Kopie_von_Kopie_von_DLO_Projekt.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "iO6v_3ep22Lb"
      },
      "outputs": [],
      "source": [
        "# https://drive.google.com/drive/folders/1hNFbb-Ksc9Qdyx0_CB7tTJFy7-UNwYwy?usp=sharing \n",
        "\n",
        "# Bildgröße 300x200 RGB\n",
        "\n",
        "from google.colab import drive\n",
        "import torchvision \n",
        "from torchvision import transforms\n",
        "\n",
        "from timeit import default_timer as timer\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "from torchsummary import summary\n",
        "\n",
        "import torch\n",
        "import torch.nn as nn\n",
        "from torch.optim import SGD\n",
        "from torch.utils.data import Dataset, random_split\n",
        "from torchvision import datasets, transforms\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QnHFd1D9Dg79",
        "outputId": "d2a3b13b-2d38-4ccf-97f8-b8919b034b40"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ],
      "source": [
        "drive.mount(\"/content/drive\", force_remount=True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "id": "AK7eqCSGDh4x"
      },
      "outputs": [],
      "source": [
        "# Config Feld\n",
        "path = f\"/content/drive/MyDrive/Colab Notebooks/data\"\n",
        "ModelPath = f'/content/drive/MyDrive/Colab Notebooks/models/pythorch.model'\n",
        "device = \"cuda\" if torch.cuda.is_available() else \"cpu\"\n",
        "TrainingMode = True\n",
        "\n",
        "valid_size = 0.2\n",
        "test_size = 0.1\n",
        "\n",
        "epochs = 50\n",
        "\n",
        "image_width = 300\n",
        "image_height = 200"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "'''dataset = datasets.ImageFolder(('/content/drive/MyDrive/Colab Notebooks/data'),transform = transforms.ToTensor())\n",
        "loader = torch.utils.data.DataLoader(dataset,\n",
        "                                     batch_size = 32,\n",
        "                                     num_workers = 0,\n",
        "                                     shuffle = False)\n",
        "mean = 0.0\n",
        "std = 0.0\n",
        "for images,_ in loader:\n",
        "  batch_samples = images.size(0)\n",
        "  images = images.view(batch_samples,images.size(1),-1)\n",
        "  mean += images.mean(2).sum(0)\n",
        "  std += images.std(2).sum(0)\n",
        "\n",
        "  \n",
        "mean = mean/len(loader.dataset)\n",
        "std /= len(loader.dataset)\n",
        "\n",
        "print(mean)\n",
        "print(std)'''"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 71
        },
        "id": "za_IavnPIRSg",
        "outputId": "92e50758-3345-44da-a242-4619e41ae865"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "\"dataset = datasets.ImageFolder(('/content/drive/MyDrive/Colab Notebooks/data'),transform = transforms.ToTensor())\\nloader = torch.utils.data.DataLoader(dataset,\\n                                     batch_size = 32,\\n                                     num_workers = 0,\\n                                     shuffle = False)\\nmean = 0.0\\nstd = 0.0\\nfor images,_ in loader:\\n  batch_samples = images.size(0)\\n  images = images.view(batch_samples,images.size(1),-1)\\n  mean += images.mean(2).sum(0)\\n  std += images.std(2).sum(0)\\n\\n  \\nmean = mean/len(loader.dataset)\\nstd /= len(loader.dataset)\\n\\nprint(mean)\\nprint(std)\""
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "id": "82OKg_XvLbtj"
      },
      "outputs": [],
      "source": [
        "#  transforms.CenterCrop(224),\n",
        "transform = transforms.Compose([transforms.Resize((image_height, image_width)),\n",
        "                                transforms.ToTensor(),\n",
        "                                transforms.Normalize((0.4592, 0.5382, 0.4111),(0.1968, 0.1375, 0.1554))])\n",
        "\n",
        "dataset = datasets.ImageFolder(path, transform=transform)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "id": "lj33ceFs_smg",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        },
