mlp.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "4c86fb3d-9047-47bd-96a7-c9a33671d5fd",
   "metadata": {},
   "source": [
    "# Import Object"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "eb72165e-471e-4db5-9110-e08e16a2768b",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn import preprocessing\n",
    "import torch\n",
    "from torch import nn\n",
    "from torch import optim\n",
    "from torch.utils.data import TensorDataset, DataLoader\n",
    "\n",
    "from models import MLP\n",
    "from train_model import train_net"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "93c714b2-e646-48eb-922b-503b0afb5226",
   "metadata": {},
   "source": [
    "# Prepare Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "d500a961-f492-4289-99f9-ebaa810b2f26",
   "metadata": {},
   "outputs": [],
   "source": [
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "6353efab-67c2-4c8d-aa75-6ed18486c1a6",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_x = pd.read_csv(\"X_train.csv\").values\n",
    "train_y = pd.read_csv(\"Y_train.csv\").values.reshape(-1)\n",
    "train_x, test_x, train_y, test_y = train_test_split(train_x,\n",
    "                                                   train_y,\n",
    "                                                   test_size=0.2)\n",
    "\n",
    "ss = preprocessing.StandardScaler()\n",
    "train_x = ss.fit_transform(train_x)\n",
    "test_x = ss.fit_transform(test_x)\n",
    "\n",
    "train_x = torch.tensor(train_x, dtype=torch.float32)\n",
    "train_y = torch.tensor(train_y, dtype=torch.int64)\n",
    "test_x = torch.tensor(test_x, dtype=torch.float32)\n",
    "test_y = torch.tensor(test_y, dtype=torch.int64)\n",
    "\n",
    "train_x = train_x.to(device)\n",
    "train_y = train_y.to(device)\n",
    "test_x = test_x.to(device)\n",
    "test_y = test_y.to(device)\n",
    "\n",
    "train_ds = TensorDataset(train_x, train_y)\n",
    "train_loader = DataLoader(train_ds, batch_size=24, shuffle=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ae3585ab-b1f9-42fd-b1c7-a1eec97d92aa",
   "metadata": {},
   "source": [
    "# Set Free Parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "4d9a3059-10ce-44c3-a405-a5bbfac5ca6a",
   "metadata": {},
   "outputs": [],
   "source": [
    "input_size = 8\n",
    "num_classes = 2\n",
    "learning_rate = 0.001\n",
    "num_epochs = 100\n",
    "dropout_ratio = 0.5"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "37760ecc-a587-4c5c-85f8-f7bec3271c97",
   "metadata": {},
   "source": [
    "# Instantiate Model, Criterion, Optimizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "c278addf-8af1-4c74-8fb5-b1b4a05ba7f0",
   "metadata": {},
   "outputs": [],
   "source": [
    "model = MLP(input_size, num_classes, dropout_ratio).to(device)\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "optimizer = optim.Adam(model.parameters(), lr=learning_rate)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b3c7dbf0-c20b-48bb-930c-1488667d4247",
   "metadata": {},
   "source": [
    "# Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "51f316a8-4b1a-4848-93dd-acf9fc82c50d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = train_net(model, criterion, optimizer,\n",
    "                  num_epochs, train_loader, test_x,\n",
    "                  test_y, device)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "41f1fd38-031f-460b-ba61-deb625bd8033",
   "metadata": {},
   "source": [
    "# Accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "71d0ce58-3be3-4b3a-baad-ea8671c06d4f",
   "metadata": {},
   "outputs": [],
   "source": [
    "pred_y = model(test_x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "b287c07f-3eca-4a3a-8625-9ad0e3ea6203",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor(0.8229, device='cuda:0')"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sum(pred_y.argmax(1) == test_y) / int(test_y.shape[0])"
   ]
  }
 ],
 "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": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}