
            "20/50, train_loss: 0.0007\n",
            "21/50, train_loss: 0.0140\n",
            "22/50, train_loss: 0.0006\n",
            "23/50, train_loss: 0.0156\n",
            "24/50, train_loss: 0.0110\n",
            "25/50, train_loss: 0.0119\n",
            "26/50, train_loss: 0.0133\n",
            "27/50, train_loss: 0.0282\n",
            "28/50, train_loss: 0.0116\n",
            "29/50, train_loss: 0.0149\n",
            "30/50, train_loss: 0.0007\n",
            "31/50, train_loss: 0.0007\n",
            "32/50, train_loss: 0.0304\n",
            "33/50, train_loss: 0.0276\n",
            "34/50, train_loss: 0.0006\n",
            "35/50, train_loss: 0.0230\n",
            "36/50, train_loss: 0.0124\n",
            "37/50, train_loss: 0.0132\n",
            "38/50, train_loss: 0.0007\n",
            "39/50, train_loss: 0.0007\n",
            "40/50, train_loss: 0.0176\n",
            "41/50, train_loss: 0.0007\n",
            "42/50, train_loss: 0.0151\n",
            "43/50, train_loss: 0.0135\n",
            "44/50, train_loss: 0.0195\n",
            "45/50, train_loss: 0.0008\n",
            "46/50, train_loss: 0.0170\n",
            "47/50, train_loss: 0.0007\n",
            "48/50, train_loss: 0.0226\n",
            "49/50, train_loss: 0.0449\n",
            "50/50, train_loss: 0.0009\n",
            "epoch 169 average loss: 0.0114\n",
            "----------\n",
            "epoch 170/400\n",
            "1/50, train_loss: 0.0277\n",
            "2/50, train_loss: 0.0146\n",
            "3/50, train_loss: 0.0138\n",
            "4/50, train_loss: 0.0006\n",
            "5/50, train_loss: 0.0007\n",
            "6/50, train_loss: 0.0131\n",
            "7/50, train_loss: 0.0328\n",
            "8/50, train_loss: 0.0007\n",
            "9/50, train_loss: 0.0014\n",
            "10/50, train_loss: 0.0113\n",
            "11/50, train_loss: 0.0006\n",
            "12/50, train_loss: 0.0006\n",
            "13/50, train_loss: 0.0007\n",
            "14/50, train_loss: 0.0008\n",
            "15/50, train_loss: 0.0121\n",
            "16/50, train_loss: 0.0007\n",
            "17/50, train_loss: 0.0120\n",
            "18/50, train_loss: 0.0006\n",
            "19/50, train_loss: 0.0119\n",
            "20/50, train_loss: 0.0007\n",
            "21/50, train_loss: 0.0007\n",
            "22/50, train_loss: 0.0010\n",
            "23/50, train_loss: 0.0006\n",
            "24/50, train_loss: 0.0005\n",
            "25/50, train_loss: 0.0185\n",
            "26/50, train_loss: 0.0252\n",
            "27/50, train_loss: 0.0255\n",
            "28/50, train_loss: 0.0005\n",
            "29/50, train_loss: 0.0135\n",
            "30/50, train_loss: 0.0230\n",
            "31/50, train_loss: 0.0010\n",
            "32/50, train_loss: 0.0007\n",
            "33/50, train_loss: 0.0167\n",
            "34/50, train_loss: 0.0006\n",
            "35/50, train_loss: 0.0007\n",
            "36/50, train_loss: 0.0145\n",
            "37/50, train_loss: 0.0118\n",
            "38/50, train_loss: 0.0009\n",
            "39/50, train_loss: 0.0007\n",
            "40/50, train_loss: 0.0198\n",
            "41/50, train_loss: 0.0005\n",
            "42/50, train_loss: 0.0006\n",
            "43/50, train_loss: 0.0007\n",
            "44/50, train_loss: 0.0005\n",
            "45/50, train_loss: 0.0009\n",
            "46/50, train_loss: 0.0209\n",
            "47/50, train_loss: 0.0136\n",
            "48/50, train_loss: 0.0006\n",
            "49/50, train_loss: 0.0009\n",
            "50/50, train_loss: 0.0115\n",
            "epoch 170 average loss: 0.0077\n",
            "current epoch: 170 current accuracy: 0.7500 current AUC: 0.6975 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 171/400\n",
            "1/50, train_loss: 0.0007\n",
            "2/50, train_loss: 0.0006\n",
            "3/50, train_loss: 0.0154\n",
            "4/50, train_loss: 0.0007\n",
            "5/50, train_loss: 0.0005\n",
            "6/50, train_loss: 0.0244\n",
            "7/50, train_loss: 0.0006\n",
            "8/50, train_loss: 0.0009\n",
            "9/50, train_loss: 0.0006\n",
            "10/50, train_loss: 0.0006\n",
            "11/50, train_loss: 0.0006\n",
            "12/50, train_loss: 0.0137\n",
            "13/50, train_loss: 0.0220\n",
            "14/50, train_loss: 0.0203\n",
            "15/50, train_loss: 0.0210\n",
            "16/50, train_loss: 0.0206\n",
            "17/50, train_loss: 0.0211\n",
            "18/50, train_loss: 0.0009\n",
            "19/50, train_loss: 0.0196\n",
            "20/50, train_loss: 0.0004\n",
            "21/50, train_loss: 0.0117\n",
            "22/50, train_loss: 0.0153\n",
            "23/50, train_loss: 0.0120\n",
            "24/50, train_loss: 0.0178\n",
            "25/50, train_loss: 0.0212\n",
            "26/50, train_loss: 0.0109\n",
            "27/50, train_loss: 0.0004\n",
            "28/50, train_loss: 0.0007\n",
            "29/50, train_loss: 0.0199\n",
            "30/50, train_loss: 0.0007\n",
            "31/50, train_loss: 0.0006\n",
            "32/50, train_loss: 0.0006\n",
            "33/50, train_loss: 0.0146\n",
            "34/50, train_loss: 0.0116\n",
            "35/50, train_loss: 0.0115\n",
            "36/50, train_loss: 0.0137\n",
            "37/50, train_loss: 0.0157\n",