        "outputId": "390f38b9-572e-47f5-a41b-063e2825a90b"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "dataset[0][0][2]\n",
        "plt.imshow(dataset[0][0][0])\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "id": "A2JECJ2VTFr2"
      },
      "outputs": [],
      "source": [
        "# https://discuss.pytorch.org/t/how-to-use-imagefolder-with-list-of-images-for-train-and-test-sets/120247\n",
        "# trainset = torch.utils.data.Subset(dataset_train, train_indices)\n",
        "# valset = torch.utils.data.Subset(dataset_val, val_indices)\n",
        "def get_data(): \n",
        "  n_val = int(np.floor(valid_size * len(dataset)))\n",
        "  n_test = int(np.floor(test_size * len(dataset)))\n",
        "  n_train = len(dataset) - n_val - n_test\n",
        "\n",
        "  train_ds, val_ds, test_ds = random_split(dataset, [n_train, n_val, n_test])\n",
        "  train_dl = torch.utils.data.DataLoader(train_ds, batch_size=32, shuffle=True)\n",
        "\n",
        "  valid_dl = torch.utils.data.DataLoader(val_ds, batch_size=32, shuffle=True)\n",
        "  test_dl = torch.utils.data.DataLoader(test_ds, batch_size=32, shuffle=True)\n",
        "  return train_dl, test_dl, valid_dl"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "id": "Fr2epBmZOsTr"
      },
      "outputs": [],
      "source": [
        "@torch.no_grad()\n",
        "def accuracy(x, y, model):\n",
        "    with torch.no_grad():\n",
        "        prediction = model(x)\n",
        "    max_values, argmaxes = prediction.max(-1)\n",
        "    is_correct = argmaxes == y\n",
        "    return is_correct.cpu().numpy().tolist()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "id": "1h9VGc6TOsCT"
      },
      "outputs": [],
      "source": [
        "@torch.no_grad()\n",
        "def loss(x, y, model):\n",
        "    prediction = model(x)\n",
        "    loss = loss_fn(prediction, y)\n",
        "    return loss.item()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "id": "sFnwWJldSB1H",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 89
        },
        "outputId": "a93d9ffb-4d24-4aad-d34d-4ba545109a30"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'\\nclass Net(nn.Module):\\n    def __init__(self):\\n        super().__init__()\\n        self.conv1 = nn.Conv2d(3, 16, kernel_size=3, padding=1)\\n        self.act1 = nn.Tanh()\\n        self.pool1 = nn.MaxPool2d(2) # halbiert um 2\\n        self.conv2 = nn.Conv2d(16, 8, kernel_size=3, padding=1)\\n        self.act2 = nn.Tanh()\\n        self.pool2 = nn.MaxPool2d(2) # halbiert um 2\\n        self.fc1 = nn.Linear(8 * int((image_width/2)/2) * int((image_height/2)/2), 32)\\n        self.act3 = nn.Tanh()\\n        self.fc2 = nn.Linear(32, 4)\\n\\n    def forward(self, x):\\n        out = self.pool1(self.act1(self.conv1(x)))\\n        out = self.pool2(self.act2(self.conv2(out)))\\n        out = out.view(-1, 8 * int((image_width/2)/2) * int((image_height/2)/2)) # <1>\\n        out = self.act3(self.fc1(out))\\n        out = self.fc2(out)\\n        return out\\n        #'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 10
        }
      ],
      "source": [
        "'''\n",
        "class Net(nn.Module):\n",
        "    def __init__(self):\n",
        "        super().__init__()\n",
        "        self.conv1 = nn.Conv2d(3, 16, kernel_size=3, padding=1)\n",
        "        self.act1 = nn.Tanh()\n",
        "        self.pool1 = nn.MaxPool2d(2) # halbiert um 2\n",
        "        self.conv2 = nn.Conv2d(16, 8, kernel_size=3, padding=1)\n",
        "        self.act2 = nn.Tanh()\n",
        "        self.pool2 = nn.MaxPool2d(2) # halbiert um 2\n",
        "        self.fc1 = nn.Linear(8 * int((image_width/2)/2) * int((image_height/2)/2), 32)\n",
        "        self.act3 = nn.Tanh()\n",
        "        self.fc2 = nn.Linear(32, 4)\n",
        "\n",
        "    def forward(self, x):\n",
        "        out = self.pool1(self.act1(self.conv1(x)))\n",
        "        out = self.pool2(self.act2(self.conv2(out)))\n",
        "        out = out.view(-1, 8 * int((image_width/2)/2) * int((image_height/2)/2)) # <1>\n",
