Oversmoothing.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Oversmoothing.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true,
      "authorship_tag": "ABX9TyO8khhQCoSLn24TaQpCQBGM",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/SirArthur7/GCN-ISI-Summer-School/blob/main/Oversmoothing.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FucLXXhrntcD",
        "outputId": "daaece75-6fc6-40e2-89e4-80f90b930927"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Looking in links: https://data.pyg.org/whl/torch-+.html\n",
            "Collecting torch-scatter\n",
            "  Downloading torch_scatter-2.0.9.tar.gz (21 kB)\n",
            "Building wheels for collected packages: torch-scatter\n",
            "  Building wheel for torch-scatter (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for torch-scatter: filename=torch_scatter-2.0.9-cp37-cp37m-linux_x86_64.whl size=279639 sha256=aee7db1b9af58e30d9ab03d484bb94a06b5aa5282dc1d702e97f157d9144c0be\n",
            "  Stored in directory: /root/.cache/pip/wheels/dd/57/a3/42ea193b77378ce634eb9454c9bc1e3163f3b482a35cdee4d1\n",
            "Successfully built torch-scatter\n",
            "Installing collected packages: torch-scatter\n",
            "Successfully installed torch-scatter-2.0.9\n",
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Looking in links: https://data.pyg.org/whl/torch-+.html\n",
            "Collecting torch-sparse\n",
            "  Downloading torch_sparse-0.6.14.tar.gz (51 kB)\n",
            "\u001b[K     |████████████████████████████████| 51 kB 197 kB/s \n",
            "\u001b[?25hRequirement already satisfied: scipy in /usr/local/lib/python3.7/dist-packages (from torch-sparse) (1.4.1)\n",
            "Requirement already satisfied: numpy>=1.13.3 in /usr/local/lib/python3.7/dist-packages (from scipy->torch-sparse) (1.21.6)\n",
            "Building wheels for collected packages: torch-sparse\n",
            "  Building wheel for torch-sparse (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for torch-sparse: filename=torch_sparse-0.6.14-cp37-cp37m-linux_x86_64.whl size=514667 sha256=66f56bddd7b442d4984212a67603d07cdf746dd1e1c183cbeda46f6cf41729e3\n",
            "  Stored in directory: /root/.cache/pip/wheels/3c/aa/62/db0259eae2abce84f1ee2cf1c531bba683aab4bf79054172f8\n",
            "Successfully built torch-sparse\n",
            "Installing collected packages: torch-sparse\n",
            "Successfully installed torch-sparse-0.6.14\n",
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Collecting torch-geometric\n",
            "  Downloading torch_geometric-2.0.4.tar.gz (407 kB)\n",
            "\u001b[K     |████████████████████████████████| 407 kB 5.2 MB/s \n",
            "\u001b[?25hRequirement already satisfied: tqdm in /usr/local/lib/python3.7/dist-packages (from torch-geometric) (4.64.0)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.7/dist-packages (from torch-geometric) (1.21.6)\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.7/dist-packages (from torch-geometric) (1.4.1)\n",
            "Requirement already satisfied: pandas in /usr/local/lib/python3.7/dist-packages (from torch-geometric) (1.3.5)\n",
            "Requirement already satisfied: jinja2 in /usr/local/lib/python3.7/dist-packages (from torch-geometric) (2.11.3)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.7/dist-packages (from torch-geometric) (2.23.0)\n",
            "Requirement already satisfied: pyparsing in /usr/local/lib/python3.7/dist-packages (from torch-geometric) (3.0.9)\n",
            "Requirement already satisfied: scikit-learn in /usr/local/lib/python3.7/dist-packages (from torch-geometric) (1.0.2)\n",
            "Requirement already satisfied: MarkupSafe>=0.23 in /usr/local/lib/python3.7/dist-packages (from jinja2->torch-geometric) (2.0.1)\n",
            "Requirement already satisfied: pytz>=2017.3 in /usr/local/lib/python3.7/dist-packages (from pandas->torch-geometric) (2022.1)\n",
            "Requirement already satisfied: python-dateutil>=2.7.3 in /usr/local/lib/python3.7/dist-packages (from pandas->torch-geometric) (2.8.2)\n",
            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.7/dist-packages (from python-dateutil>=2.7.3->pandas->torch-geometric) (1.15.0)\n",
            "Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.7/dist-packages (from requests->torch-geometric) (3.0.4)\n",
            "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.7/dist-packages (from requests->torch-geometric) (2.10)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.7/dist-packages (from requests->torch-geometric) (2022.6.15)\n",
            "Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.7/dist-packages (from requests->torch-geometric) (1.24.3)\n",
            "Requirement already satisfied: joblib>=0.11 in /usr/local/lib/python3.7/dist-packages (from scikit-learn->torch-geometric) (1.1.0)\n",
            "Requirement already satisfied: threadpoolctl>=2.0.0 in /usr/local/lib/python3.7/dist-packages (from scikit-learn->torch-geometric) (3.1.0)\n",
            "Building wheels for collected packages: torch-geometric\n",
            "  Building wheel for torch-geometric (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for torch-geometric: filename=torch_geometric-2.0.4-py3-none-any.whl size=616603 sha256=388a6afbd1827a14e722d4a92e0e1934b6eef0ddda8ff9a806fc82d17ad3f0af\n",
            "  Stored in directory: /root/.cache/pip/wheels/18/a6/a4/ca18c3051fcead866fe7b85700ee2240d883562a1bc70ce421\n",
            "Successfully built torch-geometric\n",
            "Installing collected packages: torch-geometric\n",
            "Successfully installed torch-geometric-2.0.4\n"
          ]
        }
      ],
      "source": [
        "!pip install torch-scatter -f https://data.pyg.org/whl/torch-${TORCH}+${CUDA}.html\n",
        "!pip install torch-sparse -f https://data.pyg.org/whl/torch-${TORCH}+${CUDA}.html\n",
        "!pip install torch-geometric"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from torch_geometric.datasets import Planetoid\n",
