CNN- 0.949.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "aa1f9b69",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "a1a25cf1",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = np.load('X.npy')\n",
    "Y = np.load('Y.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "442f695f",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import LabelEncoder\n",
    "Y = LabelEncoder().fit_transform(Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "4eb46ace",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = X.astype('float32')\n",
    "y = Y.astype('int64')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "d1bc9fd8",
   "metadata": {},
   "outputs": [],
   "source": [
    "X /= 255.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "70e1ceae",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.0, 1.0)"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.min(), X.max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "711b503f",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "dea82f77",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((52933, 1024), (52933,))"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape, y_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "7abec5eb",
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_example(X, y):\n",
    "    \"\"\"Plot the first 5 images and their labels in a row.\"\"\"\n",
    "    for i, (img, y) in enumerate(zip(X[:5].reshape(5, 32, 32), y[:5])):\n",
    "        plt.subplot(151 + i)\n",
    "        plt.imshow(img)\n",
    "        plt.xticks([])\n",
    "        plt.yticks([])\n",
    "        plt.title(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "d2be5e94",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 5 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_example(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "7921df69",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "import torch\n",
    "from torch import nn\n",
    "import torch.nn.functional as F\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "6cc19a9c",
   "metadata": {},
   "outputs": [],
   "source": [
    "device = 'cuda' if torch.cuda.is_available() else 'cpu'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "df87498b",
   "metadata": {},
   "outputs": [],
   "source": [
    "mnist_dim = X.shape[1]\n",
    "hidden_dim_1 = int(mnist_dim/2)\n",
    "hidden_dim_2 = int(mnist_dim/4)\n",
    "output_dim = len(np.unique(Y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "0cd7cd8b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1024, 512, 256, 36)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mnist_dim, hidden_dim_1, hidden_dim_2, output_dim"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "09f84421",
   "metadata": {},
   "outputs": [],
   "source": [
    "class ClassifierModule(nn.Module):\n",
    "    def __init__(\n",
    "            self,\n",
    "            input_dim=mnist_dim,\n",
    "            hidden_dim_1=hidden_dim_1,\n",
    "            hidden_dim_2=hidden_dim_2,\n",
    "            output_dim=output_dim,\n",
    "            dropout=0.5,\n",
    "    ):\n",
    "        super(ClassifierModule, self).__init__()\n",
    "        self.dropout = nn.Dropout(dropout)\n",
    "\n",
    "        self.hidden = nn.Linear(input_dim, hidden_dim_1)\n",
    "        self.hidden_1 = nn.Linear(hidden_dim_1, hidden_dim_2)\n",
    "        self.output = nn.Linear(hidden_dim_2, output_dim)\n",
    "\n",
    "    def forward(self, X, **kwargs):\n",
    "        X = F.relu(self.hidden(X))\n",
    "        X = F.relu(self.hidden_1(X))\n",
    "        X = self.dropout(X)\n",
    "        X = F.softmax(self.output(X), dim=-1)\n",
    "        return X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "95dad981",
