Arda Karaman CENG501 - Spring2022 - PA1 - Task 4 - NN with Pytorch.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"Copy of CENG501 - Spring2022 - PA1 - Task 4 - NN with Pytorch.ipynb","provenance":[{"file_id":"1Q9EJHlOh4B8lHg3YvSuS1aYQdRIszdtb","timestamp":1648468518246}],"collapsed_sections":[],"toc_visible":true},"kernelspec":{"name":"python3","display_name":"Python 3"},"language_info":{"name":"python"},"accelerator":"GPU","widgets":{"application/vnd.jupyter.widget-state+json":{"802771b712074afa9e6b57a341d3e2f9":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_ae890a642f1d4ecca3da4f088a5f74df","IPY_MODEL_3b6df0eef87e4225ad0d8a7916d44d16","IPY_MODEL_4fd44aa07b254503999abfd26f490ab1"],"layout":"IPY_MODEL_33f583872cac456c8ff8556699b24e55"}},"ae890a642f1d4ecca3da4f088a5f74df":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_817f73e575cf4c15aaa1808bf577dff1","placeholder":"​","style":"IPY_MODEL_99461a818a044fea95c9100760ef1e91","value":""}},"3b6df0eef87e4225ad0d8a7916d44d16":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_71bcc951ebcf4b1983f31737ed87cd8e","max":170498071,"min":0,"orientation":"horizontal","style":"IPY_MODEL_11ab6d6a9cdc4ae18da1464cebd69625","value":170498071}},"4fd44aa07b254503999abfd26f490ab1":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_c0bd3d2012ce4a1d9f02b8ee6fb8e255","placeholder":"​","style":"IPY_MODEL_ac26ae5fe7d34780a6b1a4d39ffa4331","value":" 170499072/? [00:06&lt;00:00, 32346696.15it/s]"}},"33f583872cac456c8ff8556699b24e55":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"817f73e575cf4c15aaa1808bf577dff1":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"99461a818a044fea95c9100760ef1e91":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"71bcc951ebcf4b1983f31737ed87cd8e":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"11ab6d6a9cdc4ae18da1464cebd69625":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"c0bd3d2012ce4a1d9f02b8ee6fb8e255":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"ac26ae5fe7d34780a6b1a4d39ffa4331":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}}}}},"cells":[{"cell_type":"markdown","metadata":{"id":"1X4X2hmaEmip"},"source":["# Task 4: NN with PyTorch\n","# CENG501 - Spring 2022 - PA1\n","\n","In this task, you will implement a fully-connected neural network (a.k.a. multi-layer perceptron) with PyTorch. In the previous tasks, we implemented everything from scratch (well, almost). Using PyTorch, we will see that loading widely used datasets, defining models and training them are very easy.\n"]},{"cell_type":"markdown","metadata":{"id":"oRaKrjO5JSu-"},"source":["## 1 Import the Modules\n","\n","Let us start with importing some libraries that we will use throughout the task."]