            "38/50, train_loss: 0.0128\n",
            "39/50, train_loss: 0.0007\n",
            "40/50, train_loss: 0.0007\n",
            "41/50, train_loss: 0.0171\n",
            "42/50, train_loss: 0.0006\n",
            "43/50, train_loss: 0.0007\n",
            "44/50, train_loss: 0.0121\n",
            "45/50, train_loss: 0.0225\n",
            "46/50, train_loss: 0.0007\n",
            "47/50, train_loss: 0.0128\n",
            "48/50, train_loss: 0.0249\n",
            "49/50, train_loss: 0.0009\n",
            "50/50, train_loss: 0.0006\n",
            "epoch 171 average loss: 0.0094\n",
            "----------\n",
            "epoch 172/400\n",
            "1/50, train_loss: 0.0247\n",
            "2/50, train_loss: 0.0126\n",
            "3/50, train_loss: 0.0261\n",
            "4/50, train_loss: 0.0006\n",
            "5/50, train_loss: 0.0006\n",
            "6/50, train_loss: 0.0007\n",
            "7/50, train_loss: 0.0007\n",
            "8/50, train_loss: 0.0008\n",
            "9/50, train_loss: 0.0197\n",
            "10/50, train_loss: 0.0005\n",
            "11/50, train_loss: 0.0007\n",
            "12/50, train_loss: 0.0187\n",
            "13/50, train_loss: 0.0121\n",
            "14/50, train_loss: 0.0005\n",
            "15/50, train_loss: 0.0118\n",
            "16/50, train_loss: 0.0112\n",
            "17/50, train_loss: 0.0185\n",
            "18/50, train_loss: 0.0006\n",
            "19/50, train_loss: 0.0137\n",
            "20/50, train_loss: 0.0007\n",
            "21/50, train_loss: 0.0114\n",
            "22/50, train_loss: 0.0120\n",
            "23/50, train_loss: 0.0121\n",
            "24/50, train_loss: 0.0188\n",
            "25/50, train_loss: 0.0006\n",
            "26/50, train_loss: 0.0005\n",
            "27/50, train_loss: 0.0206\n",
            "28/50, train_loss: 0.0007\n",
            "29/50, train_loss: 0.0108\n",
            "30/50, train_loss: 0.0114\n",
            "31/50, train_loss: 0.0007\n",
            "32/50, train_loss: 0.0124\n",
            "33/50, train_loss: 0.0007\n",
            "34/50, train_loss: 0.0006\n",
            "35/50, train_loss: 0.0190\n",
            "36/50, train_loss: 0.0007\n",
            "37/50, train_loss: 0.0006\n",
            "38/50, train_loss: 0.0006\n",
            "39/50, train_loss: 0.0221\n",
            "40/50, train_loss: 0.0199\n",
            "41/50, train_loss: 0.0008\n",
            "42/50, train_loss: 0.0008\n",
            "43/50, train_loss: 0.0007\n",
            "44/50, train_loss: 0.0107\n",
            "45/50, train_loss: 0.0006\n",
            "46/50, train_loss: 0.0226\n",
            "47/50, train_loss: 0.0005\n",
            "48/50, train_loss: 0.0173\n",
            "49/50, train_loss: 0.0107\n",
            "50/50, train_loss: 0.0006\n",
            "epoch 172 average loss: 0.0083\n",
            "current epoch: 172 current accuracy: 0.7083 current AUC: 0.7059 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 173/400\n",
            "1/50, train_loss: 0.0005\n",
            "2/50, train_loss: 0.0146\n",
            "3/50, train_loss: 0.0248\n",
            "4/50, train_loss: 0.0122\n",
            "5/50, train_loss: 0.0202\n",
            "6/50, train_loss: 0.0183\n",
            "7/50, train_loss: 0.0129\n",
            "8/50, train_loss: 0.0006\n",
            "9/50, train_loss: 0.0110\n",
            "10/50, train_loss: 0.0006\n",
            "11/50, train_loss: 0.0117\n",
            "12/50, train_loss: 0.0006\n",
            "13/50, train_loss: 0.0237\n",
            "14/50, train_loss: 0.0009\n",
            "15/50, train_loss: 0.0115\n",
            "16/50, train_loss: 0.0232\n",
            "17/50, train_loss: 0.0211\n",
            "18/50, train_loss: 0.0006\n",
            "19/50, train_loss: 0.0113\n",
            "20/50, train_loss: 0.0006\n",
            "21/50, train_loss: 0.0116\n",
            "22/50, train_loss: 0.0008\n",
            "23/50, train_loss: 0.0007\n",
            "24/50, train_loss: 0.0146\n",
            "25/50, train_loss: 0.0116\n",
            "26/50, train_loss: 0.0254\n",
            "27/50, train_loss: 0.0141\n",
            "28/50, train_loss: 0.0115\n",
            "29/50, train_loss: 0.0245\n",
            "30/50, train_loss: 0.0178\n",
            "31/50, train_loss: 0.0007\n",
            "32/50, train_loss: 0.0159\n",
            "33/50, train_loss: 0.0153\n",
            "34/50, train_loss: 0.0006\n",
            "35/50, train_loss: 0.0006\n",
            "36/50, train_loss: 0.0123\n",
            "37/50, train_loss: 0.0113\n",
            "38/50, train_loss: 0.0212\n",
            "39/50, train_loss: 0.0238\n",
            "40/50, train_loss: 0.0006\n",
            "41/50, train_loss: 0.0130\n",
            "42/50, train_loss: 0.0110\n",
            "43/50, train_loss: 0.0006\n",
            "44/50, train_loss: 0.0008\n",
            "45/50, train_loss: 0.0007\n",
            "46/50, train_loss: 0.0189\n",
            "47/50, train_loss: 0.0005\n",
            "48/50, train_loss: 0.0006\n",
            "49/50, train_loss: 0.0144\n",
            "50/50, train_loss: 0.0005\n",
            "epoch 173 average loss: 0.0103\n",
            "----------\n",
            "epoch 174/400\n",
            "1/50, train_loss: 0.0007\n",
            "2/50, train_loss: 0.0148\n",
            "3/50, train_loss: 0.0007\n",
