G_Mixup.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "source": [
        "dataset_name = \"REDDIT-BINARY\"\n",
        "data_path = \"./\"\n",
        "num_epochs = 400\n",
        "batch_size = 128\n",
        "learning_rate = 0.01\n",
        "num_hidden = 64\n",
        "lam_range = [0.005, 0.01]\n",
        "aug_ratio = 0.15\n",
        "aug_num = 10\n",
        "seed = 6"
      ],
      "metadata": {
        "id": "GM8kTffPCNBW"
      },
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EKb-JqXQsp_8"
      },
      "source": [
        "**Packages**"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!python --version"
      ],
      "metadata": {
        "id": "WfjhqJ7X0nKV",
        "outputId": "4212ed5c-4523-4c04-a939-73731b15bd6c",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Python 3.10.12\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "id": "KfRUCbIpspdQ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b9d01739-c879-4504-e47c-dc5414cd34b1"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Looking in links: https://pytorch-geometric.com/whl/torch-2.2.1+cu121.html\n",
            "Requirement already satisfied: torch-scatter in /usr/local/lib/python3.10/dist-packages (2.1.2+pt22cu121)\n",
            "Looking in links: https://pytorch-geometric.com/whl/torch-2.2.1+cu121.html\n",
            "Requirement already satisfied: torch-sparse in /usr/local/lib/python3.10/dist-packages (0.6.18+pt22cu121)\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.10/dist-packages (from torch-sparse) (1.11.4)\n",
            "Requirement already satisfied: numpy<1.28.0,>=1.21.6 in /usr/local/lib/python3.10/dist-packages (from scipy->torch-sparse) (1.25.2)\n",
            "Looking in links: https://pytorch-geometric.com/whl/torch-2.2.1+cu121.html\n",
            "Requirement already satisfied: torch-cluster in /usr/local/lib/python3.10/dist-packages (1.6.3+pt22cu121)\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.10/dist-packages (from torch-cluster) (1.11.4)\n",
            "Requirement already satisfied: numpy<1.28.0,>=1.21.6 in /usr/local/lib/python3.10/dist-packages (from scipy->torch-cluster) (1.25.2)\n",
            "Looking in links: https://pytorch-geometric.com/whl/torch-2.2.1+cu121.html\n",
            "Requirement already satisfied: torch-spline-conv in /usr/local/lib/python3.10/dist-packages (1.2.2+pt22cu121)\n",
            "Requirement already satisfied: torch-geometric in /usr/local/lib/python3.10/dist-packages (2.5.2)\n",
            "Requirement already satisfied: tqdm in /usr/local/lib/python3.10/dist-packages (from torch-geometric) (4.66.2)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.10/dist-packages (from torch-geometric) (1.25.2)\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.10/dist-packages (from torch-geometric) (1.11.4)\n",
            "Requirement already satisfied: fsspec in /usr/local/lib/python3.10/dist-packages (from torch-geometric) (2023.6.0)\n",
            "Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from torch-geometric) (3.1.3)\n",
            "Requirement already satisfied: aiohttp in /usr/local/lib/python3.10/dist-packages (from torch-geometric) (3.9.3)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.10/dist-packages (from torch-geometric) (2.31.0)\n",
            "Requirement already satisfied: pyparsing in /usr/local/lib/python3.10/dist-packages (from torch-geometric) (3.1.2)\n",
            "Requirement already satisfied: scikit-learn in /usr/local/lib/python3.10/dist-packages (from torch-geometric) (1.2.2)\n",
            "Requirement already satisfied: psutil>=5.8.0 in /usr/local/lib/python3.10/dist-packages (from torch-geometric) (5.9.5)\n",
            "Requirement already satisfied: aiosignal>=1.1.2 in /usr/local/lib/python3.10/dist-packages (from aiohttp->torch-geometric) (1.3.1)\n",
            "Requirement already satisfied: attrs>=17.3.0 in /usr/local/lib/python3.10/dist-packages (from aiohttp->torch-geometric) (23.2.0)\n",
            "Requirement already satisfied: frozenlist>=1.1.1 in /usr/local/lib/python3.10/dist-packages (from aiohttp->torch-geometric) (1.4.1)\n",
            "Requirement already satisfied: multidict<7.0,>=4.5 in /usr/local/lib/python3.10/dist-packages (from aiohttp->torch-geometric) (6.0.5)\n",
            "Requirement already satisfied: yarl<2.0,>=1.0 in /usr/local/lib/python3.10/dist-packages (from aiohttp->torch-geometric) (1.9.4)\n",
            "Requirement already satisfied: async-timeout<5.0,>=4.0 in /usr/local/lib/python3.10/dist-packages (from aiohttp->torch-geometric) (4.0.3)\n",
            "Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->torch-geometric) (2.1.5)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests->torch-geometric) (3.3.2)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests->torch-geometric) (3.6)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests->torch-geometric) (2.0.7)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests->torch-geometric) (2024.2.2)\n",
            "Requirement already satisfied: joblib>=1.1.1 in /usr/local/lib/python3.10/dist-packages (from scikit-learn->torch-geometric) (1.3.2)\n",
            "Requirement already satisfied: threadpoolctl>=2.0.0 in /usr/local/lib/python3.10/dist-packages (from scikit-learn->torch-geometric) (3.4.0)\n"
          ]
        }
      ],
      "source": [
        "# to install the correct version of Pytorch Geometric on Goolge Colab\n",
        "import torch\n",
        "\n",
        "def format_pytorch_version(version):\n",
        "  return version.split('+')[0]\n",
        "\n",
        "TORCH_version = torch.__version__\n",
        "TORCH = format_pytorch_version(TORCH_version)\n",
        "\n",
        "def format_cuda_version(version):\n",
        "  return 'cu' + version.replace('.', '')\n",
        "\n",
        "CUDA_version = torch.version.cuda\n",
        "CUDA = format_cuda_version(CUDA_version)\n",
        "\n",
        "!pip install torch-scatter     -f https://pytorch-geometric.com/whl/torch-{TORCH}+{CUDA}.html\n",
        "!pip install torch-sparse      -f https://pytorch-geometric.com/whl/torch-{TORCH}+{CUDA}.html\n",
        "!pip install torch-cluster     -f https://pytorch-geometric.com/whl/torch-{TORCH}+{CUDA}.html\n",
        "!pip install torch-spline-conv -f https://pytorch-geometric.com/whl/torch-{TORCH}+{CUDA}.html\n",
        "!pip install torch-geometric"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "BcgY6zvdsxgw"
      },
      "outputs": [],
      "source": [
        "from typing import List, Tuple\n",
        "from skimage.restoration import denoise_tv_chambolle\n",
        "import numpy as np\n",
        "import copy\n",
        "import random\n",
        "from math import ceil\n",
        "import cv2\n",
        "from time import time\n",
        "import logging\n",
        "import os\n",
        "import os.path as osp\n",
        "import random\n",
        "import argparse\n",
        "\n",
        "import torch\n",
        "from torch_geometric.utils import dense_to_sparse\n",
        "from torch_geometric.utils import to_dense_batch, to_dense_adj\n",
        "from torch_geometric.data import Data\n",
        "from torch_geometric.nn import GINConv, global_mean_pool, JumpingKnowledge\n",
        "from torch_geometric.nn import GraphConv, TopKPooling\n",
        "from torch_geometric.nn import GCNConv, global_mean_pool, JumpingKnowledge\n",
        "from torch_geometric.nn import SAGEConv, global_mean_pool, JumpingKnowledge\n",
        "from torch_geometric.nn import global_mean_pool as gap, global_max_pool as gmp\n",
        "from torch_geometric.nn import DenseSAGEConv, dense_diff_pool, JumpingKnowledge\n",
        "from torch_geometric.nn import GCNConv, DenseGraphConv, dense_mincut_pool, dense_diff_pool\n",
        "from torch.nn import Linear, Sequential, ReLU, BatchNorm1d as BN\n",
        "from torch.nn import Linear\n",
        "import torch.nn.functional as F\n",
        "from torch_geometric.datasets import TUDataset\n",
        "from torch_geometric.data import DataLoader\n",
        "import torch_geometric.transforms as T\n",
        "from torch_geometric.utils import degree, to_dense_adj\n",
        "from torch.autograd import Variable\n",
        "from torch.optim.lr_scheduler import StepLR\n",
        "\n",
        "import matplotlib.pyplot as plt"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def pad_graphs(max_num, sorted_node_degree, sorted_graph, index):\n",
        "    normal_node_degree = np.zeros((max_num, 1))\n",
        "    aligned_graph = np.zeros((max_num, max_num))\n",
        "\n",
        "    normal_node_degree[:index, :] = sorted_node_degree\n",
        "    aligned_graph[:index, :index] = sorted_graph\n",
        "\n",
        "    return normal_node_degree, aligned_graph\n"
      ],
      "metadata": {
        "id": "VRO1K1T5QzIz"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "id": "8bTwo-XCs4ZK"
      },
      "outputs": [],
      "source": [
        "# sorting graphs nodes by descending node degrees to align them\n",
        "def sort_graphs(graphs: List[np.ndarray], padding: bool = False, N: int = None):\n",
        "    num_nodes = []\n",
        "    for i in range(len(graphs)):\n",
        "        num_nodes.append(graphs[i].shape[0])\n",
        "\n",
        "    max_num = max(N if N is not None else -np.inf, max(num_nodes))\n",
        "\n",
        "    # aligned adjacency matrices\n",
        "    aligned_graphs = []\n",
        "    # sorted normalized node degrees (as node distributions)\n",
        "    normal_node_degrees = []\n",
        "    for i in range(len(graphs)):\n",
        "\n",
        "        node_degree = 0.5 * np.sum(graphs[i], axis=0) + 0.5 * np.sum(graphs[i], axis=1)\n",
        "        node_degree_avg = node_degree/np.sum(node_degree)\n",
        "        asc_idx = np.argsort(node_degree_avg)\n",
        "        desc_idx = asc_idx[::-1]\n",
        "\n",
        "        sorted_node_degree = node_degree[desc_idx]\n",
        "        sorted_node_degree = sorted_node_degree.reshape(-1, 1)\n",
        "\n",
        "        sorted_graph = copy.deepcopy(graphs[i])\n",
        "        sorted_graph = sorted_graph[desc_idx, :]\n",
        "        sorted_graph = sorted_graph[:, desc_idx]\n",
        "\n",
        "        if padding:\n",
        "            num_i = num_nodes[i]\n",
        "            normalized_node_degree, aligned_graph = pad_graphs(max_num, sorted_node_degree, sorted_graph, num_i)\n",
        "\n",
        "            normal_node_degrees.append(normalized_node_degree)\n",
        "            aligned_graphs.append(aligned_graph)\n",
        "\n",
        "        else:\n",
        "            normal_node_degrees.append(sorted_node_degree)\n",
        "            aligned_graphs.append(sorted_graph)\n",
        "\n",
        "        if N != 0:\n",
        "            normal_node_degrees = normal_node_degrees[:N]\n",
        "            aligned_graphs = [aligned_graph[:N, :N] for aligned_graph in aligned_graphs]\n",
        "\n",
        "    # Tuple[List[np.ndarray], List[np.ndarray], int, int]\n",
        "    return aligned_graphs, normal_node_degrees"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "id": "wm8BYhccs6v2"
      },
      "outputs": [],
      "source": [
        "def graphons_generation(two_graphons, la=0.5, num_sample=20):\n",
        "\n",
        "    first_graphon_label = two_graphons[0][0]\n",
