bayesian_neural_net.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "bayesian_neural_net.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyN7KBxqZK9D1hG4zRBk1eLa",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU",
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "2c75879a652146c28affcf17961b2598": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_7d5dffb73c7a443287a6586dc1ab9f7f",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_652a29e58bd14597b010556714f3bfca",
              "IPY_MODEL_2c8071a441c7488ea6af5103f20e06dd"
            ]
          }
        },
        "7d5dffb73c7a443287a6586dc1ab9f7f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "652a29e58bd14597b010556714f3bfca": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_c4d39de5b932473abdf94ed766e139c7",
            "_dom_classes": [],
            "description": "100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 46827520,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 46827520,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_396f3b26861e4cffb748d0874c36a201"
          }
        },
        "2c8071a441c7488ea6af5103f20e06dd": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_f83949b49d6445b08e2a53fc14d17206",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 44.7M/44.7M [48:06&lt;00:00, 16.2kB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_f1c733c3dbb1491e8f88151179280f5d"
          }
        },
        "c4d39de5b932473abdf94ed766e139c7": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "396f3b26861e4cffb748d0874c36a201": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "f83949b49d6445b08e2a53fc14d17206": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "f1c733c3dbb1491e8f88151179280f5d": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        }
      }
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/themendu/covid_from_ctscan/blob/master/bayesian_neural_net.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7ueucDH2lOes",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "#from matplotlib import colors\n",
        "#from IPython import display\n",
        "#import os\n",
        "#from PIL import Image\n",
        "from torch.utils.data.dataset import Dataset\n",
        "#from scipy.misc import imread"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2GZAGy5MGSdt",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import torch\n",
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "import torch.optim as optim\n",
        "import torchvision\n",
        "from torchvision import datasets, transforms\n",
        "from torch.utils.data import DataLoader"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "U9K2-CYClUWs",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 121
        },
        "outputId": "c7bb1821-4470-4fe5-b485-82af55159f75"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount(\"/content/gdrive\", force_remount=True)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3aietf%3awg%3aoauth%3a2.0%3aoob&response_type=code&scope=email%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdocs.test%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive.photos.readonly%20https%3a%2f%2fwww.googleapis.com%2fauth%2fpeopleapi.readonly\n",
            "\n",
            "Enter your authorization code:\n",
            "··········\n",
            "Mounted at /content/gdrive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FPROyv9jqYhA",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 84
        },
        "outputId": "38c9b15d-e294-42ff-db56-0c620610d57f"
      },
      "source": [
        "pip install torchbnn"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting torchbnn\n",
            "  Downloading https://files.pythonhosted.org/packages/68/47/4629eb37ec9e28d162f76cc44a988bb1f29c7fb2598918b29a06e342fa0d/torchbnn-1.2-py3-none-any.whl\n",
            "Installing collected packages: torchbnn\n",
            "Successfully installed torchbnn-1.2\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QUgajmEvqkw0",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 84
        },
        "outputId": "0096e570-9b49-4802-d11d-d8768560ed75"
      },
      "source": [
        "pip install torchhk"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting torchhk\n",
            "  Downloading https://files.pythonhosted.org/packages/31/06/6b0e557d979d41db6ce525b1bb6eb1c2b0c38e76ea9e43a9c668d732fb5c/torchhk-0.62-py3-none-any.whl\n",
            "Installing collected packages: torchhk\n",
            "Successfully installed torchhk-0.62\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uUw7oq1SXIVT",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#pip install Augmentor"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LoZq84OQThor",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 70
        },
        "outputId": "27877ca9-520f-4519-b258-37e58342db74"
      },
      "source": [