        "        out = self.act3(self.fc1(out))\n",
        "        out = self.fc2(out)\n",
        "        return out\n",
        "        #'''"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "class Net(nn.Module):\n",
        "    def __init__(self):\n",
        "        super().__init__()\n",
        "        self.conv1 = nn.Conv2d(3, 96, kernel_size=11, stride=4, padding=0)\n",
        "        self.act1 = nn.ReLU()\n",
        "        self.pool1 = nn.MaxPool2d(kernel_size=3, stride=2)\n",
        "        self.conv2 = nn.Conv2d(96, 256, kernel_size=5, padding=2)\n",
        "        self.act2 = nn.ReLU()\n",
        "        self.pool2 = nn.MaxPool2d(kernel_size=3, stride=2)\n",
        "        self.conv3 = nn.Conv2d(256, 384, kernel_size=3, padding=1)\n",
        "        self.act3 = nn.ReLU()\n",
        "        self.conv4 = nn.Conv2d(384, 384, kernel_size=3, padding=1)\n",
        "        self.act4 = nn.ReLU()\n",
        "        self.conv5 = nn.Conv2d(384, 256, kernel_size=3, padding=1)\n",
        "        self.act5 = nn.ReLU()\n",
        "        self.pool3 = nn.MaxPool2d(kernel_size=3, stride=2)\n",
        "        self.fc1 = nn.Linear(10240, 4096)\n",
        "        self.act6 = nn.ReLU()\n",
        "        self.fc2 = nn.Linear(4096, 4096)\n",
        "        self.act7 = nn.ReLU()\n",
        "        self.fc3 = nn.Linear(4096, 4)\n",
        "        self.softmax = nn.Softmax(1)\n",
        "\n",
        "    def forward(self, x):\n",
        "        out = self.pool1(self.act1(self.conv1(x)))\n",
        "        out = self.pool2(self.act2(self.conv2(out)))\n",
        "        out = self.act3(self.conv3(out))\n",
        "        out = self.act4(self.conv4(out))\n",
        "        out = self.pool3(self.act5(self.conv5(out)))\n",
        "        out = out.view(-1,10240) # <1>\n",
        "        out = self.act6(self.fc1(out))\n",
        "        out = self.act7(self.fc2(out))\n",
        "        out = self.fc3(out)\n",
        "        out = self.softmax(out)\n",
        "        return out\n",
        "        \n",
        "#'''"
      ],
      "metadata": {
        "id": "Kb9fu8syQo0u"
      },
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "id": "2g1_9v9COOZG"
      },
      "outputs": [],
      "source": [
        "def get_model():\n",
        "    model = Net().to(device)\n",
        "    loss_fn = nn.CrossEntropyLoss()  # Kreuzentropie\n",
        "    optimizer = SGD(model.parameters(), lr=1e-1)\n",
        "    return model, loss_fn, optimizer"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "metadata": {
        "id": "an8Tp-JPPzEh"
      },
      "outputs": [],
      "source": [
        "def train_batch(x, y, model, opt, loss_fn):\n",
        "    model.train()\n",
        "    prediction = model(x)\n",
        "    batch_loss = loss_fn(prediction, y)\n",
        "    batch_loss.backward()\n",
        "    optimizer.step()\n",
        "    optimizer.zero_grad()\n",
        "    return batch_loss.item()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "metadata": {
        "id": "PoCkjxT7qKaZ"
      },
      "outputs": [],
      "source": [
        "def is_early_stopping(val_accuracies, amount_epoch, threshold=0.01):\n",
        "\n",
        "  # Wenn man mindestens <amount_epoch> vergleichen möchte aber noch nicht \n",
        "  # so viele epochen durchlaufen hat, kann noch kein Early Stopping stattfinden\n",
        "  if len(val_accuracies) < amount_epoch: \n",
        "    return False\n",
        "\n",
        "  # nur die letzten <amount_epoch> Elemente übernehmen\n",
        "  epochs = val_accuracies[-amount_epoch:]\n",
        "\n",
        "  # Betragliche differenz zwischen den Werten für index i und i+1 bilden\n",
        "  abs_differences  = []\n",
        "  for i in range(0 , len(epochs) -1):\n",
        "    abs_differences.append(abs(epochs[i] - epochs[i + 1]))\n",
        "    \n",
        "  # Wenn alle Differenzen in der Liste kleiner sind als der Schwellwert ist Early Stopping!\n",
        "  return all(x < threshold for x in abs_differences)\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 15,