        "    first_graphon = two_graphons[0][1]\n",
        "\n",
        "    second_graphon_label = two_graphons[1][0]\n",
        "    second_graphon = two_graphons[1][1]\n",
        "\n",
        "    label = la * first_graphon_label + (1 - la) * second_graphon_label\n",
        "    new_graphon = la * first_graphon + (1 - la) * second_graphon\n",
        "\n",
        "    mixed_up_label = torch.from_numpy(label).type(torch.float32)\n",
        "\n",
        "    graphons = []\n",
        "    for i in range(num_sample):\n",
        "\n",
        "        graphon = np.random.rand(*new_graphon.shape) <= new_graphon\n",
        "        graphon = graphon.astype(np.int32)\n",
        "        graphon = np.triu(graphon)\n",
        "        graphon = graphon + graphon.T - np.diag(np.diag(graphon))\n",
        "\n",
        "        graphon = graphon[graphon.sum(axis=1) != 0]\n",
        "        graphon = graphon[:, graphon.sum(axis=0) != 0]\n",
        "\n",
        "        edge_index, _ = dense_to_sparse(torch.from_numpy(graphon))\n",
        "\n",
        "        new_graph = Data()\n",
        "        new_graph.y = mixed_up_label\n",
        "        new_graph.edge_index= edge_index\n",
        "        new_graph.num_nodes = int(torch.max(edge_index)) + 1\n",
        "        graphons.append(new_graph)\n",
        "\n",
        "    return graphons"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "id": "km6lvMa4s9F3"
      },
      "outputs": [],
      "source": [
        "# Splits the class of graohs\n",
        "def split_class(dataset):\n",
        "\n",
        "    data_y_list = []\n",
        "    for data in dataset:\n",
        "        y = tuple(data.y.tolist())\n",
        "        data_y_list.append(y)\n",
        "\n",
        "    all_graphs_list = []\n",
        "    for data in dataset:\n",
        "        graph = to_dense_adj(data.edge_index)\n",
        "        all_graphs_list.append(graph[0].numpy())\n",
        "\n",
        "    class_graphs = []\n",
        "    for class_label in set(data_y_list):\n",
        "        c_graph_list = []\n",
        "        for i in range(len(data_y_list)):\n",
        "            if data_y_list[i] == class_label:\n",
        "                c_graph_list.append(all_graphs_list[i])\n",
        "        class_graphs.append( ( np.array(class_label), c_graph_list ) )\n",
        "\n",
        "    return class_graphs"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def graph_resolution(graphs_list: List[Data]):\n",
        "    num_total_nodes = []\n",
        "    for graph in graphs_list:\n",
        "        num_total_nodes.append(graph.num_nodes)\n",
        "\n",
        "    median_num_nodes = np.median( num_total_nodes )\n",
        "    resolution = int(median_num_nodes)\n",
        "\n",
        "    return resolution"
      ],
      "metadata": {
        "id": "2n4dcBKEgSzP"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "id": "PNEZxZCktKy5"
      },
      "outputs": [],
      "source": [
        "# Estimate a graphon by USV\n",
        "def svd(sorted_graphs: List[np.ndarray], threshold: float = 0.2):\n",
        "\n",
        "    sorted_graphs =  torch.from_numpy(np.array(sorted_graphs)).float()\n",
        "\n",
        "    if sorted_graphs.size(0) < 1:\n",
        "      sum_graph = sorted_graphs[0, :, :]\n",
        "    else:\n",
        "      sum_graph = torch.mean(sorted_graphs, dim=0)\n",
        "\n",
        "    u, s, v = torch.svd(sum_graph)\n",
        "    singular_threshold = threshold * (sum_graph.size(0) ** 0.5)\n",
        "    s[torch.lt(s, singular_threshold)] = 0\n",
        "    graphon = u @ torch.diag(s) @ torch.t(v)\n",
        "    graphon[graphon > 1] = 1\n",
        "    graphon[graphon < 0] = 0\n",
        "    graphon = graphon.numpy()\n",
        "    # (r, r) graphon model\n",
        "    # np.ndarray\n",
        "    return graphon"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ob_quX2YtBJ4"
      },
      "source": [
        "**Model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "id": "30CI1m6jtDHv"
      },
      "outputs": [],
      "source": [
        "class GIN(torch.nn.Module):\n",
        "    def __init__(self, num_features=1, num_classes=1, num_hidden=64):\n",
        "        super(GIN, self).__init__()\n",
        "\n",
        "        dim = num_hidden\n",
        "\n",
        "        nn1 = Sequential(Linear(num_features, dim), ReLU(), Linear(dim, dim))\n",
        "        self.conv1 = GINConv(nn1)\n",
        "        self.bn1 = torch.nn.BatchNorm1d(dim)\n",
        "\n",
        "        nn2 = Sequential(Linear(dim, dim), ReLU(), Linear(dim, dim))\n",
        "        self.conv2 = GINConv(nn2)\n",
        "        self.bn2 = torch.nn.BatchNorm1d(dim)\n",
        "\n",
        "        nn3 = Sequential(Linear(dim, dim), ReLU(), Linear(dim, dim))\n",
        "        self.conv3 = GINConv(nn3)\n",
        "        self.bn3 = torch.nn.BatchNorm1d(dim)\n",
        "\n",
        "        nn4 = Sequential(Linear(dim, dim), ReLU(), Linear(dim, dim))\n",
        "        self.conv4 = GINConv(nn4)\n",
        "        self.bn4 = torch.nn.BatchNorm1d(dim)\n",
        "\n",
        "        nn5 = Sequential(Linear(dim, dim), ReLU(), Linear(dim, dim))\n",
        "        self.conv5 = GINConv(nn5)\n",
        "        self.bn5 = torch.nn.BatchNorm1d(dim)\n",
        "\n",
        "        self.fc1 = Linear(dim, dim)\n",
        "        self.fc2 = Linear(dim, num_classes)\n",
        "\n",
        "    def forward(self, x, edge_index, batch):\n",
        "        x = F.relu(self.conv1(x, edge_index))\n",
        "        x = self.bn1(x)\n",
        "        x = F.relu(self.conv2(x, edge_index))\n",
        "        x = self.bn2(x)\n",
        "        x = F.relu(self.conv3(x, edge_index))\n",
        "        x = self.bn3(x)\n",
        "        x = F.relu(self.conv4(x, edge_index))\n",
        "        x = self.bn4(x)\n",
        "        x = F.relu(self.conv5(x, edge_index))\n",
        "        x = self.bn5(x)\n",
        "        x = global_mean_pool(x, batch)\n",
        "        x = F.relu(self.fc1(x))\n",
        "        x = self.fc2(x)\n",
        "        return F.log_softmax(x, dim=-1)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3KLE0bLstQGw"
      },
      "source": [
        "**Mixup**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "id": "tI71T7lntUUq"
      },
      "outputs": [],
      "source": [
        "def prepare_data_x(dataset):\n",
        "\n",
        "    if dataset[0].x is None:\n",
        "        max_deg = 0\n",
        "        degrees = []\n",
        "        for data in dataset:\n",
        "            degrees += [degree(data.edge_index[0], dtype=torch.long)]\n",
        "            max_deg = max( max_deg, degrees[-1].max().item() )\n",
        "            data.num_nodes = int( torch.max(data.edge_index) ) + 1\n",
        "\n",
        "        if max_deg > 2000:\n",
        "            deg = torch.cat(degrees, dim=0).to(torch.float)\n",
        "            mean = deg.mean().item()\n",
        "            std = deg.std().item()\n",
        "            for data in dataset:\n",
        "                degrees = degree(data.edge_index[0], dtype=torch.long)\n",
        "                data.x = ( (degrees - mean) / std ).view( -1, 1 )\n",
        "\n",
        "        else:\n",
        "          for data in dataset:\n",
        "                degrees = degree(data.edge_index[0], dtype=torch.long)\n",
        "                data.x = F.one_hot(degrees, num_classes=max_deg + 1).to(torch.float)\n",
        "\n",
        "    return dataset"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "metadata": {
        "id": "rBCyAgmetV41"
      },
      "outputs": [],
      "source": [
        "# Onehots y\n",
        "def prepare_data_y(dataset):\n",
        "    for data in dataset:\n",
        "        data.y = F.one_hot(data.y, num_classes=len(dataset)).to(torch.float)[0]\n",
        "    return dataset"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "metadata": {
        "id": "tyWWhInOtXZL"
      },
      "outputs": [],
      "source": [
        "def cross_entropy_loss(input, labels, size_average=True):\n",
        "\n",
        "    assert input.size() == labels.size()\n",
        "    assert isinstance(input, Variable) and isinstance(labels, Variable)\n",
        "    loss = - torch.sum(input * labels)\n",
        "\n",
        "    return loss / input.size()[0] if size_average else loss"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 15,
      "metadata": {
        "id": "Eqw4dvketZXp"
      },
      "outputs": [],
      "source": [
        "def train(model, dataset):\n",
        "    model.train()\n",
        "    loss_total = 0\n",
        "    graph_total = 0\n",
        "    for data in dataset:\n",
        "        data = data.to(device)\n",
        "        optimizer.zero_grad()\n",
        "        output = model(data.x, data.edge_index, data.batch)\n",
        "        y = data.y.view(-1, num_classes)\n",
        "        loss = cross_entropy_loss(output, y)\n",
        "        loss.backward()\n",
        "        loss_total += loss.item() * data.num_graphs\n",
        "        graph_total += data.num_graphs\n",
        "        optimizer.step()\n",
        "    return model, loss_total / graph_total"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "metadata": {
        "id": "McBHEpWvtbGv"
      },
      "outputs": [],
      "source": [
        "def test(model, dataset):\n",
        "    model.eval()\n",
        "    accurate = 0\n",
        "    total = 0\n",
        "    loss = 0\n",
        "    for data in dataset:\n",
        "        data = data.to(device)\n",
        "        output = model(data.x, data.edge_index, data.batch)\n",
        "        pred = output.max(dim=1)[1]\n",
        "        y = data.y.view(-1, num_classes)\n",
        "        num_graphs = data.num_graphs\n",
        "        loss += cross_entropy_loss(output, y).item() * num_graphs\n",
        "        y = y.max(dim=1)[1]\n",
        "        accurate += pred.eq(y).sum().item()\n",
        "        total += num_graphs\n",
        "    return accurate / total, loss / total\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "metadata": {
        "id": "xihApNKAtetz",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "9a954b4d-9cfe-4d99-f229-9424bd637c34"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:root:This is running on: cuda\n",
            "DEBUG:fsspec.local:open file: /content/REDDIT-BINARY/REDDIT-BINARY/processed/data.pt\n",
            "INFO:root:num_features_before_gmixup: torch.Size([1246, 1])\n",
            "INFO:root:num_classes_before_gmixup: torch.Size([2000])\n",
            "INFO:root:num_features_after_gmixup: 1\n",
            "INFO:root:num_classes_after_gmixup: 2000\n",
            "/usr/local/lib/python3.10/dist-packages/torch_geometric/deprecation.py:26: UserWarning: 'data.DataLoader' is deprecated, use 'loader.DataLoader' instead\n",
            "  warnings.warn(out)\n",
            "INFO:root:Epoch: 001, Train Loss: 3.599637, Val Loss: 1.129928, Test Loss: 1.215616,  Val Acc:  0.488333, Test Acc:  0.482500\n",
            "INFO:root:Epoch: 002, Train Loss: 0.708519, Val Loss: 0.807697, Test Loss: 0.811398,  Val Acc:  0.553333, Test Acc:  0.560000\n",
            "INFO:root:Epoch: 003, Train Loss: 0.525971, Val Loss: 0.531046, Test Loss: 0.524304,  Val Acc:  0.713333, Test Acc:  0.712500\n",