   "metadata": {},
   "outputs": [],
   "source": [
    "from skorch import NeuralNetClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "b8c53257",
   "metadata": {},
   "outputs": [],
   "source": [
    "torch.manual_seed(0)\n",
    "\n",
    "net = NeuralNetClassifier(\n",
    "    ClassifierModule,\n",
    "    max_epochs=30,\n",
    "    lr=0.1,\n",
    "    device=device,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "06a18435",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  epoch    train_loss    valid_acc    valid_loss     dur\n",
      "-------  ------------  -----------  ------------  ------\n",
      "      1        \u001b[36m2.5648\u001b[0m       \u001b[32m0.5931\u001b[0m        \u001b[35m1.5213\u001b[0m  2.3415\n",
      "      2        \u001b[36m1.4606\u001b[0m       \u001b[32m0.6846\u001b[0m        \u001b[35m1.1561\u001b[0m  2.5916\n",
      "      3        \u001b[36m1.1682\u001b[0m       \u001b[32m0.7517\u001b[0m        \u001b[35m0.9086\u001b[0m  2.2726\n",
      "      4        \u001b[36m0.9525\u001b[0m       \u001b[32m0.7925\u001b[0m        \u001b[35m0.7487\u001b[0m  1.9419\n",
      "      5        \u001b[36m0.7916\u001b[0m       \u001b[32m0.8242\u001b[0m        \u001b[35m0.6318\u001b[0m  2.8519\n",
      "      6        \u001b[36m0.6697\u001b[0m       \u001b[32m0.8492\u001b[0m        \u001b[35m0.5492\u001b[0m  2.7987\n",
      "      7        \u001b[36m0.5774\u001b[0m       \u001b[32m0.8650\u001b[0m        \u001b[35m0.4879\u001b[0m  2.8008\n",
      "      8        \u001b[36m0.5061\u001b[0m       \u001b[32m0.8755\u001b[0m        \u001b[35m0.4447\u001b[0m  2.9177\n",
      "      9        \u001b[36m0.4469\u001b[0m       \u001b[32m0.8847\u001b[0m        \u001b[35m0.4136\u001b[0m  3.2855\n",
      "     10        \u001b[36m0.3958\u001b[0m       \u001b[32m0.8964\u001b[0m        \u001b[35m0.3733\u001b[0m  3.7357\n",
      "     11        \u001b[36m0.3580\u001b[0m       \u001b[32m0.8979\u001b[0m        \u001b[35m0.3690\u001b[0m  2.9977\n",
      "     12        \u001b[36m0.3271\u001b[0m       \u001b[32m0.9064\u001b[0m        \u001b[35m0.3370\u001b[0m  2.8678\n",
      "     13        \u001b[36m0.2937\u001b[0m       0.9049        \u001b[35m0.3357\u001b[0m  2.9287\n",
      "     14        \u001b[36m0.2633\u001b[0m       \u001b[32m0.9120\u001b[0m        \u001b[35m0.3115\u001b[0m  2.8322\n",
      "     15        \u001b[36m0.2404\u001b[0m       0.9112        0.3156  2.9000\n",
      "     16        \u001b[36m0.2242\u001b[0m       \u001b[32m0.9162\u001b[0m        \u001b[35m0.2945\u001b[0m  2.8465\n",
      "     17        \u001b[36m0.2076\u001b[0m       0.9136        0.2958  2.9169\n",
      "     18        \u001b[36m0.1843\u001b[0m       \u001b[32m0.9163\u001b[0m        0.2960  2.8747\n",
      "     19        \u001b[36m0.1704\u001b[0m       \u001b[32m0.9206\u001b[0m        \u001b[35m0.2779\u001b[0m  2.7785\n",
      "     20        \u001b[36m0.1602\u001b[0m       0.9204        \u001b[35m0.2756\u001b[0m  2.8231\n",
      "     21        \u001b[36m0.1441\u001b[0m       \u001b[32m0.9236\u001b[0m        \u001b[35m0.2723\u001b[0m  2.7944\n",
      "     22        \u001b[36m0.1364\u001b[0m       \u001b[32m0.9249\u001b[0m        \u001b[35m0.2627\u001b[0m  3.0555\n",
      "     23        \u001b[36m0.1275\u001b[0m       0.9242        0.2695  2.9975\n",
      "     24        \u001b[36m0.1176\u001b[0m       \u001b[32m0.9262\u001b[0m        \u001b[35m0.2530\u001b[0m  3.0422\n",
      "     25        \u001b[36m0.1108\u001b[0m       \u001b[32m0.9279\u001b[0m        \u001b[35m0.2463\u001b[0m  3.0343\n",
      "     26        \u001b[36m0.1013\u001b[0m       \u001b[32m0.9296\u001b[0m        0.2541  3.0527\n",
      "     27        \u001b[36m0.0982\u001b[0m       0.9282        0.2491  3.0163\n",