      "metadata": {
        "id": "aPp5jG4awg1L"
      },
      "outputs": [],
      "source": [
        "# Test Early stopping\n",
        "assert is_early_stopping([0.1, 0.2], 4, 0.11) == False, \"Sollte False sein, da len(val_accuracies) kleiner ist als die mindestens erfoderlichen 4 epochen zum prüfen der differenz\"\n",
        "assert is_early_stopping([0.1, 0.2, 0.3, 0.4], 4, 0.1) == False, \"Sollte False sein, da 0.1 < 0.1 false ist\"\n",
        "assert is_early_stopping([0.1, 0.1, 0.2, 0.3], 4, 0.11) == True, \"Sollte True sein, da 0.1 < 0.11 True ist\"\n",
        "assert is_early_stopping([0.1, 0.2, 0.3, 0.5, 0.7, 0.9], 4, 0.11) == False, \"Sollte False sein, da die letzten drei Größere differenzen haben als 0.11\"\n",
        "# Aus Trainingsbeispiel\n",
        "assert is_early_stopping([96.15, 100.0, 96.15, 96.15, 96.15, 96.15, 84.62], 5, 1) == False, \"Sollte False sein weil der Letzte wert nicht passt\""
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "metadata": {
        "id": "AwBLvp6CP-Z8"
      },
      "outputs": [],
      "source": [
        "model, loss_fn, optimizer = get_model()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "metadata": {
        "id": "L19rLzgotmx1"
      },
      "outputs": [],
      "source": [
        "# 3 - R, G, B \n",
        "# 224 - Width \n",
        "# 224 - Height\n",
        "#summary(Net(), (3, 300, 200))\n",
        "# Output Shape\n",
        "# [-1, 16, 224, 224] \n",
        "# First Index: Batch-Size (-1 means it doesn't matter)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "metadata": {
        "id": "4kQ7V2XiWTIe"
      },
      "outputs": [],
      "source": [
        "train_dl, test_dl, val_dl = get_data()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "5jYt2fTSPVSw",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b496acc6-53fd-4e89-d4ca-2d22698414d6"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Starting training...\n",
            "epoch: 0  train_acc: 32.06%  val_acc: 32.00%  time: 1653.8301525650002s\n",
            "epoch: 1  train_acc: 47.70%  val_acc: 44.00%  time: 502.78939960699995s\n",
            "epoch: 2  train_acc: 62.13%  val_acc: 64.00%  time: 507.374833629s\n",
            "epoch: 3  train_acc: 64.31%  val_acc: 56.00%  time: 500.79833370000006s\n",
            "epoch: 4  train_acc: 71.60%  val_acc: 56.00%  time: 500.819037406s\n",
            "epoch: 5  train_acc: 74.56%  val_acc: 80.00%  time: 503.23231246000023s\n",
            "epoch: 6  train_acc: 66.33%  val_acc: 52.00%  time: 500.1277833280001s\n",
            "epoch: 7  train_acc: 77.10%  val_acc: 76.00%  time: 500.35383976899993s\n",
            "epoch: 8  train_acc: 79.51%  val_acc: 60.00%  time: 501.1703096729998s\n",
            "epoch: 9  train_acc: 81.02%  val_acc: 60.00%  time: 503.7784219619998s\n",
            "epoch: 10  train_acc: 83.52%  val_acc: 84.00%  time: 500.44785822499944s\n",
            "epoch: 11  train_acc: 76.77%  val_acc: 60.00%  time: 501.0394723240006s\n",
            "epoch: 12  train_acc: 60.13%  val_acc: 56.00%  time: 502.3506264810003s\n",
            "epoch: 13  train_acc: 67.36%  val_acc: 72.00%  time: 503.81436637500065s\n",
            "epoch: 14  train_acc: 80.40%  val_acc: 72.00%  time: 507.69961302900083s\n",
            "epoch: 15  train_acc: 66.66%  val_acc: 60.00%  time: 506.55947083400133s\n",
            "epoch: 16  train_acc: 42.11%  val_acc: 48.00%  time: 509.79775147299915s\n",
            "epoch: 17  train_acc: 50.75%  val_acc: 56.00%  time: 504.98503094199987s\n",
            "epoch: 18  train_acc: 37.42%  val_acc: 36.00%  time: 501.84459189100016s\n",
            "epoch: 19  train_acc: 23.55%  val_acc: 20.00%  time: 501.0036248389988s\n",
            "epoch: 20  train_acc: 28.56%  val_acc: 12.00%  time: 497.83084375999897s\n",
            "epoch: 21  train_acc: 22.90%  val_acc: 24.00%  time: 497.34519630900104s\n",
            "epoch: 22  train_acc: 24.41%  val_acc: 12.00%  time: 497.45698837099917s\n",
            "epoch: 23  train_acc: 28.43%  val_acc: 20.00%  time: 495.96622986700095s\n"
          ]
        }
      ],
      "source": [
