01_4_DeepWalk_and_Node2vec_Embedding.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "01.4: DeepWalk_and_Node2vec_Embedding.ipynb",
      "provenance": [],
      "include_colab_link": true
    },
    "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": "3.8.3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/gulabpatel/Graph_Neural_Network/blob/main/01.4%3A%20DeepWalk_and_Node2vec_Embedding.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kZqgl6b04WfT"
      },
      "source": [
        "# Tutorial 11: DeepWalk and node2vec - Implementation details\n",
        "  \n",
        "\n",
        "Paper:\n",
        "* [DeepWalk: Online Learning of Social Representation](https://arxiv.org/pdf/1403.6652.pdf)  \n",
        "* [node2vec: Scalable Feature Learning for Networks](https://arxiv.org/pdf/1607.00653.pdf)  \n",
        "\n",
        "Code:\n",
        "\n",
        " * [node2vec doc](https://pytorch-geometric.readthedocs.io/en/latest/modules/nn.html?highlight=node2vec#torch_geometric.nn.models.Node2Vec)\n",
        " * [node2vec code](https://pytorch-geometric.readthedocs.io/en/latest/_modules/torch_geometric/nn/models/node2vec.html)\n",
        " * [Example on clustering](https://github.com/rusty1s/pytorch_geometric/blob/master/examples/node2vec.py)\n",
        "\n",
        "\n",
        " Video walkthrough the code : \n",
        "\n",
        " * Part 1: https://www.youtube.com/watch?v=QZQBnl1QbCQ\n",
        " * Part 2: https://www.youtube.com/watch?v=5YOcpI3dB7I"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yRBP2yPC4Wfu"
      },
      "source": [
        "# Setup"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6Rt7xXTG-ZPV"
      },
      "source": [
        "# Install required packages.\n",
        "!pip install -q torch-cluster -f https://pytorch-geometric.com/whl/torch-1.9.0+cu102.html\n",
        "!pip install -q torch-scatter -f https://pytorch-geometric.com/whl/torch-1.9.0+cu102.html\n",
        "!pip install -q torch-sparse -f https://pytorch-geometric.com/whl/torch-1.9.0+cu102.html\n",
        "!pip install -q torch-geometric"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lWYjdX-v4Wfv"
      },
      "source": [
        "from torch_geometric.nn import Node2Vec\n",
        "import os.path as osp\n",
        "import torch\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn.manifold import TSNE\n",
        "from torch_geometric.datasets import Planetoid\n",
        "from tqdm.notebook import tqdm"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WUv2LvO_4Wfx"
      },
      "source": [
        "dataset = 'Cora'\n",
        "path = osp.join('.', 'data', dataset)\n",
        "dataset = Planetoid(path, dataset)\n",
        "data = dataset[0]"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-NGSxW39Hk1-",
        "outputId": "d4b62cdd-d2cc-4e22-dd9d-44ecf3fb0613"
      },
      "source": [
        "data"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Data(edge_index=[2, 10556], test_mask=[2708], train_mask=[2708], val_mask=[2708], x=[2708, 1433], y=[2708])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KdYQQ8jG4Wfy"
      },
      "source": [
        "device = 'cuda' if torch.cuda.is_available() else 'cpu'\n",
        "model = Node2Vec(data.edge_index, embedding_dim=128, \n",
        "                 walk_length=20,                        # lenght of rw\n",
        "                 context_size=10, walks_per_node=20,\n",
        "                 num_negative_samples=1, \n",
        "                 p=200, q=1,                             # bias parameters\n",
        "                 sparse=True).to(device)"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ymvZJsxlQAmG",
        "outputId": "00e872ba-f6d5-43d5-8da5-d3fa1485fc73"
      },
      "source": [
        "model"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Node2Vec(2708, 128)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8lgxytC64Wfz"
      },
      "source": [
        "# Random walks"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bD7rJUsr4Wf0"
      },
      "source": [
        "### The data loader"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bJ470hVh4Wf1",
        "outputId": "808b0256-0235-404a-d27e-c5e57553e9c6"
      },
      "source": [
        "loader = model.loader(batch_size=128, shuffle=True, num_workers=4)"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/torch/utils/data/dataloader.py:481: UserWarning: This DataLoader will create 4 worker processes in total. Our suggested max number of worker in current system is 2, which is smaller than what this DataLoader is going to create. Please be aware that excessive worker creation might get DataLoader running slow or even freeze, lower the worker number to avoid potential slowness/freeze if necessary.\n",
            "  cpuset_checked))\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "oK-asqUT4Wf3",
        "outputId": "53f4292c-662a-42dc-f5ac-18cff704b9d7"
      },
      "source": [
        "for idx, (pos_rw, neg_rw) in enumerate(loader):\n",
        "    print(idx, pos_rw.shape, neg_rw.shape)                  #pow_rw size = batch_size*(walk_length-context_size+1)*walks_per_node"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/torch/utils/data/dataloader.py:481: UserWarning: This DataLoader will create 4 worker processes in total. Our suggested max number of worker in current system is 2, which is smaller than what this DataLoader is going to create. Please be aware that excessive worker creation might get DataLoader running slow or even freeze, lower the worker number to avoid potential slowness/freeze if necessary.\n",