        "\n",
        "datasets = Planetoid(root='/tmp/Cora', name='Cora')\n",
        "dataset = datasets[0]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hwRBDCSqnxHl",
        "outputId": "27981efe-a1a2-44f9-dc69-b3715457c6e5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "Downloading https://github.com/kimiyoung/planetoid/raw/master/data/ind.cora.x\n",
            "Downloading https://github.com/kimiyoung/planetoid/raw/master/data/ind.cora.tx\n",
            "Downloading https://github.com/kimiyoung/planetoid/raw/master/data/ind.cora.allx\n",
            "Downloading https://github.com/kimiyoung/planetoid/raw/master/data/ind.cora.y\n",
            "Downloading https://github.com/kimiyoung/planetoid/raw/master/data/ind.cora.ty\n",
            "Downloading https://github.com/kimiyoung/planetoid/raw/master/data/ind.cora.ally\n",
            "Downloading https://github.com/kimiyoung/planetoid/raw/master/data/ind.cora.graph\n",
            "Downloading https://github.com/kimiyoung/planetoid/raw/master/data/ind.cora.test.index\n",
            "Processing...\n",
            "Done!\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import torch \n",
        "\n",
        "def adj_matrix(dataset):\n",
        "  adj_mat = torch.zeros([dataset.num_nodes, dataset.num_nodes])\n",
        "\n",
        "  for i in range(dataset.num_edges):\n",
        "    s = dataset.edge_index[0][i]\n",
        "    t = dataset.edge_index[1][i]\n",
        "    adj_mat[s][t]=1\n",
        "  \n",
        "  return adj_mat\n",
        "\n",
        "def improv_adj_matrix(adj_mat):\n",
        "  adj_mat += torch.eye(adj_mat.shape[0])\n",
        "  return adj_mat\n",
        "\n",
        "def norm_degree_mat(adj_mat):\n",
        "  norm_diag = torch.diag(torch.sum(adj_mat, axis=1)**(-0.5))\n",
        "  return norm_diag\n",
        "\n",
        "def Lapl_norm(dataset):\n",
        "  A_i = improv_adj_matrix(adj_matrix(dataset))\n",
        "  D_inv = norm_degree_mat(A_i)\n",
        "  L = torch.mm(D_inv,A_i)\n",
        "  L= torch.mm(L,D_inv)\n",
        "  return L\n",
        "\n",
        "print(Lapl_norm(dataset))\n",
        "  # print(adj_mat[500,400],adj_mat[400,500])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0Etd022wnzF8",
        "outputId": "17bdbc59-fae7-48b0-a498-de0144ddb02e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "tensor([[0.2500, 0.0000, 0.0000,  ..., 0.0000, 0.0000, 0.0000],\n",
            "        [0.0000, 0.2500, 0.2041,  ..., 0.0000, 0.0000, 0.0000],\n",
            "        [0.0000, 0.2041, 0.1667,  ..., 0.0000, 0.0000, 0.0000],\n",
            "        ...,\n",
            "        [0.0000, 0.0000, 0.0000,  ..., 0.5000, 0.0000, 0.0000],\n",
            "        [0.0000, 0.0000, 0.0000,  ..., 0.0000, 0.2000, 0.2000],\n",
            "        [0.0000, 0.0000, 0.0000,  ..., 0.0000, 0.2000, 0.2000]])\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import torch\n",
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "import numpy as np\n",
        "import math\n",
        "# from torchinfo import summary\n",
        "  \n",
        "class GCN(torch.nn.Module):\n",
        "\n",
        "    def __init__(self, D_in_feat, H_feat, D_out_classes, num_layers):\n",
        "        super(GCN, self).__init__()\n",
        "        self.layers = [nn.Linear(D_in_feat, H_feat)]\n",
        "        for _ in range(num_layers-2):\n",
        "          self.layers.append(nn.Linear(H_feat,H_feat))\n",
        "        self.layers.append(nn.Linear(H_feat, D_out_classes))\n",
        "        self.layers = nn.Sequential(*self.layers)\n",
        "        \n",
        "        for i in range(num_layers-1):\n",
        "          nn.init.kaiming_normal_(self.layers[i].weight, mode='fan_in', nonlinearity='relu')\n",
        "          # nn.init.xavier_uniform_(self.layers[i].weight, gain= math.sqrt(2))        \n",
        "        nn.init.xavier_uniform_(self.layers[num_layers-1].weight)\n",
        "    \n",
        "    def forward(self, dataset):\n",
        "        x = dataset.x\n",
        "        L = Lapl_norm(dataset)\n",
        "        for i in range(len(self.layers)-1):\n",
        "          x = torch.mm(L, x)\n",
        "          # x = F.relu(self.bnorm1(self.linear1(x)))\n",
        "          x = F.relu(self.layers[i](x))\n",
        "          x = F.dropout(x, p = 0.5, training=self.training)\n",
        "        x = torch.mm(L, x)\n",
        "        x = self.layers[len(self.layers)-1](x)\n",
        "        x = F.log_softmax(x, dim=1)\n",
        "        return x\n",
        "\n",
        "model1 = GCN(dataset.num_node_features,32,7,4)\n",
        "print(model1.parameters)\n",
        "model1.forward(dataset).shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HPQaR_IBn18g",
        "outputId": "683c8b0a-e26d-406f-ca67-7caa670cec65"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<bound method Module.parameters of GCN(\n",
            "  (layers): Sequential(\n",
            "    (0): Linear(in_features=1433, out_features=32, bias=True)\n",
            "    (1): Linear(in_features=32, out_features=32, bias=True)\n",
            "    (2): Linear(in_features=32, out_features=32, bias=True)\n",
            "    (3): Linear(in_features=32, out_features=7, bias=True)\n",
            "  )\n",
            ")>\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([2708, 7])"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def train(model, criterion, dataset, optimizer, epochs=100):\n",
        "    i = 0\n",
        "    loss_and_acc = {'training_loss':[], 'validation_accuracy':[]}  \n",
        "\n",
        "    for epoch in range(epochs):        \n",
        "          model.train()\n",
        "          optimizer.zero_grad()\n",
        "          z = model(dataset)\n",
        "          x = z[dataset.train_mask]\n",
        "          y = dataset.y[dataset.train_mask]\n",
        "          # print(x.shape, y.shape)\n",
        "          loss = criterion(x, y)\n",