            "4/50, train_loss: 0.0006\n",
            "5/50, train_loss: 0.0105\n",
            "6/50, train_loss: 0.0101\n",
            "7/50, train_loss: 0.0006\n",
            "8/50, train_loss: 0.0007\n",
            "9/50, train_loss: 0.0112\n",
            "10/50, train_loss: 0.0292\n",
            "11/50, train_loss: 0.0005\n",
            "12/50, train_loss: 0.0223\n",
            "13/50, train_loss: 0.0005\n",
            "14/50, train_loss: 0.0005\n",
            "15/50, train_loss: 0.0006\n",
            "16/50, train_loss: 0.0188\n",
            "17/50, train_loss: 0.0006\n",
            "18/50, train_loss: 0.0198\n",
            "19/50, train_loss: 0.0006\n",
            "20/50, train_loss: 0.0006\n",
            "21/50, train_loss: 0.0137\n",
            "22/50, train_loss: 0.0267\n",
            "23/50, train_loss: 0.0113\n",
            "24/50, train_loss: 0.0250\n",
            "25/50, train_loss: 0.0004\n",
            "26/50, train_loss: 0.0005\n",
            "27/50, train_loss: 0.0096\n",
            "28/50, train_loss: 0.0143\n",
            "29/50, train_loss: 0.0004\n",
            "30/50, train_loss: 0.0110\n",
            "31/50, train_loss: 0.0215\n",
            "32/50, train_loss: 0.0140\n",
            "33/50, train_loss: 0.0222\n",
            "34/50, train_loss: 0.0160\n",
            "35/50, train_loss: 0.0203\n",
            "36/50, train_loss: 0.0104\n",
            "37/50, train_loss: 0.0192\n",
            "38/50, train_loss: 0.0006\n",
            "39/50, train_loss: 0.0005\n",
            "40/50, train_loss: 0.0005\n",
            "41/50, train_loss: 0.0005\n",
            "42/50, train_loss: 0.0006\n",
            "43/50, train_loss: 0.0115\n",
            "44/50, train_loss: 0.0005\n",
            "45/50, train_loss: 0.0007\n",
            "46/50, train_loss: 0.0108\n",
            "47/50, train_loss: 0.0156\n",
            "48/50, train_loss: 0.0109\n",
            "49/50, train_loss: 0.0005\n",
            "50/50, train_loss: 0.0136\n",
            "epoch 174 average loss: 0.0089\n",
            "current epoch: 174 current accuracy: 0.7083 current AUC: 0.6723 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 175/400\n",
            "1/50, train_loss: 0.0005\n",
            "2/50, train_loss: 0.0108\n",
            "3/50, train_loss: 0.0004\n",
            "4/50, train_loss: 0.0206\n",
            "5/50, train_loss: 0.0004\n",
            "6/50, train_loss: 0.0006\n",
            "7/50, train_loss: 0.0113\n",
            "8/50, train_loss: 0.0237\n",
            "9/50, train_loss: 0.0005\n",
            "10/50, train_loss: 0.0007\n",
            "11/50, train_loss: 0.0186\n",
            "12/50, train_loss: 0.0205\n",
            "13/50, train_loss: 0.0005\n",
            "14/50, train_loss: 0.0006\n",
            "15/50, train_loss: 0.0124\n",
            "16/50, train_loss: 0.0167\n",
            "17/50, train_loss: 0.0183\n",
            "18/50, train_loss: 0.0111\n",
            "19/50, train_loss: 0.0006\n",
            "20/50, train_loss: 0.0005\n",
            "21/50, train_loss: 0.0005\n",
            "22/50, train_loss: 0.0123\n",
            "23/50, train_loss: 0.0107\n",
            "24/50, train_loss: 0.0119\n",
            "25/50, train_loss: 0.0005\n",
            "26/50, train_loss: 0.0189\n",
            "27/50, train_loss: 0.0006\n",
            "28/50, train_loss: 0.0006\n",
            "29/50, train_loss: 0.0104\n",
            "30/50, train_loss: 0.0128\n",
            "31/50, train_loss: 0.0123\n",
            "32/50, train_loss: 0.0175\n",
            "33/50, train_loss: 0.0172\n",
            "34/50, train_loss: 0.0098\n",
            "35/50, train_loss: 0.0241\n",
            "36/50, train_loss: 0.0108\n",
            "37/50, train_loss: 0.0129\n",
            "38/50, train_loss: 0.0198\n",
            "39/50, train_loss: 0.0005\n",
            "40/50, train_loss: 0.0162\n",
            "41/50, train_loss: 0.0006\n",
            "42/50, train_loss: 0.0006\n",
            "43/50, train_loss: 0.0005\n",
            "44/50, train_loss: 0.0006\n",
            "45/50, train_loss: 0.0109\n",
            "46/50, train_loss: 0.0119\n",
            "47/50, train_loss: 0.0132\n",
            "48/50, train_loss: 0.0188\n",
            "49/50, train_loss: 0.0008\n",
            "50/50, train_loss: 0.0003\n",
            "epoch 175 average loss: 0.0090\n",
            "----------\n",
            "epoch 176/400\n",
            "1/50, train_loss: 0.0124\n",
            "2/50, train_loss: 0.0100\n",
            "3/50, train_loss: 0.0006\n",
            "4/50, train_loss: 0.0132\n",
            "5/50, train_loss: 0.0004\n",
            "6/50, train_loss: 0.0116\n",
            "7/50, train_loss: 0.0112\n",
            "8/50, train_loss: 0.0137\n",
            "9/50, train_loss: 0.0175\n",
            "10/50, train_loss: 0.0110\n",
            "11/50, train_loss: 0.0108\n",
            "12/50, train_loss: 0.0005\n",
            "13/50, train_loss: 0.0005\n",
            "14/50, train_loss: 0.0005\n",
            "15/50, train_loss: 0.0116\n",
            "16/50, train_loss: 0.0004\n",
            "17/50, train_loss: 0.0126\n",
            "18/50, train_loss: 0.0330\n",
            "19/50, train_loss: 0.0006\n",
            "20/50, train_loss: 0.0118\n",
            "21/50, train_loss: 0.0007\n",