},{"cell_type":"code","metadata":{"id":"v4L5nogMKyNx","executionInfo":{"status":"ok","timestamp":1650222129960,"user_tz":-180,"elapsed":8512,"user":{"displayName":"Arda Karaman","userId":"09649945184885111594"}}},"source":["import matplotlib.pyplot as plt # For plotting\n","import numpy as np              # NumPy, for working with arrays/tensors \n","import time                     # For measuring time\n","\n","# PyTorch libraries:\n","import torch\n","import torchvision\n","import torchvision.transforms as transforms\n","\n","%matplotlib inline\n","plt.rcParams['figure.figsize'] = [12, 8]\n","plt.rcParams['figure.dpi'] = 100 # 200 e.g. is really fine, but slower"],"execution_count":1,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"4BXVQOGRI4Sc"},"source":["### 1.1 Enable GPU\n","\n","From \"Edit -> Notebook Settings -> Hardware accelerator\" select GPU. With the following we will specify to PyTorch that we want to use the GPU."]},{"cell_type":"code","metadata":{"id":"97DJEyArJLcR","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1650222132562,"user_tz":-180,"elapsed":350,"user":{"displayName":"Arda Karaman","userId":"09649945184885111594"}},"outputId":"a7261d8d-6055-42a6-bbf9-88c4839ca0fc"},"source":["if torch.cuda.is_available():\n","  print(\"Cuda (GPU support) is available and enabled!\")\n","  device = torch.device(\"cuda\")\n","else:\n","  print(\"Cuda (GPU support) is not available :(\")\n","  device = torch.device(\"cpu\")"],"execution_count":2,"outputs":[{"output_type":"stream","name":"stdout","text":["Cuda (GPU support) is available and enabled!\n"]}]},{"cell_type":"markdown","metadata":{"id":"UMmi17e-JX7o"},"source":["## 2 The Dataset\n","\n","We will use torchvision.datasets to download the CIFAR10 dataset. "]},{"cell_type":"code","metadata":{"id":"MlXFbhR8Aq0m","colab":{"base_uri":"https://localhost:8080/","height":100,"referenced_widgets":["802771b712074afa9e6b57a341d3e2f9","ae890a642f1d4ecca3da4f088a5f74df","3b6df0eef87e4225ad0d8a7916d44d16","4fd44aa07b254503999abfd26f490ab1","33f583872cac456c8ff8556699b24e55","817f73e575cf4c15aaa1808bf577dff1","99461a818a044fea95c9100760ef1e91","71bcc951ebcf4b1983f31737ed87cd8e","11ab6d6a9cdc4ae18da1464cebd69625","c0bd3d2012ce4a1d9f02b8ee6fb8e255","ac26ae5fe7d34780a6b1a4d39ffa4331"]},"executionInfo":{"status":"ok","timestamp":1650222146527,"user_tz":-180,"elapsed":11647,"user":{"displayName":"Arda Karaman","userId":"09649945184885111594"}},"outputId":"3267719f-6d95-455a-b689-8ea4b27d47c8"},"source":["transform = transforms.Compose(\n","    [transforms.ToTensor(),\n","     transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))])\n","\n","batch_size = 200\n","\n","trainset = torchvision.datasets.CIFAR10(root='./data', train=True,\n","                                        download=True, transform=transform)\n","trainloader = torch.utils.data.DataLoader(trainset, batch_size=batch_size,\n","                                          shuffle=True, num_workers=2)\n","\n","testset = torchvision.datasets.CIFAR10(root='./data', train=False,\n","                                       download=True, transform=transform)\n","testloader = torch.utils.data.DataLoader(testset, batch_size=batch_size,\n","                                         shuffle=False, num_workers=2)\n","\n","classes = ('plane', 'car', 'bird', 'cat',\n","           'deer', 'dog', 'frog', 'horse', 'ship', 'truck')"],"execution_count":3,"outputs":[{"output_type":"stream","name":"stdout","text":["Downloading https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz to ./data/cifar-10-python.tar.gz\n"]},{"output_type":"display_data","data":{"text/plain":["  0%|          | 0/170498071 [00:00<?, ?it/s]"],"application/vnd.jupyter.widget-view+json":{"version_major":2,"version_minor":0,"model_id":"802771b712074afa9e6b57a341d3e2f9"}},"metadata":{}},{"output_type":"stream","name":"stdout","text":["Extracting ./data/cifar-10-python.tar.gz to ./data\n","Files already downloaded and verified\n"]}]},{"cell_type":"markdown","metadata":{"id":"0i9tug4jGGg4"},"source":["### 2.1 Visualize Samples"]},{"cell_type":"code","metadata":{"id":"aiLTjQoKBIpQ","colab":{"base_uri":"https://localhost:8080/","height":264},"executionInfo":{"status":"ok","timestamp":1650222155318,"user_tz":-180,"elapsed":2701,"user":{"displayName":"Arda Karaman","userId":"09649945184885111594"}},"outputId":"632d079e-e9f3-4d57-b00a-397de962c1a0"},"source":["dataiter = iter(trainloader)\n","classes = ['plane', 'car', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck']\n","S = 4\n","\n","for i in range(S):\n","  for j in range(S):\n","    images, labels = dataiter.next()\n","    X = np.transpose(images[0].numpy()/2+0.5, (1, 2, 0))\n","    y = labels[0]\n","\n","    plt.subplot(S, S, i*S+j+1)\n","    plt.imshow(X)\n","    plt.axis('off')\n","    plt.title(classes[y])\n","    plt.subplots_adjust(hspace = 0.5)\n","\n","plt.show()"],"execution_count":4,"outputs":[{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 16 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"kUmdAA1QJvRx"},"source":["## 4 Define and Train a Small Model\n","\n","Now, all the pieces are ready and we can define a model."]},{"cell_type":"markdown","metadata":{"id":"fJq7aT_wh9SK"},"source":["### 4.1 Model Definition\n","\n","Thanks to our modular definition, we can define models in a very simple manner as follows. Note that this is very similar to how you would define a model in PyTorch. For how to create a linear layer and add non-linearity, check out the [Linear](https://pytorch.org/docs/stable/generated/torch.nn.Linear.html) and [ReLU](https://pytorch.org/docs/stable/nn.functional.html?highlight=relu#torch.nn.functional.relu) document pages of PyTorch. The linear layer weights should have bias."]},{"cell_type":"code","metadata":{"id":"n13wrSZTiDEp","executionInfo":{"status":"ok","timestamp":1650222158649,"user_tz":-180,"elapsed":5,"user":{"displayName":"Arda Karaman","userId":"09649945184885111594"}}},"source":["import torch.nn as nn\n","import torch.nn.functional as F\n","\n","class TwoLayerNet(nn.Module):\n","    def __init__(self, D, H, C):\n","        super(TwoLayerNet, self).