        "import torchbnn as bnn\n",
        "from torchhk import transform_model"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/statsmodels/tools/_testing.py:19: FutureWarning: pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.\n",
            "  import pandas.util.testing as tm\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4U9DRwtLkSpU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NpOyNxK26bEE",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "4e0c9353-7272-4021-b672-5aa6cee0452d"
      },
      "source": [
        "device"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "device(type='cuda', index=0)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "79p4wKyiThlX",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 82,
          "referenced_widgets": [
            "2c75879a652146c28affcf17961b2598",
            "7d5dffb73c7a443287a6586dc1ab9f7f",
            "652a29e58bd14597b010556714f3bfca",
            "2c8071a441c7488ea6af5103f20e06dd",
            "c4d39de5b932473abdf94ed766e139c7",
            "396f3b26861e4cffb748d0874c36a201",
            "f83949b49d6445b08e2a53fc14d17206",
            "f1c733c3dbb1491e8f88151179280f5d"
          ]
        },
        "outputId": "5228835c-ec26-4eb2-bd70-98bcc8dfb0b3"
      },
      "source": [
        "\n",
        "import torchvision.models as models\n",
        "import torch.nn as nn\n",
        "\n",
        "model = models.resnet18(pretrained=True)\n",
        "\n",
        "#model = models.alexnet(pretrained=True).to(device)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading: \"https://download.pytorch.org/models/resnet18-5c106cde.pth\" to /root/.cache/torch/hub/checkpoints/resnet18-5c106cde.pth\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "2c75879a652146c28affcf17961b2598",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, max=46827520.0), HTML(value='')))"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dxe-DPLeXKdf",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "for param in model.parameters():\n",
        "  param.requires_grad=True\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sC0m8YK7GOS-",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model.fc=nn.Sequential(\n",
        "               nn.Linear(512,128),\n",
        "               nn.ReLU(inplace=True),\n",
        "#               nn.Dropout(p=0.2,inplace=False),\n",
        "               nn.Linear(128, 2))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BfhPNnpAlU6U",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "transform_model(model, nn.Conv2d, bnn.BayesConv2d, \n",
        "                args={\"prior_mu\":0, \"prior_sigma\":0.2, \"in_channels\" : \".in_channels\",\n",
        "                      \"out_channels\" : \".out_channels\", \"kernel_size\" : \".kernel_size\",\n",
        "                      \"stride\" : \".stride\", \"padding\" : \".padding\", \"bias\":\".bias\"\n",
        "                     }, \n",
        "                attrs={\"weight_mu\" : \".weight\"})"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yBCfKwkZpg7S",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "45c4dcdc-fb59-4262-c3a0-c33ece48978c"
      },
      "source": [
        "transform_model(model, nn.Linear, bnn.BayesLinear, \n",
        "            args={\"prior_mu\":0, \"prior_sigma\":0.2, \"in_features\" : \".in_features\",\n",
        "                  \"out_features\" : \".out_features\", \"bias\":\".bias\"\n",
        "                 }, \n",
        "            attrs={\"weight_mu\" : \".weight\"})"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/torchhk/transform.py:34: Warning:\n",
            "\n",
            "\n",
            " * Caution : The Input Model is CHANGED because inplace=True.\n",
            "\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "ResNet(\n",
              "  (conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)\n",
              "  (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "  (relu): ReLU(inplace=True)\n",
              "  (maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)\n",
              "  (layer1): Sequential(\n",
              "    (0): BasicBlock(\n",
              "      (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
              "      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      (relu): ReLU(inplace=True)\n",
              "      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
              "      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "    )\n",
              "    (1): BasicBlock(\n",
              "      (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
              "      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      (relu): ReLU(inplace=True)\n",
              "      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
              "      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "    )\n",
              "  )\n",