            "  cpuset_checked))\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "0 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "1 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "2 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "3 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "4 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "5 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "6 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "7 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "8 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "9 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "10 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "11 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "12 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "13 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "14 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "15 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "16 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "17 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "18 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "19 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "20 torch.Size([28160, 10]) torch.Size([28160, 10])\n",
            "21 torch.Size([4400, 10]) torch.Size([4400, 10])\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-svmS0Hl4Wf6",
        "outputId": "3880c8bc-88be-4b57-ed1c-afb678aa103a"
      },
      "source": [
        "idx, (pos_rw, neg_rw) = next(enumerate(loader))"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/torch/utils/data/dataloader.py:481: UserWarning: This DataLoader will create 4 worker processes in total. Our suggested max number of worker in current system is 2, which is smaller than what this DataLoader is going to create. Please be aware that excessive worker creation might get DataLoader running slow or even freeze, lower the worker number to avoid potential slowness/freeze if necessary.\n",
            "  cpuset_checked))\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bV6MTlvO4Wf7",
        "outputId": "af4c2add-da21-4b72-9ed8-d9329def4a3c"
      },
      "source": [
        "idx"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XjbfdyLR4Wf8",
        "outputId": "6b82c283-e72b-4089-e679-a50bbf0cc3a5"
      },
      "source": [
        "(pos_rw.shape, neg_rw.shape)"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(torch.Size([28160, 10]), torch.Size([28160, 10]))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DF2YzoHE4Wf-",
        "outputId": "2b2475b4-f3c7-4ae2-e7f8-a4fbd0546f79"
      },
      "source": [
        "pos_rw"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[ 670, 2606,  670,  ..., 2606,  670,  267],\n",
              "        [ 858, 2502,  858,  ..., 2502,  858, 2502],\n",
              "        [1808, 1810,  230,  ...,  465, 1403, 2317],\n",
              "        ...,\n",
              "        [1701, 1665,  962,  ..., 1621, 1424, 1621],\n",
              "        [2626, 2625, 2626,  ..., 2625, 2626, 2625],\n",
              "        [1831,  824, 1831,  ...,  824, 1831,  824]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kEZt_atO4Wf_",
        "outputId": "488d49b1-d5a0-4c1c-d317-90ae92db77b2"
      },
      "source": [
        "neg_rw"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[ 670,  111, 2484,  ...,  345,  178, 1018],\n",
              "        [ 858, 2347,  602,  ..., 1022, 2336, 2298],\n",
              "        [1808, 1770,  780,  ..., 2108, 1154,  355],\n",
              "        ...,\n",
              "        [2190,  794,  288,  ..., 2559, 2384, 2114],\n",
              "        [1998, 1139, 2648,  ..., 1724,  853, 1010],\n",
              "        [1727, 1734, 1535,  ..., 1883, 2683,   86]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lhvGJnjU4WgA"
      },
      "source": [
        "### Visualization"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9rVynzlD4WgA"
      },
      "source": [
        "import networkx as nx \n",
        "edge_tuples = [tuple(x) for x in data.edge_index.numpy().transpose()]\n",
        "G = nx.from_edgelist(edge_tuples)\n",
        "pos = nx.spring_layout(G, center=[0.5, 0.5])\n",
        "nx.set_node_attributes(G, pos, 'pos')"
      ],
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 629
        },
        "id": "c70sW-_K4WgB",
        "outputId": "8405fe0d-f456-4637-e600-3c7493d22d73"
      },
      "source": [
        "nodelist = next(enumerate(loader))[1][0][0].tolist()\n",
        "walk = nx.path_graph(len(nodelist))\n",
        "nx.set_node_attributes(walk, {idx: pos[node_id] for idx, node_id in enumerate(nodelist)}, 'pos')\n",
        "\n",
        "fig = plt.figure(figsize=(20, 10))\n",
        "ax = fig.add_subplot(1, 2, 1)\n",
        "nx.draw_networkx_nodes(G, \n",
        "                       ax=ax,\n",
        "                       pos=nx.get_node_attributes(G, 'pos'), \n",
        "                       node_size=5,\n",
        "                       alpha=0.3,\n",
        "                       node_color='b')\n",
        "nx.draw(walk, \n",
        "        node_size=40,\n",
        "        node_color='r',\n",
        "        ax=ax,\n",
        "        pos=nx.get_node_attributes(walk, 'pos'), \n",
        "        width=2,\n",
        "        edge_color='r') \n",
        "ax = fig.add_subplot(1, 2, 2)\n",
        "nx.draw(walk, \n",
        "        node_size=40,\n",
        "        node_color='r',\n",
