ImageRecognition.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "PKcBS6xfBK1Z"
   },
   "outputs": [],
   "source": [
    "import torch\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import torch.nn.functional as F\n",
    "from torch import nn\n",
    "from torchvision import datasets, transforms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "hC4ZVkTRB79P"
   },
   "outputs": [],
   "source": [
    "transform = transforms.Compose([transforms.Resize((28,28)),\n",
    "                               transforms.ToTensor(),\n",
    "                               transforms.Normalize((0.5,), (0.5,))\n",
    "                               ])\n",
    "training_dataset = datasets.MNIST(root='./data', train=True, download=True, transform=transform)\n",
    "validation_dataset = datasets.MNIST(root='./data', train=False, download=True, transform=transform)\n",
    "\n",
    "training_loader = torch.utils.data.DataLoader(training_dataset, batch_size=100, shuffle=True)\n",
    "validation_loader = torch.utils.data.DataLoader(validation_dataset, batch_size = 100, shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "jCqX-QWODId3"
   },
   "outputs": [],
   "source": [
    "def im_convert(tensor):\n",
    "  image = tensor.clone().detach().numpy()\n",
    "  image = image.transpose(1, 2, 0)\n",
    "  image = image * np.array((0.5, 0.5, 0.5)) + np.array((0.5, 0.5, 0.5))\n",
    "  image = image.clip(0, 1)\n",
    "  return image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 283
    },
    "colab_type": "code",
    "id": "8bUYaxqxGFET",
    "outputId": "4946bab5-02fb-4cb1-987b-67da8126f500"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x288 with 20 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "dataiter = iter(training_loader)\n",
    "images, labels = dataiter.next()\n",
    "fig = plt.figure(figsize=(25, 4))\n",
    "\n",
    "for idx in np.arange(20):\n",
    "  ax = fig.add_subplot(2, 10, idx+1, xticks=[], yticks=[])\n",
    "  plt.imshow(im_convert(images[idx]))\n",
    "  ax.set_title([labels[idx].item()])\n",
    "  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "ELDFsBKeHqJy"
   },
   "outputs": [],
   "source": [
    "class Classifier(nn.Module):\n",
    "    \n",
    "    def __init__(self, D_in, H1, H2, D_out):\n",
    "        super().__init__()\n",
    "        self.linear1 = nn.Linear(D_in, H1)\n",
    "        self.linear2 = nn.Linear(H1, H2)\n",
    "        self.linear3 = nn.Linear(H2, D_out)\n",
    "    def forward(self, x):\n",
    "        x = F.relu(self.linear1(x))  \n",
    "        x = F.relu(self.linear2(x))\n",
    "        x = self.linear3(x)\n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 102
    },
    "colab_type": "code",
    "id": "cOa--7ggN8Nu",
    "outputId": "4b545dc9-55b2-460e-b1b2-a2faf6ce96a2"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Classifier(\n",
       "  (linear1): Linear(in_features=784, out_features=125, bias=True)\n",
       "  (linear2): Linear(in_features=125, out_features=65, bias=True)\n",
       "  (linear3): Linear(in_features=65, out_features=10, bias=True)\n",
       ")"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = Classifier(784, 125, 65, 10)\n",
    "model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "w6wxPBg_Od3t"
   },
   "outputs": [],
   "source": [
    "criterion = nn.CrossEntropyLoss()\n",
    "optimizer = torch.optim.Adam(model.parameters(), lr = 0.0001)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 782
    },
    "colab_type": "code",
    "id": "zsd3HPylP9UE",
    "outputId": "af979d67-6107-4d46-879c-6cd7745751df"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch : 1\n",
      "training loss: 0.9650, acc 76.6400 \n",