              "  (layer2): Sequential(\n",
              "    (0): BasicBlock(\n",
              "      (conv1): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
              "      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      (relu): ReLU(inplace=True)\n",
              "      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
              "      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      (downsample): Sequential(\n",
              "        (0): Conv2d(64, 128, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
              "        (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "    )\n",
              "    (1): BasicBlock(\n",
              "      (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
              "      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      (relu): ReLU(inplace=True)\n",
              "      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
              "      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "    )\n",
              "  )\n",
              "  (layer3): Sequential(\n",
              "    (0): BasicBlock(\n",
              "      (conv1): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
              "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      (relu): ReLU(inplace=True)\n",
              "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
              "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      (downsample): Sequential(\n",
              "        (0): Conv2d(128, 256, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
              "        (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "    )\n",
              "    (1): BasicBlock(\n",
              "      (conv1): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
              "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      (relu): ReLU(inplace=True)\n",
              "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
              "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "    )\n",
              "  )\n",
              "  (layer4): Sequential(\n",
              "    (0): BasicBlock(\n",
              "      (conv1): Conv2d(256, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
              "      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      (relu): ReLU(inplace=True)\n",
              "      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
              "      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      (downsample): Sequential(\n",
              "        (0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
              "        (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "    )\n",
              "    (1): BasicBlock(\n",
              "      (conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
              "      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      (relu): ReLU(inplace=True)\n",
              "      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
              "      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "    )\n",
              "  )\n",
              "  (avgpool): AdaptiveAvgPool2d(output_size=(1, 1))\n",
              "  (fc): Sequential(\n",
              "    (0): BayesLinear(prior_mu=0, prior_sigma=0.2, in_features=512, out_features=128, bias=True)\n",
              "    (1): ReLU(inplace=True)\n",
              "    (2): BayesLinear(prior_mu=0, prior_sigma=0.2, in_features=128, out_features=2, bias=True)\n",
              "  )\n",
              ")"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hscGno9dNO7s",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "transform_model(model, nn.BatchNorm2d,bnn.modules.batchnorm.BayesBatchNorm2d, \n",
        "            args={\"prior_mu\":0, \"prior_sigma\":0.2, \"num_features\":\".num_features\",\"momentum\":0.1\n",
        "                 }, \n",
        "            attrs={\"weight_mu\" : \".weight\"})"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tZCTeLu1lU88",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model=model.to(device)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VWGRRfJcUA16",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#import Augmentor\n",
        "#p = Augmentor.Pipeline(source_directory=\"/content/gdrive/My Drive/corona/bayesian/ct/\",output_directory=\"/content/gdrive/My Drive/corona/bay_aug/ct/\")\n",
        "#p.random_distortion(probability=0.7, grid_width=4, grid_height=4, magnitude=8)\n",
        "#p.rotate(probability=0.7, max_left_rotation=5, max_right_rotation=5)\n",
        "#p.zoom_random(probability=0.7, percentage_area=0.95)\n",
        "#p.flip_top_bottom(probability=0.7)\n",
        "#p.rotate_random_90(probability=0.7)\n",
        "#p.random_erasing(probability=0.7,rectangle_area=0.2)\n",
        "\n",