            "INFO:root:Epoch: 004, Train Loss: 0.497297, Val Loss: 0.510006, Test Loss: 0.509026,  Val Acc:  0.756667, Test Acc:  0.745000\n",
            "INFO:root:Epoch: 005, Train Loss: 0.443183, Val Loss: 0.715951, Test Loss: 0.717914,  Val Acc:  0.675000, Test Acc:  0.667500\n",
            "INFO:root:Epoch: 006, Train Loss: 0.533867, Val Loss: 0.510529, Test Loss: 0.516276,  Val Acc:  0.730000, Test Acc:  0.727500\n",
            "INFO:root:Epoch: 007, Train Loss: 0.439648, Val Loss: 0.464850, Test Loss: 0.470418,  Val Acc:  0.770000, Test Acc:  0.762500\n",
            "INFO:root:Epoch: 008, Train Loss: 0.409546, Val Loss: 0.450600, Test Loss: 0.454376,  Val Acc:  0.778333, Test Acc:  0.780000\n",
            "INFO:root:Epoch: 009, Train Loss: 0.459372, Val Loss: 0.593749, Test Loss: 0.595989,  Val Acc:  0.748333, Test Acc:  0.757500\n",
            "INFO:root:Epoch: 010, Train Loss: 0.474121, Val Loss: 1.609940, Test Loss: 1.608932,  Val Acc:  0.498333, Test Acc:  0.492500\n",
            "INFO:root:Epoch: 011, Train Loss: 0.452816, Val Loss: 0.619864, Test Loss: 0.615992,  Val Acc:  0.726667, Test Acc:  0.730000\n",
            "INFO:root:Epoch: 012, Train Loss: 0.433372, Val Loss: 0.512150, Test Loss: 0.511767,  Val Acc:  0.758333, Test Acc:  0.755000\n",
            "INFO:root:Epoch: 013, Train Loss: 0.402750, Val Loss: 1.937699, Test Loss: 1.959990,  Val Acc:  0.488333, Test Acc:  0.482500\n",
            "INFO:root:Epoch: 014, Train Loss: 0.441742, Val Loss: 0.446983, Test Loss: 0.456733,  Val Acc:  0.803333, Test Acc:  0.787500\n",
            "INFO:root:Epoch: 015, Train Loss: 0.436839, Val Loss: 0.460476, Test Loss: 0.459384,  Val Acc:  0.783333, Test Acc:  0.775000\n",
            "INFO:root:Epoch: 016, Train Loss: 0.358573, Val Loss: 0.491629, Test Loss: 0.494214,  Val Acc:  0.783333, Test Acc:  0.775000\n",
            "INFO:root:Epoch: 017, Train Loss: 0.399394, Val Loss: 0.413657, Test Loss: 0.416875,  Val Acc:  0.828333, Test Acc:  0.812500\n",
            "INFO:root:Epoch: 018, Train Loss: 0.356320, Val Loss: 0.495454, Test Loss: 0.517621,  Val Acc:  0.756667, Test Acc:  0.742500\n",
            "INFO:root:Epoch: 019, Train Loss: 0.348816, Val Loss: 0.387404, Test Loss: 0.392748,  Val Acc:  0.831667, Test Acc:  0.812500\n",
            "INFO:root:Epoch: 020, Train Loss: 0.321215, Val Loss: 0.392027, Test Loss: 0.385466,  Val Acc:  0.841667, Test Acc:  0.835000\n",
            "INFO:root:Epoch: 021, Train Loss: 0.331412, Val Loss: 0.514842, Test Loss: 0.511365,  Val Acc:  0.778333, Test Acc:  0.780000\n",
            "INFO:root:Epoch: 022, Train Loss: 0.353166, Val Loss: 0.505989, Test Loss: 0.529527,  Val Acc:  0.765000, Test Acc:  0.742500\n",
            "INFO:root:Epoch: 023, Train Loss: 0.310041, Val Loss: 0.341728, Test Loss: 0.354502,  Val Acc:  0.856667, Test Acc:  0.842500\n",
            "INFO:root:Epoch: 024, Train Loss: 0.366078, Val Loss: 0.549753, Test Loss: 0.567974,  Val Acc:  0.710000, Test Acc:  0.697500\n",
            "INFO:root:Epoch: 025, Train Loss: 0.460288, Val Loss: 0.521013, Test Loss: 0.533893,  Val Acc:  0.766667, Test Acc:  0.752500\n",
            "INFO:root:Epoch: 026, Train Loss: 0.363168, Val Loss: 0.432051, Test Loss: 0.444027,  Val Acc:  0.823333, Test Acc:  0.817500\n",
            "INFO:root:Epoch: 027, Train Loss: 0.331849, Val Loss: 0.752711, Test Loss: 0.744924,  Val Acc:  0.728333, Test Acc:  0.732500\n",
            "INFO:root:Epoch: 028, Train Loss: 0.319635, Val Loss: 0.371236, Test Loss: 0.392558,  Val Acc:  0.858333, Test Acc:  0.852500\n",
            "INFO:root:Epoch: 029, Train Loss: 0.301429, Val Loss: 0.352649, Test Loss: 0.369739,  Val Acc:  0.828333, Test Acc:  0.817500\n",
            "INFO:root:Epoch: 030, Train Loss: 0.274324, Val Loss: 0.427462, Test Loss: 0.452546,  Val Acc:  0.841667, Test Acc:  0.832500\n",
            "INFO:root:Epoch: 031, Train Loss: 0.297739, Val Loss: 0.787261, Test Loss: 0.825691,  Val Acc:  0.618333, Test Acc:  0.607500\n",
            "INFO:root:Epoch: 032, Train Loss: 0.308234, Val Loss: 0.322140, Test Loss: 0.333306,  Val Acc:  0.876667, Test Acc:  0.865000\n",
            "INFO:root:Epoch: 033, Train Loss: 0.281737, Val Loss: 0.326263, Test Loss: 0.345622,  Val Acc:  0.865000, Test Acc:  0.857500\n",
            "INFO:root:Epoch: 034, Train Loss: 0.271645, Val Loss: 0.483537, Test Loss: 0.467950,  Val Acc:  0.826667, Test Acc:  0.830000\n",
            "INFO:root:Epoch: 035, Train Loss: 0.271806, Val Loss: 0.303377, Test Loss: 0.301307,  Val Acc:  0.875000, Test Acc:  0.867500\n",
            "INFO:root:Epoch: 036, Train Loss: 0.307635, Val Loss: 0.327081, Test Loss: 0.324603,  Val Acc:  0.883333, Test Acc:  0.882500\n",
            "INFO:root:Epoch: 037, Train Loss: 0.284607, Val Loss: 0.610569, Test Loss: 0.636041,  Val Acc:  0.761667, Test Acc:  0.755000\n",
            "INFO:root:Epoch: 038, Train Loss: 0.274308, Val Loss: 1.252561, Test Loss: 1.257478,  Val Acc:  0.528333, Test Acc:  0.527500\n",
            "INFO:root:Epoch: 039, Train Loss: 0.277191, Val Loss: 0.327456, Test Loss: 0.332182,  Val Acc:  0.866667, Test Acc:  0.862500\n",
            "INFO:root:Epoch: 040, Train Loss: 0.262832, Val Loss: 0.333816, Test Loss: 0.343801,  Val Acc:  0.863333, Test Acc:  0.857500\n",
            "INFO:root:Epoch: 041, Train Loss: 0.253108, Val Loss: 0.336312, Test Loss: 0.349522,  Val Acc:  0.868333, Test Acc:  0.862500\n",
            "INFO:root:Epoch: 042, Train Loss: 0.262402, Val Loss: 0.357753, Test Loss: 0.385116,  Val Acc:  0.855000, Test Acc:  0.840000\n",
            "INFO:root:Epoch: 043, Train Loss: 0.295414, Val Loss: 0.319669, Test Loss: 0.308351,  Val Acc:  0.868333, Test Acc:  0.865000\n",
            "INFO:root:Epoch: 044, Train Loss: 0.303673, Val Loss: 0.798674, Test Loss: 0.829636,  Val Acc:  0.721667, Test Acc:  0.705000\n",
            "INFO:root:Epoch: 045, Train Loss: 0.278739, Val Loss: 0.311528, Test Loss: 0.321511,  Val Acc:  0.863333, Test Acc:  0.857500\n",
            "INFO:root:Epoch: 046, Train Loss: 0.275728, Val Loss: 0.273366, Test Loss: 0.280976,  Val Acc:  0.890000, Test Acc:  0.877500\n",
            "INFO:root:Epoch: 047, Train Loss: 0.267396, Val Loss: 0.385537, Test Loss: 0.397368,  Val Acc:  0.826667, Test Acc:  0.817500\n",
            "INFO:root:Epoch: 048, Train Loss: 0.260477, Val Loss: 0.296335, Test Loss: 0.299152,  Val Acc:  0.883333, Test Acc:  0.890000\n",
            "INFO:root:Epoch: 049, Train Loss: 0.264447, Val Loss: 0.769220, Test Loss: 0.767735,  Val Acc:  0.700000, Test Acc:  0.707500\n",
            "INFO:root:Epoch: 050, Train Loss: 0.315198, Val Loss: 0.369531, Test Loss: 0.368207,  Val Acc:  0.840000, Test Acc:  0.840000\n",
            "INFO:root:Epoch: 051, Train Loss: 0.301917, Val Loss: 0.348741, Test Loss: 0.362703,  Val Acc:  0.848333, Test Acc:  0.852500\n",
            "INFO:root:Epoch: 052, Train Loss: 0.310753, Val Loss: 1.555194, Test Loss: 1.534618,  Val Acc:  0.605000, Test Acc:  0.612500\n",
            "INFO:root:Epoch: 053, Train Loss: 0.266836, Val Loss: 0.365456, Test Loss: 0.375400,  Val Acc:  0.851667, Test Acc:  0.842500\n",
            "INFO:root:Epoch: 054, Train Loss: 0.253927, Val Loss: 0.786935, Test Loss: 0.810923,  Val Acc:  0.633333, Test Acc:  0.625000\n",
            "INFO:root:Epoch: 055, Train Loss: 0.233500, Val Loss: 0.303273, Test Loss: 0.303427,  Val Acc:  0.890000, Test Acc:  0.887500\n",
            "INFO:root:Epoch: 056, Train Loss: 0.254479, Val Loss: 0.728362, Test Loss: 0.622541,  Val Acc:  0.785000, Test Acc:  0.785000\n",
            "INFO:root:Epoch: 057, Train Loss: 0.301217, Val Loss: 0.373987, Test Loss: 0.384851,  Val Acc:  0.860000, Test Acc:  0.860000\n",
            "INFO:root:Epoch: 058, Train Loss: 0.264802, Val Loss: 0.413404, Test Loss: 0.417065,  Val Acc:  0.846667, Test Acc:  0.847500\n",
            "INFO:root:Epoch: 059, Train Loss: 0.258966, Val Loss: 0.291623, Test Loss: 0.304031,  Val Acc:  0.866667, Test Acc:  0.852500\n",
            "INFO:root:Epoch: 060, Train Loss: 0.247062, Val Loss: 0.283187, Test Loss: 0.294535,  Val Acc:  0.900000, Test Acc:  0.900000\n",
            "INFO:root:Epoch: 061, Train Loss: 0.250018, Val Loss: 0.314393, Test Loss: 0.320041,  Val Acc:  0.880000, Test Acc:  0.875000\n",
            "INFO:root:Epoch: 062, Train Loss: 0.257670, Val Loss: 0.297371, Test Loss: 0.315483,  Val Acc:  0.873333, Test Acc:  0.865000\n",
            "INFO:root:Epoch: 063, Train Loss: 0.276669, Val Loss: 0.367562, Test Loss: 0.385550,  Val Acc:  0.841667, Test Acc:  0.837500\n",
            "INFO:root:Epoch: 064, Train Loss: 0.318396, Val Loss: 0.353116, Test Loss: 0.348826,  Val Acc:  0.856667, Test Acc:  0.865000\n",
            "INFO:root:Epoch: 065, Train Loss: 0.264124, Val Loss: 0.301366, Test Loss: 0.314283,  Val Acc:  0.880000, Test Acc:  0.880000\n",
            "INFO:root:Epoch: 066, Train Loss: 0.280596, Val Loss: 0.279681, Test Loss: 0.298620,  Val Acc:  0.901667, Test Acc:  0.890000\n",
            "INFO:root:Epoch: 067, Train Loss: 0.279498, Val Loss: 0.286255, Test Loss: 0.279108,  Val Acc:  0.886667, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 068, Train Loss: 0.267760, Val Loss: 0.981481, Test Loss: 0.992850,  Val Acc:  0.621667, Test Acc:  0.617500\n",
            "INFO:root:Epoch: 069, Train Loss: 0.254120, Val Loss: 0.549822, Test Loss: 0.570015,  Val Acc:  0.886667, Test Acc:  0.870000\n",
            "INFO:root:Epoch: 070, Train Loss: 0.248100, Val Loss: 0.277733, Test Loss: 0.271308,  Val Acc:  0.896667, Test Acc:  0.905000\n",
            "INFO:root:Epoch: 071, Train Loss: 0.258141, Val Loss: 0.599967, Test Loss: 0.625027,  Val Acc:  0.730000, Test Acc:  0.715000\n",
            "INFO:root:Epoch: 072, Train Loss: 0.366901, Val Loss: 0.827209, Test Loss: 0.856120,  Val Acc:  0.655000, Test Acc:  0.642500\n",
            "INFO:root:Epoch: 073, Train Loss: 0.327623, Val Loss: 0.319839, Test Loss: 0.323956,  Val Acc:  0.878333, Test Acc:  0.882500\n",
            "INFO:root:Epoch: 074, Train Loss: 0.281779, Val Loss: 0.366427, Test Loss: 0.366100,  Val Acc:  0.846667, Test Acc:  0.842500\n",
            "INFO:root:Epoch: 075, Train Loss: 0.258615, Val Loss: 0.383298, Test Loss: 0.378615,  Val Acc:  0.853333, Test Acc:  0.860000\n",
            "INFO:root:Epoch: 076, Train Loss: 0.272238, Val Loss: 1.212157, Test Loss: 1.192778,  Val Acc:  0.563333, Test Acc:  0.552500\n",
            "INFO:root:Epoch: 077, Train Loss: 0.276038, Val Loss: 0.334981, Test Loss: 0.325972,  Val Acc:  0.881667, Test Acc:  0.887500\n",
            "INFO:root:Epoch: 078, Train Loss: 0.309545, Val Loss: 0.380904, Test Loss: 0.384250,  Val Acc:  0.845000, Test Acc:  0.842500\n",
            "INFO:root:Epoch: 079, Train Loss: 0.285720, Val Loss: 0.389454, Test Loss: 0.415643,  Val Acc:  0.840000, Test Acc:  0.825000\n",
            "INFO:root:Epoch: 080, Train Loss: 0.271115, Val Loss: 1.179285, Test Loss: 1.203148,  Val Acc:  0.541667, Test Acc:  0.535000\n",
            "INFO:root:Epoch: 081, Train Loss: 0.261100, Val Loss: 0.290266, Test Loss: 0.296571,  Val Acc:  0.895000, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 082, Train Loss: 0.266004, Val Loss: 0.281717, Test Loss: 0.288172,  Val Acc:  0.901667, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 083, Train Loss: 0.254208, Val Loss: 0.302257, Test Loss: 0.294344,  Val Acc:  0.885000, Test Acc:  0.880000\n",
            "INFO:root:Epoch: 084, Train Loss: 0.246922, Val Loss: 0.330745, Test Loss: 0.336156,  Val Acc:  0.868333, Test Acc:  0.860000\n",
            "INFO:root:Epoch: 085, Train Loss: 0.295952, Val Loss: 0.644150, Test Loss: 0.631819,  Val Acc:  0.811667, Test Acc:  0.812500\n",
            "INFO:root:Epoch: 086, Train Loss: 0.306438, Val Loss: 0.385572, Test Loss: 0.385988,  Val Acc:  0.833333, Test Acc:  0.835000\n",
            "INFO:root:Epoch: 087, Train Loss: 0.306429, Val Loss: 2.276003, Test Loss: 2.325509,  Val Acc:  0.485000, Test Acc:  0.482500\n",
            "INFO:root:Epoch: 088, Train Loss: 0.310486, Val Loss: 0.490283, Test Loss: 0.508569,  Val Acc:  0.800000, Test Acc:  0.787500\n",
            "INFO:root:Epoch: 089, Train Loss: 0.298114, Val Loss: 0.483314, Test Loss: 0.510417,  Val Acc:  0.761667, Test Acc:  0.745000\n",
            "INFO:root:Epoch: 090, Train Loss: 0.274942, Val Loss: 0.413829, Test Loss: 0.402110,  Val Acc:  0.850000, Test Acc:  0.847500\n",
            "INFO:root:Epoch: 091, Train Loss: 0.280778, Val Loss: 0.359312, Test Loss: 0.381522,  Val Acc:  0.845000, Test Acc:  0.837500\n",
            "INFO:root:Epoch: 092, Train Loss: 0.281516, Val Loss: 0.348297, Test Loss: 0.359752,  Val Acc:  0.863333, Test Acc:  0.857500\n",
            "INFO:root:Epoch: 093, Train Loss: 0.277143, Val Loss: 1.459944, Test Loss: 1.476302,  Val Acc:  0.536667, Test Acc:  0.527500\n",
            "INFO:root:Epoch: 094, Train Loss: 0.251405, Val Loss: 0.276357, Test Loss: 0.286177,  Val Acc:  0.890000, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 095, Train Loss: 0.245524, Val Loss: 0.267990, Test Loss: 0.271864,  Val Acc:  0.898333, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 096, Train Loss: 0.242812, Val Loss: 0.421872, Test Loss: 0.439395,  Val Acc:  0.830000, Test Acc:  0.820000\n",
            "INFO:root:Epoch: 097, Train Loss: 0.255326, Val Loss: 0.532742, Test Loss: 0.554091,  Val Acc:  0.801667, Test Acc:  0.790000\n",
            "INFO:root:Epoch: 098, Train Loss: 0.255308, Val Loss: 0.366325, Test Loss: 0.383180,  Val Acc:  0.845000, Test Acc:  0.837500\n",
            "INFO:root:Epoch: 099, Train Loss: 0.405898, Val Loss: 1.121692, Test Loss: 1.153589,  Val Acc:  0.538333, Test Acc:  0.535000\n",
            "INFO:root:Epoch: 100, Train Loss: 0.485181, Val Loss: 0.587221, Test Loss: 0.611443,  Val Acc:  0.668333, Test Acc:  0.655000\n",
            "INFO:root:Epoch: 101, Train Loss: 0.462856, Val Loss: 0.563321, Test Loss: 0.575036,  Val Acc:  0.741667, Test Acc:  0.735000\n",
            "INFO:root:Epoch: 102, Train Loss: 0.387601, Val Loss: 0.426308, Test Loss: 0.429173,  Val Acc:  0.826667, Test Acc:  0.812500\n",
            "INFO:root:Epoch: 103, Train Loss: 0.408244, Val Loss: 0.405703, Test Loss: 0.415898,  Val Acc:  0.853333, Test Acc:  0.840000\n",
            "INFO:root:Epoch: 104, Train Loss: 0.369716, Val Loss: 0.366696, Test Loss: 0.358852,  Val Acc:  0.841667, Test Acc:  0.835000\n",
            "INFO:root:Epoch: 105, Train Loss: 0.351853, Val Loss: 0.362353, Test Loss: 0.355003,  Val Acc:  0.861667, Test Acc:  0.862500\n",
            "INFO:root:Epoch: 106, Train Loss: 0.332174, Val Loss: 0.335245, Test Loss: 0.327803,  Val Acc:  0.883333, Test Acc:  0.882500\n",
            "INFO:root:Epoch: 107, Train Loss: 0.325699, Val Loss: 0.344252, Test Loss: 0.339074,  Val Acc:  0.871667, Test Acc:  0.865000\n",
            "INFO:root:Epoch: 108, Train Loss: 0.316075, Val Loss: 0.325510, Test Loss: 0.306184,  Val Acc:  0.881667, Test Acc:  0.882500\n",
            "INFO:root:Epoch: 109, Train Loss: 0.309040, Val Loss: 0.334119, Test Loss: 0.319494,  Val Acc:  0.873333, Test Acc:  0.867500\n",
            "INFO:root:Epoch: 110, Train Loss: 0.305358, Val Loss: 0.523281, Test Loss: 0.507381,  Val Acc:  0.745000, Test Acc:  0.750000\n",
            "INFO:root:Epoch: 111, Train Loss: 0.303539, Val Loss: 0.329460, Test Loss: 0.311399,  Val Acc:  0.885000, Test Acc:  0.882500\n",
            "INFO:root:Epoch: 112, Train Loss: 0.301090, Val Loss: 0.333888, Test Loss: 0.324402,  Val Acc:  0.878333, Test Acc:  0.875000\n",
            "INFO:root:Epoch: 113, Train Loss: 0.351445, Val Loss: 0.443406, Test Loss: 0.424693,  Val Acc:  0.813333, Test Acc:  0.820000\n",
            "INFO:root:Epoch: 114, Train Loss: 0.338064, Val Loss: 0.493352, Test Loss: 0.472568,  Val Acc:  0.756667, Test Acc:  0.755000\n",
            "INFO:root:Epoch: 115, Train Loss: 0.301658, Val Loss: 0.331794, Test Loss: 0.315049,  Val Acc:  0.883333, Test Acc:  0.880000\n",
            "INFO:root:Epoch: 116, Train Loss: 0.320712, Val Loss: 0.334083, Test Loss: 0.310582,  Val Acc:  0.865000, Test Acc:  0.867500\n",
            "INFO:root:Epoch: 117, Train Loss: 0.306354, Val Loss: 0.611030, Test Loss: 0.587107,  Val Acc:  0.673333, Test Acc:  0.680000\n",
            "INFO:root:Epoch: 118, Train Loss: 0.315091, Val Loss: 0.593901, Test Loss: 0.613574,  Val Acc:  0.830000, Test Acc:  0.817500\n",
            "INFO:root:Epoch: 119, Train Loss: 0.350868, Val Loss: 0.667154, Test Loss: 0.663100,  Val Acc:  0.700000, Test Acc:  0.702500\n",
            "INFO:root:Epoch: 120, Train Loss: 0.323014, Val Loss: 0.325443, Test Loss: 0.325944,  Val Acc:  0.881667, Test Acc:  0.875000\n",
            "INFO:root:Epoch: 121, Train Loss: 0.347283, Val Loss: 0.350891, Test Loss: 0.352286,  Val Acc:  0.860000, Test Acc:  0.850000\n",
            "INFO:root:Epoch: 122, Train Loss: 0.354389, Val Loss: 0.383998, Test Loss: 0.383399,  Val Acc:  0.840000, Test Acc:  0.837500\n",
            "INFO:root:Epoch: 123, Train Loss: 0.328070, Val Loss: 0.386618, Test Loss: 0.377843,  Val Acc:  0.865000, Test Acc:  0.867500\n",
            "INFO:root:Epoch: 124, Train Loss: 0.314753, Val Loss: 0.344263, Test Loss: 0.335826,  Val Acc:  0.881667, Test Acc:  0.872500\n",
            "INFO:root:Epoch: 125, Train Loss: 0.299527, Val Loss: 0.347019, Test Loss: 0.327112,  Val Acc:  0.876667, Test Acc:  0.880000\n",
            "INFO:root:Epoch: 126, Train Loss: 0.280364, Val Loss: 0.346499, Test Loss: 0.338230,  Val Acc:  0.873333, Test Acc:  0.867500\n",
            "INFO:root:Epoch: 127, Train Loss: 0.280387, Val Loss: 0.306713, Test Loss: 0.293569,  Val Acc:  0.886667, Test Acc:  0.877500\n",
            "INFO:root:Epoch: 128, Train Loss: 0.275514, Val Loss: 0.315895, Test Loss: 0.307403,  Val Acc:  0.885000, Test Acc:  0.880000\n",
            "INFO:root:Epoch: 129, Train Loss: 0.267942, Val Loss: 0.385365, Test Loss: 0.364549,  Val Acc:  0.850000, Test Acc:  0.852500\n",
            "INFO:root:Epoch: 130, Train Loss: 0.258131, Val Loss: 0.318644, Test Loss: 0.310543,  Val Acc:  0.885000, Test Acc:  0.880000\n",
            "INFO:root:Epoch: 131, Train Loss: 0.335640, Val Loss: 0.616909, Test Loss: 0.649930,  Val Acc:  0.738333, Test Acc:  0.720000\n",
            "INFO:root:Epoch: 132, Train Loss: 0.341686, Val Loss: 0.312436, Test Loss: 0.304000,  Val Acc:  0.886667, Test Acc:  0.880000\n",
            "INFO:root:Epoch: 133, Train Loss: 0.290601, Val Loss: 0.318700, Test Loss: 0.317814,  Val Acc:  0.866667, Test Acc:  0.857500\n",
            "INFO:root:Epoch: 134, Train Loss: 0.267944, Val Loss: 0.328453, Test Loss: 0.329786,  Val Acc:  0.883333, Test Acc:  0.882500\n",
            "INFO:root:Epoch: 135, Train Loss: 0.273429, Val Loss: 0.362968, Test Loss: 0.363666,  Val Acc:  0.865000, Test Acc:  0.855000\n",
            "INFO:root:Epoch: 136, Train Loss: 0.283363, Val Loss: 0.305915, Test Loss: 0.306351,  Val Acc:  0.896667, Test Acc:  0.890000\n",
            "INFO:root:Epoch: 137, Train Loss: 0.282509, Val Loss: 0.327869, Test Loss: 0.321210,  Val Acc:  0.886667, Test Acc:  0.880000\n",
            "INFO:root:Epoch: 138, Train Loss: 0.327435, Val Loss: 0.480571, Test Loss: 0.500708,  Val Acc:  0.793333, Test Acc:  0.777500\n",
            "INFO:root:Epoch: 139, Train Loss: 0.300217, Val Loss: 0.477589, Test Loss: 0.488399,  Val Acc:  0.780000, Test Acc:  0.770000\n",
            "INFO:root:Epoch: 140, Train Loss: 0.319502, Val Loss: 0.315556, Test Loss: 0.299316,  Val Acc:  0.875000, Test Acc:  0.867500\n",
            "INFO:root:Epoch: 141, Train Loss: 0.310540, Val Loss: 0.401405, Test Loss: 0.406678,  Val Acc:  0.846667, Test Acc:  0.845000\n",
            "INFO:root:Epoch: 142, Train Loss: 0.297364, Val Loss: 0.316549, Test Loss: 0.322433,  Val Acc:  0.876667, Test Acc:  0.865000\n",
            "INFO:root:Epoch: 143, Train Loss: 0.302568, Val Loss: 0.336041, Test Loss: 0.324898,  Val Acc:  0.876667, Test Acc:  0.875000\n",
            "INFO:root:Epoch: 144, Train Loss: 0.299494, Val Loss: 0.398266, Test Loss: 0.392901,  Val Acc:  0.861667, Test Acc:  0.860000\n",
            "INFO:root:Epoch: 145, Train Loss: 0.292690, Val Loss: 0.726188, Test Loss: 0.726961,  Val Acc:  0.650000, Test Acc:  0.650000\n",
            "INFO:root:Epoch: 146, Train Loss: 0.287342, Val Loss: 0.470504, Test Loss: 0.455688,  Val Acc:  0.803333, Test Acc:  0.810000\n",
            "INFO:root:Epoch: 147, Train Loss: 0.279911, Val Loss: 0.343033, Test Loss: 0.346123,  Val Acc:  0.875000, Test Acc:  0.867500\n",
            "INFO:root:Epoch: 148, Train Loss: 0.294239, Val Loss: 0.328743, Test Loss: 0.327333,  Val Acc:  0.878333, Test Acc:  0.867500\n",
            "INFO:root:Epoch: 149, Train Loss: 0.282393, Val Loss: 0.315010, Test Loss: 0.311859,  Val Acc:  0.888333, Test Acc:  0.880000\n",
            "INFO:root:Epoch: 150, Train Loss: 0.291378, Val Loss: 0.331757, Test Loss: 0.331591,  Val Acc:  0.885000, Test Acc:  0.877500\n",
            "INFO:root:Epoch: 151, Train Loss: 0.292802, Val Loss: 0.307805, Test Loss: 0.295634,  Val Acc:  0.888333, Test Acc:  0.887500\n",
            "INFO:root:Epoch: 152, Train Loss: 0.280287, Val Loss: 0.342751, Test Loss: 0.341237,  Val Acc:  0.876667, Test Acc:  0.870000\n",
            "INFO:root:Epoch: 153, Train Loss: 0.288975, Val Loss: 0.332324, Test Loss: 0.319591,  Val Acc:  0.878333, Test Acc:  0.882500\n",
            "INFO:root:Epoch: 154, Train Loss: 0.291597, Val Loss: 0.605402, Test Loss: 0.592054,  Val Acc:  0.745000, Test Acc:  0.755000\n",
            "INFO:root:Epoch: 155, Train Loss: 0.299905, Val Loss: 0.318828, Test Loss: 0.300190,  Val Acc:  0.890000, Test Acc:  0.892500\n",
            "INFO:root:Epoch: 156, Train Loss: 0.292821, Val Loss: 0.303635, Test Loss: 0.285655,  Val Acc:  0.895000, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 157, Train Loss: 0.268251, Val Loss: 0.356854, Test Loss: 0.346236,  Val Acc:  0.860000, Test Acc:  0.857500\n",
            "INFO:root:Epoch: 158, Train Loss: 0.257865, Val Loss: 0.318431, Test Loss: 0.308632,  Val Acc:  0.890000, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 159, Train Loss: 0.276034, Val Loss: 0.494630, Test Loss: 0.486311,  Val Acc:  0.780000, Test Acc:  0.780000\n",
            "INFO:root:Epoch: 160, Train Loss: 0.252310, Val Loss: 0.283777, Test Loss: 0.269375,  Val Acc:  0.900000, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 161, Train Loss: 0.243845, Val Loss: 0.298428, Test Loss: 0.292188,  Val Acc:  0.895000, Test Acc:  0.892500\n",
            "INFO:root:Epoch: 162, Train Loss: 0.241158, Val Loss: 0.294143, Test Loss: 0.291686,  Val Acc:  0.895000, Test Acc:  0.890000\n",
            "INFO:root:Epoch: 163, Train Loss: 0.260743, Val Loss: 0.365475, Test Loss: 0.368457,  Val Acc:  0.860000, Test Acc:  0.850000\n",
            "INFO:root:Epoch: 164, Train Loss: 0.255436, Val Loss: 0.307803, Test Loss: 0.292504,  Val Acc:  0.895000, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 165, Train Loss: 0.257888, Val Loss: 0.350404, Test Loss: 0.357423,  Val Acc:  0.876667, Test Acc:  0.870000\n",
