NN_Episode12.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "NN_Episode12.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "QfRzyjrJSMdM",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import torch \n",
        "import torchvision\n",
        "import torch.nn as nn\n",
        "from torchvision import transforms, datasets\n",
        "from torch.utils.tensorboard import SummaryWriter\n",
        "\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "from tqdm import tqdm\n",
        "import matplotlib.pyplot as plt\n",
        "from google.colab import output\n",
        "\n",
        "%load_ext tensorboard"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "W7_CZ27llagu",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# device = torch.device('cpu')\n",
        "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pUOV8Xj_c1TW",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def matplotlib_imshow(img, one_channel=False):\n",
        "    if one_channel:\n",
        "        img = img.mean(dim=0)\n",
        "    img = img / 2 + 0.5\n",
        "    npimg = img.cpu().numpy()\n",
        "    if one_channel:\n",
        "        plt.imshow(npimg, cmap=\"Greys\")\n",
        "    else:\n",
        "        plt.imshow(np.transpose(npimg, (1, 2, 0)))\n",
        "\n",
        "def class_preds_fig(N, x, y, out):\n",
        "    fig = plt.figure(figsize=(6, 24))\n",
        "    for idx in np.arange(N):\n",
        "        ax = fig.add_subplot(N, 1, idx+1, xticks=[], yticks=[])\n",
        "        matplotlib_imshow(x[idx], one_channel=True)\n",
        "        ax.set_title(\"{0}, {1:.2f}%\\n\".format(y[idx], torch.exp(out[idx][y[idx]]) * 100.0),\n",
        "                color=(\"green\" if torch.argmax(out[idx])==y[idx] else \"red\"),\n",
        "                y=0.85, fontsize=18)\n",
        "    return fig"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RYe3Jor9tWo8",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class ConvNet(nn.Module):\n",
        "\n",
        "    def __init__(self, input_shape, output_shape, output_act, name, seed=43):\n",
        "        super(ConvNet, self).__init__()\n",
        "        torch.manual_seed(seed)\n",
        "        torch.cuda.manual_seed(seed)\n",
        "\n",
        "        self.layer1 = nn.Sequential(\n",
        "            nn.Conv2d(input_shape, 16, kernel_size=5, stride=1, padding=2),\n",
        "            nn.ReLU(),\n",
        "            nn.MaxPool2d(kernel_size=2, stride=2))\n",
        "        self.layer2 = nn.Sequential(\n",
        "            nn.Conv2d(16, 32, kernel_size=5, stride=1, padding=2),\n",
        "            nn.ReLU(),\n",
        "            nn.MaxPool2d(kernel_size=2, stride=2))\n",
        "        self.fc = nn.Linear(7*7*32, output_shape)\n",
        "\n",
        "        self.output_act = output_act\n",
        "\n",
        "        self.writer = SummaryWriter('runs/{}'.format(name))\n",
        "    \n",
        "    def forward(self, x):\n",
        "        out = self.layer1(x)\n",
        "        out = self.layer2(out)\n",
        "        out = out.reshape(out.size(0), -1)\n",
        "        out = self.fc(out)\n",
        "        if self.output_act:\n",
        "            out = self.output_act(out)\n",
        "        return out\n",
        "\n",
        "    def test(self, epoch, criterion, testdata):\n",
        "        mean_loss = 0.0\n",
        "        accuracy = 0.0\n",
        "        total = 0.0\n",
        "\n",
        "        with torch.no_grad():\n",
        "            for x, y in testdata:\n",