        "#p.sample(1000)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oRIM-5CfUBmh",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#p1 = Augmentor.Pipeline(source_directory=\"/content/gdrive/My Drive/corona/bayesian/xray/\",output_directory=\"/content/gdrive/My Drive/corona/bay_aug/xray/\")\n",
        "\n",
        "#p1.random_distortion(probability=0.4, grid_width=4, grid_height=4, magnitude=8)\n",
        "#p1.rotate(probability=0.4, max_left_rotation=5, max_right_rotation=5)\n",
        "#p1.zoom_random(probability=0.4, percentage_area=0.95)\n",
        "#p1.flip_top_bottom(probability=0.5)\n",
        "\n",
        "#p1.sample(1000)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pBnfUHS7Yye-",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#p1 = Augmentor.Pipeline(source_directory=\"/content/gdrive/My Drive/lp/dataset/neg/\",output_directory=\"/content/gdrive/My Drive/corona/bay_aug/ood/\")\n",
        "\n",
        "#p1.random_distortion(probability=0.7, grid_width=4, grid_height=4, magnitude=8)\n",
        "#p1.rotate(probability=0.7, max_left_rotation=5, max_right_rotation=5)\n",
        "#p1.zoom_random(probability=0.7, percentage_area=0.95)\n",
        "#p1.flip_top_bottom(probability=0.7)\n",
        "\n",
        "#p1.sample(1000)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "htGrhDFtXCBR",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from torchvision import transforms\n",
        "TRANSFORM_IMG = transforms.Compose([\n",
        "    transforms.Resize((224,224)),\n",
        "    transforms.ToTensor(),transforms.Normalize(mean=[0.485, 0.456, 0.406],\n",
        "              std=[0.229, 0.224, 0.225])])\n",
        "path3=\"/content/gdrive/My Drive/corona/bay_aug/\"\n",
        "\n",
        "\n",
        "\n",
        "data=torchvision.datasets.ImageFolder(path3,transform=TRANSFORM_IMG)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kMNNo4rFXCEJ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 168
        },
        "outputId": "9675e48d-bbc8-47bb-b87a-d3e4ddd672aa"
      },
      "source": [
        "data"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Dataset ImageFolder\n",
              "    Number of datapoints: 2000\n",
              "    Root location: /content/gdrive/My Drive/corona/bay_aug/\n",
              "    StandardTransform\n",
              "Transform: Compose(\n",
              "               Resize(size=(224, 224), interpolation=PIL.Image.BILINEAR)\n",
              "               ToTensor()\n",
              "               Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])\n",
              "           )"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 38
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Tcw-1rUKXCG-",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "6f367697-03bc-4513-f313-bb637dbf76c3"
      },
      "source": [
        "data.class_to_idx"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'ct': 0, 'ood': 1}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 39
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eKozF7q_XB7t",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "train,test =torch.utils.data.random_split(data, [1900,100])\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uCUi8KT0mUvm",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "del data\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lDV2Nx_eIiRI",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "dataloaders = {\n",
        "    'train':\n",
        "    torch.utils.data.DataLoader(train,\n",
        "                                batch_size=64,\n",
        "                                shuffle=True,\n",
        "                                num_workers=0,\n",
        "                                drop_last=True),  \n",
        "    'validation':\n",
        "    torch.utils.data.DataLoader(test,\n",
        "                                batch_size=1,\n",
        "                                shuffle=False,\n",
        "                                num_workers=0,\n",
        "                                drop_last=False)  \n",
        "}"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uIOkTCPxIiT_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#trainiter = iter(dataloaders['train'])\n",
        "#features, labels = next(trainiter)\n",
        "#features.shape, labels.shape"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kgxLsFDUmhfY",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#del features\n",
        "#del labels\n",
        "#del trainiter"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CGuBrYvhoG2x",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def precision(y_true:torch.Tensor, y_pred:torch.Tensor, is_training=False) -> torch.Tensor:\n",
        "\n",
        "    assert y_true.ndim == 1\n",
        "    assert y_pred.ndim == 1 or y_pred.ndim == 2\n",
        "    \n",
        "    if y_pred.ndim == 2:\n",
        "        y_pred = y_pred.argmax(dim=1)\n",
        "        \n",
        "    \n",
