{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "a2220df6",
   "metadata": {},
   "source": [
    "# Import Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "7249bea4",
   "metadata": {},
   "outputs": [],
   "source": [
    "import gradio as gr\n",
    "import torch\n",
    "import torch.nn.functional as F\n",
    "from facenet_pytorch import MTCNN, InceptionResnetV1\n",
    "import numpy as np\n",
    "from PIL import Image\n",
    "import cv2\n",
    "from pytorch_grad_cam import GradCAM\n",
    "from pytorch_grad_cam.utils.model_targets import ClassifierOutputTarget\n",
    "from pytorch_grad_cam.utils.image import show_cam_on_image\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d25e1c5d",
   "metadata": {},
   "source": [
    "# Download and Load Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "237fbf44",
   "metadata": {},
   "outputs": [],
   "source": [
    "DEVICE = 'cuda:0' if torch.cuda.is_available() else 'cpu'\n",
    "\n",
    "mtcnn = MTCNN(\n",
    "    select_largest=False,\n",
    "    post_process=False,\n",
    "    device=DEVICE\n",
    ").to(DEVICE).eval()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "f3ef2b4f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "InceptionResnetV1(\n",
       "  (conv2d_1a): BasicConv2d(\n",
       "    (conv): Conv2d(3, 32, kernel_size=(3, 3), stride=(2, 2), bias=False)\n",
       "    (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    (relu): ReLU()\n",
       "  )\n",
       "  (conv2d_2a): BasicConv2d(\n",
       "    (conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), bias=False)\n",
       "    (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    (relu): ReLU()\n",
       "  )\n",
       "  (conv2d_2b): BasicConv2d(\n",
       "    (conv): Conv2d(32, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "    (bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    (relu): ReLU()\n",
       "  )\n",
       "  (maxpool_3a): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
       "  (conv2d_3b): BasicConv2d(\n",
       "    (conv): Conv2d(64, 80, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "    (bn): BatchNorm2d(80, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    (relu): ReLU()\n",
       "  )\n",
       "  (conv2d_4a): BasicConv2d(\n",
       "    (conv): Conv2d(80, 192, kernel_size=(3, 3), stride=(1, 1), bias=False)\n",
       "    (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    (relu): ReLU()\n",
       "  )\n",
       "  (conv2d_4b): BasicConv2d(\n",
       "    (conv): Conv2d(192, 256, kernel_size=(3, 3), stride=(2, 2), bias=False)\n",
       "    (bn): BatchNorm2d(256, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    (relu): ReLU()\n",
       "  )\n",
       "  (repeat_1): Sequential(\n",
       "    (0): Block35(\n",
       "      (branch0): BasicConv2d(\n",
       "        (conv): Conv2d(256, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (branch1): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(256, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (branch2): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(256, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (2): BasicConv2d(\n",
       "          (conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (conv2d): Conv2d(96, 256, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "    (1): Block35(\n",
       "      (branch0): BasicConv2d(\n",
       "        (conv): Conv2d(256, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (branch1): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(256, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (branch2): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(256, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (2): BasicConv2d(\n",
       "          (conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (conv2d): Conv2d(96, 256, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "    (2): Block35(\n",
       "      (branch0): BasicConv2d(\n",
       "        (conv): Conv2d(256, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (branch1): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(256, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (branch2): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(256, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (2): BasicConv2d(\n",
       "          (conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (conv2d): Conv2d(96, 256, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "    (3): Block35(\n",
       "      (branch0): BasicConv2d(\n",
       "        (conv): Conv2d(256, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (branch1): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(256, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (branch2): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(256, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (2): BasicConv2d(\n",
       "          (conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (conv2d): Conv2d(96, 256, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "    (4): Block35(\n",
       "      (branch0): BasicConv2d(\n",