        "                x, y = x.to(device), y.to(device)\n",
        "                out = self(x)\n",
        "                loss = criterion(out, y)\n",
        "                total += y.size(0)\n",
        "                mean_loss += loss.item() * y.size(0)\n",
        "                accuracy += (torch.argmax(out, dim=1) == y).sum()\n",
        "\n",
        "        self.writer.add_scalar('test loss', mean_loss/total, epoch)\n",
        "        self.writer.add_scalar('test acc', accuracy/total, epoch)\n",
        "\n",
        "        return mean_loss/total, accuracy/total\n",
        "\n",
        "    def learn(self, criterion, epochs, traindata, testdata, \n",
        "            r_lambda=0, optimizer=None, lr_schedulers=None, \n",
        "            nesterov=False, eta=0.001, verbos=1):\n",
        "\n",
        "        dataiter = iter(traindata)\n",
        "        images, _ = dataiter.next()\n",
        "        self.writer.add_graph(self, images.to(device))\n",
        "\n",
        "        if optimizer:\n",
        "            optimizer = optimizer[0](self.parameters(), lr=eta,\n",
        "                                     weight_decay=r_lambda, **optimizer[1])\n",
        "        else:\n",
        "            optimizer = torch.optim.SGD(self.parameters(), lr=eta,\n",
        "                                        momentum=0.9, weight_decay=r_lambda,\n",
        "                                        nesterov=nesterov)\n",
        "\n",
        "        l_bar = '{percentage:3.0f}%|'\n",
        "        r_bar = '|{n_fmt}/{total_fmt}[{elapsed}<{remaining},{rate_fmt}{postfix}]'\n",
        "        t = tqdm(range(epochs), disable=(not verbos), position=0, leave=True,\n",
        "                 bar_format= l_bar+ '{bar}' + r_bar +' {desc}')\n",
        "        fig_r = 0\n",
        "        test_loss, test_acc = 0, 0\n",
        "        train_loss, train_acc = 0, 0\n",
        "        for i in t:\n",
        "            mean_loss = 0.0\n",
        "            mean_accuracy = 0.0\n",
        "            total = 0.0\n",
        "\n",
        "            for step, (x, y) in enumerate(traindata):\n",
        "                x, y = x.to(device), y.to(device)\n",
        "                out = self(x)\n",
        "                loss = criterion(out, y)\n",
        "                \n",
        "                optimizer.zero_grad()\n",
        "                loss.backward()\n",
        "                optimizer.step()\n",
        "                \n",
        "                accuracy = (torch.argmax(out, dim=1) == y).sum()\n",
        "\n",
        "                if verbos == 1:\n",
        "                    if i>0:\n",
        "                        t.set_description(\"train loss:{:.8f} train acc:{:.2f} test loss:{:.8f} test acc:{:.2f}\" \n",
        "                                    .format(loss.item(), \n",
        "                                            accuracy/(y.size(0)*1.),\n",
        "                                            test_loss,\n",
        "                                            test_acc))\n",
        "                    else:\n",
        "                        t.set_description(\"train loss:{} train acc:{:.2f}\"\n",
        "                                    .format(loss.item(),\n",
        "                                            accuracy/(y.size(0)*1.)))\n",
        "                    t.refresh()\n",
        "\n",
        "                total += y.size(0)\n",
        "                mean_loss += loss.item() * y.size(0)\n",