            "22/50, train_loss: 0.0224\n",
            "23/50, train_loss: 0.0109\n",
            "24/50, train_loss: 0.0204\n",
            "25/50, train_loss: 0.0006\n",
            "26/50, train_loss: 0.0095\n",
            "27/50, train_loss: 0.0005\n",
            "28/50, train_loss: 0.0180\n",
            "29/50, train_loss: 0.0112\n",
            "30/50, train_loss: 0.0098\n",
            "31/50, train_loss: 0.0005\n",
            "32/50, train_loss: 0.0005\n",
            "33/50, train_loss: 0.0406\n",
            "34/50, train_loss: 0.0106\n",
            "35/50, train_loss: 0.0217\n",
            "36/50, train_loss: 0.0006\n",
            "37/50, train_loss: 0.0006\n",
            "38/50, train_loss: 0.0005\n",
            "39/50, train_loss: 0.0116\n",
            "40/50, train_loss: 0.0007\n",
            "41/50, train_loss: 0.0232\n",
            "42/50, train_loss: 0.0159\n",
            "43/50, train_loss: 0.0109\n",
            "44/50, train_loss: 0.0241\n",
            "45/50, train_loss: 0.0198\n",
            "46/50, train_loss: 0.0005\n",
            "47/50, train_loss: 0.0004\n",
            "48/50, train_loss: 0.0217\n",
            "49/50, train_loss: 0.0130\n",
            "50/50, train_loss: 0.0005\n",
            "epoch 176 average loss: 0.0101\n",
            "current epoch: 176 current accuracy: 0.7500 current AUC: 0.6975 best accuracy: 0.7917 at epoch 60\n",
            "----------\n",
            "epoch 177/400\n",
            "1/50, train_loss: 0.0105\n",
            "2/50, train_loss: 0.0005\n",
            "3/50, train_loss: 0.0006\n",
            "4/50, train_loss: 0.0005\n",
            "5/50, train_loss: 0.0103\n",
            "6/50, train_loss: 0.0182\n",
            "7/50, train_loss: 0.0176\n",
            "8/50, train_loss: 0.0004\n",
            "9/50, train_loss: 0.0122\n",
            "10/50, train_loss: 0.0005\n",
            "11/50, train_loss: 0.0100\n",
            "12/50, train_loss: 0.0267\n",
            "13/50, train_loss: 0.0115\n",
            "14/50, train_loss: 0.0005\n",
            "15/50, train_loss: 0.0107\n",
            "16/50, train_loss: 0.0124\n",
            "17/50, train_loss: 0.0161\n",
            "18/50, train_loss: 0.0178\n",
            "19/50, train_loss: 0.0100\n",
            "20/50, train_loss: 0.0097\n",
            "21/50, train_loss: 0.0136\n",
            "22/50, train_loss: 0.0165\n",
            "23/50, train_loss: 0.0176\n",
            "24/50, train_loss: 0.0005\n",
            "25/50, train_loss: 0.0152\n",
            "26/50, train_loss: 0.0007\n",
            "27/50, train_loss: 0.0092\n",
            "28/50, train_loss: 0.0006\n",
            "29/50, train_loss: 0.0179\n",
            "30/50, train_loss: 0.0171\n",
            "31/50, train_loss: 0.0185\n",
            "32/50, train_loss: 0.0287\n",
            "33/50, train_loss: 0.0005\n",
            "34/50, train_loss: 0.0005\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZIsk3v3gKTrB"
      },
      "source": [
        "**Evaluation Metrics**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-LyP68KcHELg",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 464
        },
        "outputId": "e31cf99d-0808-4852-9386-5cba0a96009f"
      },
      "source": [
        "from enum import Enum\n",
        "from sklearn.metrics import (\n",
        "    classification_report, confusion_matrix,\n",
        "    ConfusionMatrixDisplay\n",
        ")\n",
        "class Diagnosis(Enum):\n",
        "    Liver_Cirrhosis = 0\n",
        "    Lymphoma = 1\n",
        "    \n",
        "\n",
        "%matplotlib inline\n",
        "# Load best model\n",
        "model.load_state_dict(torch.load(\"/content/drive/My Drive/Spleen_AI/best_metric_model_classification3d.pth\"))\n",
        "model.to(device)\n",
        "model.eval()\n",
        "\n",
        "with torch.no_grad():\n",
        "    y_pred = torch.tensor([], dtype=torch.float32, device=device)\n",
        "    y = torch.tensor([], dtype=torch.long, device=device)\n",
        "\n",
        "    for val_data in val_loader:\n",
        "        val_images, val_labels = val_data[\"image\"].to(\n",
        "            device), val_data[\"label\"].to(device),\n",
        "\n",
        "        outputs = model(val_images)\n",
        "        y_pred = torch.cat([y_pred, outputs.argmax(dim=1)], dim=0)\n",
        "        y = torch.cat([y, val_labels], dim=0)\n",
        "\n",
        "print(classification_report(\n",
        "    y.cpu().numpy(),\n",
        "    y_pred.cpu().numpy(),\n",
        "    target_names=[d.name for d in Diagnosis]))\n",
        "\n",
        "cm = confusion_matrix (\n",
        "    y.cpu().numpy(),\n",
        "    y_pred.cpu().numpy(),\n",
        "    normalize='true',\n",
        ")\n",
        "cmm = ConfusionMatrixMetric (include_background=True, metric_name=[\"sensitivity\",\"specificity\",\"negative predictive value\"])\n",
        "\n",
        "disp = ConfusionMatrixDisplay(\n",
        "    confusion_matrix=cm,\n",
        "    display_labels=[d.name for d in Diagnosis],\n",
        ")\n",
        "disp.plot(ax=plt.subplots(1, 1, facecolor='white')[1])"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "                 precision    recall  f1-score   support\n",
            "\n",
            "Liver_Cirrhosis       0.60      0.43      0.50         7\n",