       "        (conv): Conv2d(256, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (branch1): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(256, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (branch2): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(256, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (2): BasicConv2d(\n",
       "          (conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (conv2d): Conv2d(96, 256, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "  )\n",
       "  (mixed_6a): Mixed_6a(\n",
       "    (branch0): BasicConv2d(\n",
       "      (conv): Conv2d(256, 384, kernel_size=(3, 3), stride=(2, 2), bias=False)\n",
       "      (bn): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "    (branch1): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(256, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(192, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (2): BasicConv2d(\n",
       "        (conv): Conv2d(192, 256, kernel_size=(3, 3), stride=(2, 2), bias=False)\n",
       "        (bn): BatchNorm2d(256, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "    )\n",
       "    (branch2): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
       "  )\n",
       "  (repeat_2): Sequential(\n",
       "    (0): Block17(\n",
       "      (branch0): BasicConv2d(\n",
       "        (conv): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (branch1): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(128, 128, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (2): BasicConv2d(\n",
       "          (conv): Conv2d(128, 128, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (conv2d): Conv2d(256, 896, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "    (1): Block17(\n",
       "      (branch0): BasicConv2d(\n",
       "        (conv): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (branch1): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(128, 128, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (2): BasicConv2d(\n",
       "          (conv): Conv2d(128, 128, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (conv2d): Conv2d(256, 896, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "    (2): Block17(\n",
       "      (branch0): BasicConv2d(\n",
       "        (conv): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (branch1): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(128, 128, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (2): BasicConv2d(\n",
       "          (conv): Conv2d(128, 128, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (conv2d): Conv2d(256, 896, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "    (3): Block17(\n",
       "      (branch0): BasicConv2d(\n",
       "        (conv): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (branch1): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(128, 128, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (2): BasicConv2d(\n",
       "          (conv): Conv2d(128, 128, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (conv2d): Conv2d(256, 896, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "    (4): Block17(\n",
       "      (branch0): BasicConv2d(\n",
       "        (conv): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (branch1): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(128, 128, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (2): BasicConv2d(\n",
       "          (conv): Conv2d(128, 128, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (conv2d): Conv2d(256, 896, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "    (5): Block17(\n",
       "      (branch0): BasicConv2d(\n",
       "        (conv): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (branch1): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(128, 128, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (2): BasicConv2d(\n",
       "          (conv): Conv2d(128, 128, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (conv2d): Conv2d(256, 896, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "    (6): Block17(\n",
       "      (branch0): BasicConv2d(\n",
       "        (conv): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (branch1): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(128, 128, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (2): BasicConv2d(\n",
       "          (conv): Conv2d(128, 128, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (conv2d): Conv2d(256, 896, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "    (7): Block17(\n",
       "      (branch0): BasicConv2d(\n",
       "        (conv): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (branch1): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(128, 128, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (2): BasicConv2d(\n",
       "          (conv): Conv2d(128, 128, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (conv2d): Conv2d(256, 896, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "    (8): Block17(\n",
       "      (branch0): BasicConv2d(\n",
       "        (conv): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (branch1): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(128, 128, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (2): BasicConv2d(\n",
       "          (conv): Conv2d(128, 128, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (conv2d): Conv2d(256, 896, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "    (9): Block17(\n",