        "          print(type(loss))\n",
        "          loss.backward()\n",
        "          optimizer.step()\n",
        "          loss_and_acc['training_loss'].append(loss.data.item())\n",
        "          print(epoch,\"\\t\",loss.data.item())\n",
        "            \n",
        "    correct = 0\n",
        "    model.eval()\n",
        "    z = model(dataset)\n",
        "    valid_z = z[dataset.val_mask]\n",
        "    valid_y = dataset.y[dataset.val_mask]\n",
        "    _, label = torch.max(valid_z, 1)\n",
        "    correct += (label == valid_y).sum().item()\n",
        "    print(valid_y.shape, \"\\t\", correct)  \n",
        "    accuracy = 100 * (correct / len(valid_y))\n",
        "    loss_and_acc['validation_accuracy'].append(accuracy)\n",
        "    \n",
        "    return loss_and_acc"
      ],
      "metadata": {
        "id": "vHsv0lwMn40T"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "criterion = nn.CrossEntropyLoss()\n",
        "epochs = 200\n",
        "learning_rate = 0.01\n",
        "input_dim = dataset.num_node_features\n",
        "hidden_dim = 16\n",
        "output_dim = 7\n",
        "num_layer = 256\n",
        "\n",
        "model_norm  = GCN(input_dim, hidden_dim, output_dim, num_layer)\n",
        "optimizer = torch.optim.Adam(model_norm.parameters(), lr=learning_rate, weight_decay = 5e-4)\n",
        "training_results_Norm=train(model_norm, criterion, dataset, optimizer, epochs)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "ICUIDNOCn7l7",
        "outputId": "0f914c95-6783-44cd-fca8-829e2dc2a1cd"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<class 'torch.Tensor'>\n",
            "0 \t 1095.2098388671875\n",
            "<class 'torch.Tensor'>\n",
            "1 \t 677.9274291992188\n",
            "<class 'torch.Tensor'>\n",
            "2 \t 214.1556854248047\n",
            "<class 'torch.Tensor'>\n",
            "3 \t 2.4839682579040527\n",
            "<class 'torch.Tensor'>\n",
            "4 \t 2.9291012287139893\n",
            "<class 'torch.Tensor'>\n",
            "5 \t 2.117098808288574\n",
            "<class 'torch.Tensor'>\n",
            "6 \t 1.9469804763793945\n",
            "<class 'torch.Tensor'>\n",
            "7 \t 1.9933924674987793\n",
            "<class 'torch.Tensor'>\n",
            "8 \t 110.6950912475586\n",
            "<class 'torch.Tensor'>\n",
            "9 \t 1.9327902793884277\n",
            "<class 'torch.Tensor'>\n",
            "10 \t 1.9459433555603027\n",
            "<class 'torch.Tensor'>\n",
            "11 \t 1.9856561422348022\n",
            "<class 'torch.Tensor'>\n",
            "12 \t 3.9179251194000244\n",
            "<class 'torch.Tensor'>\n",
            "13 \t 1.944671630859375\n",
            "<class 'torch.Tensor'>\n",
            "14 \t 1.9431904554367065\n",
            "<class 'torch.Tensor'>\n",
            "15 \t 1.9816641807556152\n",
            "<class 'torch.Tensor'>\n",
            "16 \t 2.000216484069824\n",
            "<class 'torch.Tensor'>\n",
            "17 \t 1.9400174617767334\n",
            "<class 'torch.Tensor'>\n",
            "18 \t 1.9418877363204956\n",
            "<class 'torch.Tensor'>\n",
            "19 \t 1.9384303092956543\n",
            "<class 'torch.Tensor'>\n",
            "20 \t 1.9322808980941772\n",
            "<class 'torch.Tensor'>\n",
            "21 \t 1.945523738861084\n",
            "<class 'torch.Tensor'>\n",
            "22 \t 1.9301973581314087\n",
            "<class 'torch.Tensor'>\n",
            "23 \t 1.9378607273101807\n",
            "<class 'torch.Tensor'>\n",
            "24 \t 1.9329838752746582\n",
            "<class 'torch.Tensor'>\n",
            "25 \t 1.935915231704712\n",
            "<class 'torch.Tensor'>\n",
            "26 \t 1.9411581754684448\n",
            "<class 'torch.Tensor'>\n",
            "27 \t 1.944161295890808\n",
            "<class 'torch.Tensor'>\n",
            "28 \t 1.942112684249878\n",
            "<class 'torch.Tensor'>\n",
            "29 \t 1.9350864887237549\n",
            "<class 'torch.Tensor'>\n",
            "30 \t 1.9341251850128174\n",
            "<class 'torch.Tensor'>\n",
            "31 \t 1.9414620399475098\n",
            "<class 'torch.Tensor'>\n",
            "32 \t 1.9372425079345703\n",
            "<class 'torch.Tensor'>\n",
            "33 \t 1.9473109245300293\n",
            "<class 'torch.Tensor'>\n",
            "34 \t 1.986440658569336\n",
            "<class 'torch.Tensor'>\n",
            "35 \t 1.9503884315490723\n",
            "<class 'torch.Tensor'>\n",
            "36 \t 1.933774471282959\n",
            "<class 'torch.Tensor'>\n",
            "37 \t 1.937339186668396\n",
            "<class 'torch.Tensor'>\n",
            "38 \t 1.9505730867385864\n",
            "<class 'torch.Tensor'>\n",
            "39 \t 1.9296588897705078\n",
            "<class 'torch.Tensor'>\n",
            "40 \t 1.9346634149551392\n",
            "<class 'torch.Tensor'>\n",
            "41 \t 1.9431273937225342\n",
            "<class 'torch.Tensor'>\n",
            "42 \t 1.9473811388015747\n",
            "<class 'torch.Tensor'>\n",
            "43 \t 1.935114860534668\n",
            "<class 'torch.Tensor'>\n",
            "44 \t 1.9321415424346924\n",
            "<class 'torch.Tensor'>\n",
            "45 \t 1.9364053010940552\n",
            "<class 'torch.Tensor'>\n",
            "46 \t 1.9335747957229614\n",
            "<class 'torch.Tensor'>\n",
            "47 \t 1.9356026649475098\n",
            "<class 'torch.Tensor'>\n",
            "48 \t 1.9365166425704956\n",
            "<class 'torch.Tensor'>\n",
            "49 \t 2.221137523651123\n",
            "<class 'torch.Tensor'>\n",
            "50 \t 1.9598612785339355\n",
            "<class 'torch.Tensor'>\n",
            "51 \t 2.0020439624786377\n",
            "<class 'torch.Tensor'>\n",
            "52 \t 1.945700764656067\n",
            "<class 'torch.Tensor'>\n",
            "53 \t 1.9323804378509521\n",
            "<class 'torch.Tensor'>\n",