__init__()\n","        torch.manual_seed(501)        \n","        self.fc1 = nn.Linear(D, H)\n","        self.fc2 = nn.Linear(H, C)\n","        ###########################################################\n","        # @TODO: Create two linear layers                         #\n","        #  (1) One with size D x H and stored in self.fc1         #\n","        #  (2) Second with size H x C and stored in self.fc2      #\n","        ###########################################################\n","        pass\n","        ###########################################################\n","        #                         END OF YOUR CODE                #\n","        ###########################################################\n","\n","    def forward(self, x):\n","      ###########################################################\n","      # @TODO: Feedforward x through the layers. Be careful     #\n","      # about the shape of x. It needs to be reshaped to        #\n","      # (batchsize, 3072) first. Needs to return x. Apply relu  #\n","      # nonlinearity after the first layer.                     #\n","      ###########################################################\n","      x = x.view(200, -1)\n","      x = self.fc1(x)\n","      x = F.relu(x)\n","      x = self.fc2(x)\n","      pass\n","      ###########################################################\n","      #                         END OF YOUR CODE                #\n","      ###########################################################\n","      return x"],"execution_count":5,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"0r8pT7Y9FmDH"},"source":["### 4.2 Define Loss Function and the Optimizer\n","\n","We will create an instance of our model and \"define\" which loss function we want to use. We will also state our choice for the optimizer here.\n","\n","For more information, check the PyTorch docs: [CrossEntropyLoss](https://pytorch.org/docs/stable/generated/torch.nn.CrossEntropyLoss.html) and [SGD](https://pytorch.org/docs/stable/optim.html#torch.optim.SGD)."]},{"cell_type":"code","metadata":{"id":"uEn3vti5Fmhp","executionInfo":{"status":"ok","timestamp":1650222163358,"user_tz":-180,"elapsed":1436,"user":{"displayName":"Arda Karaman","userId":"09649945184885111594"}}},"source":["import torch.optim as optim\n","\n","D = 3072 # dimensionality\n","C = 10 # num of classes\n","H = 200 # number of hidden neurons\n","\n","model = TwoLayerNet(D, H, C)\n","\n","criterion = nn.CrossEntropyLoss()\n","optimizer = optim.SGD(model.parameters(), lr=1e-2, weight_decay=1e-1)"],"execution_count":6,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"CvpiVF5uiBE9"},"source":["### 4.3 Train the Model\n","\n","We will train our the cuda device for training our model. Since our model is too small, we may not see any benefit of the GPU. In fact, the transfer between the CPU and the GPU might exceed the benefit we get from the speedup we obtain from the GPU."]