      "     28        \u001b[36m0.0885\u001b[0m       0.9273        0.2576  3.2160\n",
      "     29        \u001b[36m0.0831\u001b[0m       0.9274        0.2490  3.0709\n",
      "     30        \u001b[36m0.0798\u001b[0m       0.9287        0.2520  3.1212\n"
     ]
    }
   ],
   "source": [
    "net.fit(X_train, y_train);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "c61fd3d9",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import accuracy_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "bf76b2b4",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = net.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "a77c02fa",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.932275432133749"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy_score(y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "f5976b7b",
   "metadata": {},
   "outputs": [],
   "source": [
    "error_mask = y_pred != y_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "5a507cf2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 5 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_example(X_test[error_mask], y_pred[error_mask])  #SOME BAD PREDICTION"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "79073e51",
   "metadata": {},
   "source": [
    "# CNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "a70db259",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "32.0"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "1024/32\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "62801da9",
   "metadata": {},
   "outputs": [],
   "source": [
    "class Cnn(nn.Module):\n",
    "    def __init__(self, dropout=0.5):\n",
    "        super(Cnn, self).__init__()\n",
    "        self.conv1 = nn.Conv2d(1, 32, kernel_size=3)\n",
    "        self.conv2 = nn.Conv2d(32, 64, kernel_size=3)\n",
    "        self.conv2_drop = nn.Dropout2d(p=dropout)\n",
    "        self.fc1 = nn.Linear(2304, 100) # 2304 = number channels * width * height\n",
    "        self.fc2 = nn.Linear(100, 36)\n",
    "        self.fc1_drop = nn.Dropout(p=dropout)\n",
    "\n",
    "    def forward(self, x):\n",
    "        x = torch.relu(F.max_pool2d(self.conv1(x), 2))\n",
    "        x = torch.relu(F.max_pool2d(self.conv2_drop(self.conv2(x)), 2))\n",
    "        \n",
    "        # flatten over channel, height and width = 2304\n",
    "        x = x.view(-1, x.size(1) * x.size(2) * x.size(3))\n",
    "        \n",
    "        x = torch.relu(self.fc1_drop(self.fc1(x)))\n",
    "        x = torch.softmax(self.fc2(x), dim=-1)\n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "9f3264b2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "144.0"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "2304/16"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "858f55ab",
   "metadata": {},
   "outputs": [],
   "source": [
    "DEVICE = 'cuda' if torch.cuda.is_available() else 'cpu'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "915bd922",
   "metadata": {},
   "outputs": [],
   "source": [
    "torch.manual_seed(0)\n",
    "\n",
    "cnn = NeuralNetClassifier(\n",
    "    Cnn,\n",
    "    max_epochs=25,\n",
    "    lr=0.0002,\n",
    "    optimizer=torch.optim.Adam,\n",
    "    device=DEVICE,\n",
    "    iterator_train__num_workers=4,\n",
    "    iterator_valid__num_workers=4,\n",
    "    #callbacks=callbacks,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "352297d9",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X.reshape(70578,1,32,32), y, test_size=0.25, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "09957383",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  epoch    train_loss    valid_acc    valid_loss      dur\n",
      "-------  ------------  -----------  ------------  -------\n",