            "54 \t 1.9964972734451294\n",
            "<class 'torch.Tensor'>\n",
            "55 \t 1.920566201210022\n",
            "<class 'torch.Tensor'>\n",
            "56 \t 1.927891492843628\n",
            "<class 'torch.Tensor'>\n",
            "57 \t 1.9466736316680908\n",
            "<class 'torch.Tensor'>\n",
            "58 \t 1.9402865171432495\n",
            "<class 'torch.Tensor'>\n",
            "59 \t 1.9310154914855957\n",
            "<class 'torch.Tensor'>\n",
            "60 \t 1.934807300567627\n",
            "<class 'torch.Tensor'>\n",
            "61 \t 1.942234992980957\n",
            "<class 'torch.Tensor'>\n",
            "62 \t 1.9360193014144897\n",
            "<class 'torch.Tensor'>\n",
            "63 \t 1.9280390739440918\n",
            "<class 'torch.Tensor'>\n",
            "64 \t 1.9409364461898804\n",
            "<class 'torch.Tensor'>\n",
            "65 \t 1.9417517185211182\n",
            "<class 'torch.Tensor'>\n",
            "66 \t 1.9261990785598755\n",
            "<class 'torch.Tensor'>\n",
            "67 \t 1.9390535354614258\n",
            "<class 'torch.Tensor'>\n",
            "68 \t 1.9341973066329956\n",
            "<class 'torch.Tensor'>\n",
            "69 \t 1.941686749458313\n"
          ]
        },
        {
          "output_type": "error",
          "ename": "KeyboardInterrupt",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-7-838c22355854>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      9\u001b[0m \u001b[0mmodel_norm\u001b[0m  \u001b[0;34m=\u001b[0m \u001b[0mGCN\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput_dim\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mhidden_dim\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moutput_dim\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnum_layer\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     10\u001b[0m \u001b[0moptimizer\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0moptim\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mAdam\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel_norm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mparameters\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlr\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mlearning_rate\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mweight_decay\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m5e-4\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 11\u001b[0;31m \u001b[0mtraining_results_Norm\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mtrain\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel_norm\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcriterion\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdataset\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moptimizer\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mepochs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;32m<ipython-input-6-4d53e87dc0c0>\u001b[0m in \u001b[0;36mtrain\u001b[0;34m(model, criterion, dataset, optimizer, epochs)\u001b[0m\n\u001b[1;32m      6\u001b[0m           \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m           \u001b[0moptimizer\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mzero_grad\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 8\u001b[0;31m           \u001b[0mz\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdataset\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      9\u001b[0m           \u001b[0mx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mz\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mdataset\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain_mask\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     10\u001b[0m           \u001b[0my\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdataset\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0my\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mdataset\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain_mask\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m_call_impl\u001b[0;34m(self, *input, **kwargs)\u001b[0m\n\u001b[1;32m   1108\u001b[0m         if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks or _global_backward_hooks\n\u001b[1;32m   1109\u001b[0m                 or _global_forward_hooks or _global_forward_pre_hooks):\n\u001b[0;32m-> 1110\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mforward_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1111\u001b[0m         \u001b[0;31m# Do not call functions when jit is used\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1112\u001b[0m         \u001b[0mfull_backward_hooks\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnon_full_backward_hooks\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m<ipython-input-5-7af65f1c210a>\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, dataset)\u001b[0m\n\u001b[1;32m     23\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdataset\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     24\u001b[0m         \u001b[0mx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdataset\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 25\u001b[0;31m         \u001b[0mL\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mLapl_norm\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdataset\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     26\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlayers\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     27\u001b[0m           \u001b[0mx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmm\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mL\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m<ipython-input-4-97a0cec1b4e2>\u001b[0m in \u001b[0;36mLapl_norm\u001b[0;34m(dataset)\u001b[0m\n\u001b[1;32m     23\u001b[0m   \u001b[0mD_inv\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnorm_degree_mat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mA_i\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     24\u001b[0m   \u001b[0mL\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmm\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mD_inv\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mA_i\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 25\u001b[0;31m   \u001b[0mL\u001b[0m\u001b[0;34m=\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmm\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mL\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mD_inv\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     26\u001b[0m   \u001b[0;32mreturn\u001b[0m \u001b[0mL\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     27\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from matplotlib import pyplot as plt\n",
