04_Learning_Methods_on_Graphs.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## **Learning Methods on Graphs**\n",
    "- Implement our first graph neural network"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **Simple GCN layer with the Cora citation dataset**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Data(edge_index=[2, 10556], test_mask=[2708], train_mask=[2708], val_mask=[2708], x=[2708, 1433], y=[2708])\n"
     ]
    }
   ],
   "source": [
    "from torch_geometric.datasets import Planetoid\n",
    "dataset = Planetoid(root='/tmp/Cora', name='Cora')\n",
    "CORA = dataset[0]\n",
    "print(CORA)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### **Two-layer GCN implimentation**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn.functional as F\n",
    "from torch_geometric.nn import GCNConv"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Graph Convolutional Network(GCN) Layer : https://pytorch-geometric.readthedocs.io/en/latest/modules/nn.html#torch_geometric.nn.conv.GCNConv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Net(torch.nn.Module):\n",
    "    def __init__(self):\n",
    "        super(Net, self).__init__()\n",
    "        self.conv1 = GCNConv(dataset.num_node_features, 16)\n",
    "        self.conv2 = GCNConv(16, dataset.num_classes)\n",
    "\n",
    "    def forward(self, data):\n",
    "        x, edge_index = data.x, data.edge_index\n",
    "\n",
    "        x = self.conv1(x, edge_index)\n",
    "        x = F.relu(x)\n",
    "        x = F.dropout(x, training=self.training)\n",
    "        x = self.conv2(x, edge_index)\n",
    "\n",
    "        return F.log_softmax(x, dim=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
    "model = Net().to(device)\n",
    "data = dataset[0].to(device)\n",
    "optimizer = torch.optim.Adam(model.parameters(), lr=0.01, weight_decay=5e-4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "device(type='cuda')"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "device"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Net(\n",
       "  (conv1): GCNConv(1433, 16)\n",
       "  (conv2): GCNConv(16, 7)\n",
       ")"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Data(edge_index=[2, 10556], test_mask=[2708], train_mask=[2708], val_mask=[2708], x=[2708, 1433], y=[2708])"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Adam (\n",
       "Parameter Group 0\n",
       "    amsgrad: False\n",
       "    betas: (0.9, 0.999)\n",
       "    eps: 1e-08\n",
       "    lr: 0.01\n",
       "    weight_decay: 0.0005\n",
       ")"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "optimizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Net(\n",
       "  (conv1): GCNConv(1433, 16)\n",
       "  (conv2): GCNConv(16, 7)\n",
       ")"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.train()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "- Epoch 0 / Loss : 0.005529883783310652\n",
      "- Epoch 20 / Loss : 0.0068445513024926186\n",
      "- Epoch 40 / Loss : 0.008773078210651875\n",
      "- Epoch 60 / Loss : 0.008899000473320484\n",
      "- Epoch 80 / Loss : 0.008681017905473709\n",
      "- Epoch 100 / Loss : 0.008523937314748764\n",
      "- Epoch 120 / Loss : 0.008361315354704857\n",
      "- Epoch 140 / Loss : 0.008198386989533901\n",
      "- Epoch 160 / Loss : 0.008046497590839863\n",
      "- Epoch 180 / Loss : 0.007908221334218979\n"
     ]
    }
   ],
   "source": [
    "for epoch in range(200):\n",
    "    optimizer.zero_grad()\n",
    "    out = model(data)\n",
    "    loss = F.nll_loss(out[data.train_mask], data.y[data.train_mask])\n",
    "    if epoch % 20 == 0: print(\"- Epoch {} / Loss : {}\".format(epoch, loss))\n",
    "    loss.backward()\n",
    "    optimizer.step()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Net(\n",
       "  (conv1): GCNConv(1433, 16)\n",
       "  (conv2): GCNConv(16, 7)\n",
       ")"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.eval()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([3, 4, 4,  ..., 0, 3, 3], device='cuda:0')\n"
     ]
    }
   ],
   "source": [
    "_, pred = model(data).max(dim=1)\n",
    "print(pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([3, 4, 4,  ..., 3, 3, 3], device='cuda:0')"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.8120\n"
     ]
    }
   ],
   "source": [
    "correct = float (pred[data.test_mask].eq(data.y[data.test_mask]).sum().item())\n",
    "acc = correct / data.test_mask.sum().item()\n",
    "print('Accuracy: {:.4f}'.format(acc))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "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": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}