        "print('Starting training...')\n",
        "\n",
        "time0 = timer()\n",
        "\n",
        "\n",
        "\n",
        "arrPlotX = []\n",
        "train_losses, train_accuracies = [], []\n",
        "val_losses, val_accuracies = [], []\n",
        "epoch_time = []\n",
        "\n",
        "if TrainingMode:\n",
        "    for epoch in range(epochs):\n",
        "        time1 = timer()\n",
        "        train_epoch_losses, train_epoch_accuracies = [], []\n",
        "\n",
        "        for ix, batch in enumerate(iter(train_dl)):\n",
        "            x, y = batch\n",
        "            \n",
        "            batch_loss = train_batch(x, y, model, optimizer, loss_fn)  # Mini-Batch\n",
        "            train_epoch_losses.append(batch_loss)\n",
        "        train_epoch_loss = np.array(train_epoch_losses).mean()\n",
        "\n",
        "        for ix, batch in enumerate(iter(train_dl)):\n",
        "            x, y = batch\n",
        "            is_correct = accuracy(x, y, model)\n",
        "            train_epoch_accuracies.extend(is_correct)\n",
        "        train_epoch_accuracy = np.mean(train_epoch_accuracies)\n",
        "\n",
        "        val_is_correct = None\n",
        "        validation_loss = None\n",
        "\n",
        "        for ix, batch in enumerate(iter(val_dl)):\n",
        "            x, y = batch\n",
        "            val_is_correct = accuracy(x, y, model)\n",
        "            validation_loss = loss(x, y, model)\n",
        "\n",
        "        val_epoch_accuracy = np.mean(val_is_correct)\n",
        "        arrPlotX.append(epoch)\n",
        "        train_losses.append(train_epoch_loss)\n",
        "        train_accuracies.append(train_epoch_accuracy)\n",
        "        val_losses.append(validation_loss)\n",
        "        val_accuracies.append(val_epoch_accuracy)\n",
        "\n",
        "        time2 = timer() - time1\n",
        "        epoch_time.append(time2)\n",
        "        print(f\"epoch: {epoch}  train_acc: {100 * train_epoch_accuracy:.2f}%  val_acc: {100 * val_epoch_accuracy:.2f}%  time: {time2}s\")\n",
        "        # Early stopping\n",
        "        if is_early_stopping(val_accuracies, 5, 0.01):\n",
        "            print(f'Early Stopping...')\n",
        "            break\n",
        "\n",
        "        \n",
        "\n",
        "   # torch.save(model.state_dict(), ModelPath)\n",
        "\n",
        "    plt.plot(arrPlotX, train_accuracies)\n",
        "    plt.plot(arrPlotX, val_accuracies)\n",
        "    plt.savefig(\"accuracies.png\")\n",
        "else:\n",
        "    val_is_correct = None\n",
        "    for ix, batch in enumerate(iter(test_dl)):\n",
        "        x, y = batch\n",
        "        val_is_correct = accuracy(x, y, model)\n",
        "        validation_loss = loss(x, y, model)\n",
        "\n",
        "    for ix, var in enumerate(iter(val_is_correct)):\n",
        "        if var is False:\n",
        "            print(test_dl.dataset.__getitem__(ix))\n",
        "    val_epoch_accuracy = np.mean(val_is_correct)\n",
        "    print(f\"test_acc: {100 * val_epoch_accuracy:.2f}%\")\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "rkGk1lUysqSQ"
      },
      "outputs": [],
      "source": [
        "val_is_correct = None\n",
        "for ix, batch in enumerate(iter(test_dl)):\n",
        "    x, y = batch\n",
        "    val_is_correct = accuracy(x, y, model)\n",
        "    validation_loss = loss(x, y, model)\n",
        "\n",
        "for ix, var in enumerate(iter(val_is_correct)):\n",
        "    if var is False:\n",
        "        print(test_dl.dataset.__getitem__(ix))\n",
        "val_epoch_accuracy = np.mean(val_is_correct)\n",
        "print(f\"test_acc: {100 * val_epoch_accuracy:.2f}%\")"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "collapsed_sections": [],
      "name": "Kopie von Kopie von DLO-Projekt.ipynb",
      "provenance": [],
      "toc_visible": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    },
    "gpuClass": "standard"
  },
  "nbformat": 4,
  "nbformat_minor": 0
}