        "\n",
        "x = [i+1 for i in range(200)] \n",
        "y = training_results_Norm['training_loss']\n",
        "\n",
        "print(training_results_Norm['validation_accuracy'])\n",
        "plt.figure(figsize=(32, 8), dpi=200)\n",
        "plt.scatter(x, y, c =\"blue\")\n",
        " \n",
        "plt.show()"
      ],
      "metadata": {
        "id": "UB2iKhfNn8Ux"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "PATH = \"state_dict_model.pt\" \n",
        "torch.save(model_norm.state_dict(), PATH)\n",
        "\n",
        "model = GCN(input_dim, hidden_dim, output_dim, num_layer)\n",
        "model.load_state_dict(torch.load(PATH))\n",
        "\n",
        "model.eval()\n",
        "z = model(dataset)\n",
        "test_z = z[dataset.test_mask]\n",
        "test_y = dataset.y[dataset.test_mask]\n",
        "\n",
        "correct = 0\n",
        "_, label = torch.max(test_z, 1)\n",
        "correct += (label == test_y).sum().item()\n",
        "print(test_y.shape, \"\\t\", correct)\n",
        "test_acc = 100 * (correct / len(test_y))\n",
        "print(test_acc)\n",
        "\n",
        "from sklearn.manifold import TSNE\n",
        "\n",
        "colors = []\n",
        "for i in range(len(test_z)):\n",
        "  if test_y[i]==0:\n",
        "    colors.append(\"orange\");\n",
        "  elif test_y[i]==1:\n",
        "    colors.append(\"red\");\n",
        "  elif test_y[i]==2:\n",
        "    colors.append(\"green\");\n",
        "  elif test_y[i]==3:\n",
        "    colors.append(\"violet\");\n",
        "  elif test_y[i]==4:\n",
        "    colors.append(\"limegreen\");\n",
        "  elif test_y[i]==5:\n",
        "    colors.append(\"blue\");\n",
        "  elif test_y[i]==6:\n",
        "    colors.append(\"yellow\");\n",
        "\n",
        "tsne = TSNE(n_components=2, random_state=0, perplexity=10)\n",
        "tsne_results = tsne.fit_transform(test_z.detach().numpy())\n",
        "\n",
        "plt.figure(figsize=(6, 4))\n",
        "plt.scatter(tsne_results[:,0], tsne_results[:,1], c= colors)"
      ],
      "metadata": {
        "id": "M8p7Bu0In-6y"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "datasets2 = Planetoid(root='/tmp/CiteSeer', name='CiteSeer')\n",
        "dataset2 = datasets2[0]\n",
        "\n",
        "criterion = nn.CrossEntropyLoss()\n",
        "epochs = 200\n",
        "learning_rate = 0.01\n",
        "input_dim = dataset2.num_node_features\n",
        "hidden_dim = 16\n",
        "output_dim = 6\n",
        "num_layer = 1024\n",
        "\n",
        "model_norm  = GCN(input_dim, hidden_dim, output_dim, num_layer)\n",
        "optimizer = torch.optim.Adam(model_norm.parameters(), lr=learning_rate, weight_decay = 5e-4)\n",
        "training_results_Norm=train(model_norm, criterion, dataset2, optimizer, epochs)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qRriyDCmgck5",
        "outputId": "82b9cc82-6b7f-42b9-c57a-928cc420b83c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "Downloading https://github.com/kimiyoung/planetoid/raw/master/data/ind.citeseer.x\n",
            "Downloading https://github.com/kimiyoung/planetoid/raw/master/data/ind.citeseer.tx\n",
            "Downloading https://github.com/kimiyoung/planetoid/raw/master/data/ind.citeseer.allx\n",
            "Downloading https://github.com/kimiyoung/planetoid/raw/master/data/ind.citeseer.y\n",
            "Downloading https://github.com/kimiyoung/planetoid/raw/master/data/ind.citeseer.ty\n",
            "Downloading https://github.com/kimiyoung/planetoid/raw/master/data/ind.citeseer.ally\n",
            "Downloading https://github.com/kimiyoung/planetoid/raw/master/data/ind.citeseer.graph\n",
            "Downloading https://github.com/kimiyoung/planetoid/raw/master/data/ind.citeseer.test.index\n",
            "Processing...\n",
            "Done!\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<class 'torch.Tensor'>\n",
            "0 \t 6.325865602938307e+17\n",
            "<class 'torch.Tensor'>\n",
            "1 \t 17436581888.0\n",
            "<class 'torch.Tensor'>\n",
            "2 \t 29.958698272705078\n",
            "<class 'torch.Tensor'>\n",
            "3 \t 8.119136810302734\n",
            "<class 'torch.Tensor'>\n",
            "4 \t 6.974127769470215\n",
            "<class 'torch.Tensor'>\n",
            "5 \t 119.38877868652344\n",
            "<class 'torch.Tensor'>\n",
            "6 \t 2.886352300643921\n",
            "<class 'torch.Tensor'>\n",
            "7 \t 1.903768539428711\n",
            "<class 'torch.Tensor'>\n",
            "8 \t 3.193369150161743\n",
            "<class 'torch.Tensor'>\n",
            "9 \t 1.7944297790527344\n",
            "<class 'torch.Tensor'>\n",
            "10 \t 1.8422412872314453\n",
            "<class 'torch.Tensor'>\n",
            "11 \t 1.854284405708313\n",
            "<class 'torch.Tensor'>\n",
            "12 \t 1.9477477073669434\n",
            "<class 'torch.Tensor'>\n",
            "13 \t 2.5784192085266113\n",
            "<class 'torch.Tensor'>\n",
            "14 \t 1.9470703601837158\n",
            "<class 'torch.Tensor'>\n",
            "15 \t 1.8208383321762085\n",
            "<class 'torch.Tensor'>\n",
            "16 \t 2.261073589324951\n",
            "<class 'torch.Tensor'>\n",
            "17 \t 1.8063828945159912\n",