        "        ax=ax,\n",
        "        pos=nx.get_node_attributes(walk, 'pos'), \n",
        "        width=2,\n",
        "        edge_color='r') "
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/torch/utils/data/dataloader.py:481: UserWarning: This DataLoader will create 4 worker processes in total. Our suggested max number of worker in current system is 2, which is smaller than what this DataLoader is going to create. Please be aware that excessive worker creation might get DataLoader running slow or even freeze, lower the worker number to avoid potential slowness/freeze if necessary.\n",
            "  cpuset_checked))\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x720 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "H94QVPGg4WgE"
      },
      "source": [
        "# Training"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wjUkTTjP4WgF"
      },
      "source": [
        "### Model definition"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "J4oatqbj4WgF",
        "outputId": "0ea828c4-ff95-490e-d0c6-e85a343d3cbb"
      },
      "source": [
        "model = Node2Vec(data.edge_index, embedding_dim=128, walk_length=20,\n",
        "                 context_size=10, walks_per_node=10,\n",
        "                 num_negative_samples=1, p=1, q=1, sparse=True).to(device)\n",
        "\n",
        "loader = model.loader(batch_size=128, shuffle=True, num_workers=4)\n",
        "optimizer = torch.optim.SparseAdam(list(model.parameters()), lr=0.01)"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/torch/utils/data/dataloader.py:481: UserWarning: This DataLoader will create 4 worker processes in total. Our suggested max number of worker in current system is 2, which is smaller than what this DataLoader is going to create. Please be aware that excessive worker creation might get DataLoader running slow or even freeze, lower the worker number to avoid potential slowness/freeze if necessary.\n",
            "  cpuset_checked))\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vPghn1WS4WgG"
      },
      "source": [
        "### Training function"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HYYLOtI64WgH"
      },
      "source": [
        "def train():\n",
        "    model.train()\n",
        "    total_loss = 0\n",
        "    for pos_rw, neg_rw in tqdm(loader):\n",
        "        optimizer.zero_grad()\n",
        "        loss = model.loss(pos_rw.to(device), neg_rw.to(device))\n",
        "        loss.backward()\n",
        "        optimizer.step()\n",
        "        total_loss += loss.item()\n",
        "    return total_loss / len(loader)"
      ],
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zIh8mGs04WgI"
      },
      "source": [
        "### Test function"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GrMzbm3b4WgI",
        "outputId": "9c7f1c4e-4267-4a03-9e5a-4610fc6dc052"
      },
      "source": [
        "data"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Data(edge_index=[2, 10556], test_mask=[2708], train_mask=[2708], val_mask=[2708], x=[2708, 1433], y=[2708])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LwYRXxxZ4WgJ",
        "outputId": "c6b1437c-dd43-4b88-dd2b-a51a0d214dcb"
      },
      "source": [
        "data.y"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([3, 4, 4,  ..., 3, 3, 3])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XUo1a1A44WgK",
        "outputId": "863ccaa1-f9db-4a42-d27c-3c76d39d4b95"
      },
      "source": [
        "data.y.unique()"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([0, 1, 2, 3, 4, 5, 6])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 20
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "f0m5TC444WgL"
      },
      "source": [
        "@torch.no_grad()\n",
        "def test():\n",
        "    model.eval()\n",
        "    z = model()\n",
        "    acc = model.test(z[data.train_mask], data.y[data.train_mask],\n",
        "                     z[data.test_mask], data.y[data.test_mask],\n",
        "                     max_iter=150)\n",
        "    return acc"
      ],
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "l82sUefb4WgM"
      },
      "source": [
        "### Training"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JJVYAiT44WgN"
      },
      "source": [
        "for epoch in range(1, 201):\n",
        "    loss = train()\n",
        "    acc = test()\n",
        "    print(f'Epoch: {epoch:02d}, Loss: {loss:.4f}, Acc: {acc:.4f}')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "z1_59eBo4WgQ"
      },
      "source": [
        "### Visualization"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kLJ2Okv44WgQ"
      },
      "source": [
        "@torch.no_grad()\n",
        "def plot_points(colors):\n",
        "    model.eval()\n",
        "    z = model(torch.arange(data.num_nodes, device=device))\n",
        "    z = TSNE(n_components=2).fit_transform(z.cpu().numpy())\n",
        "    y = data.y.cpu().numpy()\n",
        "\n",
        "    plt.figure(figsize=(8, 8))\n",
        "    for i in range(dataset.num_classes):\n",
        "        plt.scatter(z[y == i, 0], z[y == i, 1], s=20, color=colors[i])\n",
        "    plt.axis('off')\n",
        "    plt.show()\n",
        "\n",
        "colors = [\n",
        "    '#ffc0cb', '#bada55', '#008080', '#420420', '#7fe5f0', '#065535',\n",
        "    '#ffd700'\n",
        "]\n",
        "plot_points(colors)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0YOKDY-oH3YV"
      },
      "source": [
        "---------------------------------"
      ]
    }
  ]
}