        "                mean_accuracy += accuracy\n",
        "\n",
        "                fig_r += 1\n",
        "                if fig_r % 200 == 0:\n",
        "                    self.writer.add_figure('Class Prediction Figure',\n",
        "                                        class_preds_fig(8, x, y, out),\n",
        "                                        global_step=i)\n",
        "                    fig_r = 0\n",
        "\n",
        "            self.writer.add_scalar('train loss', mean_loss/total, i)\n",
        "            self.writer.add_scalar('train acc', mean_accuracy/total, i)\n",
        "            \n",
        "            train_loss, train_acc = mean_loss/total, mean_accuracy/total\n",
        "            test_loss, test_acc = self.test(i, criterion, testdata)\n",
        "\n",
        "            if verbos == 1:\n",
        "                t.set_description(\"train loss:{:.8f} train acc:{:.2f} test loss:{:.8f} test acc:{:.2f}\" \n",
        "                            .format(train_loss, train_acc, test_loss, test_acc))\n",
        "                t.refresh()\n",
        "\n",
        "        self.writer.close()\n",
        "\n",
        "    def weight_init(self, m):\n",
        "        if isinstance(m, nn.Linear):\n",
        "            nn.init.xavier_uniform_(m.weight, gain=nn.init.calculate_gain('relu'))\n",
        "            nn.init.zeros_(m.bias)\n",
        "\n",
        "    def reset(self):\n",
        "        self.apply(self.weight_init)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ewcRulUJo-34",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "RMSprop = (torch.optim.RMSprop, {'alpha':0.99, 'eps':1e-08,\n",
        "                                 'momentum':0, 'centered':False})\n",
        "Adam = (torch.optim.Adam, {'betas':(0.9, 0.999), 'eps':1e-08,\n",
        "                           'amsgrad':False})\n",
        "Adadelta = (torch.optim.Adadelta, {'rho':0.9, 'eps':1e-06})\n",
        "Adagrad = (torch.optim.Adagrad, {'lr_decay':0, 'initial_accumulator_value':0,\n",
        "                                'eps':1e-10})"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nHnRrcOJk-1Q",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "9292fce5-b887-406b-a3ec-f7f66c58cd6b"
      },
      "source": [
        "epochs = 10\n",
        "batch_size = 128\n",
        "\n",
        "transform = transforms.Compose([transforms.ToTensor(),\n",
        "                            transforms.Normalize((0.5), (0.5))])\n",
        "\n",
        "traindata = datasets.MNIST(root='../../data/',\n",
        "                            train=True, \n",
        "                            transform=transform,\n",
        "                            download=True)\n",
        "testdata = datasets.MNIST(root='../../data/',\n",
        "                            train=False, \n",
        "                            transform=transform)\n",
        "\n",
        "traindata = torch.utils.data.DataLoader(dataset=traindata, \n",
        "                                    batch_size=batch_size, \n",
        "                                    shuffle=True)\n",
        "testdata = torch.utils.data.DataLoader(dataset=testdata,\n",
        "                                    shuffle=False)\n",
        "\n",
        "mynet = ConvNet(1, 10, nn.LogSoftmax(dim=1), 'BS_128').to(device)\n",
        "mynet.reset()\n",
        "mynet.learn(nn.NLLLoss(), epochs, traindata, testdata, optimizer=RMSprop,\n",