      "validation loss: 0.4183, validation acc 89.3300 \n",
      "epoch : 2\n",
      "training loss: 0.3733, acc 89.5717 \n",
      "validation loss: 0.3208, validation acc 90.9700 \n",
      "epoch : 3\n",
      "training loss: 0.3116, acc 91.0167 \n",
      "validation loss: 0.2792, validation acc 91.9300 \n",
      "epoch : 4\n",
      "training loss: 0.2776, acc 91.9667 \n",
      "validation loss: 0.2550, validation acc 92.7500 \n",
      "epoch : 5\n",
      "training loss: 0.2526, acc 92.6617 \n",
      "validation loss: 0.2384, validation acc 93.1500 \n",
      "epoch : 6\n",
      "training loss: 0.2322, acc 93.2500 \n",
      "validation loss: 0.2191, validation acc 93.6100 \n",
      "epoch : 7\n",
      "training loss: 0.2153, acc 93.7450 \n",
      "validation loss: 0.2052, validation acc 94.1300 \n",
      "epoch : 8\n",
      "training loss: 0.2003, acc 94.2467 \n",
      "validation loss: 0.1927, validation acc 94.2300 \n",
      "epoch : 9\n",
      "training loss: 0.1878, acc 94.5650 \n",
      "validation loss: 0.1835, validation acc 94.5200 \n",
      "epoch : 10\n",
      "training loss: 0.1765, acc 94.8283 \n",
      "validation loss: 0.1721, validation acc 94.8200 \n",
      "epoch : 11\n",
      "training loss: 0.1670, acc 95.1683 \n",
      "validation loss: 0.1686, validation acc 94.9800 \n",
      "epoch : 12\n",
      "training loss: 0.1566, acc 95.4250 \n",
      "validation loss: 0.1586, validation acc 94.9900 \n",
      "epoch : 13\n",
      "training loss: 0.1485, acc 95.7033 \n",
      "validation loss: 0.1531, validation acc 95.3900 \n",
      "epoch : 14\n",
      "training loss: 0.1408, acc 95.8850 \n",
      "validation loss: 0.1519, validation acc 95.1100 \n",
      "epoch : 15\n",
      "training loss: 0.1338, acc 96.1267 \n",
      "validation loss: 0.1407, validation acc 95.6700 \n"
     ]
    }
   ],
   "source": [
    "epochs = 15\n",
    "running_loss_history = []\n",
    "running_corrects_history = []\n",
    "val_running_loss_history = []\n",
    "val_running_corrects_history = []\n",
    "\n",
    "for e in range(epochs):\n",
    "  \n",
    "  running_loss = 0.0\n",
    "  running_corrects = 0.0\n",
    "  val_running_loss = 0.0\n",
    "  val_running_corrects = 0.0\n",
    "  \n",
    "  for inputs, labels in training_loader:\n",
    "    inputs = inputs.view(inputs.shape[0], -1)\n",
    "    outputs = model(inputs)\n",
    "    loss = criterion(outputs, labels)\n",
    "    \n",
    "    optimizer.zero_grad()\n",
    "    loss.backward()\n",
    "    optimizer.step()\n",
    "    \n",
    "    _, preds = torch.max(outputs, 1)\n",
    "    running_loss += loss.item()\n",
    "    running_corrects += torch.sum(preds == labels.data)\n",
    "\n",
    "  else:\n",
    "    with torch.no_grad():\n",
    "      for val_inputs, val_labels in validation_loader:\n",
    "        val_inputs = val_inputs.view(val_inputs.shape[0], -1)\n",
    "        val_outputs = model(val_inputs)\n",
    "        val_loss = criterion(val_outputs, val_labels)\n",
    "        \n",
    "        _, val_preds = torch.max(val_outputs, 1)\n",
    "        val_running_loss += val_loss.item()\n",
    "        val_running_corrects += torch.sum(val_preds == val_labels.data)\n",
    "      \n",
    "    epoch_loss = running_loss/len(training_loader)\n",
    "    epoch_acc = running_corrects.float()/ len(training_loader)\n",
    "    running_loss_history.append(epoch_loss)\n",
    "    running_corrects_history.append(epoch_acc)\n",
    "    \n",
    "    val_epoch_loss = val_running_loss/len(validation_loader)\n",
    "    val_epoch_acc = val_running_corrects.float()/ len(validation_loader)\n",
    "    val_running_loss_history.append(val_epoch_loss)\n",