       "      (branch0): BasicConv2d(\n",
       "        (conv): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (branch1): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(128, 128, kernel_size=(1, 7), stride=(1, 1), padding=(0, 3), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (2): BasicConv2d(\n",
       "          (conv): Conv2d(128, 128, kernel_size=(7, 1), stride=(1, 1), padding=(3, 0), bias=False)\n",
       "          (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (conv2d): Conv2d(256, 896, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "  )\n",
       "  (mixed_7a): Mixed_7a(\n",
       "    (branch0): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(896, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(256, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(256, 384, kernel_size=(3, 3), stride=(2, 2), bias=False)\n",
       "        (bn): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "    )\n",
       "    (branch1): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(896, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(256, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), bias=False)\n",
       "        (bn): BatchNorm2d(256, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "    )\n",
       "    (branch2): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(896, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(256, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(256, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (2): BasicConv2d(\n",
       "        (conv): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), bias=False)\n",
       "        (bn): BatchNorm2d(256, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "    )\n",
       "    (branch3): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
       "  )\n",
       "  (repeat_3): Sequential(\n",
       "    (0): Block8(\n",
       "      (branch0): BasicConv2d(\n",
       "        (conv): Conv2d(1792, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (branch1): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(1792, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(192, 192, kernel_size=(1, 3), stride=(1, 1), padding=(0, 1), bias=False)\n",
       "          (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (2): BasicConv2d(\n",
       "          (conv): Conv2d(192, 192, kernel_size=(3, 1), stride=(1, 1), padding=(1, 0), bias=False)\n",
       "          (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (conv2d): Conv2d(384, 1792, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "    (1): Block8(\n",
       "      (branch0): BasicConv2d(\n",
       "        (conv): Conv2d(1792, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (branch1): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(1792, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(192, 192, kernel_size=(1, 3), stride=(1, 1), padding=(0, 1), bias=False)\n",
       "          (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (2): BasicConv2d(\n",
       "          (conv): Conv2d(192, 192, kernel_size=(3, 1), stride=(1, 1), padding=(1, 0), bias=False)\n",
       "          (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (conv2d): Conv2d(384, 1792, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "    (2): Block8(\n",
       "      (branch0): BasicConv2d(\n",
       "        (conv): Conv2d(1792, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (branch1): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(1792, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(192, 192, kernel_size=(1, 3), stride=(1, 1), padding=(0, 1), bias=False)\n",
       "          (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (2): BasicConv2d(\n",
       "          (conv): Conv2d(192, 192, kernel_size=(3, 1), stride=(1, 1), padding=(1, 0), bias=False)\n",
       "          (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (conv2d): Conv2d(384, 1792, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "    (3): Block8(\n",
       "      (branch0): BasicConv2d(\n",
       "        (conv): Conv2d(1792, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (branch1): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(1792, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(192, 192, kernel_size=(1, 3), stride=(1, 1), padding=(0, 1), bias=False)\n",
       "          (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (2): BasicConv2d(\n",
       "          (conv): Conv2d(192, 192, kernel_size=(3, 1), stride=(1, 1), padding=(1, 0), bias=False)\n",
       "          (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (conv2d): Conv2d(384, 1792, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "    (4): Block8(\n",
       "      (branch0): BasicConv2d(\n",
       "        (conv): Conv2d(1792, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (branch1): Sequential(\n",
       "        (0): BasicConv2d(\n",
       "          (conv): Conv2d(1792, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "          (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (1): BasicConv2d(\n",
       "          (conv): Conv2d(192, 192, kernel_size=(1, 3), stride=(1, 1), padding=(0, 1), bias=False)\n",
       "          (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "        (2): BasicConv2d(\n",