        "            r_lambda=1e-4, eta=0.001)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "100%|██████████|10/10[05:04<00:00,30.42s/it] train loss:0.01210315 train acc:1.00 test loss:0.02295778 test acc:0.99: \n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kWGUaGabVRtw",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "%tensorboard --logdir runs"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Zny87ACDuxYp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def conv_shape(input_size, padding, dilation, kernel, stride):\n",
        "    return int((input_size + (2*padding) - dilation*(kernel-1) - 1)/stride+1)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UWGBrxSADiyt",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "ca22d946-cec8-4d0b-df1f-ff6d59ce1e1b"
      },
      "source": [
        "conv_shape(14, 0, 1, 2, 2)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "7"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 22
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0f-nZO6q-OxU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "filters, layers = [], []\n",
        "for i in mynet.children():\n",
        "    if isinstance(i, nn.Sequential):\n",
        "        for j in i.children():\n",
        "            if isinstance(j, nn.Conv2d):\n",
        "                filters.append(j.weight)\n",
        "            layers.append(j)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YO9q1-C4-qI0",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 483
        },
        "outputId": "20d828f1-5369-4476-dbf0-7c0f4b31cb08"
      },
      "source": [
        "plt.figure(figsize=(20, 17))\n",
        "for i, filter in enumerate(filters[1]):\n",
        "    plt.subplot(8, 8, i+1)\n",
        "    plt.imshow(filter[0, :, :].detach().cpu(), cmap='gray')\n",
        "    plt.axis('off')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x1224 with 32 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4dqxYepWCBRJ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "dataiter = iter(testdata)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1Lom-woSCwEG",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "3b82920b-a538-4e88-d63e-e7a1b112cbf4"
      },
      "source": [
        "img, _ = dataiter.next()\n",
        "results = [layers[0](img.to(device))]\n",
        "\n",
        "for i in range(1, len(layers)):\n",
        "    results.append(layers[i](results[-1]))\n",
        "outputs = results\n",
        "\n",
        "for num_layer in range(len(outputs)):\n",
        "    plt.figure(figsize=(30, 30))\n",
        "    layer_viz = outputs[num_layer][0, :, :, :]\n",
        "    layer_viz = layer_viz.data\n",
        "    print(layer_viz.size())\n",
        "    for i, filter in enumerate(layer_viz):\n",
        "        plt.subplot(8, 8, i + 1)\n",
        "        plt.imshow(filter.cpu(), cmap='gray')\n",
        "        plt.axis(\"off\")\n",
        "    plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "torch.Size([16, 28, 28])\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 2160x2160 with 16 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "torch.Size([16, 28, 28])\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 2160x2160 with 16 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "torch.Size([16, 14, 