            "<class 'torch.Tensor'>\n",
            "18 \t 1.8205515146255493\n",
            "<class 'torch.Tensor'>\n",
            "19 \t 1.8020488023757935\n",
            "<class 'torch.Tensor'>\n",
            "20 \t 2.264711856842041\n",
            "<class 'torch.Tensor'>\n",
            "21 \t 1.8218979835510254\n",
            "<class 'torch.Tensor'>\n",
            "22 \t 1.9179496765136719\n",
            "<class 'torch.Tensor'>\n",
            "23 \t 30.053802490234375\n",
            "<class 'torch.Tensor'>\n",
            "24 \t 1.7936972379684448\n",
            "<class 'torch.Tensor'>\n",
            "25 \t 2.743823289871216\n",
            "<class 'torch.Tensor'>\n",
            "26 \t 1.8026220798492432\n",
            "<class 'torch.Tensor'>\n",
            "27 \t 1.8099106550216675\n",
            "<class 'torch.Tensor'>\n",
            "28 \t 8.234309196472168\n",
            "<class 'torch.Tensor'>\n",
            "29 \t 1.7993494272232056\n",
            "<class 'torch.Tensor'>\n",
            "30 \t 2.3620502948760986\n",
            "<class 'torch.Tensor'>\n",
            "31 \t 2.879871129989624\n",
            "<class 'torch.Tensor'>\n",
            "32 \t 1.7828028202056885\n",
            "<class 'torch.Tensor'>\n",
            "33 \t 1.7779786586761475\n",
            "<class 'torch.Tensor'>\n",
            "34 \t 1.8010761737823486\n",
            "<class 'torch.Tensor'>\n",
            "35 \t 4.316243648529053\n",
            "<class 'torch.Tensor'>\n",
            "36 \t 1.7816678285598755\n",
            "<class 'torch.Tensor'>\n",
            "37 \t 1.7947393655776978\n",
            "<class 'torch.Tensor'>\n",
            "38 \t 12.201647758483887\n",
            "<class 'torch.Tensor'>\n",
            "39 \t 1.7884001731872559\n",
            "<class 'torch.Tensor'>\n",
            "40 \t 1.7715429067611694\n",
            "<class 'torch.Tensor'>\n",
            "41 \t 1.7894883155822754\n",
            "<class 'torch.Tensor'>\n",
            "42 \t 4.830505847930908\n",
            "<class 'torch.Tensor'>\n",
            "43 \t 1.775356411933899\n",
            "<class 'torch.Tensor'>\n",
            "44 \t 1.8077188730239868\n",
            "<class 'torch.Tensor'>\n",
            "45 \t 1.798171877861023\n",
            "<class 'torch.Tensor'>\n",
            "46 \t 1.8022894859313965\n",
            "<class 'torch.Tensor'>\n",
            "47 \t 1.7979358434677124\n",
            "<class 'torch.Tensor'>\n",
            "48 \t 1.798776626586914\n",
            "<class 'torch.Tensor'>\n",
            "49 \t 196.7509765625\n",
            "<class 'torch.Tensor'>\n",
            "50 \t 1.7963731288909912\n",
            "<class 'torch.Tensor'>\n",
            "51 \t 2.2015979290008545\n",
            "<class 'torch.Tensor'>\n",
            "52 \t 1.796288013458252\n",
            "<class 'torch.Tensor'>\n",
            "53 \t 1.815017580986023\n",
            "<class 'torch.Tensor'>\n",
            "54 \t 1.7887104749679565\n",
            "<class 'torch.Tensor'>\n",
            "55 \t 1.7956007719039917\n",
            "<class 'torch.Tensor'>\n",
            "56 \t 1.7967915534973145\n",
            "<class 'torch.Tensor'>\n",
            "57 \t 1.7890664339065552\n",
            "<class 'torch.Tensor'>\n",
            "58 \t 1.7765774726867676\n",
            "<class 'torch.Tensor'>\n",
            "59 \t 1.7978826761245728\n",
            "<class 'torch.Tensor'>\n",
            "60 \t 1.8131871223449707\n",
            "<class 'torch.Tensor'>\n",
            "61 \t 34.034461975097656\n",
            "<class 'torch.Tensor'>\n",
            "62 \t 16.143495559692383\n",
            "<class 'torch.Tensor'>\n",
            "63 \t 1.7890758514404297\n",
            "<class 'torch.Tensor'>\n",
            "64 \t 1.7896785736083984\n",
            "<class 'torch.Tensor'>\n",
            "65 \t 1.9037038087844849\n",
            "<class 'torch.Tensor'>\n",
            "66 \t 1.7910774946212769\n",
            "<class 'torch.Tensor'>\n",
            "67 \t 1.811935544013977\n",
            "<class 'torch.Tensor'>\n",
            "68 \t 1.7811235189437866\n",
            "<class 'torch.Tensor'>\n",
            "69 \t 1.800903081893921\n",
            "<class 'torch.Tensor'>\n",
            "70 \t 1.7867274284362793\n",
            "<class 'torch.Tensor'>\n",
            "71 \t 1.7662370204925537\n",
            "<class 'torch.Tensor'>\n",
            "72 \t 1.7943611145019531\n",
            "<class 'torch.Tensor'>\n",
            "73 \t 1.7907408475875854\n",
            "<class 'torch.Tensor'>\n",
            "74 \t 1.7929717302322388\n",
            "<class 'torch.Tensor'>\n",
            "75 \t 1.789615273475647\n",
            "<class 'torch.Tensor'>\n",
            "76 \t 1.7858999967575073\n",
            "<class 'torch.Tensor'>\n",
            "77 \t 1.7890757322311401\n",
            "<class 'torch.Tensor'>\n",
            "78 \t 1.7915406227111816\n",
            "<class 'torch.Tensor'>\n",
            "79 \t 1.8028976917266846\n",
            "<class 'torch.Tensor'>\n",
            "80 \t 2.3547120094299316\n",
            "<class 'torch.Tensor'>\n",
            "81 \t 1.7881008386611938\n",
            "<class 'torch.Tensor'>\n",
            "82 \t 11.433979034423828\n",
            "<class 'torch.Tensor'>\n",
            "83 \t 1.78612220287323\n",
            "<class 'torch.Tensor'>\n",
            "84 \t 1.8525652885437012\n",
            "<class 'torch.Tensor'>\n",
            "85 \t 1.7956664562225342\n",
            "<class 'torch.Tensor'>\n",
            "86 \t 1.790250301361084\n",
            "<class 'torch.Tensor'>\n",
            "87 \t 1.7939318418502808\n",
            "<class 'torch.Tensor'>\n",
            "88 \t 1.7979623079299927\n",
            "<class 'torch.Tensor'>\n",
            "89 \t 1.780409574508667\n",
            "<class 'torch.Tensor'>\n",
            "90 \t 1.784773588180542\n",
            "<class 'torch.Tensor'>\n",
            "91 \t 1.7903519868850708\n",
            "<class 'torch.Tensor'>\n",