        "    tp = (y_true * y_pred).sum().to(torch.float32)\n",
        "    fp = ((1-y_true) * y_pred).sum().to(torch.float32)\n",
        "    \n",
        "    epsilon = 1e-7\n",
        "    precision = tp / (tp + fp + epsilon)\n",
        "  \n",
        "    return precision"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XQ6o98btkaJR",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#model=torch.load('/content/gdrive/My Drive/corona/bayesian_aug.pt')\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Cl5Pt3AhksJU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#for param in model.parameters():\n",
        "#  param.requires_grad=True\n",
        "#model=model.to(device)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tQLJXLyqlVAE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "\n",
        "criterion = nn.CrossEntropyLoss()\n",
        "kl_loss = bnn.BKLLoss(reduction='mean', last_layer_only=False)\n",
        "factor = 1\n",
        "\n",
        "optimizer = optim.Adam(model.parameters(),lr=0.0000005, betas=(0.9, 0.999), eps=1e-08, weight_decay=0.3, amsgrad=False)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "K_c1tmPTnY1q",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def train_model(model, criterion, optimizer, num_epochs,kl_loss):\n",
        "    losses_train_list=[]\n",
        "    losses_test_list=[]\n",
        "    prec_train_list=[]\n",
        "    prec_test_list=[]\n",
        "    kl_train_list=[]\n",
        "    kl_test_list=[]\n",
        "    for epoch in range(num_epochs):\n",
        "        print('Epoch {}/{}'.format(epoch+1, num_epochs))\n",
        "        print('-' * 10)\n",
        "\n",
        "        for phase in ['train', 'validation']:\n",
        "            if phase == 'train':\n",
        "                model.train()\n",
        "            else:\n",
        "                model.eval()\n",
        "\n",
        "            running_loss = 0.0\n",
        "            running_corrects = 0\n",
        "            running_kl=0.0\n",
        "\n",
        "            for inputs, labels in dataloaders[phase]:\n",
        "                inputs = inputs.to(device)\n",
        "                labels = labels.to(device)\n",
        "\n",
        "                outputs = model(inputs)\n",
        "                loss = criterion(outputs, labels)\n",
        "                kl=kl_loss(model)\n",
        "                cost=loss+(factor*kl)\n",
        "\n",
        "                if phase == 'train':\n",
        "                    optimizer.zero_grad()\n",
        "                    cost.backward()\n",
        "                    optimizer.step()\n",
        "                _, preds = torch.max(outputs, 1)\n",
        "                running_loss += loss.item() * inputs.size(0)\n",
        "                running_corrects += torch.sum(preds == labels.data)\n",
        "                running_kl+=kl.item()\n",
        "\n",
        "            if phase=='train':\n",
        "                           epoch_loss = running_loss /1900\n",
        "                           epoch_acc = running_corrects.double() /1900\n",
        "                           epoch_kl=running_kl/1900\n",
        "                           print('train precision score is : {}'.format(precision(labels.data.to('cpu'), outputs.data.to('cpu'))))\n",
        "                           losses_train_list.append(epoch_loss)\n",
        "                           prec_train_list.append(precision(labels.data.to('cpu'), outputs.data.to('cpu')))\n",
        "                           kl_train_list.append(epoch_kl)\n",
        "\n",
        "\n",
        "            else: \n",
        "                           epoch_loss = running_loss /100\n",
        "                           epoch_acc = running_corrects.double() /100\n",
        "                           epoch_kl=running_kl/100\n",
        "                           print('valid precision score is : {}'.format(precision(labels.data.to('cpu'), outputs.data.to('cpu'))))\n",
        "                           losses_test_list.append(epoch_loss)\n",
        "                           prec_test_list.append(precision(labels.data.to('cpu'), outputs.data.to('cpu')))\n",
        "                           kl_test_list.append(epoch_kl)\n",
        "\n",
        "            \n",
        "\n",
        "            print('{} loss: {:.4f}, acc: {:.6f}, kl:{:.6f}'.format(phase,\n",
        "                                                        epoch_loss,\n",
        "                                                        epoch_acc,\n",
        "                                                        epoch_kl))\n",
        "\n",
        "\n",