14])\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABpMAAAGMCAYAAAA2p1qsAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3be5AV5H0+8He5rFwFIqCgRMRLGDXxgrc2WmM0vWpaU23iVIszmZixJpp0JtZbjI2dVP9oNUmtaXVqRjODqUarqMTGG7bipSghaga8oIjclYvAslx29/ffz2bi6/CsHM5h9/P5d5/vfl9Y3nPOnofT1tPTUwAAAAAAAOCDDGj2AQAAAAAAAGhdyiQAAAAAAACqlEkAAAAAAABUKZMAAAAAAACoUiYBAAAAAABQpUwCAAAAAACgatCHffGoo47qSb7ZL3/5y492GvgAl1xySZS/8cYb2xp0lJ1i2rRp0b3ad999G3UU+rEhQ4ZE+f/4j/9o6XvV1tYW3av+bMKECQ3NL168OMqXUsq7774bz7Si0aNHR/m1a9e27L1ypxpn0KAPffn9W7Zv396gk/Q9PT09LXunSnGvEuPHj4/y69evj/JbtmyJ8n3J4MGDo/zWrVv71L1qb29v1FF2qYEDB8YzU6ZMifKrVq2K8uk97Eu2bt0a5Vv5+epb3/pWdKdeffXVRh3lIxkxYkSU781jw7Bhw6L8n/7pn0b53jxX3XPPPVF+zZo18Y5d4eSTT47y3/zmN1v2TpXSd14Dpu8NpK/nSill8+bNUf6VV16Jd7Bjas9VPpkEAAAAAABAlTIJAAAAAACAKmUSAAAAAAAAVcokAAAAAAAAqpRJAAAAAAAAVCmTAAAAAAAAqFImAQAAAAAAUKVMAgAAAAAAoEqZBAAAAAAAQJUyCQAAAAAAgCplEgAAAAAAAFVtPT091S9+5jOfqX/xA8yePfsjH6gRRo4cGeU3bNjQoJO877nnnovyN998c7xj27ZtUf4nP/lJvGNX+N3f/d0o/9RTT7U16Cg7RVtbW3SvoBX09PS4V7vA97///Sh/0kknxTsWLVoU5efNmxflZ82aFeVLKeWFF16IZ1rR5MmTo/wbb7zRsvdq6tSp0Z1auHBho47y/w0YkP8fqL/8y7+M8meddVa8Y9KkSVF+48aNUf6dd96J8qWU8r3vfS/Kz507N97RijxX7RrpY9369evjHQMHDozyvbkn7Bj3qjWl97CUUrq7u6P8mWeeGeW3bt0a5Uvp3fscfUEr36v+eqda0TnnnBPPbNq0Kcrff//98Y5d4ZRTTonyjz32WMveqVJa815NmzYtnpkyZUqUnzhxYrxj8+bNUT59z/j888+P8v1Z7bnKJ5MAAAAAAACoUiYBAAAAAABQpUwCAAAAAACgSpkEAAAAAABAlTIJAAAAAACAKmUSAAAAAAAAVcokAAAAAAAAqpRJAAAAAAAAVCmTAAAAAAAAqFImAQAAAAAAUKVMAgAAAAAAoEqZBAAAAAAAQNWgZh8gdfnll8czHR0dUf6ll16Kd6xcuTLK33PPPVH+tttui/KllHL11VfHM7CjxowZE+WvuuqqeMfRRx8d5Z955pko35vHE9hR55xzTjxz8cUXR/n/+Z//iXecddZZUX7fffeN8l1dXVEedlR7e3s8M3bs2Cj/9NNPxzuuvPLKKH/GGWdE+SlTpkT5Uko5/PDDo/zcuXPjHfRfGzZsiPJr166Nd6T36vzzz4/yBx98cJSHRrvgggui/CuvvBLveOKJJ6L8McccE+UnT54c5Usp5eabb45noBFOOOGEeOb444+P8jfddFOUnzFjRpQvpXevG+mfDj300Hhm+fLlUf6uu+6Kd6TOO++8KN+be9Wb93b6Mp9MAgAAAAAAoEqZBAAAAAAAQJUyCQAAAAAAgCplEgAAAAAAAFXKJAAAAAAAAKqUSQAAAAAAAFQpkwAAAAAAAKhSJgEAAAAAAFClTAIAAAAAAKBKmQQAAAAAAECVMgkAAAAAAICqQc0+QOrXv/51PHPfffdF+dGjR8c71q1bF+VnzJgR5a+77rooX0opL7/8cjxD/3ThhRfGMwcffHCUT+9IKaUsXLgwyr/11ltRvr29PcqXUsrWrVvjGfqnyZMnxzPpY/3ll18e7zjmmGOi/BVXXBHlv/rVr0Z52FGdnZ3xzA033NCAk3w0CxYsiPLnnXdevGPlypXxDOyod999t+E7TjvttCh/0EEHNegksGucffbZUf4///M/4x1PPPFElP/iF78Y74BGePrpp+OZb3/721F+2rRp8Y5TTz01yg8YkP1//t68jl20aFE8Q/90xx13NHxHb94TOf3006P8SSedFOVHjRoV5fltPpkEAAAAAABAlTIJAAAAAACAKmUSAAAAAAAAVcokAAAAAAAAqpRJAAAAAAAAVCmTAAAAAAAAqFImAQAAAAAAUKVMAgAAAAAAoEqZBAAAAAAAQJUyCQAAAAAAgCplEgAAAAAAAFXKJAAAAAAAAKoGNfsAqfvuuy+eOe2006L8I488Eu8YNmxYlP/kJz8Z70hNmDCh4TvoGxYsWBDPvPjii1F++/bt8Y5nnnkmyu+5555Rfty4cVG+lFKWLl0az9A//cM//EOzj/CBfvrTn0b5MWPGRPnVq1dHeWglAwcOjGe6urqifHt7e5R/4YUXonwpvXvOhVbyzW9+M8qvWLGiQSeBXG9+x5g9e3aUv+mmm+IdqcGDB0f5V199tUEnob9buXJlPPNHf/RHUf7666+Pd/zgBz+I8ps3b453wO7s9NNPj2dGjRrVgJO8b/369fHM+PHjo/yqVaviHbsTn0wCAAAAAACgSpkEAAAAAABAlTIJAAAAAACAKmUSAAAAAAAAVcokAAAAAAAAqpRJAAAAAAAAVCmTAAAAAAAAqFImAQAAAAAAUKVMAgAAAAAAoEqZBAAAAAAAQJUyCQAAAAAAgKpBzT7ArvDII49E+QED8o7tyCOPjPJz5syJ8tOnT4/ypZTyz//8z/EM/dPjjz/e7CN8oD333DPKT5kyJcpv3749ypdSyvLly6N8d3d3vAN21PXXXx/PHHDAAVH+zDPPjHfA7qqnpyeeaW9vj/IdHR1RfsGCBVG+lN49v0GjjB07Np5J/w2vWLEi3gGNsnr16njm7//+7xtwkt90xx13RPm5c+dG+YsuuijKw476sz/7s2Yf4QONGDEiyu+zzz5R3nMbu7s333wznnnggQd2/kH+j+HDh8czgwb1i/pkh/lkEgAAAAAAAFXKJAAAAAAAAKqUSQAAAAAAAFQpkwAAAAAAAKhSJgEAAAAAAFClTAIAAAAAAKBKmQQAAAAAAECVMgkAAAAAAIAqZRIAAAAAAABVyiQAAAAAAACqlEkAAAAAAABUKZMAAAAAAACoGtTsA+wKn/rUp6L8AQccEO+47777ovy1114b5b/97W9HeegLBg4cGOU7Ozuj/Lp166J8KaV0d3fHM7Cjxo0bF+UvuOCCeMerr74a5dPnN9idDRs2LJ7ZuHFjlG9ra4vyW7ZsifLQaoYOHRrPvPTSSw04CfQdX/ziF+OZyZMnR/mnnnoqyj/33HNRHnZ3Q4YMifLt7e0NOgnsGn/wB38Q5RcsWNCgk/Te4MGD45lly5Y14CS7L59MAgAAAAAAoEqZBAAAAAAAQJUyCQAAAAAAgCplEgAAAAAAAFXKJAAAAAAAAKqUSQAAAAAAAFQpkwAAAAAAAKhSJgEAAAAAAFClTAIAAAAAAKBKmQQAAAAAAECVMgkAAAAAAICqQc0+wK7Q0dER5efPnx/vmDp1apR/7LHH4h2wOxsxYkQ8M2bMmCj/7rvvRvnVq1dHeWi0mTNnRvn0+a2UUj7xiU/EM9BfbNq0KZ5pb2+P8kOGDIny7733XpSHRps8eXKU7829Aj7cZZddFs+0tbVF+W984xvxDuhP0t/F3nnnnQadBHK9eV+gq6sryr/22mvxjtTo0aOj/Lp16xp0kv7DJ5MAAAAAAACoUiYBAAAAAABQpUwCAAAAAACgSpkEAAAAAABAlTIJAAAAAACAKmUSAAAAAAAAVcokAAAAAAAAqpRJAAAAAAAAVCmTAAAAAAAAqFImAQAAAAAAUKVMAgAAAAAAoGpQsw+QOuecc+KZZcuWRfnZs2fHO77yla9E+VtuuSXe0WjDhw+PZzZt2tTwHfQNvfnZjxgxIsovWrQo3gGNcvTRR8czRx11VJS/8cYb4x391ZAhQ6L8uHHj4h177LFHPENrGTZsWDwzYED2f7O6urriHa3oYx/7WJTvzeuAJUuWxDO0ntWrVzf7CNDy5syZE+X322+/eMell14az/QFQ4cOjfK9eb5655134hl23IQJE+KZ5cuXR/l999033rFly5Yo39HREe/oCyZPnhzPfOELX9j5B+E3fPKTn4xnevN+eWrs2LFRft26dQ06Se+NHj06nmnFP0eNTyYBAAAAAABQpUwCAAAAAACgSpkEAAAAAABAlTIJAAAAAACAKmUSAAAAAAAAVcokAAAAAAAAqpRJAAAAAAAAVCmTAAAAAAAAqFImAQAAAAAAUKVMAgAAAAAAoEqZBAAAAAAAQJUyCQAAAAAAgKpBzT5Aau+9945nZsyY0YCT/Kbf+73fi/K33HJLg07Se1u2bGn4js9+9rPxzP3339+Ak/B/TZs2LZ7ZtGlTlF+wYEG8Y+XKlfEMtIopU6bEM7feemuU/9u//dt4x9e+9rUoP3To0Cj/4osvRvlSSjn11FOj/MEHHxzvSJ/jHnnkkXjHk08+Gc/QWnrzmiN9bTNv3rwov2rVqijfG+PHj49n2traovyGDRviHVdccUU8Qyb9OZZSytVXXx3lP/7xj8c70ufDO++8M8oPHDgwypdSysiRI6N8V1dXvKM394S+IX3duNdee8U7fvrTn0b59HfEt99+O8qXUkpHR0eU37ZtW7xj//33j/Kf+9zn4h0//OEP4xl23OWXXx7PXHzxxVF+6dKl8Y4zzjgjyqfvb/Tm/c+HHnooyvf09MQ7uru7o/ybb74Z75g8eXI8Q+buu++OZ0444YQo/9xzz8U7Xn311Sj/X//1X1H+5ZdfjvKllDJ8+PAo35v3gh577LEo39nZGe946aWX4pkP4pNJAAAAAAAAVCmTAAAAAAAAqFImAQAAAAAAUKVMAgAAAAAAoEqZBAAAAAAAQJUyCQAAAAAAgCplEgAAAAAAAFXKJAAAAAAAAKqUSQAAAAAAAFQpkwAAAAAAAKhSJgEAAAAAAFA1qNkHSN14443xzNlnnx3l77rrrnjHT37ykyh/wgknRPmRI0dG+VJKmTFjRpSfNWtWvGPJkiVRvqurK96R/jzSnzelHH744fHMV7/61Sg/ZcqUeMevf/3rKP+LX/wiyq9bty7Kl1LK/vvvH+UPPvjgeMfAgQOj/OrVq+MdP//5z+MZMsOGDYtn0sftvffeO96xYMGCKD969Ogof9hhh0X5Ukppa2uL8vfcc0+8I30u2bx5c7yD3d8rr7wSz2zfvj3Kf/zjH29ovpRSXn/99Sj/s5/9LN6RPud2dnbGO371q1/FM2R6enrimTlz5kT5Bx98MN7Rm3+Tid78TtKb142woxYvXhzlL7/88nhH+norfQ24du3aKF9KKe3t7VF+5cqV8Y70cW7jxo3xDhrrr//6r+OZo446KsqfdNJJ8Y6ZM2fGM33BmDFjonxvHhvOOOOMKN+b1zPknnnmmSj/3nvvxTs+97nPRflnn302yi9btizKl1LK/Pnz45lGGzFiRDxz4okn7pTdPpkEAAAAAABAlTIJAAAAAACAKmUSAAAAAAAAVcokAAAAAAAAqpRJAAAAAAAAVCmTAAAAAAAAqFImAQAAAAAAUKVMAgAAAAAAoEqZBAAAAAAAQJUyCQAAAAAAgCplEgAAAAAAAFXKJAAAAAAAAKoGNfsAu8LQoUMbvuPhhx9u+I7U2LFjm32E3zJp0qSGz5x99tnxjv7uxRdfjGdee+21KN+bn/2nP/3pKH/KKadE+TfeeCPKl1LK4sWLo/zWrVvjHYsWLYryL7/8cryjs7MzniFz++23N/sIH2jlypXNPkKfNXny5GYfgY/owgsvbPYRoOXdeuutUX6fffaJd6SPp5s3b47y3d3dUb6UUlavXh3lR48eHe9oa2uL8mvXro130JqmT58e5XvzWj6dWb58eZQfM2ZMlC8l/zffm7s7ePDgKJ/+LkbjzZ49O5757Gc/G+XnzZsX70ifq9asWRPlH3300ShfSv7+Z2/uVEdHR5RPH0t6s4PWdMQRRzT7CH3Wxo0b45mFCxfulN0+mQQAAAAAAECVMgkAAAAAAIAqZRIAAAAAAABVyiQAAAAAAACqlEkAAAAAAABUKZMAAAAAAACoUiYBAAAAAABQpUwCAAAAAACgSpkEAAAAAABAlTIJAAAAAACAKmUSAAAAAAAAVYOafYBd4fbbb2/2EZpi0qRJUX7JkiUNOslH25H+Oci98MIL8cxf/dVfNeAkH82QIUOi/B577BHv6OjoiPLbtm2LdwAA9BUrVqxo9hGaYt26dc0+AruR5cuXR/n0d5JSShk2bFiU37p1a5R/6623onwppXR1dUX5zZs3xzuWLl0az9BaTj311Hhm6tSpUf6www6LdxxzzDFRfsqUKVF+zZo1Ub6UUu68884o/8tf/jLesSuccsopzT4C9DmrV6/eKd/HJ5MAAAAAAACoUiYBAAAAAABQpUwCAAAAAACgSpkEAAAAAABAlTIJAAAAAACAKmUSAAAAAAAAVcokAAAAAAAAqpRJAAAAAAAAVCmTAAAAAAAAqFImAQAAAAAAUKVMAgAAAAAAoEqZBAAAAAAAQNWgD/vixIkTd9U5aIAlS5Y0+wg7xZw5c5p9BHYTnZ2dDc3TuoYOHRrlDzzwwAadhP7shBNOaPYRdpq/+Iu/iPJvv/12g05Cf/a///u/zT4C0McMHjy42Uf4QB0dHQ3fkf7us3jx4gadpPda9ee3bdu2Zh+haUaMGNHsI3yg9LVpb17LPvzww/FMq2nVn9+0adOafYSd6ktf+lKUnzVrVoNOQn+2devWnfJ9fDIJAAAAAACAKmUSAAAAAAAAVcokAAAAAAAAqpRJAAAAAAAAVCmTAAAAAAAAqFImAQAAAAAAUKVMAgAAAAAAoEqZBAAAAAAAQJUyCQAAAAAAgCplEgAAAAAAAFXKJAAAAAAAAKraenp6mn0GAAAAAAAAWpRPJgEAAAAAAFClTAIAAAAAAKBKmQQAAAAAAECVMgkAAAAAAIAqZRIAAAAAAABVyiQAAAAAAACqlEkAAAAAAABUKZMAAAAAAACoUiYBAAAAAABQpUwCAAAAAACgSpkEAAAAAABAlTIJAAAAAACAKmUSAAAAAAAAVcokAAAAAAAAqpRJAAAAAAAAVCmTAAAAAAAAqFImAQAAAAAAUKVMAgAAAAAAoEqZBAAAAAAAQJUyCQAAAAAAgCplEgAAAAAAAFXKJAAAAAAAAKqUSQAAAAAAAFQpkwAAAAAAAKhSJgEAAAAAAFClTAIAAAAAAKBKmQQAAAAAAECVMgkAAAAAAIAqZRIAAAAAAABVyiQAAAAAAACqlEkAAAAAAABUKZMAAAAAAACoUiYBAAAAAABQNejDvviFL3yhJ/lmhxxyyEc7DXyA9vb2KP/d7363rUFH2Sna2tqie9Wq9ttvvyi/fv36eMeGDRviGXbM1VdfHeX/7u/+zr2Cnaynp6dl79WAAQOiO9XT03+v4ODBg6P8kCFDonx/fi78zGc+E+Uff/zxlr1TpZRy8sknRxflySefbNRRdqm2tvzHMm7cuCifPgatXr06yvdnrfxcVYrXgIlRo0ZF+d78/saOaeV7NWbMmOhOrVu3rlFHgR3WyneqlFJOOeWU6F6l74PCjhg+fHiUv+eeez7wXvlkEgAAAAAAAFXKJAAAAAAAAKqUSQAAAAAAAFQpkwAAAAAAAKhSJgEAAAAAAFClTAIAAAAAAKBKmQQAAAAAAECVMgkAAAAAAIAqZRIAAAAAAABVyiQAAAAAAACqlEkAAAAAAABUDfqwL95777276hxQdfrppzf7CH3eeeedF8/su+++Uf62226LdwwePDjKr1mzJt7RX6V/t0D/0tPT0+wj7BR77rlnlB806ENfGn+gSZMmRfn58+fHO/qr559/vtlH2Kk6OjqafYSmaG9vj2fSuzt9+vQov2nTpihfSinXXXddPAM7asKECVE+vSOllPL2229H+f333z/KL168OMrTmtatW9fsI+w2jjvuuCi/devWKL9y5cooX0opy5cvj2dovNmzZ0f5vvK7GK3lwAMP3CnfxyeTAAAAAAAAqFImAQAAAAAAUKVMAgAAAAAAoEqZBAAAAAAAQJUyCQAAAAAAgCplEgAAAAAAAFXKJAAAAAAAAKqUSQAAAAAAAFQpkwAAAAAAAKhSJgEAAAAAAFClTAIAAAAAAKBKmQQAAAAAAEDVoGYfILXXXnvFM93d3VH+uOOOi3ece+65UX7p0qVR/rLLLovy9G/Dhg2L8n/zN38T73jggQei/MqVK+Md06dPj/IHHnhglL/66qujPCT23HPPeOZTn/pUlD/66KPjHem96ujoiPL/9E//FOVLKeXee++NZ+h/pk6dGs+8+eabUf69996Ld1xzzTVRft68eVH+tttui/LQakaOHBnPvPbaa1H++eefj/K9+d3quuuui2dgR51//vlRftasWfGOUaNGRflDDz00yi9evDjKQyNdfvnlUf6KK66Idzz99NNR/h//8R+j/DvvvBPlS8nv+fr16+Md9A1tbW3xTE9PT5RP/82XUspXvvKVKD9//vwof9JJJ0X5UvK/q/TvaXfjk0kAAAAAAABUKZMAAAAAAACoUiYBAAAAAABQpUwCAAAAAACgSpkEAAAAAABAlTIJAAAAAACAKmUSAAAAAAAAVcokAAAAAAAAqpRJAAAAAAAAVCmTAAAAAAAAqFImAQAAAAAAUDWo2QfYY489ovy//du/xTsWL14c5bdt2xbv+MQnPtHQPCQ6Ozuj/BNPPBHvePHFF+OZ1Pnnnx/l16xZ05iDQC9MmzYtnhk7dmyU//d///d4xw9+8IN4JjFu3LiGfn/6r+7u7ngmfT7szb39+te/HuVnz54d5W+77bYoX0opw4YNi/IdHR3xDvqvCy+8MMp3dXXFO9Lf+ebNmxflDznkkChfSilDhw6N8ps3b4530DccdNBB8Uz6fLV8+fJ4x9q1a6P89ddfH+UfeeSRKF9K7957of8599xz45nvfe97UX7mzJnxjs9//vNRfvLkyVH+xBNPjPKllHLnnXfGM/QNAwcOjPIXXXRRvGNXvJd9yy23RPlnn322QSd5X/oex6pVqxp0ktbgk0kAAAAAAABUKZMAAAAAAACoUiYBAAAAAABQpUwCAAAAAACgSpkEAAAAAABAlTIJAAAAAACAKmUSAAAAAAAAVcokAAAAAAAAqpRJAAAAAAAAVCmTAAAAAAAAqFImAQAAAAAAUKVMAgAAAAAAoGpQsw+wZcuWKL9x48Z4x/bt26P8bbfdFu+44447ovzee+8d74Ad1d3dHeX/9V//Nd4xZMiQKP+d73wn3pE+PvzLv/xLvAMa5fHHH2/2EZpizJgx8cyaNWuifFdXV7yD3d8rr7zS8B0rVqyIZ772ta9F+Ztuuinekero6Gj4DvqvF154IcqvW7euQSd53yWXXBLlN2zYEO/YvHlzPEP/1Jt/K7NmzYryK1eujHf86Ec/ivLp661t27ZFedhRS5cujWeOPPLIKD9//vx4R+r222+P8jNmzGjQSeiL0sfs119/vUEned99990XzyxevDjKDx8+PMr35v34dGbVqlXxjt2JTyYBAAAAAABQpUwCAAAAAACgSpkEAAAAAABAlTIJAAAAAACAKmUSAAAAAAAAVcokAAAAAAAAqpRJAAAAAAAAVCmTAAAAAAAAqFImAQAAAAAAUKVMAgAAAAAAoEqZBAAAAAAAQNWgZh8gNX369IbvGDlyZDxzxBFHRPnnn38+3gGNsmDBgnjm9NNPj/IHHXRQvOPuu++O8o8++mi8A3ZnAwbk/yeku7u7ASd536pVq+KZrq6uBpwEckceeWQ805vnUNidPfvss80+wm857LDDovzDDz/coJNAKUuXLm34jvT9h1JKmThxYpSfM2dOvAMa4fHHH2/2ET7Q/fffH+XHjx8f5W+++eYoD4kHH3yw2UfYKY499tgo39nZGe/41a9+Fc/0ZT6ZBAAAAAAAQJUyCQAAAAAAgCplEgAAAAAAAFXKJAAAAAAAAKqUSQAAAAAAAFQpkwAAAAAAAKhSJgEAAAAAAFClTAIAAAAAAKBKmQQAAAAAAECVMgkAAAAAAIAqZRIAAAAAAABVyiQAAAAAAACqBjX7ALvCsGHDovzEiRPjHRs3bozy69evj/IDBuS9X3d3dzwDO+qSSy6J8h0dHfGOW2+9NZ6B/qS9vT2e6ezsbMBJ3rdly5aGfn9I/Pmf/3mU37BhQ7zj0UcfjWeADzdz5swo39bWFuUvuOCCKA+t5tprr41nxowZE+U///nPxztgd3XkkUfGM4cffniU/+EPfxjvgP5m2rRpUX7r1q1RftmyZVG+lN69n9mX+WQSAAAAAAAAVcokAAAAAAAAqpRJAAAAAAAAVCmTAAAAAAAAqFImAQAAAAAAUKVMAgAAAAAAoEqZBAAAAAAAQJUyCQAAAAAAgCplEgAAAAAAAFXKJAAAAAAAAKqUSQAAAAAAAFQNavYBdoV99tknyo8aNSre8dxzz0X5tra2KN/d3R3lIXHRRRfFM11dXVH+xz/+cbwD+pv0uaGzs7NBJ3nfHnvsEeU3b97coJNAbuzYsVH+Zz/7WYNOAv3XxRdfHM+kv79997vfjXdAK7nyyiuj/KRJk+Idl156aTwD/cXNN98czyxdujTK33DDDfEO6G/Wr18f5bdv3x7l33rrrSjPb/PJJAAAAAAAAKqUSQAAAAAAAFQpkwAAAAAAAKhSJgEAAAAAAFClTAIAAAAAAKBKmQQAAAAAAECVMgkAAAAAAIAqZRIAAAAAAABVyiQAAAAAAACqlEkAAGk4f7MAAAkMSURBVAAAAABUKZMAAAAAAACoGtTsA6SmTJkSzxx66KFRfv78+fGOVE9PT8N3wI668sor45lZs2ZF