    "    val_running_corrects_history.append(val_epoch_acc)\n",
    "    print('epoch :', (e+1))\n",
    "    print('training loss: {:.4f}, acc {:.4f} '.format(epoch_loss, epoch_acc.item()))\n",
    "    print('validation loss: {:.4f}, validation acc {:.4f} '.format(val_epoch_loss, val_epoch_acc.item()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 366
    },
    "colab_type": "code",
    "id": "qq4-LouHVQwp",
    "outputId": "5bd1f8dd-1f2c-4355-85f9-47d4206f2850"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7fd5a31a3d00>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(running_loss_history, label='training loss')\n",
    "plt.plot(val_running_loss_history, label='validation loss')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 364
    },
    "colab_type": "code",
    "id": "R1A-fOideMnx",
    "outputId": "de77cb7f-fa4d-4ad9-fec0-6baaf274cb37"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7fd5a3b7ab80>"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(running_corrects_history, label='training accuracy')\n",
    "plt.plot(val_running_corrects_history, label='validation accuracy')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Bdd5aONjGO5_"
   },
   "outputs": [],
   "source": [
    "import PIL.ImageOps"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 368
    },
    "colab_type": "code",
    "id": "gUPLBiV0gDd8",
    "outputId": "8b38d0a7-1791-4ef2-8994-dc9ea07e2aa3"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fd5a03ad8e0>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import requests\n",
    "from PIL import Image\n",
    "\n",
    "url = 'https://images.homedepot-static.com/productImages/007164ea-d47e-4f66-8d8c-fd9f621984a2/svn/architectural-mailboxes-house-letters-numbers-3585b-5-64_1000.jpg'\n",
    "response = requests.get(url, stream = True)\n",
    "img = Image.open(response.raw)\n",
    "plt.imshow(img)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 364
    },
    "colab_type": "code",
    "id": "v24yJujOERx8",
    "outputId": "f3c6a6ed-ca57-4e42-8e3b-aa3405b59bdc"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fd5a0304310>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "img = PIL.ImageOps.invert(img)\n",
    "img = img.convert('1')\n",
    "img = transform(img) \n",
    "plt.imshow(im_convert(img))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "Gd_XSROHFzsd",
    "outputId": "534f000c-1119-4fda-9021-ff8843280f6e"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5\n"
     ]
    }
   ],
   "source": [
    "img = img.view(img.shape[0], -1)\n",
    "output = model(img)\n",
    "_, pred = torch.max(output, 1)\n",
    "print(pred.item())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 303
    },
    "colab_type": "code",
    "id": "UE7haDIMG2cE",
    "outputId": "14ae08d1-b70f-4b6b-b47d-e631ab8a203f"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x288 with 20 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "dataiter = iter(validation_loader)\n",
    "images, labels = dataiter.next()\n",
    "images_ = images.view(images.shape[0], -1)\n",
    "output = model(images_)\n",
    "_, preds = torch.max(output, 1)\n",
    "\n",
    "fig = plt.figure(figsize=(25, 4))\n",
    "\n",
    "for idx in np.arange(20):\n",
    "  ax = fig.add_subplot(2, 10, idx+1, xticks=[], yticks=[])\n",
    "  plt.imshow(im_convert(images[idx]))\n",
    "  ax.set_title(\"{} ({})\".format(str(preds[idx].item()), str(labels[idx].item())), color=(\"green\" if preds[idx]==labels[idx] else \"red\"))"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "name": "MNIST.ipynb",
   "provenance": [],
   "version": "0.3.2"
  },
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.8.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}