},{"cell_type":"code","metadata":{"id":"5Flgr2C6Xb2s","executionInfo":{"status":"ok","timestamp":1650222167032,"user_tz":-180,"elapsed":1029,"user":{"displayName":"Arda Karaman","userId":"09649945184885111594"}}},"source":["\n","def train(model, criterion, optimizer, epochs, dataloader, verbose=True):\n","  \"\"\"\n","    Define the trainer function. We can use this for training any model.\n","    The parameter names are self-explanatory.\n","\n","    Returns: the loss history.\n","  \"\"\"\n","  loss_history = [] \n","  for epoch in range(epochs):\n","    for i, data in enumerate(dataloader, 0):    \n","      \n","      # Our batch:\n","      inputs, labels = data\n","      inputs = inputs.to(device)\n","      labels = labels.to(device)\n","\n","      # zero the gradients as PyTorch accumulates them\n","      optimizer.zero_grad()\n","\n","      # Obtain the scores\n","      outputs = model(inputs)\n","\n","      # Calculate loss\n","      loss = criterion(outputs.to(device), labels)\n","\n","      # Backpropagate\n","      loss.backward()\n","\n","      # Update the weights\n","      optimizer.step()\n","\n","      loss_history.append(loss.item())\n","    \n","    if verbose: print(f'Epoch {epoch} / {epochs}: avg. loss of last 5 iterations {np.sum(loss_history[:-6:-1])/5}')\n","\n","  return loss_history"],"execution_count":7,"outputs":[]},{"cell_type":"code","metadata":{"id":"8Gg_Xp3vrg42","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1650222288688,"user_tz":-180,"elapsed":118284,"user":{"displayName":"Arda Karaman","userId":"09649945184885111594"}},"outputId":"a05e7e66-ab3b-46c1-abf2-431c56acf0b4"},"source":["model = model.to(device)\n","epochs = 10\n","loss_history = train(model, criterion, optimizer, epochs, trainloader)"],"execution_count":8,"outputs":[{"output_type":"stream","name":"stdout","text":["Epoch 0 / 10: avg. loss of last 5 iterations 1.9326148271560668\n","Epoch 1 / 10: avg. loss of last 5 iterations 1.8747639894485473\n","Epoch 2 / 10: avg. loss of last 5 iterations 1.8535312175750733\n","Epoch 3 / 10: avg. loss of last 5 iterations 1.8426478147506713\n","Epoch 4 / 10: avg. loss of last 5 iterations 1.8215257167816161\n","Epoch 5 / 10: avg. loss of last 5 iterations 1.790655755996704\n","Epoch 6 / 10: avg. loss of last 5 iterations 1.8052138090133667\n","Epoch 7 / 10: avg. loss of last 5 iterations 1.7893161535263062\n","Epoch 8 / 10: avg. loss of last 5 iterations 1.7860467672348022\n","Epoch 9 / 10: avg. loss of last 5 iterations 1.8112120151519775\n"]}]},{"cell_type":"markdown","metadata":{"id":"DILBn6y92Ceg"},"source":["### 4.4 The Loss Curve\n","\n","Let us visualize the loss curve."]