            "INFO:root:Epoch: 166, Train Loss: 0.272770, Val Loss: 0.302627, Test Loss: 0.301098,  Val Acc:  0.886667, Test Acc:  0.875000\n",
            "INFO:root:Epoch: 167, Train Loss: 0.260669, Val Loss: 0.302932, Test Loss: 0.299540,  Val Acc:  0.888333, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 168, Train Loss: 0.236218, Val Loss: 0.281277, Test Loss: 0.269372,  Val Acc:  0.901667, Test Acc:  0.900000\n",
            "INFO:root:Epoch: 169, Train Loss: 0.254758, Val Loss: 0.644797, Test Loss: 0.620059,  Val Acc:  0.698333, Test Acc:  0.705000\n",
            "INFO:root:Epoch: 170, Train Loss: 0.265342, Val Loss: 0.298652, Test Loss: 0.283865,  Val Acc:  0.896667, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 171, Train Loss: 0.257955, Val Loss: 0.330974, Test Loss: 0.328849,  Val Acc:  0.888333, Test Acc:  0.882500\n",
            "INFO:root:Epoch: 172, Train Loss: 0.248804, Val Loss: 0.297778, Test Loss: 0.291034,  Val Acc:  0.880000, Test Acc:  0.875000\n",
            "INFO:root:Epoch: 173, Train Loss: 0.267625, Val Loss: 0.295753, Test Loss: 0.298937,  Val Acc:  0.891667, Test Acc:  0.887500\n",
            "INFO:root:Epoch: 174, Train Loss: 0.247292, Val Loss: 0.276082, Test Loss: 0.279633,  Val Acc:  0.901667, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 175, Train Loss: 0.251591, Val Loss: 0.284757, Test Loss: 0.276772,  Val Acc:  0.893333, Test Acc:  0.900000\n",
            "INFO:root:Epoch: 176, Train Loss: 0.252635, Val Loss: 0.279410, Test Loss: 0.283531,  Val Acc:  0.896667, Test Acc:  0.897500\n",
            "INFO:root:Epoch: 177, Train Loss: 0.241838, Val Loss: 0.272442, Test Loss: 0.272289,  Val Acc:  0.903333, Test Acc:  0.900000\n",
            "INFO:root:Epoch: 178, Train Loss: 0.239055, Val Loss: 0.290370, Test Loss: 0.285396,  Val Acc:  0.895000, Test Acc:  0.887500\n",
            "INFO:root:Epoch: 179, Train Loss: 0.234813, Val Loss: 0.288616, Test Loss: 0.285735,  Val Acc:  0.898333, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 180, Train Loss: 0.227478, Val Loss: 0.266000, Test Loss: 0.272645,  Val Acc:  0.901667, Test Acc:  0.900000\n",
            "INFO:root:Epoch: 181, Train Loss: 0.256480, Val Loss: 0.277375, Test Loss: 0.284036,  Val Acc:  0.893333, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 182, Train Loss: 0.235408, Val Loss: 0.275810, Test Loss: 0.271895,  Val Acc:  0.900000, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 183, Train Loss: 0.247455, Val Loss: 0.274317, Test Loss: 0.279484,  Val Acc:  0.898333, Test Acc:  0.897500\n",
            "INFO:root:Epoch: 184, Train Loss: 0.236500, Val Loss: 0.302925, Test Loss: 0.295546,  Val Acc:  0.888333, Test Acc:  0.890000\n",
            "INFO:root:Epoch: 185, Train Loss: 0.233519, Val Loss: 0.361950, Test Loss: 0.359817,  Val Acc:  0.858333, Test Acc:  0.865000\n",
            "INFO:root:Epoch: 186, Train Loss: 0.238726, Val Loss: 0.265010, Test Loss: 0.262870,  Val Acc:  0.906667, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 187, Train Loss: 0.232052, Val Loss: 0.267777, Test Loss: 0.270413,  Val Acc:  0.901667, Test Acc:  0.900000\n",
            "INFO:root:Epoch: 188, Train Loss: 0.232163, Val Loss: 0.305536, Test Loss: 0.307295,  Val Acc:  0.878333, Test Acc:  0.880000\n",
            "INFO:root:Epoch: 189, Train Loss: 0.252005, Val Loss: 0.279657, Test Loss: 0.280389,  Val Acc:  0.898333, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 190, Train Loss: 0.241807, Val Loss: 0.264368, Test Loss: 0.270363,  Val Acc:  0.895000, Test Acc:  0.887500\n",
            "INFO:root:Epoch: 191, Train Loss: 0.250274, Val Loss: 0.273961, Test Loss: 0.278304,  Val Acc:  0.896667, Test Acc:  0.892500\n",
            "INFO:root:Epoch: 192, Train Loss: 0.248900, Val Loss: 0.269573, Test Loss: 0.270597,  Val Acc:  0.900000, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 193, Train Loss: 0.252687, Val Loss: 0.326586, Test Loss: 0.320108,  Val Acc:  0.873333, Test Acc:  0.875000\n",
            "INFO:root:Epoch: 194, Train Loss: 0.236740, Val Loss: 0.301562, Test Loss: 0.297641,  Val Acc:  0.883333, Test Acc:  0.882500\n",
            "INFO:root:Epoch: 195, Train Loss: 0.233646, Val Loss: 0.318771, Test Loss: 0.317763,  Val Acc:  0.878333, Test Acc:  0.877500\n",
            "INFO:root:Epoch: 196, Train Loss: 0.221517, Val Loss: 0.276201, Test Loss: 0.285503,  Val Acc:  0.895000, Test Acc:  0.890000\n",
            "INFO:root:Epoch: 197, Train Loss: 0.374089, Val Loss: 0.396314, Test Loss: 0.419967,  Val Acc:  0.831667, Test Acc:  0.832500\n",
            "INFO:root:Epoch: 198, Train Loss: 0.418590, Val Loss: 0.553750, Test Loss: 0.576830,  Val Acc:  0.723333, Test Acc:  0.707500\n",
            "INFO:root:Epoch: 199, Train Loss: 0.355567, Val Loss: 0.398197, Test Loss: 0.416963,  Val Acc:  0.835000, Test Acc:  0.825000\n",
            "INFO:root:Epoch: 200, Train Loss: 0.355015, Val Loss: 0.371514, Test Loss: 0.369213,  Val Acc:  0.843333, Test Acc:  0.837500\n",
            "INFO:root:Epoch: 201, Train Loss: 0.311640, Val Loss: 0.350800, Test Loss: 0.346359,  Val Acc:  0.860000, Test Acc:  0.855000\n",
            "INFO:root:Epoch: 202, Train Loss: 0.318689, Val Loss: 0.368288, Test Loss: 0.369008,  Val Acc:  0.845000, Test Acc:  0.842500\n",
            "INFO:root:Epoch: 203, Train Loss: 0.323759, Val Loss: 0.316018, Test Loss: 0.324105,  Val Acc:  0.885000, Test Acc:  0.877500\n",
            "INFO:root:Epoch: 204, Train Loss: 0.298419, Val Loss: 0.313711, Test Loss: 0.316336,  Val Acc:  0.891667, Test Acc:  0.892500\n",
            "INFO:root:Epoch: 205, Train Loss: 0.284349, Val Loss: 0.321587, Test Loss: 0.321394,  Val Acc:  0.876667, Test Acc:  0.872500\n",
            "INFO:root:Epoch: 206, Train Loss: 0.285099, Val Loss: 0.329524, Test Loss: 0.334990,  Val Acc:  0.878333, Test Acc:  0.870000\n",
            "INFO:root:Epoch: 207, Train Loss: 0.277482, Val Loss: 0.315424, Test Loss: 0.317773,  Val Acc:  0.890000, Test Acc:  0.880000\n",
            "INFO:root:Epoch: 208, Train Loss: 0.274932, Val Loss: 0.309715, Test Loss: 0.310552,  Val Acc:  0.893333, Test Acc:  0.882500\n",
            "INFO:root:Epoch: 209, Train Loss: 0.278092, Val Loss: 0.321290, Test Loss: 0.321788,  Val Acc:  0.895000, Test Acc:  0.890000\n",
            "INFO:root:Epoch: 210, Train Loss: 0.255401, Val Loss: 0.320518, Test Loss: 0.318555,  Val Acc:  0.886667, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 211, Train Loss: 0.261133, Val Loss: 0.306094, Test Loss: 0.311778,  Val Acc:  0.900000, Test Acc:  0.890000\n",
            "INFO:root:Epoch: 212, Train Loss: 0.261664, Val Loss: 0.329108, Test Loss: 0.313111,  Val Acc:  0.886667, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 213, Train Loss: 0.267186, Val Loss: 0.304800, Test Loss: 0.305389,  Val Acc:  0.888333, Test Acc:  0.887500\n",
            "INFO:root:Epoch: 214, Train Loss: 0.262126, Val Loss: 0.298502, Test Loss: 0.302050,  Val Acc:  0.895000, Test Acc:  0.890000\n",
            "INFO:root:Epoch: 215, Train Loss: 0.259646, Val Loss: 0.408541, Test Loss: 0.408663,  Val Acc:  0.850000, Test Acc:  0.855000\n",
            "INFO:root:Epoch: 216, Train Loss: 0.255197, Val Loss: 0.430085, Test Loss: 0.440216,  Val Acc:  0.820000, Test Acc:  0.825000\n",
            "INFO:root:Epoch: 217, Train Loss: 0.244052, Val Loss: 0.284873, Test Loss: 0.293263,  Val Acc:  0.895000, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 218, Train Loss: 0.240671, Val Loss: 0.452076, Test Loss: 0.452454,  Val Acc:  0.818333, Test Acc:  0.827500\n",
            "INFO:root:Epoch: 219, Train Loss: 0.267364, Val Loss: 0.283271, Test Loss: 0.291623,  Val Acc:  0.900000, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 220, Train Loss: 0.249525, Val Loss: 0.395490, Test Loss: 0.407689,  Val Acc:  0.823333, Test Acc:  0.822500\n",
            "INFO:root:Epoch: 221, Train Loss: 0.250287, Val Loss: 0.306041, Test Loss: 0.320198,  Val Acc:  0.893333, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 222, Train Loss: 0.241150, Val Loss: 0.310445, Test Loss: 0.320090,  Val Acc:  0.893333, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 223, Train Loss: 0.237457, Val Loss: 0.317178, Test Loss: 0.320170,  Val Acc:  0.873333, Test Acc:  0.870000\n",
            "INFO:root:Epoch: 224, Train Loss: 0.250670, Val Loss: 0.302587, Test Loss: 0.313931,  Val Acc:  0.898333, Test Acc:  0.890000\n",
            "INFO:root:Epoch: 225, Train Loss: 0.258115, Val Loss: 0.345264, Test Loss: 0.352804,  Val Acc:  0.866667, Test Acc:  0.867500\n",
            "INFO:root:Epoch: 226, Train Loss: 0.239023, Val Loss: 0.270103, Test Loss: 0.283320,  Val Acc:  0.905000, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 227, Train Loss: 0.244996, Val Loss: 0.511727, Test Loss: 0.518397,  Val Acc:  0.788333, Test Acc:  0.792500\n",
            "INFO:root:Epoch: 228, Train Loss: 0.241039, Val Loss: 0.275866, Test Loss: 0.285004,  Val Acc:  0.908333, Test Acc:  0.897500\n",
            "INFO:root:Epoch: 229, Train Loss: 0.219928, Val Loss: 0.304148, Test Loss: 0.309513,  Val Acc:  0.888333, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 230, Train Loss: 0.227082, Val Loss: 0.270743, Test Loss: 0.287205,  Val Acc:  0.903333, Test Acc:  0.890000\n",
            "INFO:root:Epoch: 231, Train Loss: 0.220895, Val Loss: 0.278683, Test Loss: 0.292501,  Val Acc:  0.898333, Test Acc:  0.887500\n",
            "INFO:root:Epoch: 232, Train Loss: 0.224146, Val Loss: 0.280680, Test Loss: 0.293810,  Val Acc:  0.900000, Test Acc:  0.887500\n",
            "INFO:root:Epoch: 233, Train Loss: 0.239804, Val Loss: 0.304202, Test Loss: 0.312820,  Val Acc:  0.888333, Test Acc:  0.890000\n",
            "INFO:root:Epoch: 234, Train Loss: 0.254457, Val Loss: 0.463184, Test Loss: 0.495380,  Val Acc:  0.796667, Test Acc:  0.772500\n",
            "INFO:root:Epoch: 235, Train Loss: 0.231890, Val Loss: 0.390843, Test Loss: 0.398532,  Val Acc:  0.853333, Test Acc:  0.847500\n",
            "INFO:root:Epoch: 236, Train Loss: 0.218539, Val Loss: 0.298600, Test Loss: 0.316868,  Val Acc:  0.890000, Test Acc:  0.880000\n",
            "INFO:root:Epoch: 237, Train Loss: 0.218338, Val Loss: 0.274039, Test Loss: 0.285115,  Val Acc:  0.906667, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 238, Train Loss: 0.216257, Val Loss: 0.312237, Test Loss: 0.324929,  Val Acc:  0.885000, Test Acc:  0.875000\n",
            "INFO:root:Epoch: 239, Train Loss: 0.226854, Val Loss: 0.265722, Test Loss: 0.281138,  Val Acc:  0.901667, Test Acc:  0.892500\n",
            "INFO:root:Epoch: 240, Train Loss: 0.223522, Val Loss: 0.264080, Test Loss: 0.276840,  Val Acc:  0.896667, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 241, Train Loss: 0.212846, Val Loss: 0.300343, Test Loss: 0.321154,  Val Acc:  0.891667, Test Acc:  0.877500\n",
            "INFO:root:Epoch: 242, Train Loss: 0.221202, Val Loss: 0.376286, Test Loss: 0.407488,  Val Acc:  0.843333, Test Acc:  0.822500\n",
            "INFO:root:Epoch: 243, Train Loss: 0.219787, Val Loss: 0.303796, Test Loss: 0.319225,  Val Acc:  0.898333, Test Acc:  0.887500\n",
            "INFO:root:Epoch: 244, Train Loss: 0.236158, Val Loss: 0.368191, Test Loss: 0.376642,  Val Acc:  0.871667, Test Acc:  0.875000\n",
            "INFO:root:Epoch: 245, Train Loss: 0.237366, Val Loss: 0.275447, Test Loss: 0.299172,  Val Acc:  0.885000, Test Acc:  0.865000\n",