            "92 \t 1.7969439029693604\n",
            "<class 'torch.Tensor'>\n",
            "93 \t 1.7870005369186401\n",
            "<class 'torch.Tensor'>\n",
            "94 \t 1.7928749322891235\n",
            "<class 'torch.Tensor'>\n",
            "95 \t 1.7944754362106323\n",
            "<class 'torch.Tensor'>\n",
            "96 \t 1.7870224714279175\n",
            "<class 'torch.Tensor'>\n",
            "97 \t 1.7951109409332275\n",
            "<class 'torch.Tensor'>\n",
            "98 \t 1.7857376337051392\n",
            "<class 'torch.Tensor'>\n",
            "99 \t 1.7955158948898315\n",
            "<class 'torch.Tensor'>\n",
            "100 \t 1.796602725982666\n",
            "<class 'torch.Tensor'>\n",
            "101 \t 1.7886561155319214\n",
            "<class 'torch.Tensor'>\n",
            "102 \t 1.7842495441436768\n",
            "<class 'torch.Tensor'>\n",
            "103 \t 1.7848845720291138\n",
            "<class 'torch.Tensor'>\n",
            "104 \t 1.7939492464065552\n",
            "<class 'torch.Tensor'>\n",
            "105 \t 2.2523019313812256\n",
            "<class 'torch.Tensor'>\n",
            "106 \t 1.7924343347549438\n",
            "<class 'torch.Tensor'>\n",
            "107 \t 1.8525813817977905\n",
            "<class 'torch.Tensor'>\n",
            "108 \t 2.215118169784546\n",
            "<class 'torch.Tensor'>\n",
            "109 \t 1.7854293584823608\n",
            "<class 'torch.Tensor'>\n",
            "110 \t 1.7836774587631226\n",
            "<class 'torch.Tensor'>\n",
            "111 \t 1.7891854047775269\n",
            "<class 'torch.Tensor'>\n",
            "112 \t 1.786780595779419\n",
            "<class 'torch.Tensor'>\n",
            "113 \t 1.790057897567749\n",
            "<class 'torch.Tensor'>\n",
            "114 \t 1.7916251420974731\n",
            "<class 'torch.Tensor'>\n",
            "115 \t 1.7906630039215088\n",
            "<class 'torch.Tensor'>\n",
            "116 \t 1.813353419303894\n",
            "<class 'torch.Tensor'>\n",
            "117 \t 1.7935371398925781\n",
            "<class 'torch.Tensor'>\n",
            "118 \t 1.7890591621398926\n",
            "<class 'torch.Tensor'>\n",
            "119 \t 1.7743948698043823\n",
            "<class 'torch.Tensor'>\n",
            "120 \t 1.7886751890182495\n",
            "<class 'torch.Tensor'>\n",
            "121 \t 3.2993016242980957\n",
            "<class 'torch.Tensor'>\n",
            "122 \t 1.795501708984375\n",
            "<class 'torch.Tensor'>\n",
            "123 \t 1.792702555656433\n",
            "<class 'torch.Tensor'>\n",
            "124 \t 1.779544711112976\n",
            "<class 'torch.Tensor'>\n",
            "125 \t 1.7887440919876099\n",
            "<class 'torch.Tensor'>\n",
            "126 \t 1.8616981506347656\n",
            "<class 'torch.Tensor'>\n",
            "127 \t 1.7862412929534912\n",
            "<class 'torch.Tensor'>\n",
            "128 \t 1.785725474357605\n",
            "<class 'torch.Tensor'>\n",
            "129 \t 1.7891156673431396\n",
            "<class 'torch.Tensor'>\n",
            "130 \t 1.7874308824539185\n",
            "<class 'torch.Tensor'>\n",
            "131 \t 1.787855863571167\n",
            "<class 'torch.Tensor'>\n",
            "132 \t 1.9624178409576416\n",
            "<class 'torch.Tensor'>\n",
            "133 \t 1.7999156713485718\n",
            "<class 'torch.Tensor'>\n",
            "134 \t 1.7815607786178589\n",
            "<class 'torch.Tensor'>\n",
            "135 \t 1.7944916486740112\n",
            "<class 'torch.Tensor'>\n",
            "136 \t 1.7824509143829346\n",
            "<class 'torch.Tensor'>\n",
            "137 \t 1.7921664714813232\n",
            "<class 'torch.Tensor'>\n",
            "138 \t 1.8336421251296997\n",
            "<class 'torch.Tensor'>\n",
            "139 \t 1.7846887111663818\n",
            "<class 'torch.Tensor'>\n",
            "140 \t 2.6852757930755615\n",
            "<class 'torch.Tensor'>\n",
            "141 \t 1.7852052450180054\n",
            "<class 'torch.Tensor'>\n",
            "142 \t 1.7877150774002075\n",
            "<class 'torch.Tensor'>\n",
            "143 \t 1.7822022438049316\n",
            "<class 'torch.Tensor'>\n",
            "144 \t 1.7823783159255981\n",
            "<class 'torch.Tensor'>\n",
            "145 \t 1.7792110443115234\n",
            "<class 'torch.Tensor'>\n",
            "146 \t 1.7919198274612427\n",
            "<class 'torch.Tensor'>\n",
            "147 \t 1.791054129600525\n",
            "<class 'torch.Tensor'>\n",
            "148 \t 1.7867439985275269\n",
            "<class 'torch.Tensor'>\n",
            "149 \t 1.8213342428207397\n",
            "<class 'torch.Tensor'>\n",
            "150 \t 1.784691572189331\n",
            "<class 'torch.Tensor'>\n",
            "151 \t 1.7945334911346436\n",
            "<class 'torch.Tensor'>\n",
            "152 \t 1.7897409200668335\n",
            "<class 'torch.Tensor'>\n",
            "153 \t 1.794987440109253\n",
            "<class 'torch.Tensor'>\n",
            "154 \t 1.7877261638641357\n",
            "<class 'torch.Tensor'>\n",
            "155 \t 1.7980823516845703\n",
            "<class 'torch.Tensor'>\n",
            "156 \t 1.769713044166565\n",
            "<class 'torch.Tensor'>\n",
            "157 \t 1.7865264415740967\n",
            "<class 'torch.Tensor'>\n",
            "158 \t 1.779041051864624\n",
            "<class 'torch.Tensor'>\n",
            "159 \t 1.7788931131362915\n",
            "<class 'torch.Tensor'>\n",
            "160 \t 1.7880467176437378\n",
            "<class 'torch.Tensor'>\n",
            "161 \t 1.7884658575057983\n",
            "<class 'torch.Tensor'>\n",
            "162 \t 1.789850115776062\n",
            "<class 'torch.Tensor'>\n",
            "163 \t 1.803398609161377\n",
            "<class 'torch.Tensor'>\n",
            "164 \t 2.3444740772247314\n",
            "<class 'torch.Tensor'>\n",