      "      1        \u001b[36m2.8813\u001b[0m       \u001b[32m0.6141\u001b[0m        \u001b[35m1.6206\u001b[0m  25.2352\n",
      "      2        \u001b[36m1.7045\u001b[0m       \u001b[32m0.7670\u001b[0m        \u001b[35m0.9489\u001b[0m  26.0941\n",
      "      3        \u001b[36m1.2740\u001b[0m       \u001b[32m0.8278\u001b[0m        \u001b[35m0.6850\u001b[0m  28.6138\n",
      "      4        \u001b[36m1.0532\u001b[0m       \u001b[32m0.8571\u001b[0m        \u001b[35m0.5456\u001b[0m  26.9171\n",
      "      5        \u001b[36m0.9133\u001b[0m       \u001b[32m0.8748\u001b[0m        \u001b[35m0.4627\u001b[0m  26.7618\n",
      "      6        \u001b[36m0.8311\u001b[0m       \u001b[32m0.8911\u001b[0m        \u001b[35m0.4095\u001b[0m  27.0445\n",
      "      7        \u001b[36m0.7611\u001b[0m       \u001b[32m0.8953\u001b[0m        \u001b[35m0.3818\u001b[0m  26.8944\n",
      "      8        \u001b[36m0.7174\u001b[0m       \u001b[32m0.9029\u001b[0m        \u001b[35m0.3470\u001b[0m  25.4342\n",
      "      9        \u001b[36m0.6787\u001b[0m       \u001b[32m0.9102\u001b[0m        \u001b[35m0.3257\u001b[0m  25.7061\n",
      "     10        \u001b[36m0.6430\u001b[0m       \u001b[32m0.9159\u001b[0m        \u001b[35m0.3053\u001b[0m  25.8426\n",
      "     11        \u001b[36m0.6108\u001b[0m       \u001b[32m0.9200\u001b[0m        \u001b[35m0.2911\u001b[0m  25.5205\n",
      "     12        \u001b[36m0.5944\u001b[0m       \u001b[32m0.9238\u001b[0m        \u001b[35m0.2748\u001b[0m  25.1723\n",
      "     13        \u001b[36m0.5740\u001b[0m       \u001b[32m0.9276\u001b[0m        \u001b[35m0.2600\u001b[0m  25.5823\n",
      "     14        \u001b[36m0.5550\u001b[0m       \u001b[32m0.9293\u001b[0m        \u001b[35m0.2538\u001b[0m  25.3695\n",
      "     15        \u001b[36m0.5392\u001b[0m       \u001b[32m0.9317\u001b[0m        \u001b[35m0.2442\u001b[0m  25.3345\n",
      "     16        \u001b[36m0.5263\u001b[0m       \u001b[32m0.9338\u001b[0m        \u001b[35m0.2358\u001b[0m  24.8898\n",
      "     17        \u001b[36m0.5067\u001b[0m       \u001b[32m0.9354\u001b[0m        \u001b[35m0.2232\u001b[0m  25.2163\n",
      "     18        \u001b[36m0.4956\u001b[0m       \u001b[32m0.9375\u001b[0m        \u001b[35m0.2208\u001b[0m  25.3298\n",
      "     19        \u001b[36m0.4926\u001b[0m       \u001b[32m0.9380\u001b[0m        \u001b[35m0.2160\u001b[0m  25.5248\n",
      "     20        \u001b[36m0.4738\u001b[0m       \u001b[32m0.9414\u001b[0m        \u001b[35m0.2063\u001b[0m  25.0147\n",
      "     21        \u001b[36m0.4647\u001b[0m       0.9401        \u001b[35m0.2024\u001b[0m  25.3116\n",
      "     22        \u001b[36m0.4559\u001b[0m       0.9412        \u001b[35m0.2011\u001b[0m  25.2367\n",
      "     23        \u001b[36m0.4506\u001b[0m       \u001b[32m0.9455\u001b[0m        \u001b[35m0.1922\u001b[0m  25.2545\n",
      "     24        \u001b[36m0.4379\u001b[0m       \u001b[32m0.9466\u001b[0m        \u001b[35m0.1899\u001b[0m  25.4966\n",
      "     25        \u001b[36m0.4299\u001b[0m       0.9466        \u001b[35m0.1868\u001b[0m  25.2290\n"
     ]
    }
   ],
   "source": [
    "cnn.fit(X_train, y=y_train);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "2d155d15",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9492774156984981\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.597489539748954"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred_cnn = cnn.predict(X_test)\n",
    "print(accuracy_score(y_test, y_pred_cnn))\n",
    "accuracy_score(y_test[error_mask], y_pred_cnn[error_mask])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3efb1aaf",
   "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": "3.9.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}