            "INFO:root:Epoch: 246, Train Loss: 0.223818, Val Loss: 0.310485, Test Loss: 0.322487,  Val Acc:  0.878333, Test Acc:  0.870000\n",
            "INFO:root:Epoch: 247, Train Loss: 0.221413, Val Loss: 0.273005, Test Loss: 0.290573,  Val Acc:  0.891667, Test Acc:  0.877500\n",
            "INFO:root:Epoch: 248, Train Loss: 0.216597, Val Loss: 0.268253, Test Loss: 0.289513,  Val Acc:  0.895000, Test Acc:  0.882500\n",
            "INFO:root:Epoch: 249, Train Loss: 0.219302, Val Loss: 0.287830, Test Loss: 0.306376,  Val Acc:  0.886667, Test Acc:  0.872500\n",
            "INFO:root:Epoch: 250, Train Loss: 0.220438, Val Loss: 0.394722, Test Loss: 0.419901,  Val Acc:  0.850000, Test Acc:  0.850000\n",
            "INFO:root:Epoch: 251, Train Loss: 0.211919, Val Loss: 0.278025, Test Loss: 0.292451,  Val Acc:  0.896667, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 252, Train Loss: 0.227161, Val Loss: 0.333910, Test Loss: 0.344125,  Val Acc:  0.873333, Test Acc:  0.867500\n",
            "INFO:root:Epoch: 253, Train Loss: 0.216925, Val Loss: 0.280909, Test Loss: 0.292801,  Val Acc:  0.898333, Test Acc:  0.882500\n",
            "INFO:root:Epoch: 254, Train Loss: 0.221867, Val Loss: 0.271696, Test Loss: 0.268159,  Val Acc:  0.905000, Test Acc:  0.900000\n",
            "INFO:root:Epoch: 255, Train Loss: 0.215252, Val Loss: 0.278429, Test Loss: 0.296390,  Val Acc:  0.898333, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 256, Train Loss: 0.210485, Val Loss: 0.269382, Test Loss: 0.276034,  Val Acc:  0.891667, Test Acc:  0.890000\n",
            "INFO:root:Epoch: 257, Train Loss: 0.222432, Val Loss: 0.246097, Test Loss: 0.258231,  Val Acc:  0.908333, Test Acc:  0.897500\n",
            "INFO:root:Epoch: 258, Train Loss: 0.213792, Val Loss: 0.247666, Test Loss: 0.254913,  Val Acc:  0.910000, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 259, Train Loss: 0.207106, Val Loss: 0.251477, Test Loss: 0.262408,  Val Acc:  0.906667, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 260, Train Loss: 0.205796, Val Loss: 0.253062, Test Loss: 0.263666,  Val Acc:  0.910000, Test Acc:  0.900000\n",
            "INFO:root:Epoch: 261, Train Loss: 0.212521, Val Loss: 0.346806, Test Loss: 0.362440,  Val Acc:  0.870000, Test Acc:  0.867500\n",
            "INFO:root:Epoch: 262, Train Loss: 0.206283, Val Loss: 0.249285, Test Loss: 0.252657,  Val Acc:  0.915000, Test Acc:  0.910000\n",
            "INFO:root:Epoch: 263, Train Loss: 0.203762, Val Loss: 0.273560, Test Loss: 0.289783,  Val Acc:  0.888333, Test Acc:  0.877500\n",
            "INFO:root:Epoch: 264, Train Loss: 0.200849, Val Loss: 0.242726, Test Loss: 0.252302,  Val Acc:  0.910000, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 265, Train Loss: 0.198501, Val Loss: 0.258708, Test Loss: 0.267131,  Val Acc:  0.905000, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 266, Train Loss: 0.197095, Val Loss: 0.239810, Test Loss: 0.247844,  Val Acc:  0.913333, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 267, Train Loss: 0.192229, Val Loss: 0.270722, Test Loss: 0.283089,  Val Acc:  0.883333, Test Acc:  0.872500\n",
            "INFO:root:Epoch: 268, Train Loss: 0.205554, Val Loss: 0.285311, Test Loss: 0.282629,  Val Acc:  0.895000, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 269, Train Loss: 0.200512, Val Loss: 0.579237, Test Loss: 0.621057,  Val Acc:  0.805000, Test Acc:  0.800000\n",
            "INFO:root:Epoch: 270, Train Loss: 0.220323, Val Loss: 0.311215, Test Loss: 0.312261,  Val Acc:  0.885000, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 271, Train Loss: 0.208649, Val Loss: 0.261215, Test Loss: 0.271725,  Val Acc:  0.901667, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 272, Train Loss: 0.210116, Val Loss: 0.239028, Test Loss: 0.246921,  Val Acc:  0.910000, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 273, Train Loss: 0.217726, Val Loss: 0.296836, Test Loss: 0.320172,  Val Acc:  0.881667, Test Acc:  0.860000\n",
            "INFO:root:Epoch: 274, Train Loss: 0.229394, Val Loss: 0.335279, Test Loss: 0.338384,  Val Acc:  0.870000, Test Acc:  0.870000\n",
            "INFO:root:Epoch: 275, Train Loss: 0.212918, Val Loss: 0.244635, Test Loss: 0.243470,  Val Acc:  0.911667, Test Acc:  0.905000\n",
            "INFO:root:Epoch: 276, Train Loss: 0.196383, Val Loss: 0.272542, Test Loss: 0.274973,  Val Acc:  0.910000, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 277, Train Loss: 0.201719, Val Loss: 0.227865, Test Loss: 0.233503,  Val Acc:  0.908333, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 278, Train Loss: 0.193673, Val Loss: 0.252107, Test Loss: 0.260740,  Val Acc:  0.906667, Test Acc:  0.892500\n",
            "INFO:root:Epoch: 279, Train Loss: 0.190627, Val Loss: 0.238315, Test Loss: 0.243080,  Val Acc:  0.915000, Test Acc:  0.907500\n",
            "INFO:root:Epoch: 280, Train Loss: 0.191793, Val Loss: 0.262544, Test Loss: 0.273181,  Val Acc:  0.896667, Test Acc:  0.887500\n",
            "INFO:root:Epoch: 281, Train Loss: 0.196119, Val Loss: 0.339139, Test Loss: 0.340398,  Val Acc:  0.871667, Test Acc:  0.880000\n",
            "INFO:root:Epoch: 282, Train Loss: 0.209068, Val Loss: 0.239298, Test Loss: 0.249496,  Val Acc:  0.905000, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 283, Train Loss: 0.196004, Val Loss: 0.254337, Test Loss: 0.256709,  Val Acc:  0.905000, Test Acc:  0.900000\n",
            "INFO:root:Epoch: 284, Train Loss: 0.192920, Val Loss: 0.240585, Test Loss: 0.246021,  Val Acc:  0.916667, Test Acc:  0.910000\n",
            "INFO:root:Epoch: 285, Train Loss: 0.194833, Val Loss: 0.256537, Test Loss: 0.264343,  Val Acc:  0.915000, Test Acc:  0.907500\n",
            "INFO:root:Epoch: 286, Train Loss: 0.198378, Val Loss: 0.245134, Test Loss: 0.256221,  Val Acc:  0.905000, Test Acc:  0.897500\n",
            "INFO:root:Epoch: 287, Train Loss: 0.207587, Val Loss: 0.272636, Test Loss: 0.274515,  Val Acc:  0.893333, Test Acc:  0.890000\n",
            "INFO:root:Epoch: 288, Train Loss: 0.205335, Val Loss: 0.309481, Test Loss: 0.325254,  Val Acc:  0.878333, Test Acc:  0.867500\n",
            "INFO:root:Epoch: 289, Train Loss: 0.192206, Val Loss: 0.229152, Test Loss: 0.233709,  Val Acc:  0.908333, Test Acc:  0.897500\n",
            "INFO:root:Epoch: 290, Train Loss: 0.195456, Val Loss: 0.326338, Test Loss: 0.338646,  Val Acc:  0.881667, Test Acc:  0.877500\n",
            "INFO:root:Epoch: 291, Train Loss: 0.203200, Val Loss: 0.234704, Test Loss: 0.240953,  Val Acc:  0.903333, Test Acc:  0.892500\n",
            "INFO:root:Epoch: 292, Train Loss: 0.182522, Val Loss: 0.241002, Test Loss: 0.250781,  Val Acc:  0.905000, Test Acc:  0.892500\n",
            "INFO:root:Epoch: 293, Train Loss: 0.195639, Val Loss: 0.232947, Test Loss: 0.235805,  Val Acc:  0.911667, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 294, Train Loss: 0.188474, Val Loss: 0.324341, Test Loss: 0.327858,  Val Acc:  0.883333, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 295, Train Loss: 0.190347, Val Loss: 0.283716, Test Loss: 0.295875,  Val Acc:  0.888333, Test Acc:  0.880000\n",
            "INFO:root:Epoch: 296, Train Loss: 0.193985, Val Loss: 0.258039, Test Loss: 0.261589,  Val Acc:  0.901667, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 297, Train Loss: 0.193119, Val Loss: 0.262924, Test Loss: 0.273838,  Val Acc:  0.895000, Test Acc:  0.882500\n",
            "INFO:root:Epoch: 298, Train Loss: 0.200102, Val Loss: 0.239018, Test Loss: 0.246245,  Val Acc:  0.910000, Test Acc:  0.905000\n",
            "INFO:root:Epoch: 299, Train Loss: 0.196481, Val Loss: 0.632599, Test Loss: 0.635939,  Val Acc:  0.736667, Test Acc:  0.747500\n",
            "INFO:root:Epoch: 300, Train Loss: 0.195734, Val Loss: 0.254181, Test Loss: 0.260506,  Val Acc:  0.903333, Test Acc:  0.892500\n",
            "INFO:root:Epoch: 301, Train Loss: 0.185653, Val Loss: 0.244621, Test Loss: 0.246057,  Val Acc:  0.911667, Test Acc:  0.907500\n",
            "INFO:root:Epoch: 302, Train Loss: 0.181859, Val Loss: 0.233991, Test Loss: 0.235720,  Val Acc:  0.910000, Test Acc:  0.905000\n",
            "INFO:root:Epoch: 303, Train Loss: 0.182113, Val Loss: 0.233093, Test Loss: 0.234235,  Val Acc:  0.916667, Test Acc:  0.915000\n",
            "INFO:root:Epoch: 304, Train Loss: 0.178161, Val Loss: 0.224620, Test Loss: 0.228150,  Val Acc:  0.906667, Test Acc:  0.897500\n",
            "INFO:root:Epoch: 305, Train Loss: 0.184212, Val Loss: 0.229339, Test Loss: 0.231162,  Val Acc:  0.908333, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 306, Train Loss: 0.175625, Val Loss: 0.249177, Test Loss: 0.252477,  Val Acc:  0.905000, Test Acc:  0.897500\n",
            "INFO:root:Epoch: 307, Train Loss: 0.176753, Val Loss: 0.239146, Test Loss: 0.247742,  Val Acc:  0.905000, Test Acc:  0.897500\n",
            "INFO:root:Epoch: 308, Train Loss: 0.178695, Val Loss: 0.249436, Test Loss: 0.254844,  Val Acc:  0.905000, Test Acc:  0.897500\n",
            "INFO:root:Epoch: 309, Train Loss: 0.176756, Val Loss: 0.230917, Test Loss: 0.234662,  Val Acc:  0.911667, Test Acc:  0.905000\n",
            "INFO:root:Epoch: 310, Train Loss: 0.180717, Val Loss: 0.229291, Test Loss: 0.237824,  Val Acc:  0.910000, Test Acc:  0.900000\n",
            "INFO:root:Epoch: 311, Train Loss: 0.168598, Val Loss: 0.242997, Test Loss: 0.247166,  Val Acc:  0.903333, Test Acc:  0.897500\n",
            "INFO:root:Epoch: 312, Train Loss: 0.172543, Val Loss: 0.230328, Test Loss: 0.234601,  Val Acc:  0.915000, Test Acc:  0.910000\n",
            "INFO:root:Epoch: 313, Train Loss: 0.171864, Val Loss: 0.233815, Test Loss: 0.236189,  Val Acc:  0.910000, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 314, Train Loss: 0.174431, Val Loss: 0.226263, Test Loss: 0.226623,  Val Acc:  0.910000, Test Acc:  0.900000\n",
            "INFO:root:Epoch: 315, Train Loss: 0.171136, Val Loss: 0.228609, Test Loss: 0.234137,  Val Acc:  0.911667, Test Acc:  0.907500\n",
            "INFO:root:Epoch: 316, Train Loss: 0.177682, Val Loss: 0.230256, Test Loss: 0.233122,  Val Acc:  0.908333, Test Acc:  0.897500\n",
            "INFO:root:Epoch: 317, Train Loss: 0.166978, Val Loss: 0.234849, Test Loss: 0.243343,  Val Acc:  0.903333, Test Acc:  0.892500\n",
            "INFO:root:Epoch: 318, Train Loss: 0.183147, Val Loss: 0.236447, Test Loss: 0.240003,  Val Acc:  0.908333, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 319, Train Loss: 0.171980, Val Loss: 0.240747, Test Loss: 0.247874,  Val Acc:  0.913333, Test Acc:  0.910000\n",
            "INFO:root:Epoch: 320, Train Loss: 0.164571, Val Loss: 0.255328, Test Loss: 0.263528,  Val Acc:  0.898333, Test Acc:  0.892500\n",
            "INFO:root:Epoch: 321, Train Loss: 0.170267, Val Loss: 0.245479, Test Loss: 0.261172,  Val Acc:  0.900000, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 322, Train Loss: 0.178217, Val Loss: 0.225652, Test Loss: 0.235480,  Val Acc:  0.911667, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 323, Train Loss: 0.178568, Val Loss: 0.236890, Test Loss: 0.228624,  Val Acc:  0.905000, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 324, Train Loss: 0.187107, Val Loss: 0.240461, Test Loss: 0.249585,  Val Acc:  0.906667, Test Acc:  0.897500\n",
            "INFO:root:Epoch: 325, Train Loss: 0.176361, Val Loss: 0.229121, Test Loss: 0.236422,  Val Acc:  0.910000, Test Acc:  0.905000\n",
            "INFO:root:Epoch: 326, Train Loss: 0.179769, Val Loss: 0.261052, Test Loss: 0.264180,  Val Acc:  0.908333, Test Acc:  0.907500\n",