+XvvvTfe0ReMGTMmnlm7dm0DTsKuNnLkyHhm9OjRUX7JkiXxjvRcGzZsiHdAI5x77rnxzKpVqxpwkt80fvz4KL8rztRX/P7v/36zj8BOMH369Hjm9ddfj/IzZ86Md0Arueyyy6L8Qw89FO/4xS9+Ec/A7urss8+O8tu2bYt3nHfeefEMtIohQ4bEM52dnVF+6tSp8Y6tW7dG+WXLlsU7+Gh8MgkAAAAAAIAqZRIAAAAAAABVyiQAAAAAAACqlEkAAAAAAABUKZMAAAAAAACoUiYBAAAAAABQpUwCAAAAAACgSpkEAAAAAABAlTIJAAAAAACAKmUSAAAAAAAAVcokAAAAAAAAqpRJAAAAAAAAVA1q9gFS5557bjyzaNGiKL9kyZJ4B43z1ltvNfsIfd6YMWPimbPPPjvKf/nLX453tKKhQ4dG+Y6Ojgad5KNZuHBhs4/Q540aNSqeSe9ib56vNmzYEM9AKzj++OPjma9//etR/pxzzol3HH300VH+W9/6VpTvzZkWL14c5bdu3RrveOWVV6L8e++9F+848cQT4xka75prrony6R0ppZTvf//7UT59fFi/fn2ULyW/V+lrxlJKGTFiRJT/nd/5nXjHXXfdFc+QOeWUU+KZ9Gf/5JNPxjugP9l7772j/A033BDvSJ8X+quTTz45ntm0aVMDTsL/1ZvX2SNHjozy9957b7yDxnnjjTd2yvfxySQAAAAAAACqlEkAAAAAAABUKZMAAAAAAACoUiYBAAAAAABQpUwCAAAAAACgSpkEAAAAAABAlTIJAAAAAACAKmUSAAAAAAAAVcokAAAAAAAAqpRJAAAAAAAAVCmTAAAAAAAAqBrU7AOkTjvttHhm0KDsjzl8+PB4x+zZs6P8ggUL4h2pww8/PMqfeeaZ8Y7x48dH+dWrV8c7Xn/99XiGzMKFC+OZI444ogEn+U377LNPlF+xYkWDTvK+zZs3N3zHrjBv3rxmH6HPW7t2bTxz5ZVXRvkvf/nL8Y7BgwdH+fRx++WXX47ypZSy7777RvmhQ4fGO9J/8735+T3zzDPxDDuuu7u74TtmzJixS2Za6fu3srvvvjvKf+Mb32jQSfi/nnrqqSh/6aWXxjt+/vOfR/njjjsuyo8aNSrKl5L//tab14xr1qyJ8hMnTox3jBs3Lp4hs2TJknjmgQceiPI33XRTvCN9vTVhwoQoP3fu3Ci/q6T3ff369Q06Cb119NFHxzPLly+P8hs2bIh3nHXWWVF+7NixUf7ggw+O8qWUsv/++0f5YcOGxTs6Ozuj/KZNm+Id11xzTTxDpjf/vtL35NM7UkopkydPjvLLli2L8i+99FKULyV/nfnHf/zH8Y6enp4o/+qrr8Y7brjhhnjmg/hkEgAAAAAAAFXKJAAAAAAAAKqUSQAAAAAAAFQpkwAAAAAAAKhSJgEAAAAAAFClTAIAAAAAAKBKmQQAAAAAAECVMgkAAAAAAIAqZRIAAAAAAABVyiQAAAAAAACqlEkAAAAAAABUKZMAAAAAAACoauvp6al/sa2t/sUm+fGPfxzPnHjiiVF+4sSJ8Y6hQ4dG+dWrV0f5t99+O8qXUsq6deui/Lhx4+IdGzZsiPILFy6MdzzzzDNR/kc/+lFbvGQXasV7NXXq1Hjm05/+dJR/6KGH4h1btmyJ8u3t7VF+xYoVUb4v+c53vhPlr7nmGvdqF2hry/6a//AP/zDekd73dMfgwYOjfCml3HDDDVG+N48nXV1d8UwqfQzasmVLy96rVrxTY8aMiWfWrl0b5ffbb794x/jx46P84sWLo/yAAfn//Vq/fn2U32uvveIdnZ2dUT79WZSSv4b/7//+75a9U6WUcuyxx0b3au7cuY06Sr+XPt+WUsqH/d7crB3Dhg2Ld3R0dET5np6elr5Xrfh81Rt/8id/EuUffPDBeEf6Gm3IkCFRPn1voJT8TNu2bYt3tKJWvleteKd68x7dhAkTovzKlSvjHelj8AEHHBDvSG3evDnKP/nkkw06ya7VyneqlFIGDBgQ3av09UBvHHnkkfHMl770pSh/7LHHxjsmTZoU5QcNGhTl0/fjSynlYx/7WJSfM2dOvGPRokVRft68efGO9HeLpUuXfuC98skkAAAAAAAAqpRJAAAAAAAAVCmTAAAAAAAAqFImAQAAAAAAUKVMAgAAAAAAoEqZBAAAAAAAQJUyCQAAAAAAgCplEgAAAAAAAFXKJAAAAAAAAKqUSQAAAAAAAFQpkwAAAAAAAKhq6+npqX+xra3+xd3I+PHjo/yQIUPiHYccckiUHzx4cJRfvXp1lC+llLVr10b5119/Pd6xK5x++ulRfubMmW0NOspO0Vfu1a7Q1pb9KNN8d3d3lO9Lrr322ih/1VVXuVewk/X09LTsvXKnaAUjR46M8u+9917L3qlSSjn22GOjezV37txGHQV2WCs/V5Xi+YrdUyvfK3eK3VEr36lSShkwYEB0rz7svfpmGjhwYJTv6upq0EnojQMPPDDKv/baax94r3wyCQAAAAAAgCplEgAAAAAAAFXKJAAAAAAAAKqUSQAAAAAAAFQpkwAAAAAAAKhSJgEAAAAAAFClTAIAAAAAAKBKmQQAAAAAAECVMgkAAAAAAIAqZRIAAAAAAABVyiQAAAAAAACqlEkAAAAAAABUDWr2AXaFVatWNXzHW2+91fAd/dU777zT7CPQJD09PQ3N92ePPfZYlL/qqqsadBIA+GDHH398s4+wU40bN67ZRwCgAUaNGtXsIwAtrK+8V9XV1dXsI/ARvPnmmzvl+/hkEgAAAAAAAFXKJAAAAAAAAKqUSQAAAAAAAFQpkwAAAAAAAKhSJgEAAAAAAFClTAIAAAAAAKBKmQQAAAAAAECVMgkAAAAAAIAqZRIAAAAAAABVyiQAAAAAAACqlEkAAAAAAABUtfX09DT7DAAAAAAAALQon0wCAAAAAACgSpkEAAAAAABAlTIJAAAAAACAKmUSAAAAAAAAVcokAAAAAAAAqpRJAAAAAAAAVP0/Lv+88fB/WY4AAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 2160x2160 with 16 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "torch.Size([32, 14, 14])\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 2160x2160 with 32 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "torch.Size([32, 14, 14])\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 2160x2160 with 32 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "torch.Size([32, 7, 7])\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 2160x2160 with 32 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}