       "          (conv): Conv2d(192, 192, kernel_size=(3, 1), stride=(1, 1), padding=(1, 0), bias=False)\n",
       "          (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "          (relu): ReLU()\n",
       "        )\n",
       "      )\n",
       "      (conv2d): Conv2d(384, 1792, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "  )\n",
       "  (block8): Block8(\n",
       "    (branch0): BasicConv2d(\n",
       "      (conv): Conv2d(1792, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU()\n",
       "    )\n",
       "    (branch1): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(1792, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(192, 192, kernel_size=(1, 3), stride=(1, 1), padding=(0, 1), bias=False)\n",
       "        (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "      (2): BasicConv2d(\n",
       "        (conv): Conv2d(192, 192, kernel_size=(3, 1), stride=(1, 1), padding=(1, 0), bias=False)\n",
       "        (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "        (relu): ReLU()\n",
       "      )\n",
       "    )\n",
       "    (conv2d): Conv2d(384, 1792, kernel_size=(1, 1), stride=(1, 1))\n",
       "  )\n",
       "  (avgpool_1a): AdaptiveAvgPool2d(output_size=1)\n",
       "  (dropout): Dropout(p=0.6, inplace=False)\n",
       "  (last_linear): Linear(in_features=1792, out_features=512, bias=False)\n",
       "  (last_bn): BatchNorm1d(512, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (logits): Linear(in_features=512, out_features=1, bias=True)\n",
       ")"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = InceptionResnetV1(\n",
    "    pretrained=\"vggface2\",\n",
    "    classify=True,\n",
    "    num_classes=1,\n",
    "    device=DEVICE\n",
    ")\n",
    "\n",
    "checkpoint = torch.load(\"resnetinceptionv1_epoch_32.pth\", map_location=torch.device('cpu'))\n",
    "model.load_state_dict(checkpoint['model_state_dict'])\n",
    "model.to(DEVICE)\n",
    "model.eval()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a499194a",
   "metadata": {},
   "source": [
    "# Model Inference "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "376e6cd6",
   "metadata": {},
   "outputs": [],
   "source": [
    "def predict(input_image:Image.Image):\n",
    "    \"\"\"Predict the label of the input_image\"\"\"\n",
    "    face = mtcnn(input_image)\n",
    "    if face is None:\n",
    "        raise Exception('No face detected')\n",
    "    face = face.unsqueeze(0) # add the batch dimension\n",
    "    face = F.interpolate(face, size=(256, 256), mode='bilinear', align_corners=False)\n",
    "    \n",
    "    # convert the face into a numpy array to be able to plot it\n",
    "    prev_face = face.squeeze(0).permute(1, 2, 0).cpu().detach().int().numpy()\n",
    "    prev_face = prev_face.astype('uint8')\n",
    "\n",
    "    face = face.to(DEVICE)\n",
    "    face = face.to(torch.float32)\n",
    "    face = face / 255.0\n",
    "    face_image_to_plot = face.squeeze(0).permute(1, 2, 0).cpu().detach().int().numpy()\n",
    "\n",
    "    target_layers=[model.block8.branch1[-1]]\n",
    "    use_cuda = True if torch.cuda.is_available() else False\n",
    "    cam = GradCAM(model=model, target_layers=target_layers, use_cuda=use_cuda)\n",
    "    targets = [ClassifierOutputTarget(0)]\n",
    "\n",
    "    grayscale_cam = cam(input_tensor=face, targets=targets, eigen_smooth=True)\n",
    "    grayscale_cam = grayscale_cam[0, :]\n",
    "    visualization = show_cam_on_image(face_image_to_plot, grayscale_cam, use_rgb=True)\n",
    "    face_with_mask = cv2.addWeighted(prev_face, 1, visualization, 0.5, 0)\n",
    "\n",
    "    with torch.no_grad():\n",
    "        output = torch.sigmoid(model(face).squeeze(0))\n",
    "        prediction = \"real\" if output.item() < 0.5 else \"fake\"\n",
    "        \n",
    "        real_prediction = 1 - output.item()\n",
    "        fake_prediction = output.item()\n",
    "        \n",
    "        confidences = {\n",
    "            'real': real_prediction,\n",
    "            'fake': fake_prediction\n",
    "        }\n",
    "    return confidences, face_with_mask\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "14f47b5a",
   "metadata": {},
   "source": [
    "# Gradio Interface"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "d62177b5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Running on local URL:  http://127.0.0.1:7860\n",
      "\n",
      "To create a public link, set `share=True` in `launch()`.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div><iframe src=\"http://127.0.0.1:7860/\" width=\"100%\" height=\"500\" allow=\"autoplay; camera; microphone; clipboard-read; clipboard-write;\" frameborder=\"0\" allowfullscreen></iframe></div>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "interface = gr.Interface(\n",
    "    fn=predict,\n",
    "    inputs=[\n",
    "        gr.inputs.Image(label=\"Input Image\", type=\"pil\")\n",
    "    ],\n",
    "    outputs=[\n",
    "        gr.outputs.Label(label=\"Class\"),\n",
    "        gr.outputs.Image(label=\"Face with Explainability\", type=\"pil\")\n",
    "    ],\n",
    ").launch()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0c0b293c",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}