        "    return model,losses_train_list,losses_test_list,prec_train_list,prec_test_list,kl_train_list,kl_test_list"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vEy_mHPYnY5O",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model_trained1,a1,b1,c1,d1,e1,f1= train_model(model, criterion, optimizer, num_epochs=3,kl_loss=kl_loss)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3pjO9PamnY9p",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "  import matplotlib.pyplot as plt\n",
        "plt.figure(figsize=(8, 6))\n",
        "plt.plot(a1,label='train_loss')\n",
        "plt.plot(b1,label='test_loss')\n",
        "plt.plot(e1,label='kl_train')\n",
        "plt.plot(f1,label='kl_test')\n",
        "\n",
        "plt.xlabel(\"epochs\")\n",
        "plt.ylabel(\"value\")\n",
        "plt.title(\"train versus val plot\")\n",
        "\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ORhLckA3n0Di",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#torch.save(model_trained1,'/content/gdrive/My Drive/corona/bayesian_aug.pt')\n",
        "#resnet18,88% accuracy, test loss=1.5,kl_test_loss=0.3,kl_train_loss=0.03"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "y_Ae7XcHjqo_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def get_uncertainty_per_image(model, input_image, T, normalized=False):\n",
        "    input_image = input_image.unsqueeze(0)\n",
        "    input_images = input_image.repeat(T, 1, 1, 1)\n",
        "\n",
        "    net_out, _ = model(input_images)\n",
        "    pred = torch.mean(net_out, dim=0).cpu().detach().numpy()\n",
        "    if normalized:\n",
        "        prediction = F.softplus(net_out)\n",
        "        p_hat = prediction / torch.sum(prediction, dim=1).unsqueeze(1)\n",
        "    else:\n",
        "        p_hat = F.softmax(net_out, dim=1)\n",
        "    p_hat = p_hat.detach().cpu().numpy()\n",
        "    p_bar = np.mean(p_hat, axis=0)\n",
        "\n",
        "    temp = p_hat - np.expand_dims(p_bar, 0)\n",
        "    epistemic = np.dot(temp.T, temp) / T\n",
        "    epistemic = np.diag(epistemic)\n",
        "\n",
        "    aleatoric = np.diag(p_bar) - (np.dot(p_hat.T, p_hat) / T)\n",
        "    aleatoric = np.diag(aleatoric)\n",
        "\n",
        "    return pred, epistemic, aleatoric\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RvIMDFZo2ApN",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 285
        },
        "outputId": "2a36e254-5767-43e1-963a-f860e992a4c5"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "abc=plt.imread('/content/gdrive/My Drive/corona/bay_aug/ct/ct_original_6_3.jpg_f726454a-a2b2-456e-8ae6-aa3ee8c8ed9a.jpg')\n",
        "plt.imshow(abc)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7f0f79fb3f60>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 81
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GwZ83XYGVZdC",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def get_uncertainty_per_image(model, path, T, normalized=False):\n",
        "    image = Image.open(path)\n",
        "    image = loader(image).float()\n",
        "    image = image.unsqueeze(0)\n",
        "    input_image=image.to(device)\n",
        "    input_images = input_image.repeat(T, 1, 1, 1)\n",
        "\n",
        "    net_out = model(input_images)\n",
        "    pred = torch.mean(net_out, dim=0).cpu().detach().numpy()\n",
        "    if normalized:\n",
        "        prediction = F.softplus(net_out)\n",
        "        p_hat = prediction / torch.sum(prediction, dim=1).unsqueeze(1)\n",
        "    else:\n",
        "        p_hat = F.softmax(net_out, dim=1)\n",
        "    p_hat = p_hat.detach().cpu().numpy()\n",
        "    p_bar = np.mean(p_hat, axis=0)\n",
        "\n",
        "    temp = p_hat - np.expand_dims(p_bar, 0)\n",
        "    epistemic = np.dot(temp.T, temp) / T\n",
        "    epistemic = np.diag(epistemic)\n",
        "\n",
        "    aleatoric = np.diag(p_bar) - (np.dot(p_hat.T, p_hat) / T)\n",
        "    aleatoric = np.diag(aleatoric)\n",
        "    return pred, epistemic, aleatoric\n",
        " \n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fDxYRPc1VZZ0",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 67
        },
        "outputId": "2f13b47b-a60e-42b7-8844-f27008e79344"
      },
      "source": [
        "from torch.autograd import Variable\n",
        "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n",
        "\n",
        "get_uncertainty_per_image(model,'/content/gdrive/My Drive/corona/bay_aug/ct/ct_original_6_3.jpg_f726454a-a2b2-456e-8ae6-aa3ee8c8ed9a.jpg',10)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([ 21.65228 , -27.855183], dtype=float32),\n",
              " array([0., 0.], dtype=float32),\n",
              " array([0.000000e+00, 3.156332e-22], dtype=float32))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 79
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "W2r7_rHOVZl1",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Kg1516PEVZgl",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z8Njuds50F-W",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}