            "165 \t 1.7895821332931519\n",
            "<class 'torch.Tensor'>\n",
            "166 \t 1.8010716438293457\n",
            "<class 'torch.Tensor'>\n",
            "167 \t 1.8875458240509033\n",
            "<class 'torch.Tensor'>\n",
            "168 \t 1.7857966423034668\n",
            "<class 'torch.Tensor'>\n",
            "169 \t 1.7944655418395996\n",
            "<class 'torch.Tensor'>\n",
            "170 \t 1.7877442836761475\n",
            "<class 'torch.Tensor'>\n",
            "171 \t 1.9691646099090576\n",
            "<class 'torch.Tensor'>\n",
            "172 \t 1.7978087663650513\n",
            "<class 'torch.Tensor'>\n",
            "173 \t 1.7986338138580322\n",
            "<class 'torch.Tensor'>\n",
            "174 \t 1.794582724571228\n",
            "<class 'torch.Tensor'>\n",
            "175 \t 1.7911574840545654\n",
            "<class 'torch.Tensor'>\n",
            "176 \t 1.7972358465194702\n",
            "<class 'torch.Tensor'>\n",
            "177 \t 1.7979782819747925\n",
            "<class 'torch.Tensor'>\n",
            "178 \t 1.785856008529663\n",
            "<class 'torch.Tensor'>\n",
            "179 \t 1.788887619972229\n",
            "<class 'torch.Tensor'>\n",
            "180 \t 1.7922879457473755\n",
            "<class 'torch.Tensor'>\n",
            "181 \t 2.1137354373931885\n",
            "<class 'torch.Tensor'>\n",
            "182 \t 1.7877731323242188\n",
            "<class 'torch.Tensor'>\n",
            "183 \t 1.7785550355911255\n",
            "<class 'torch.Tensor'>\n",
            "184 \t 1.7878003120422363\n",
            "<class 'torch.Tensor'>\n",
            "185 \t 1.788486123085022\n",
            "<class 'torch.Tensor'>\n",
            "186 \t 1.8246418237686157\n",
            "<class 'torch.Tensor'>\n",
            "187 \t 1.7827188968658447\n",
            "<class 'torch.Tensor'>\n",
            "188 \t 1.7877498865127563\n",
            "<class 'torch.Tensor'>\n",
            "189 \t 1.7913490533828735\n",
            "<class 'torch.Tensor'>\n",
            "190 \t 1.7917588949203491\n",
            "<class 'torch.Tensor'>\n",
            "191 \t 1.7926334142684937\n",
            "<class 'torch.Tensor'>\n",
            "192 \t 1.7832911014556885\n",
            "<class 'torch.Tensor'>\n",
            "193 \t 1.7875808477401733\n",
            "<class 'torch.Tensor'>\n",
            "194 \t 1.7859946489334106\n",
            "<class 'torch.Tensor'>\n",
            "195 \t 1.7853714227676392\n",
            "<class 'torch.Tensor'>\n",
            "196 \t 1.7901026010513306\n",
            "<class 'torch.Tensor'>\n",
            "197 \t 1.7866202592849731\n",
            "<class 'torch.Tensor'>\n",
            "198 \t 1.8012667894363403\n",
            "<class 'torch.Tensor'>\n",
            "199 \t 1.7887696027755737\n",
            "torch.Size([500]) \t 65\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from matplotlib import pyplot as plt\n",
        "\n",
        "x = [i+1 for i in range(200)] \n",
        "y = training_results_Norm['training_loss']\n",
        "\n",
        "print(training_results_Norm['validation_accuracy'])\n",
        "plt.figure(figsize=(32, 8), dpi=200)\n",
        "plt.scatter(x, y, c =\"blue\")\n",
        " \n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 327
        },
        "id": "Ic7n0u7TghuS",
        "outputId": "7020ebc4-ab7a-4444-ae83-ae897bbcf7f0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[13.0]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 6400x1600 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "PATH = \"state_dict_model2.pt\" \n",
        "torch.save(model_norm.state_dict(), PATH)\n",
        "\n",
        "model = GCN(input_dim, hidden_dim, output_dim, num_layer)\n",
        "model.load_state_dict(torch.load(PATH))\n",
        "\n",
        "model.eval()\n",
        "z = model(dataset2)\n",
        "test_z = z[dataset2.test_mask]\n",
        "test_y = dataset2.y[dataset2.test_mask]\n",
        "\n",
        "correct = 0\n",
        "_, label = torch.max(test_z, 1)\n",
        "correct += (label == test_y).sum().item()\n",
        "print(test_y.shape, \"\\t\", correct)\n",
        "test_acc = 100 * (correct / len(test_y))\n",
        "print(test_acc)\n",
        "\n",
        "from sklearn.manifold import TSNE\n",
        "\n",
        "colors = []\n",
        "for i in range(len(test_z)):\n",
        "  if test_y[i]==0:\n",
        "    colors.append(\"orange\");\n",
        "  elif test_y[i]==1:\n",
        "    colors.append(\"red\");\n",
        "  elif test_y[i]==2:\n",
        "    colors.append(\"green\");\n",
        "  elif test_y[i]==3:\n",
        "    colors.append(\"yellow\");\n",
        "  elif test_y[i]==4:\n",
        "    colors.append(\"limegreen\");\n",
        "  elif test_y[i]==5:\n",
        "    colors.append(\"blue\");\n",
        "\n",
        "tsne = TSNE(n_components=2, random_state=0, perplexity=10)\n",
        "tsne_results = tsne.fit_transform(test_z.detach().numpy())\n",
        "\n",
        "plt.figure(figsize=(6, 4))\n",
        "plt.scatter(tsne_results[:,0], tsne_results[:,1], c= colors)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 406
        },
        "id": "KcoxTdTxgkzp",
        "outputId": "855c13a7-1b2d-4b1c-e0e7-77cf0631a091"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "torch.Size([1000]) \t 122\n",
            "12.2\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/manifold/_t_sne.py:783: FutureWarning: The default initialization in TSNE will change from 'random' to 'pca' in 1.2.\n",
            "  FutureWarning,\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/manifold/_t_sne.py:793: FutureWarning: The default learning rate in TSNE will change from 200.0 to 'auto' in 1.2.\n",
            "  FutureWarning,\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.collections.PathCollection at 0x7fba51f808d0>"
            ]
          },
          "metadata": {},
          "execution_count": 9
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}