            "INFO:root:Epoch: 327, Train Loss: 0.168451, Val Loss: 0.230447, Test Loss: 0.236174,  Val Acc:  0.905000, Test Acc:  0.892500\n",
            "INFO:root:Epoch: 328, Train Loss: 0.175002, Val Loss: 0.236078, Test Loss: 0.243949,  Val Acc:  0.905000, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 329, Train Loss: 0.172080, Val Loss: 0.232256, Test Loss: 0.242636,  Val Acc:  0.913333, Test Acc:  0.897500\n",
            "INFO:root:Epoch: 330, Train Loss: 0.167413, Val Loss: 0.219696, Test Loss: 0.227567,  Val Acc:  0.916667, Test Acc:  0.912500\n",
            "INFO:root:Epoch: 331, Train Loss: 0.166117, Val Loss: 0.247731, Test Loss: 0.260210,  Val Acc:  0.913333, Test Acc:  0.905000\n",
            "INFO:root:Epoch: 332, Train Loss: 0.158008, Val Loss: 0.239712, Test Loss: 0.257636,  Val Acc:  0.910000, Test Acc:  0.890000\n",
            "INFO:root:Epoch: 333, Train Loss: 0.168944, Val Loss: 0.251358, Test Loss: 0.266683,  Val Acc:  0.913333, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 334, Train Loss: 0.163591, Val Loss: 0.216714, Test Loss: 0.222972,  Val Acc:  0.913333, Test Acc:  0.905000\n",
            "INFO:root:Epoch: 335, Train Loss: 0.170876, Val Loss: 0.247613, Test Loss: 0.259582,  Val Acc:  0.905000, Test Acc:  0.890000\n",
            "INFO:root:Epoch: 336, Train Loss: 0.177407, Val Loss: 0.233712, Test Loss: 0.242948,  Val Acc:  0.920000, Test Acc:  0.920000\n",
            "INFO:root:Epoch: 337, Train Loss: 0.179472, Val Loss: 0.219283, Test Loss: 0.222409,  Val Acc:  0.911667, Test Acc:  0.910000\n",
            "INFO:root:Epoch: 338, Train Loss: 0.171639, Val Loss: 0.256891, Test Loss: 0.269314,  Val Acc:  0.903333, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 339, Train Loss: 0.173870, Val Loss: 0.244570, Test Loss: 0.257264,  Val Acc:  0.901667, Test Acc:  0.890000\n",
            "INFO:root:Epoch: 340, Train Loss: 0.164943, Val Loss: 0.224090, Test Loss: 0.234699,  Val Acc:  0.910000, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 341, Train Loss: 0.166993, Val Loss: 0.253648, Test Loss: 0.256353,  Val Acc:  0.910000, Test Acc:  0.907500\n",
            "INFO:root:Epoch: 342, Train Loss: 0.172999, Val Loss: 0.252920, Test Loss: 0.263940,  Val Acc:  0.893333, Test Acc:  0.880000\n",
            "INFO:root:Epoch: 343, Train Loss: 0.164413, Val Loss: 0.218364, Test Loss: 0.224547,  Val Acc:  0.920000, Test Acc:  0.912500\n",
            "INFO:root:Epoch: 344, Train Loss: 0.163622, Val Loss: 0.230361, Test Loss: 0.238885,  Val Acc:  0.915000, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 345, Train Loss: 0.163366, Val Loss: 0.238496, Test Loss: 0.244621,  Val Acc:  0.906667, Test Acc:  0.897500\n",
            "INFO:root:Epoch: 346, Train Loss: 0.173504, Val Loss: 0.231516, Test Loss: 0.240327,  Val Acc:  0.916667, Test Acc:  0.907500\n",
            "INFO:root:Epoch: 347, Train Loss: 0.159843, Val Loss: 0.270137, Test Loss: 0.277985,  Val Acc:  0.900000, Test Acc:  0.900000\n",
            "INFO:root:Epoch: 348, Train Loss: 0.179031, Val Loss: 0.240506, Test Loss: 0.248474,  Val Acc:  0.910000, Test Acc:  0.907500\n",
            "INFO:root:Epoch: 349, Train Loss: 0.174164, Val Loss: 0.239854, Test Loss: 0.246206,  Val Acc:  0.910000, Test Acc:  0.897500\n",
            "INFO:root:Epoch: 350, Train Loss: 0.166907, Val Loss: 0.229351, Test Loss: 0.237181,  Val Acc:  0.920000, Test Acc:  0.910000\n",
            "INFO:root:Epoch: 351, Train Loss: 0.171196, Val Loss: 0.233730, Test Loss: 0.241241,  Val Acc:  0.913333, Test Acc:  0.905000\n",
            "INFO:root:Epoch: 352, Train Loss: 0.160189, Val Loss: 0.238518, Test Loss: 0.242744,  Val Acc:  0.911667, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 353, Train Loss: 0.162230, Val Loss: 0.229248, Test Loss: 0.234172,  Val Acc:  0.905000, Test Acc:  0.897500\n",
            "INFO:root:Epoch: 354, Train Loss: 0.169092, Val Loss: 0.249721, Test Loss: 0.259170,  Val Acc:  0.898333, Test Acc:  0.890000\n",
            "INFO:root:Epoch: 355, Train Loss: 0.187786, Val Loss: 0.244226, Test Loss: 0.252416,  Val Acc:  0.910000, Test Acc:  0.910000\n",
            "INFO:root:Epoch: 356, Train Loss: 0.185981, Val Loss: 0.312908, Test Loss: 0.337303,  Val Acc:  0.875000, Test Acc:  0.862500\n",
            "INFO:root:Epoch: 357, Train Loss: 0.197183, Val Loss: 0.291645, Test Loss: 0.290583,  Val Acc:  0.888333, Test Acc:  0.890000\n",
            "INFO:root:Epoch: 358, Train Loss: 0.192659, Val Loss: 0.272525, Test Loss: 0.272806,  Val Acc:  0.898333, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 359, Train Loss: 0.181045, Val Loss: 0.240656, Test Loss: 0.243247,  Val Acc:  0.906667, Test Acc:  0.900000\n",
            "INFO:root:Epoch: 360, Train Loss: 0.182751, Val Loss: 0.232157, Test Loss: 0.236518,  Val Acc:  0.913333, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 361, Train Loss: 0.172166, Val Loss: 0.245077, Test Loss: 0.248184,  Val Acc:  0.911667, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 362, Train Loss: 0.176966, Val Loss: 0.265684, Test Loss: 0.276702,  Val Acc:  0.896667, Test Acc:  0.885000\n",
            "INFO:root:Epoch: 363, Train Loss: 0.165708, Val Loss: 0.237799, Test Loss: 0.240896,  Val Acc:  0.903333, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 364, Train Loss: 0.164589, Val Loss: 0.241126, Test Loss: 0.243458,  Val Acc:  0.905000, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 365, Train Loss: 0.167597, Val Loss: 0.234396, Test Loss: 0.234544,  Val Acc:  0.916667, Test Acc:  0.915000\n",
            "INFO:root:Epoch: 366, Train Loss: 0.170230, Val Loss: 0.224181, Test Loss: 0.223987,  Val Acc:  0.918333, Test Acc:  0.915000\n",
            "INFO:root:Epoch: 367, Train Loss: 0.165076, Val Loss: 0.277082, Test Loss: 0.275904,  Val Acc:  0.900000, Test Acc:  0.900000\n",
            "INFO:root:Epoch: 368, Train Loss: 0.164514, Val Loss: 0.220022, Test Loss: 0.218048,  Val Acc:  0.916667, Test Acc:  0.912500\n",
            "INFO:root:Epoch: 369, Train Loss: 0.164689, Val Loss: 0.230363, Test Loss: 0.231972,  Val Acc:  0.916667, Test Acc:  0.907500\n",
            "INFO:root:Epoch: 370, Train Loss: 0.161058, Val Loss: 0.232663, Test Loss: 0.237549,  Val Acc:  0.918333, Test Acc:  0.910000\n",
            "INFO:root:Epoch: 371, Train Loss: 0.164333, Val Loss: 0.234057, Test Loss: 0.241604,  Val Acc:  0.906667, Test Acc:  0.892500\n",
            "INFO:root:Epoch: 372, Train Loss: 0.164079, Val Loss: 0.236458, Test Loss: 0.240453,  Val Acc:  0.913333, Test Acc:  0.910000\n",
            "INFO:root:Epoch: 373, Train Loss: 0.170601, Val Loss: 0.227186, Test Loss: 0.233899,  Val Acc:  0.913333, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 374, Train Loss: 0.160367, Val Loss: 0.241592, Test Loss: 0.249873,  Val Acc:  0.906667, Test Acc:  0.897500\n",
            "INFO:root:Epoch: 375, Train Loss: 0.165938, Val Loss: 0.271175, Test Loss: 0.282191,  Val Acc:  0.896667, Test Acc:  0.887500\n",
            "INFO:root:Epoch: 376, Train Loss: 0.157337, Val Loss: 0.246754, Test Loss: 0.254504,  Val Acc:  0.910000, Test Acc:  0.905000\n",
            "INFO:root:Epoch: 377, Train Loss: 0.167377, Val Loss: 0.235678, Test Loss: 0.236823,  Val Acc:  0.918333, Test Acc:  0.915000\n",
            "INFO:root:Epoch: 378, Train Loss: 0.161178, Val Loss: 0.229888, Test Loss: 0.236934,  Val Acc:  0.910000, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 379, Train Loss: 0.160957, Val Loss: 0.223095, Test Loss: 0.226976,  Val Acc:  0.921667, Test Acc:  0.917500\n",
            "INFO:root:Epoch: 380, Train Loss: 0.159089, Val Loss: 0.230370, Test Loss: 0.237938,  Val Acc:  0.916667, Test Acc:  0.907500\n",
            "INFO:root:Epoch: 381, Train Loss: 0.162212, Val Loss: 0.240524, Test Loss: 0.249173,  Val Acc:  0.901667, Test Acc:  0.887500\n",
            "INFO:root:Epoch: 382, Train Loss: 0.153725, Val Loss: 0.251413, Test Loss: 0.261107,  Val Acc:  0.900000, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 383, Train Loss: 0.156576, Val Loss: 0.219522, Test Loss: 0.224800,  Val Acc:  0.918333, Test Acc:  0.910000\n",
            "INFO:root:Epoch: 384, Train Loss: 0.158867, Val Loss: 0.234018, Test Loss: 0.239285,  Val Acc:  0.918333, Test Acc:  0.917500\n",
            "INFO:root:Epoch: 385, Train Loss: 0.160626, Val Loss: 0.224222, Test Loss: 0.231185,  Val Acc:  0.913333, Test Acc:  0.905000\n",
            "INFO:root:Epoch: 386, Train Loss: 0.151926, Val Loss: 0.248290, Test Loss: 0.259669,  Val Acc:  0.908333, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 387, Train Loss: 0.162540, Val Loss: 0.234011, Test Loss: 0.245236,  Val Acc:  0.921667, Test Acc:  0.912500\n",
            "INFO:root:Epoch: 388, Train Loss: 0.160777, Val Loss: 0.233144, Test Loss: 0.244992,  Val Acc:  0.916667, Test Acc:  0.910000\n",
            "INFO:root:Epoch: 389, Train Loss: 0.174711, Val Loss: 0.228612, Test Loss: 0.239179,  Val Acc:  0.918333, Test Acc:  0.907500\n",
            "INFO:root:Epoch: 390, Train Loss: 0.165317, Val Loss: 0.244364, Test Loss: 0.255774,  Val Acc:  0.916667, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 391, Train Loss: 0.165994, Val Loss: 0.227381, Test Loss: 0.231956,  Val Acc:  0.918333, Test Acc:  0.915000\n",
            "INFO:root:Epoch: 392, Train Loss: 0.172841, Val Loss: 0.262858, Test Loss: 0.267291,  Val Acc:  0.906667, Test Acc:  0.895000\n",
            "INFO:root:Epoch: 393, Train Loss: 0.159941, Val Loss: 0.227588, Test Loss: 0.229595,  Val Acc:  0.906667, Test Acc:  0.905000\n",
            "INFO:root:Epoch: 394, Train Loss: 0.154849, Val Loss: 0.228048, Test Loss: 0.230546,  Val Acc:  0.908333, Test Acc:  0.897500\n",
            "INFO:root:Epoch: 395, Train Loss: 0.158721, Val Loss: 0.225345, Test Loss: 0.231617,  Val Acc:  0.911667, Test Acc:  0.902500\n",
            "INFO:root:Epoch: 396, Train Loss: 0.149573, Val Loss: 0.229630, Test Loss: 0.238441,  Val Acc:  0.921667, Test Acc:  0.910000\n",
            "INFO:root:Epoch: 397, Train Loss: 0.168031, Val Loss: 0.219736, Test Loss: 0.222519,  Val Acc:  0.921667, Test Acc:  0.917500\n",
            "INFO:root:Epoch: 398, Train Loss: 0.160650, Val Loss: 0.236953, Test Loss: 0.240698,  Val Acc:  0.915000, Test Acc:  0.915000\n",
            "INFO:root:Epoch: 399, Train Loss: 0.161263, Val Loss: 0.250960, Test Loss: 0.261130,  Val Acc:  0.900000, Test Acc:  0.885000\n"
          ]
        }
      ],
      "source": [
        "logger = logging.getLogger()\n",
        "logger.setLevel(logging.DEBUG)\n",
        "formatter = logging.Formatter('%(asctime)s - %(levelname)s: - %(message)s', datefmt='%Y-%m-%d')\n",
        "\n",
        "torch.manual_seed(seed)\n",
        "\n",
        "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
        "logger.info(f\"This is running on: {device}\")\n",
        "\n",
        "if not (dataset_name == \"IMDB-BINARY\" or dataset_name == \"IMDB-MULTI\" or\n",
        "        dataset_name == \"REDDIT-BINARY\" or dataset_name == \"REDDIT-MULTI-5K\" or\n",
        "        dataset_name == \"REDDIT-MULTI-12K\"):\n",
        "  raise ValueError(\"Data type is not correct\")\n",
        "\n",
        "dataset = TUDataset(data_path+dataset_name, name=dataset_name)\n",
        "\n",
        "dataset = prepare_data_y(list(dataset))\n",
        "\n",
        "train_nums = int(len(dataset) * 0.7)\n",
        "train_val_nums = int(len(dataset) * 0.8)\n",
        "\n",
        "random.seed(seed)\n",
        "random.shuffle( dataset )\n",
        "\n",