            "       Lymphoma       0.79      0.88      0.83        17\n",
            "\n",
            "       accuracy                           0.75        24\n",
            "      macro avg       0.69      0.66      0.67        24\n",
            "   weighted avg       0.73      0.75      0.74        24\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x7f968dc96a10>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXgAAAEHCAYAAACk6V2yAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3df1xUVf4/8NeFARL8hYi/AH8gijIMggyIH39rReI6tiaCrqXyUdYiS+1jWX4XNS1p0yzFVi1b103FMlvdLbFM8UdRiIiCrMEmqKAm/gYRgZnz/YP15gg4wwjMD17Px+M+HnPmnnvum0Hfczj33HMlIYQAERHZHDtzB0BERI2DCZ6IyEYxwRMR2SgmeCIiG8UET0RkoxTmDoBqUji7wKFNO3OHQfWga6EzdwhUT61u3cWVK1dMPj58hAuuXtMaVbd95+FITk42+VymYoK3QA5t2qF7zDxzh0H1cNf/jrlDoHpqt3LnIx1/9ZoWaXu7GlU39Hemf5E8CiZ4IiITCAA6WPZfbkzwREQmEBCoFMYN0ZgLEzwRkYnYgyciskECAloLX+mFCZ6IyEQ6MMETEdkcAUDLBE9EZJvYgyciskECQCXH4ImIbI+A4BANEZFNEoDWsvM7EzwRkSmq72S1bEzwREQmkaCFZO4gHooJnojIBAKAjkM0RES2RwCosPBHajDBExGZSCc4RENEZHOq72RlgicisjkCErQcoiEisk0coiEiskECEiqEvbnDeCjL/vuCiMhCVd/oZGfUZozk5GT4+vrCx8cHCQkJNfafO3cOI0aMQFBQEAICAvD1118bbJMJnojIRNr/3uxkaDPYjlaLuLg47NmzBzk5Odi2bRtycnL06ixbtgwTJ07E8ePHkZSUhBdeeMFgu0zwREQmEEKCVtgZtRmSlpYGHx8feHt7w9HREdHR0di1a5deHUmScOvWLQDAzZs30aVLF4PtcgyeiMhEOiOnSRYXF0OtVsvl2NhYxMbGyuWioiJ4eXnJZU9PT/z00096bSxevBhPPvkk1qxZg9u3b2Pfvn0Gz8sET0Rkgup58MYNgri7uyM9Pf2Rzrdt2zZMmzYNr7zyClJTU/Hss88iOzsbdnZ1x8AET0RkAgEJlaJhUqiHhwfOnz8vlwsLC+Hh4aFXZ+PGjUhOTgYADBw4EOXl5bhy5Qo6dOhQZ7scgyciMpFWSEZthoSEhCAvLw/5+fmoqKhAUlISNBqNXp2uXbviu+++AwD8+9//Rnl5Odzd3R/aLnvwREQmaMg7WRUKBRITExEeHg6tVouYmBgolUrEx8dDrVZDo9Fg5cqVmDlzJlatWgVJkrBp0yZI0sO/PJjgiYhMpDNihkw1w+sKR0REICIiQu+9N998U37t5+eH77//vj7hMcETEZmiPhdZAW1jhlInJngiIhMIGDe+bk5M8EREJhAC9ZhFU9WosdSFCZ6IyCSS0Tc6mQsTPBGRCQRg1DIE5sQET0RkIj7wg4jIBglIfOAHEZEtEqjPRVbzsOzoiIgslnFrvZsTEzwRkQkE6nMnq3kwwRMRmYg9eCIiGySExB48EZGt4jx4IiIbVP3AD3tzh/FQTPBERCaovsjKMXgiIpvEO1mJiGwQ72QlIrJhOvbgiYhsjxCw+Ad+WPbXDxGRhRKQUKWzN2ozRnJyMnx9feHj44OEhIQa++fOnYvAwEAEBgaid+/eaNu2rcE22YMnIjJRQ93JqtVqERcXh2+//Raenp4ICQmBRqOBn5+fXGfVqlXy6zVr1uD48eMG22WCpwYxuOs5LBh6BPaSwBc5ffHxsf611nui5y94P+IbTNz+DE5d7gBVx1+xeMRBAIAkAWt/UuO7M95NGXqz1eLELbT/eyEkncCt4W64oemkt7/Vwatw23YBVa4OAICbT7ZHyYj2eOxUCdp/WiTXc7hYjl9f7I4yteEepS1pyGmSaWlp8PHxgbd39b/96Oho7Nq1Sy/B32/btm1YsmSJwXYbbYimZcuWNd5bt24dNm/e3FinBABs3rwZ/v7+UKlUCAoKwooVKwAA8fHx2Ldvn1FtdO/eHVeuXHmkOC5cuIAJEyY8UhvWwk7SYeHww5i1+3fQbIlGRO//oKfrtRr1nB0qMKVfFk5c6iC/l3e1HSZun4BnkiYidtcYLBpxEPaSrinDb550Au6bzuPiqz1x7s990TL1OhwK79SoVhrWFoXL+6BweR+UjGgPAChXtpLfu7DQB8LRDndUrZv6J7AA1UsVGLMVFxdDrVbL24YNG/RaKioqgpeXl1z29PREUVHRgycEAJw9exb5+fkYOXKkwQibtAc/a9asBmmnqqoKCkXN0Pfs2YP3338f33zzDbp06YK7d+/KXyhvvvlmrW1ptVrY29vrlRtCly5dsGPHjgZpy9KpOl7G+RttUHir+j/517k+GOFdgF+OtdOr91JYGjZmBGF6UKb8XnmVg/zaSaGFsPDFm2yF0y9lqOzohKoOTgCA0jBXuBy7iRueLerVjkvaDZT1aw3h1Dwv5xn7TFZ3d3ekp6c3yDmTkpIwYcIEvbxVlyb9rSxevBgrVqzA6dOnERoaKr9fUFAAlUoFADh27BiGDRuG4OBghIeH4+LFiwCA4cOHY86cOVCr1fjggw9qbX/58uVYsWIFunTpAgBwcnLCzJkzAQDTpk2TE2737t3x2muvoX///vj8889rlIHqMa7+/ftDpVLh9OnTAIBr167h6aefRkBAAMLCwnDy5EkAwMGDB+WLH0FBQSgpKUFBQQH8/f0BAKdOnUJoaCgCAwMREBCAvLy8Bv1cza2jy21cLHWRy7+WuqBjy9t6dfq6F6NTy1IcKuhW43hVx1+xa3IS/jFpO948MNTi1/ewBYprFahyc5TLVe0cobheWaOey9Eb8Fzwb3R8Px/2Vytq7G+Veh2lA10bNVZLJQRQqbM3ajPEw8MD58+fl8uFhYXw8PCotW5SUhImTZpkVIxm+Z/Up08fVFRUID8/HwCwfft2REVFobKyErNnz8aOHTtw7NgxxMTEYOHChfJxFRUVSE9PxyuvvFJru9nZ2QgODjYqBjc3N2RkZCA6OrrWcvv27ZGRkYHnn39eHuZZtGgRgoKCcPLkSbz99tt47rnnAAArVqzA2rVrkZmZicOHD6NFC/1e0Lp16/Dyyy8jMzMT6enp8PT0rBHPhg0b5D/fqspu19hvzSQIvDr4B/z5yP/Uuj/r144YtzUaUZ9NwEz1cTjaVzVxhFSb2/3b4Oz7ShQm9MUdVSt0XHdWb7/99Uo4ni9HWUBzHJ757UYnYzZDQkJCkJeXh/z8fFRUVCApKQkajaZGvdOnT+P69esYOHCgUTGaras0ceJEbN++HcBvCf7nn39GdnY2nnjiCQQGBmLZsmUoLCyUj4mKimqw8z/Y1oPl8ePHAwCCg4NRUFAAADhy5AieffZZAMDIkSNx9epV3Lp1C4MGDcK8efOwevVq3Lhxo8bw0cCBA/H222/jnXfewdmzZ2t8AQBAbGws0tPTkZ6eDoWzS439luzX2y7ofF+PvWPL2/j1vh69i2MFerldw6bxu/HN1E/Rr9OvSByzB8oOl/XaOXPdFWUVCvRyqzl+Tw2rqp0jFPf1yBXXKuSLqffoWikAh+oUcWuEGxzzy/T2t/zpOm6r2wCK5juspoNk1GaIQqFAYmIiwsPD0bdvX0ycOBFKpRLx8fHYvXu3XC8pKQnR0dGQJOM+c7PNoomKikJkZCTGjx8PSZLQq1cvZGVlQalUIjU1tdZjXFwenviUSiWOHTtm1MWHB9t6sOzkVD02aW9vj6qqh/coFyxYgDFjxuDrr7/GoEGDsHfvXjz22GPy/smTJ2PAgAH46quvEBERgfXr1xsVo7XI/rUDura9AY/Wt3C51AURvf+D+Xsfl/eXVjhh8MfT5fJff78LK74fiFOXO8Cj9S1cKmkJrbBD51Yl6OF6A0W3Wpnjx2hW7no7w+HSXSgu30VVOwe0/PE6fo3rrlfH/noltP9N+i7HbqKyy2N6+1v+cB1Xo7o0VcgWp6EXG4uIiEBERITeew9eO1y8eHG92jRbgu/Zsyfs7e2xdOlSuffs6+uL4uJipKamYuDAgaisrERubi6USqVRbb7++uuYP38+vvrqK3Tq1AkVFRXYvHkzZsyY0SAxDxkyBFu2bMGf/vQnpKSkoH379mjdujV++eUXqFQqqFQqHD16FKdPn0ZgYKB83JkzZ+Dt7Y2XXnoJ586dw8mTJ20qwWuFHd46OAQbNP+CnZ3Alzl98Mu1dnhxQBpOXXbHgfwedR7bv/NFzPjdcVTp7KATEpYeHIob5fW70EcmsJdwZZonOr/zS/U0yWFuqPRsAdcdF3G3hzPKgtugzd5iuGTchLAHdC4KXJ712/UTRfFdKK5VorxvzdlyzUmzfeBHWVmZ3ljzvHnzatSJiorC/Pnz5bF4R0dH7NixAy+99BJu3ryJqqoqzJkzx+gEHxERgV9//RWPP/44hBCQJAkxMTEN8wOh+tszJiYGAQEBcHZ2xt/+9jcAwPvvv48DBw7Azs4OSqUSo0ePli8OA8Bnn32Gv//973BwcECnTp3wxhtvNFhMluLw2W44fFb/AmriT6G11p3+5Tj59T9/9sU/f/Zt1NiodmWBbVAW2EbvvesTOsuvr0V3wbXo2nvoVe5OOJvo36jxWTwjx9fNSRJCCHMHQfpadPZC95iaX4hkue7615xDTpat3cqdjzR10bVPBwzfGGlU3fMv/9Rg0yTrg3eyEhGZgA/8aCRvvfWWPF/9nsjISL0plUREjY0JvhEsXLiQyZyIzIoP/CAismHGLlVgLkzwRESmEByiISKySQJAla6ZzoMnIrJlHIMnIrJhggmeiMg28SIrEZENErzISkRkqyRoeZGViMg2cQyeiMgGcS0aIiJbJarH4S0ZEzwRkYmMnUVjrn4+EzwRkQkEjB+DN1eCt+xLwEREFkuCVmfcZozk5GT4+vrCx8cHCQkJtdb57LPP4OfnB6VSicmTJxtskz14IiITNdQsGq1Wi7i4OHz77bfw9PRESEgINBoN/Pz85Dp5eXlYvnw5vv/+e7i6uuLy5csG22UPnojIBEJUJ3hjNkPS0tLg4+MDb29vODo6Ijo6Grt27dKr89FHHyEuLg6urq4AgA4dOhhslwmeiMhEuv8+eNvQZkhRURG8vLzksqenJ4qKivTq5ObmIjc3F4MGDUJYWBiSk5MNtsshGiIiExk7TbK4uBhqtVoux8bGIjY2tl7nqqqqQl5eHlJSUlBYWIihQ4ciKysLbdu2rfMYJngiIhMISNAZuVSBu7s70tPT69zv4eGB8+fPy+XCwkJ4eHjo1fH09MSAAQPg4OCAHj16oHfv3sjLy0NISEid7XKIhojIRMLIzZCQkBDk5eUhPz8fFRUVSEpKgkaj0avz9NNPIyUlBQBw5coV5Obmwtvb+6HtMsETEZmiAS+yKhQKJCYmIjw8HH379sXEiROhVCoRHx+P3bt3AwDCw8Ph5uYGPz8/jBgxAu+++y7c3Nwe2q4khKXfbNv8tOjshe4x88wdBtXDXf875g6B6qndyp0PHTYx5LGeHuj6ziyj6rZO2PVI5zIVx+CJiExktatJzp49G5JUd/CrV69ulICIiKyFpY9/1Jng75/SQ0RE+oQAhLU+8GPq1Kl65bKyMjg7Ozd6QERE1sLSe/AGv35SU1Ph5+eHPn36AABOnDiBF154odEDIyKyeA01T7KRGEzwc+bMwd69e+XpOP369cOhQ4caPTAiIstm3BRJc16INWoWzf1rJACAvb19owRDRGRVLHyIxmCC9/Lywg8//ABJklBZWYkPPvgAffv2bYrYiIgsl7D8aZIGh2jWrVuHtWvXoqioCF26dEFmZibWrl3bFLEREVk2IRm3mYnBHnz79u2xZcuWpoiFiMi6WPgQjcEe/JkzZzB27Fi4u7ujQ4cOGDduHM6cOdMUsRERWTZrn0UzefJkTJw4ERcvXsSFCxcQGRmJSZMmNUVsRESWS8Dih2gMJviysjI8++yzUCgUUCgUmDJlCsrLy5siNiIii1b92D7Dm7nUOQZ/7do1AMDo0aORkJCA6OhoSJKE7du3IyIioskCJCKyWDrLnkVTZ4IPDg6GJEm4t5rw+vXr5X2SJGH58uWNHx0RkQWTLPwia50JPj8/vynjICKyLma+gGoMo+5kzc7ORk5Ojt7Y+3PPPddoQRERWT7zXkA1hsEEv2TJEqSkpCAnJwcRERHYs2cPBg8ezARPRGThPXiDs2h27NiB7777Dp06dcJf//pXnDhxAjdv3myK2IiILJuFz4M32INv0aIF7OzsoFAocOvWLXTo0AHnz59vitiIiCyXgMXPojHYg1er1bhx4wZmzpyJ4OBg9O/fHwMHDmyK2IiILJokjNuMkZycDF9fX/j4+CAhIaHG/k2bNsHd3R2BgYEIDAzExx9/bLBNgz34Dz/8EAAwa9YsPPXUU7h16xYCAgKMi5iIyJY10PCLVqtFXFwcvv32W3h6eiIkJAQajQZ+fn569aKiopCYmGh0u3Um+IyMjDoPysjIQP/+/Y0+CRER1S0tLQ0+Pj7w9vYGAERHR2PXrl01Enx91ZngX3nllToPkiQJ+/fvf6QTU90cLt2G5/IfzB0G1cPeC5nmDoHqKXRlm0duw9jhl+LiYqjVarkcGxuL2NhYuVxUVKT3YCVPT0/89NNPNdr54osvcOjQIfTu3RurVq2q8TCmB9WZ4A8cOGBc5EREzZWR8+Dd3d2Rnp7+SKcaO3YsJk2aBCcnJ6xfvx5Tp0412NE2eJGViIhqIQDojNwM8PDw0JudWFhYCA8PD706bm5ucHJyAgDMmDEDx44dM9guEzwRkYkaahZNSEgI8vLykJ+fj4qKCiQlJUGj0ejVuXjxovx69+7dRj061ailCoiIqBYNNItGoVAgMTER4eHh0Gq1iImJgVKpRHx8PNRqNTQaDVavXo3du3dDoVCgXbt22LRpk8F2JSEevlqxEAJbtmzBmTNnEB8fj3PnzuHSpUsIDQ1tmJ+MamgttcMAaZS5w6B64EVW6xP6uzaPNC7u5OUFz5fnGlXXNWnrI4/Bm8LgEM0LL7yA1NRUbNu2DQDQqlUrxMXFNXpgRESWzNjhGXMuKWxwiOann35CRkYGgoKCAACurq6oqKho9MCIiCyehS9VYDDBOzg4QKvVQpKqf5Di4mLY2fHaLBGRpT/ww2Cmfumll/D73/8ely9fxsKFCzF48GC88cYbTREbEZFls/bVJP/whz8gODgY3333HYQQ+Mc//mHU9BwiIptm5vF1YxhM8OfOnYOzszPGjh2r917Xrl0bNTAiIotn7Ql+zJgx8sO3y8vLkZ+fD19fX5w6daop4iMislzWnuCzsrL0yhkZGfISwkREzZnVD9E8qH///rWuckZE1OxYe4J/77335Nc6nQ4ZGRno0qVLowZFRGTxbOEia0lJyW+VFQqMGTMGzzzzTKMGRURkFaw5wWu1WpSUlGDFihVNFQ8RkfWw1gRfVVUFhUKB77//vinjISKyChKseIgmNDQUGRkZCAwMhEajQWRkJFxcXOT948ePb5IAiYgskgAkIx7mYU4Gx+DLy8vh5uaG/fv3y/PhJUligicistYe/OXLl/Hee+/B399fTuz33Ft4jIioWbPWBK/ValFaWorangfCBE9EZMVj8J07d0Z8fHxTxkJEZF2sNcEbeJIfEVHzZgUXWetcD/67775ryjiIiKxPA64Hn5ycDF9fX/j4+CAhIaHOel988QUkSTLqGa91Jvh27doZFxURUTPVUM9k1Wq1iIuLw549e5CTk4Nt27YhJyenRr2SkhJ88MEHGDBggFHx8dl7RESmaqAefFpaGnx8fODt7Q1HR0dER0dj165dNer96U9/wmuvvYbHHnvMqPCY4ImITGFscjciwRcVFcHLy0sue3p6oqioSK9ORkYGzp8/jzFjxhgdYr2XCyYiov8uVWBk3eLiYqjVarkcGxuL2NhYo8+l0+kwb948bNq0qV4xMsETEZnI2Fk07u7uD70o6uHhgfPnz8vlwsJCeHh4yOWSkhJkZ2dj+PDhAIBLly5Bo9Fg9+7del8cD+IQDRGRqRpoiCYkJAR5eXnIz89HRUUFkpKSoNFo5P1t2rTBlStXUFBQgIKCAoSFhRlM7gATPBGR6RoowSsUCiQmJiI8PBx9+/bFxIkToVQqER8fj927d5scHodoiIhM0cBPdIqIiEBERITee2+++WatdVNSUoxqkwmeiMhUFn7DPxM8EZGJLH2pAiZ4IiITWe1qkkRE9BD1WGfGXJjgiYhMxQRPRGR7rPqh20REZAATPBGRDRKApLPsDM8ET0RkImOHaMz1NcAET0RkKsvuwDPBExGZij14IiJbxR48EZENauDFxhoDEzwRkQkkcC0aIiLbJSy7C88ET0RkIg7RULOgHn4Ls5ZegL2dwJ5t7fBZYke9/f4DSjHrzQvw7nsHbz/fDUe+agsA8FbewezlhXBppYVWKyFpdQcc3O1qjh+h2Tl6oBXW/ckDWp2E0ZOuImr2Zb39lwsd8O6crrh90x46nYSYNy4gdFQJqiqBVf/XFf/JagFtlYTHI68h+oFjmwUrWGzMKh7Z17JlyyY5z/Dhwx/6YFyqnZ2dQNzbRfh/f+iBmcN9MWLcDXTtVa5Xp7jIESvneOHAl/rJ++4dO7z7clfEjuiDhX/wxh+XXIBLa21Tht8sabXA2jc8sWzLGXyUchoHdrnibK6TXp2tH3TE0LE38OG3uXj9LwVIfN0LAHDon21ReVfC+v0/IzH5Z3z99/a4dN7RHD+G2Uk64zZzYQ+eHplvUBkuFDji0rnqBJGyqy0Ght/EubzH5Dq/FlYnAN0D/9iLzvyWVK796oCbVxRo41aF27fsGz/wZuzn487o0v0uOnerAAAMH3cdqXvboFvv33rikgSUlVT/Hm7fske7jpXy++VldtBWARXldlA46uDcsnl+KVv6RVar6MHfT6fToVevXiguLpbLPj4+KC4uxrRp0/D8888jLCwM3t7eSElJQUxMDPr27Ytp06bJbbRs2RJz586FUqnEqFGj5LYA4PPPP0doaCh69+6Nw4cPAwDKy8sxffp0qFQqBAUF4cCBAwCATZs24emnn8YTTzyB7t27IzExEe+99x6CgoIQFhaGa9euAQA++ugjhISEoF+/fnjmmWdQVlbWRJ9W03DrVIniC7/14K5cdED7zpX1bsc3sAwKR4GLBc2zN9iUrl5ygHuX335H7TtX4spFB706U165hP07XfGHYD/86VlvxL1VCAAY8rsbeMxZh0mB/pgS4ocJs4rR2rUZJniB6ousxmxmYnUJ3s7ODlOmTMGWLVsAAPv27UO/fv3g7u4OALh+/TpSU1OxatUqaDQazJ07F6dOnUJWVhYyMzMBALdv34ZarcapU6cwbNgwLFmyRG6/qqoKaWlpeP/99+X3165dC0mSkJWVhW3btmHq1KkoL68egsjOzsbOnTtx9OhRLFy4EM7Ozjh+/DgGDhyIzZs3AwDGjx+Po0eP4sSJE+jbty82btzYZJ+XtWjXoRLz15zDyrleEEIydzgEIOUfrnhi4jVsOZaDpX8/gz/P7gadDvj5uAvs7AW2Hs/G5p/+jS/WuePi2eb5pSwJ4zZjJCcnw9fXFz4+PkhISKixf926dVCpVAgMDMTgwYORk5NjsE2rS/AAEBMTIyfPTz75BNOnT5f3jR07FpIkQaVSoWPHjlCpVLCzs4NSqURBQQGA6i+JqKgoAMCUKVNw5MgR+fjx48cDAIKDg+X6R44cwZQpUwAAffr0Qbdu3ZCbmwsAGDFiBFq1agV3d3e0adMGY8eOBQCoVCr5+OzsbAwZMgQqlQpbtmzBqVOnavxMGzZsgFqthlqtRiXuNtAn1TSqe4MVcrm23uDDOLfU4s2/52NTQiecznBpjBDpAdV/df32O6rtr67kbe0wdOwNAICfugwVdyXcuqbAgS/bQj2iBAoHoG37KviF3EbuCecmjd9iCCM3A7RaLeLi4rBnzx7k5ORg27ZtNRL45MmT5Y7qq6++innz5hls1yoTvJeXFzp27Ij9+/cjLS0No0ePlvc5OVWP6drZ2cmv75WrqqpqbU+Sfusx3jvG3t6+zvr3e/Ac95//3vHTpk1DYmIisrKysGjRIrn3f7/Y2Fikp6cjPT0dDnCqsd+S/ZzpDI8eFejodRcKBx2Gj7uBH79pY9SxCgcd4jcW4LvPXeWZNdT4fAPLUJTvhEvnHFFZISFllyvCnrylV6eDRyUyj7QCAJzLc0LFXTu0cauCu0clMo9UT3woL7PD6QwXePnU/Ddt6+498KMhevBpaWnw8fGBt7c3HB0dER0djV27dunVad26tfz69u3benmrLlaZ4AFgxowZmDJlCiIjI2FvX78LcjqdDjt27AAAbN26FYMHD35o/SFDhshDQrm5uTh37hx8fX2NPl9JSQk6d+6MyspKuR1botNKWLvQA29vPYOPDv6MQ/9si7O5j+G5+ZcQ9uRNAEDvfmX4ND0HQ8fexMvvFGLDgdMAgKFjb0IVVoonoq7hw29/xoff/gxv5R1z/jjNgr0CiHurEG9M9sbMYX0wdOwNdPctx9/+3Ampe6sTSeyiIuzZ0g6zHvfF8ue74/9WnYMkAZrpV1BeZo+Zw30xe3RvPBl1Fd5+zS/BGz3+LgSKi4vlv9DVajU2bNig11RRURG8vLzksqenJ4qKimqccu3atejZsydeffVVrF692mCIVjGLpqysDJ6ennJ53rx5mD17NqZPn643PGMsFxcXpKWlYdmyZejQoQO2b9/+0PovvPACnn/+eahUKigUCmzatEmv527I0qVLMWDAALi7u2PAgAEoKSmpd8yW7uj+1ji6v7Xee5vf7SS/zj3hjClqvxrH7d/piv07Oe/dHEJHlSB01Gm996a+ekl+3a33Xaza/Z8ax7Vw0eH/bSho7PCsgrGzaNzd3RtkCnZcXBzi4uKwdetWLFu2DH/7298eHp8QFn6vbR3S09Mxd+5ceaZLfbRs2RKlpaWNEFXDaC21wwBplLnDoHrYeyHT3CFQPYX+rs0jJd1WbT3Rf8+ful4AAA9fSURBVMjLRtW9fXH7Q8+VmpqKxYsXY+/evQCA5cuXAwBef/31WuvrdDq4urri5s2bDz2vVQ7RJCQk4JlnnpE/BCKiJicA6IRxmwEhISHIy8tDfn4+KioqkJSUBI1Go1cnLy9Pfv3VV1+hV69eBtu1iiGaBy1YsAALFiww+XhL7r0TkRVpoPEPhUKBxMREhIeHQ6vVIiYmBkqlEvHx8VCr1dBoNEhMTMS+ffvg4OAAV1dXg8MzgBUP0dgyDtFYHw7RWJ9HHqJp44ng/3nJqLqlxZ+ZZRkUq+zBExFZBAvvHzPBExGZQlj+WjRM8EREJqi+0Yk9eCIi28QePBGRbWIPnojIFlnBE52Y4ImITCIgGXETkzkxwRMRmYpDNERENojTJImIbBh78ERENsqy8zsTPBGRqThNkojIFgkAWiZ4IiKbI0GwB09EZLOY4ImIbJSxCV5q3DDqwgRPRGQKAeMXG7NvzEDqxgRPRGQijsETEdkkAegs+1ZWO3MHQERklQSqx+CN2YyQnJwMX19f+Pj4ICEhocb+9957D35+fggICMCoUaNw9uxZg20ywRMRmUpn5GaAVqtFXFwc9uzZg5ycHGzbtg05OTl6dYKCgpCeno6TJ09iwoQJePXVVw22ywRPRGQiSQijNkPS0tLg4+MDb29vODo6Ijo6Grt27dKrM2LECDg7OwMAwsLCUFhYaLBdJngiIlMZOURTXFwMtVotbxs2bNBrpqioCF5eXnLZ09MTRUVFdZ5248aNGD16tMHweJGViMgUQgBa4y6yuru7Iz09vUFO++mnnyI9PR0HDx40WJcJnojIVA00TdLDwwPnz5+Xy4WFhfDw8KhRb9++fXjrrbdw8OBBODk5GWyXQzRERKZqoFk0ISEhyMvLQ35+PioqKpCUlASNRqNX5/jx4/jjH/+I3bt3o0OHDkaFxx48EZEpBIAGeiarQqFAYmIiwsPDodVqERMTA6VSifj4eKjVamg0GsyfPx+lpaWIjIwEAHTt2hW7d+9+eLsNEh0RUbMjANFwNzpFREQgIiJC770333xTfr1v3756t8kET0RkKi5VQERkgwSMnkVjLkzwRESmYg+eiMgWGb/OjLkwwRMRmULA4leTZIInIjIVe/BERDaKCZ6IyAYJAaHVmjuKh2KCJyIyVQPdydpYmOCJiEzFIRoiIhskLP+ZrEzwRESmYg+eiMgW8SIrEZFtasDlghsLEzwRkakacLngxsAET0RkAgFAsAdPRGSDRMM+8KMxMMETEZmIPXiqN0c3O1zvnm/uMBpFcXEx3N3dzR1Ggwv9XRtzh9AobPX3BQAFBQWPdPz/hIfiypVfjKrbvn37RzqXqSQhLHwiJ9kUtVqN9PR0c4dBRuLvy7rZmTsAIiJqHEzwREQ2igmemlRsbKy5Q6B64O/LunEMnojIRrEHT0Rko5jgiYhsFBM8kRVq2bJlk5xn+PDhnCZpxZjgbVxtiWDdunXYvHlzo5538+bN8Pf3h0qlQlBQEFasWAEAiI+Px759+4xqo3v37rhy5cojxXHhwgVMmDDhkdogslqCbJqLi0ujtV1ZWVnr+19//bUICgoSRUVFQgghysvLxYYNGx7aVlVVVY1yt27dRHFxccMEa2Pu/71qtVrh4+MjLl++LJd79uwpLl++LKZOnSpmzZolBgwYIHr06CEOHDggpk+fLvr06SOmTp2q196cOXOEn5+fGDlypNzWsGHDxKuvvipCQkJEr169xKFDh4QQQty5c0dMmzZN+Pv7i8DAQLF//34hhBB//etfxbhx48Tjjz8uunXrJtasWSNWrlwpAgMDxYABA8TVq1eFEEJs2LBBqNVqERAQIMaPHy9u377dFB9bs8MefDO0ePFirFixAqdPn0ZoaKj8fkFBAVQqFQDg2LFjGDZsGIKDgxEeHo6LFy8CqP6Tfc6cOVCr1fjggw9qbX/58uVYsWIFunTpAgBwcnLCzJkzAQDTpk3Djh07AFT30F977TX0798fn3/+eY0yAKxZswb9+/eHSqXC6dOnAQDXrl3D008/jYCAAISFheHkyZMAgIMHDyIwMBCBgYEICgpCSUkJCgoK4O/vDwA4deoUQkNDERgYiICAAOTl5TXo52oudnZ2mDJlCrZs2QIA2LdvH/r16ycvMXD9+nWkpqZi1apV0Gg0mDt3Lk6dOoWsrCxkZmYCAG7fvg21Wo1Tp05h2LBhWLJkidx+VVUV0tLS8P7778vvr127FpIkISsrC9u2bcPUqVNRXl4OAMjOzsbOnTtx9OhRLFy4EM7Ozjh+/DgGDhwo/+U4fvx4HD16FCdOnEDfvn2xcePGJvu8mhMm+GasT58+qKioQH5+9bo327dvR1RUFCorKzF79mzs2LEDx44dQ0xMDBYuXCgfV1FRgfT0dLzyyiu1tpudnY3g4GCjYnBzc0NGRgaio6NrLbdv3x4ZGRl4/vnn5WGeRYsWISgoCCdPnsTbb7+N5557DgCwYsUKrF27FpmZmTh8+DBatGihd65169bh5ZdfRmZmJtLT0+Hp6VmPT8uyxcTEyMnzk08+wfTp0+V9Y8eOhSRJUKlU6NixI1QqFezs7KBUKuX1WOzs7BAVFQUAmDJlCo4cOSIfP378eABAcHCwXP/IkSOYMmUKgOp/R926dUNubi4AYMSIEWjVqhXc3d3Rpk0bjB07FgCgUqnk47OzszFkyBCoVCps2bIFp06dapwPppljgm/mJk6ciO3btwP4LcH//PPPyM7OxhNPPIHAwEAsW7YMhYWF8jH3EkFDeLCtB8t1JZdnn30WADBy5EhcvXoVt27dwqBBgzBv3jysXr0aN27cgEKhv5bewIED8fbbb+Odd97B2bNna3wBWDMvLy907NgR+/fvR1paGkaPHi3vc3JyAlCdxO+9vleuqqqqtT1Jkmocb29vX2f9+z14jvvPf+/4adOmITExEVlZWVi0aJHc+6eGxQTfzEVFReGzzz5Dbm4uJElCr169IISAUqlEZmYmMjMzkZWVhW+++UY+xsXF5aFtKpVKHDt2zKjzP9jWg+X6JJcFCxbg448/xp07dzBo0CB5SOeeyZMnY/fu3WjRogUiIiKwf/9+o2K0FjNmzMCUKVMQGRkJe3v7eh2r0+nkobOtW7di8ODBD60/ZMgQeUgoNzcX586dg6+vr9HnKykpQefOnVFZWSm3Qw2PCb6Z69mzJ+zt7bF06VK59+zr64vi4mKkpqYCACorK+v1J/Trr7+O+fPn49KlSwCqh3Q+/vjjBov5/uSSkpKC9u3bo3Xr1vjll1+gUqnw2muvISQkpEaCP3PmDLy9vfHSSy9h3Lhx8ti9NSorK4Onp6e8vffee9BoNCgtLdUbnjGWi4sL0tLS4O/vj/379yM+Pv6h9V944QXodDqoVCpERUVh06ZNej13Q5YuXYoBAwZg0KBB6NOnT73jJSOZ+yovNS5JkoSHh4e8rVy5UixatEi8++67cp13331XABD5+fnye8ePHxdDhgwRAQEBws/PT54FM2zYMHH06FGD5/3kk0+EUqkUfn5+QqlUipUrVwohhJg6dar4/PPPhRCixiyZh5WPHj0qhg0bJoQQ4urVq2LcuHFCpVKJAQMGiBMnTgghhHjxxReFUqkUKpVKREdHi/LycpGfny+USqUQQojly5cLPz8/0a9fPxEeHi7P6LAVR48eFYMHDzbp2MacbUXmw7VoiGxAQkIC/vKXv2DLli0Gh1dq07JlS5SWljZCZGROTPBERDaKj+wjk7311lvyfPV7IiMj9aZUEpH5sAdPRGSjOIuGiMhGMcETEdkoJniySvb29ggMDIS/vz8iIyNRVlZmclv3r48zY8YM5OTk1Fk3JSUFP/zwQ73PUdfKmMasmFnfpYHvrTVExARPVqlFixbIzMxEdnY2HB0dsW7dOr39xtxSX5uPP/4Yfn5+de43NcETmQMTPFm9IUOG4D//+Q9SUlIwZMgQaDQa+Pn5QavVYv78+QgJCUFAQADWr18PABBC4MUXX4Svry8ef/xxXL58WW7r/gdcJCcno3///ujXrx9GjRqFgoICrFu3DqtWrUJgYCAOHz6M4uJiPPPMMwgJCUFISAi+//57AMDVq1fx5JNPQqlUYsaMGTBmLsPTTz+N4OBgKJVKbNiwQW/f3LlzoVQqMWrUKBQXFwMAfvnlFzz11FMIDg7GkCFDaty5S8Q7Wckq3bvzsrKyUmg0GvHhhx+KAwcOCGdnZ3HmzBkhhBDr168XS5cuFUJUr0kfHBwszpw5I7744gvx+OOPi6qqKlFUVCTatGkj3117707dy5cvC09PT7mte3e9PngX8KRJk8Thw4eFEEKcPXtW9OnTRwghxOzZs8WSJUuEEEL861//EgBqXdv+/rt1752jrKxMKJVKceXKFSGEEADEp59+KoQQYsmSJSIuLk4IIcTIkSNFbm6uEEKIH3/8UYwYMaLWGKn54jx4skp37txBYGAggOoe/P/+7//ihx9+QGhoKHr06AEA+Oabb3Dy5El5fP3mzZvIy8vDoUOHMGnSJNjb26NLly4YOXJkjfZ//PFHDB06VG6rXbt2tcaxb98+vTH7W7duobS0FIcOHcLOnTsBAGPGjIGrq6vBn2n16tX48ssvAQDnz59HXl4e3NzcaizlO378eJSWluKHH35AZGSkfPzdu3cNnoOaFyZ4skr3xuAfdP9qlEIIrFmzBuHh4Xp1vv766waLQ6fT4ccff8Rjjz32SO2kpKRg3759SE1NhbOzM4YPH17nErqSJEGn06Ft27a1fgZE93AMnmxWeHg4/vKXv6CyshJA9bK2t2/fxtChQ7F9+3ZotVpcvHgRBw4cqHFsWFgYDh06JD8M5dq1awCAVq1aoaSkRK735JNPYs2aNXL5XsIdOnQotm7dCgDYs2cPrl+//tBYb968CVdXVzg7O+P06dP48ccf5X21LeXbunVr9OjRQ76TWAiBEydO1O8DIpvHBE82a8aMGfDz80P//v3h7++PP/7xj6iqqsLvf/979OrVC35+fnjuuecwcODAGse6u7tjw4YNGD9+PPr16ycPkYwdOxZffvmlfJF19erVSE9PR0BAAPz8/OTZPIsWLcKhQ4egVCqxc+dOdO3a9aGxPvXUU6iqqkLfvn2xYMEChIWFyfvqWsp3y5Yt2LhxI/r16welUoldu3Y11EdHNoJLFRAR2Sj24ImIbBQTPBGRjWKCJyKyUUzwREQ2igmeiMhGMcETEdkoJngiIhv1/wFaHukI7ZTbDQAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Nm0LXsvlgp_N"