        "resolution = graph_resolution(dataset[: train_nums])\n",
        "\n",
        "dataset_before_gmixup = prepare_data_x( dataset )\n",
        "\n",
        "logger.info(f\"num_features_before_gmixup: {dataset_before_gmixup[0].x.shape}\" )\n",
        "logger.info(f\"num_classes_before_gmixup: {dataset_before_gmixup[0].y.shape}\"  )\n",
        "\n",
        "# GMixup\n",
        "class_graphs = split_class(dataset[:train_nums])\n",
        "graphons = []\n",
        "for label, graphs in class_graphs:\n",
        "\n",
        "  align_graphs_list, normalized_node_degree = sort_graphs(graphs, padding=True, N=resolution)\n",
        "\n",
        "  graphon = svd(align_graphs_list)\n",
        "  graphons.append((label, graphon))\n",
        "\n",
        "num_sample = int( train_nums * aug_ratio / aug_num )\n",
        "lams = np.random.uniform(low=lam_range[0], high=lam_range[1], size=(aug_num,))\n",
        "\n",
        "random.seed(seed)\n",
        "new_graph = []\n",
        "for lam in lams:\n",
        "  two_graphons = random.sample(graphons, 2)\n",
        "  new_graph += graphons_generation(two_graphons, la=lam, num_sample=num_sample)\n",
        "\n",
        "dataset = new_graph + dataset\n",
        "train_nums = train_nums + len( new_graph )\n",
        "train_val_nums = train_val_nums + len( new_graph )\n",
        "\n",
        "dataset_after_gmixup = prepare_data_x( dataset )\n",
        "\n",
        "num_features = dataset_after_gmixup[0].x.shape[1]\n",
        "num_classes = dataset_after_gmixup[0].y.shape[0]\n",
        "logger.info(f\"num_features_after_gmixup: {num_features}\" )\n",
        "logger.info(f\"num_classes_after_gmixup: {num_classes}\"  )\n",
        "\n",
        "random.shuffle(dataset_after_gmixup)\n",
        "train_dataset = dataset_after_gmixup[:train_nums]\n",
        "random.shuffle(train_dataset)\n",
        "val_dataset = dataset_after_gmixup[train_nums:train_nums+train_val_nums]\n",
        "test_dataset = dataset_after_gmixup[train_val_nums:]\n",
        "\n",
        "\n",
        "train_dataset = DataLoader(train_dataset, batch_size=batch_size, shuffle=True)\n",
        "val_dataset = DataLoader(val_dataset, batch_size=batch_size)\n",
        "test_dataset = DataLoader(test_dataset, batch_size=batch_size)\n",
        "\n",
        "\n",
        "model = GIN(num_features=num_features, num_classes=num_classes, num_hidden=num_hidden).to(device)\n",
        "optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate, weight_decay=5e-4)\n",
        "lr_scheduler = StepLR(optimizer, step_size=100, gamma=0.5)\n",
        "\n",
        "train_l = []\n",
        "test_l = []\n",
        "val_l = []\n",
        "val_a = []\n",
        "test_a = []\n",
        "model.to(device)\n",
        "\n",
        "for epoch in range(1, num_epochs):\n",
        "        model, train_loss = train(model, train_dataset)\n",
        "        val_acc, val_loss = test(model, val_dataset)\n",
        "        test_acc, test_loss = test(model, test_dataset)\n",
        "        lr_scheduler.step()\n",
        "\n",
        "        train_l.append(train_loss)\n",
        "        test_l.append(test_loss)\n",
        "        val_l.append(val_loss)\n",
        "        test_a.append(test_acc)\n",
        "        val_a.append(val_acc)\n",
        "\n",
        "        logger.info('Epoch: {:03d}, Train Loss: {:.6f}, Val Loss: {:.6f}, Test Loss: {:.6f},  Val Acc: {: .6f}, Test Acc: {: .6f}'.format(\n",
        "            epoch, train_loss, val_loss, test_loss, val_acc, test_acc))\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.plot(train_l[:], label=\"train loss\")\n",
        "plt.plot(test_l[:], label=\"test loss\")\n",
        "plt.plot(val_l[:], label=\"validation loss\")\n",
        "plt.legend()\n",
        "plt.show()\n",
        "\n",
        "plt.plot(test_a[:], label=\"test acc\")\n",
        "plt.plot(val_a[:], label=\"validation acc\")\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "K9INjeZrzfYq",
        "outputId": "01b45957-542e-4992-c849-c97da461e586"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "DEBUG:matplotlib.font_manager:findfont: Matching sans\\-serif:style=normal:variant=normal:weight=normal:stretch=normal:size=10.0.\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono-Bold.ttf', name='DejaVu Sans Mono', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerif-Bold.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerif.ttf', name='DejaVu Serif', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerif-BoldItalic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/STIXNonUniBolIta.ttf', name='STIXNonUnicode', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerifDisplay.ttf', name='DejaVu Serif Display', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/STIXSizTwoSymBol.ttf', name='STIXSizeTwoSym', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/STIXSizFourSymReg.ttf', name='STIXSizeFourSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/STIXGeneralItalic.ttf', name='STIXGeneral', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/STIXSizOneSymBol.ttf', name='STIXSizeOneSym', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans-BoldOblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=700, stretch='normal', size='scalable')) = 1.335\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansDisplay.ttf', name='DejaVu Sans Display', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono-Oblique.ttf', name='DejaVu Sans Mono', style='oblique', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/cmmi10.ttf', name='cmmi10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/STIXSizThreeSymReg.ttf', name='STIXSizeThreeSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/STIXSizFourSymBol.ttf', name='STIXSizeFourSym', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono-BoldOblique.ttf', name='DejaVu Sans Mono', style='oblique', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans-Bold.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 0.33499999999999996\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/cmsy10.ttf', name='cmsy10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/STIXNonUniIta.ttf', name='STIXNonUnicode', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans.ttf', name='DejaVu Sans', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 0.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/STIXSizThreeSymBol.ttf', name='STIXSizeThreeSym', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/cmex10.ttf', name='cmex10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/STIXNonUniBol.ttf', name='STIXNonUnicode', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/cmtt10.ttf', name='cmtt10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/STIXNonUni.ttf', name='STIXNonUnicode', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/STIXGeneralBol.ttf', name='STIXGeneral', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSansMono.ttf', name='DejaVu Sans Mono', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/STIXSizTwoSymReg.ttf', name='STIXSizeTwoSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/STIXSizOneSymReg.ttf', name='STIXSizeOneSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/STIXGeneral.ttf', name='STIXGeneral', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/STIXSizFiveSymReg.ttf', name='STIXSizeFiveSym', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/cmss10.ttf', name='cmss10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/STIXGeneralBolIta.ttf', name='STIXGeneral', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans-Oblique.ttf', name='DejaVu Sans', style='oblique', variant='normal', weight=400, stretch='normal', size='scalable')) = 1.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/cmr10.ttf', name='cmr10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSerif-Italic.ttf', name='DejaVu Serif', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/cmb10.ttf', name='cmb10', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/share/fonts/truetype/liberation/LiberationSerif-Italic.ttf', name='Liberation Serif', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/share/fonts/truetype/liberation/LiberationMono-Bold.ttf', name='Liberation Mono', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/share/fonts/truetype/liberation/LiberationSans-Italic.ttf', name='Liberation Sans', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/share/fonts/truetype/liberation/LiberationSans-Regular.ttf', name='Liberation Sans', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/share/fonts/truetype/liberation/LiberationMono-BoldItalic.ttf', name='Liberation Mono', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/share/fonts/truetype/liberation/LiberationSerif-Regular.ttf', name='Liberation Serif', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/share/fonts/truetype/liberation/LiberationSans-BoldItalic.ttf', name='Liberation Sans', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/share/fonts/truetype/liberation/LiberationSansNarrow-Regular.ttf', name='Liberation Sans Narrow', style='normal', variant='normal', weight=400, stretch='condensed', size='scalable')) = 10.25\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/share/fonts/truetype/liberation/LiberationMono-Italic.ttf', name='Liberation Mono', style='italic', variant='normal', weight=400, stretch='normal', size='scalable')) = 11.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/share/fonts/truetype/liberation/LiberationSansNarrow-Italic.ttf', name='Liberation Sans Narrow', style='italic', variant='normal', weight=400, stretch='condensed', size='scalable')) = 11.25\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/share/fonts/truetype/liberation/LiberationMono-Regular.ttf', name='Liberation Mono', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/share/fonts/truetype/liberation/LiberationSerif-Bold.ttf', name='Liberation Serif', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/share/fonts/truetype/humor-sans/Humor-Sans.ttf', name='Humor Sans', style='normal', variant='normal', weight=400, stretch='normal', size='scalable')) = 10.05\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/share/fonts/truetype/liberation/LiberationSans-Bold.ttf', name='Liberation Sans', style='normal', variant='normal', weight=700, stretch='normal', size='scalable')) = 10.335\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/share/fonts/truetype/liberation/LiberationSerif-BoldItalic.ttf', name='Liberation Serif', style='italic', variant='normal', weight=700, stretch='normal', size='scalable')) = 11.335\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/share/fonts/truetype/liberation/LiberationSansNarrow-BoldItalic.ttf', name='Liberation Sans Narrow', style='italic', variant='normal', weight=700, stretch='condensed', size='scalable')) = 11.535\n",
            "DEBUG:matplotlib.font_manager:findfont: score(FontEntry(fname='/usr/share/fonts/truetype/liberation/LiberationSansNarrow-Bold.ttf', name='Liberation Sans Narrow', style='normal', variant='normal', weight=700, stretch='condensed', size='scalable')) = 10.535\n",
            "DEBUG:matplotlib.font_manager:findfont: Matching sans\\-serif:style=normal:variant=normal:weight=normal:stretch=normal:size=10.0 to DejaVu Sans ('/usr/local/lib/python3.10/dist-packages/matplotlib/mpl-data/fonts/ttf/DejaVuSans.ttf') with score of 0.050000.\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "gpuType": "T4",
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}