      },
      "source": [
        "files = [{\"img\": img} for img in images]\n",
        "post_transforms = Compose([\n",
        "        #Activationsd(keys=\"pred\", sigmoid=True),\n",
        "        #AsDiscreted(keys=\"pred\", threshold_values=True),\n",
        "        Invertd(\n",
        "            keys=\"y_pred\",  # invert the `pred` data field, also support multiple fields\n",
        "            transform=train_transforms,\n",
        "            loader=val_loader,\n",
        "            orig_keys=\"files\",  # get the previously applied pre_transforms information on the `img` data field,\n",
        "                              # then invert `pred` based on this information. we can use same info\n",
        "                              # for multiple fields, also support different orig_keys for different fields\n",
        "            #meta_keys=\"pred_meta_dict\",  # key field to save inverted meta data, every item maps to `keys`\n",
        "            #orig_meta_keys=\"img_meta_dict\",  # get the meta data from `img_meta_dict` field when inverting,\n",
        "                                             # for example, may need the `affine` to invert `Spacingd` transform,\n",
        "                                             # multiple fields can use the same meta data to invert\n",
        "            #meta_key_postfix=\"meta_dict\",  # if `meta_keys=None`, use \"{keys}_{meta_key_postfix}\" as the meta key,\n",
        "                                           # if `orig_meta_keys=None`, use \"{orig_keys}_{meta_key_postfix}\",\n",
        "                                           # otherwise, no need this arg during inverting\n",
        "            nearest_interp=True,  # change to use \"nearest\" mode in interpolation when inverting\n",
        "            to_tensor=True,  # convert to PyTorch Tensor after inverting\n",
        "        ),\n",
        "        #SaveImaged(keys=\"pred\", meta_keys=\"pred_meta_dict\", output_dir=\"./out\", output_postfix=\"seg\", resample=False),\n",
        "    ])\n",
        "\n",
        "test_ds = ImageDataset(\n",
        "    image_files=files, labels=labels, transform=post_transforms)\n",
        "test_loader = DataLoader(test_ds, batch_size=1, num_workers=4)\n",
        "itera = iter(test_loader)\n",
        "\n",
        "\n",
        "def get_next_im():\n",
        "    test_data = next(itera)\n",
        "    return test_data[0].to(device), test_data[1].unsqueeze(0).to(device)\n",
        "\n",
        "\n",
        "def plot_occlusion_heatmap(im, heatmap):\n",
        "    plt.subplots(1, 2)\n",
        "    plt.subplot(1, 2, 1)\n",
        "    plt.imshow(np.squeeze(im.cpu()))\n",
        "    plt.colorbar()\n",
        "    plt.subplot(1, 2, 2)\n",
        "    plt.imshow(heatmap)\n",
        "    plt.colorbar()\n",
        "    plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vfFYizJOg1TG"
      },
      "source": [
        "# Get a random image and its corresponding label\n",
        "img, label = get_next_im()\n",
        "\n",
        "# Get the occlusion sensitivity map\n",
        "occ_sens = monai.visualize.OcclusionSensitivity(\n",
        "    nn_module=model, mask_size=12, n_batch=10, stride=12)\n",
        "# Get the CAM\n",
        "cam = monai.visualize.GradCAM(\n",
        "    nn_module=model, target_layers=\"class_layers.relu\"\n",
        ")\n",
        "# Only get a single slice to save time.\n",
        "# For the other dimensions (channel, width, height), use\n",
        "# -1 to use 0 and img.shape[x]-1 for min and max, respectively\n",
        "depth_slice = img.shape[2] // 2\n",
        "\n",
        "occ_sens_b_box = [-1, -1, depth_slice, depth_slice, -1, -1, -1, -1]\n",
        "\n",
        "occ_result, _ = occ_sens(x=img, b_box=occ_sens_b_box)\n",
        "occ_result = occ_result[..., label.item()]\n",
        "cam_result = cam(x=img, class_idx=None)\n",
        "cam_result = cam_result[..., depth_slice]\n",
        "\n",
        "\n",
        "fig, axes = plt.subplots(1, 3, figsize=(25, 15), facecolor='white')\n",
        "\n",
        "for i, im in enumerate([img[:, :, depth_slice, ...], occ_result, cam_result]):\n",
        "    cmap = 'gray' if i == 0 else 'jet'\n",
        "    ax = axes[i]\n",
        "    if isinstance(im, torch.Tensor):\n",
        "        im = im.cpu().detach()    \n",
        "    #im_show = ax.imshow(im[0][0], cmap=cmap)\n",
        "    im_show = ax.imshow(np.squeeze(im[0][0]), cmap=cmap)\n",
        "    ax.axis('off')\n",
        "    fig.colorbar(im_show, ax=ax)\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}