},{"cell_type":"code","metadata":{"id":"bNZBSrzXiVVe","colab":{"base_uri":"https://localhost:8080/","height":279},"executionInfo":{"status":"ok","timestamp":1650222301059,"user_tz":-180,"elapsed":17,"user":{"displayName":"Arda Karaman","userId":"09649945184885111594"}},"outputId":"6a5e3ae6-9f0e-461e-f3ff-6558dc8c76cb"},"source":["plt.plot(loss_history)\n","plt.xlabel('Iteration number')\n","plt.ylabel('Loss value')\n","plt.show()"],"execution_count":9,"outputs":[{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"iVBORw0KGgoAAAANSUhEUgAAAYIAAAEGCAYAAABo25JHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO2dd5gUVdaHf2fyMAxxhjQMDCBIjkMQFFAECSqmFcxxMbCrqKuLYdeAq6ifcTEsKkbUdRXXwEpSgiAZyTkMSB7ykCae74+u6qmuruquDtVh6rzPM89U37pVdaqq+54bTiBmhiAIguBcEqItgCAIghBdRBEIgiA4HFEEgiAIDkcUgSAIgsMRRSAIguBwkqItQKBkZWVxXl5etMUQBEGIK5YvX36ImbON9sWdIsjLy8OyZcuiLYYgCEJcQUQ7zfbJ1JAgCILDEUUgCILgcEQRCIIgOBxRBIIgCA5HFIEgCILDEUUgCILgcEQRCIIgOBzHKIJN+4vwf9M34cipkmiLIgiCEFM4RhHsOHQSE2Zvxf7jZ6MtiiAIQkxhmyIgolwimk1E64loHRHdb1BnOBGtJqKVRLSMiM63S56MVJcT9amSMrsuIQiCEJfYGWKiDMBDzLyCiDIBLCeimcy8XlPnJwDfMTMTUUcAXwJobYcwqiI4WSyKQBAEQYttIwJm3sfMK5TtIgAbAOTo6pzkylyZGQBsy5tZXR0RiCIQBEHwICJrBESUB6ALgMUG+64koo0ApgK43eT4UcrU0bLCwsKgZMgQRSAIgmCI7YqAiKoD+BrAGGY+od/PzN8wc2sAVwAYZ3QOZp7IzPnMnJ+dbRhF1S/VU9SpofKgjhcEQaiq2KoIiCgZLiUwmZmn+KrLzPMANCeiLDtkyUhNBCAjAkEQBD12Wg0RgPcBbGDmV0zqnKPUAxF1BZAK4LAd8iQlJiA1KUEUgSAIgg47RwR9ANwE4CLFPHQlEQ0loruJ6G6lztUA1hLRSgBvAhihWTwOO8VlFfhp40G7Ti8IghCX2GY+yszzAZCfOi8AeMEuGYzYevBkJC8nCIIQ8zjGsxgAalVLjrYIgiAIMYejFMHQDg2RVT0l2mIIgiDEFI5SBCmJCSgtt20JQhAEIS5xlCJITiSUlldEWwxBEISYwlGKIDEhAadLynFUQlELgiC4cZQi+GWLKzzFa7M2R1kSQRCE2MFRimBoh4YAXM5lgiAIggtHtYj39m8BAEhK9OneIAiC4CgcpQiICNVTk1BSJgvGgiAIKo5SBIArMc0HCwqiLYYgCELM4DhFoCJmpIIgCC4cqwjmbzkUbREEQRBiAscqghembYy2CIIgCDGB4xTBFZ0bAQBOlUheAkEQBMCBiqCakrv4tKSsFARBAOBARVA/Mw0AcFjCTAiCIABwoCK4q1/zaIsgCIIQUzhOEaQlJ7q3u42bif8s+13yGAuC4Ggcpwi0HD5Vgoe/Wo1bJi2JtiiCIAhRw9GKQGXZzqPRFkEQBCFq2KYIiCiXiGYT0XoiWkdE9xvUuYGIVhPRGiL6lYg62SWPIAiCYEySjecuA/AQM68gokwAy4loJjOv19TZAaAfMx8loiEAJgLoaaNMgiAIgg7bFAEz7wOwT9kuIqINAHIArNfU+VVzyCIAje2SRxAEQTAmImsERJQHoAuAxT6q3QHgR5PjRxHRMiJaVlhYGH4BBUEQHIztioCIqgP4GsAYZj5hUudCuBTBX432M/NEZs5n5vzs7OyQZbqqa07I5xAEQagq2KoIiCgZLiUwmZmnmNTpCOA9AMOZ+bCd8qiM6itOZYIgCCp2Wg0RgPcBbGDmV0zqNAEwBcBNzByxjPLpGqcyQRAEp2On1VAfADcBWENEK5WyxwA0AQBmfgfA3wHUBfCWS2+gjJnzbZQJgCgCQRAELXZaDc0H4DNLPDPfCeBOu2QwIy1FFIEgCIKKIz2Lq6fYORASBEGILxypCBISfA5UBEEQHIUjFQEArHv6kmiLIAiCEBM4VhFkpHpOD537hKEvmyAIQpXHsYpAT3FZRbRFEARBiAqiCARBEByOKAINh04WR1sEQRCEiCOKQMPOw6ejLYIgCELEEUWggZkx+rMV6PDkdBSXlUdbHEEQhIjgaEXww5/P9/hcXsGYunofiorL8M6c7VGSShAEIbI4WhG0z6np8flkcZl7+1RJmb66IAhClcTRikDP379d594mcT4WBMEhOF4RXNAyy72959gZ9zb5jpcnCIJQZXC8Ivjkjp6G5RKOSBAEp+B4RWCGTA0JguAURBGYkCCaQBAEhyCKwAQSRSAIgkMQRWCCqAFBEJyCKAIAObXSvcpkQCAIglMQRQDg0zu9LYdkjUAQBKdgmyIgolwimk1E64loHRHdb1CnNREtJKJiIvqLXbL4o1Z6slfZKzM347qJi/DrtkMoOlsaBakEQRAig51Z3MsAPMTMK4goE8ByIprJzOs1dY4AuA/AFTbK4Rez3v/C7YexcPthXHhuNj64rUeEpRIEQYgMto0ImHkfM69QtosAbACQo6tzkJmXAohql5v8PIVN+4siI4ggCEIUiMgaARHlAegCYHGQx48iomVEtKywsDCcogEAEv2sB4gpqSAIVRnbFQERVQfwNYAxzHwimHMw80Rmzmfm/Ozs7PAKCGsLw0VnS9H/pdlY+fuxsF9fEAQhmtiqCIgoGS4lMJmZp9h5rVBgsM/9RMCKXcdQcPg0Xp6xKUJSCYIgRAY7rYYIwPsANjDzK3ZdJxwkJ/p+DLuPnkF5RQUAIFGi0QmCUMWw02qoD4CbAKwhopVK2WMAmgAAM79DRA0ALANQA0AFEY0B0DbYKaRgSU5MwIZnBmPzgSIMf3OBYZ3bP1wGwP96giAIQrxhmyJg5vnwE6mBmfcDaGyXDIGQnpKITrm1/NaTEYEgCFUN8SwOkBnrD0RbBEEQhLAiiiAIjp4qibYIgiAIYUMUgY7f/jbQb50u42Zi5+FTEZBGEATBfkQR6LAabK7fS3PsFUQQBCFCiCLQ4S/chJa1e47bJ4ggCEKEEEWgIxCboB2HZHpIEIT4RxSBjkDyEPj2RxYEQYgP/CoCIqpPRO8T0Y/K57ZEdIf9okWHQBTBvmNnMP7HjSguK8ePa/aBWVSDIAjxhxWHsg8BfADgceXzZgD/hit8RJUjEMfh53/cCAB4Z+42AMCkW/NxUev6doglCIJgG1amhrKY+UsAFQDAzGUAym2VKoqEEkHi2GnJZCYIQvxhRRGcIqK6UKbEiagXgCprLhNKruJPFu0EABw+WRwucQRBEGzHiiJ4EMB3AFoQ0QIAHwP4s61SRZFQFMFvu45hxrr96PbsLPy67VAYpRIEQbAPv4pASTfZD0BvAHcBaMfMq+0WLFqEGlNuacERAMCa3eEbNG3aX4QLXvxZQlsIgmALfheLiehmXVFXIgIzf2yTTFEl1LSUquFQOKNVvzVnK34/cgZzNxfiii45/g8QBEEIACtWQ90122kABgBYAdcUkaBDNSClgFzTgC0HivDLlkO4/fxm4RdKEATBB34VATN7rAcQUS0AX9gmUYyy4ZnBaPP3aX7rVShDgkBHBJdNmI+zpRWiCARBiDjBeBafAlClW6sJ13fB05e38yhLsPikKqeGvDXBzxsPIG/sVBQYhKY4W1qhHO/tlCZ+aoIg2ImVNYLvUTnjkQCgLYAv7RQq2lzasRF+2uCZgCbQFJUEYNfh0ygpL8c59TIBAN+u3AsA+O33o8jLynDXXb+3MjMns/loQrJkCoJgB1bWCP5Ps10GYCcz77ZJnphB3wu3mqLyw18LALga7b4vzQYAFIwfBqDSNLWiwvOY2ZsOurcrmJEQ4PqCIAhCKFgxH52r+VvgBCUAeAeUIyK8NqKz5eONmnK1R1+h0zLa6aAKmQYSBCHCmI4IiKgIxgE2CQAzcw3bpIoBqqd6P5q0ZOtLKt+v3udVpo4IfM35s8EjP3DirOXrCoIgBIppy8bMmcxcw+Av04oSIKJcIppNROuJaB0R3W9Qh4joDSLaSkSriahrqDcULno1r4PXR1ofAehZvvOoV9nczYUAjEYExtsqi3ccCVoOQRAEf1ju4hJRPSJqov5ZOKQMwEPM3BZALwCjiaitrs4QAC2Vv1EA3rYqj90QEYZ39nTeat0guEFQWblrUaCwyBWDaOP+IuSNnYotB4oAeA67xEJIEIRIYyUfweVEtAXADgBzARQA+NHfccy8TwlPAWYuArABgN4tdjiAj9nFIgC1iKhhYLcQOfKyMrDtuaEBH1d0tszjs7qg/P3qfVhacMSj8dePFgRBEOzGitXQOLh69LOYuQsRXQjgxkAuQkR5ALoAWKzblQPgd83n3UqZxwQ7EY2Ca8SAJk2sDEbCx/nnZGH+1soAclath7SUM2PhtsNe5W/8tAVv/LTFo0zUgCAIkcbK1FApMx8GkEBECcw8G0C+1QsQUXUAXwMYw8wn/NU3gpknMnM+M+dnZ2cHc4qgmXRrd6x5alBI5zh+phTXvbvIUl2rI4KnvluHvLFTQxFLEAQBgLURwTGlMZ8HYDIRHYTLu9gvRJQMlxKYzMxTDKrsAZCr+dxYKYsZUpISkJIUWmrnAS/PtVyXK/zXmbZ2v3t6SRAEIVSstHDDAZwG8ACAaQC2AbjM30HkirHwPoANzPyKSbXvANysWA/1AnCcmb3tLh2EkfmonhembYyAJIIgOAUrI4K7APybmfcA+CiAc/cBcBOANUS0Uil7DEATAGDmdwD8D8BQAFvhUja3BXD+Kok4lAmCEGmsKIJMADOI6AhcSev/w8wH/BwDZp4PYwdbbR0GMNqKoE5BH3Tutg+WREkSF4dOFoMZyM5MjaocgiDYh5UQE08zczu4GuyGAOYS0SzbJXMo+hHB7E2F0REEwOrdx5D/7Cx0/4e8bkGoygSyCnoQwH4AhwHUs0ccwcoawQ6DMNa3frAE+c+Gt8G+fMKCsJ5PEITYxEoY6nsBXAsgG8B/APyRmdfbLZhTCdafbE4URw6CIMQ3VtYIcuHyAVjpt6YQMvuPn0X9GmmG+0LNpywIgmCElTWCR0UJRI7hb4Y2HTNy4sIwSRIcR06VYOP+oPwGBUGIEqF5SjmUJ4a1cW93aVIrYtdVg9Zp0VsZLdruO1Lp0oIjHhnRws2wN37B4Nd+se38giCEH1EEQXDnBc3x71G9bL/Oqt+PeXwe98N6nC0t9ygL1O/gD+8sxNA37Guo9x2X3AmCEG9YiT6aQUQJynYrJRppsv2ixTYJQQSfC5TjZ0q9yopLPWNQMDOOn/auF272HT9j+zUEQYgOVkYE8wCkEVEOgBlweQt/aKdQ8YCdakCd7vl8yS6vfe/M2+bxuYJd0U31vDBto0dk07LyCvR9cXbQMp33/M9YvN07gqogCPGPFUVAzHwawFUA3mLmPwBoZ69YsY9qwKNXCE3rVgv53KdLXNM/P67d77Xv7Tl6RcDYXnjSsN4rMze7Px8sKsauI6dDkmvj/qKQjo9F5mw6iLV7jkdbjLhl68Ei7D0mo8V4x5IiIKLzANwAQI17nGifSPEFA7i3fwvk1EoHEJ6RQrsnp3utBZhen4H7v1ipK/MeIZwsLvMq88WpAOsDkc+tPGn+DsM8D1Z5deZm3PrBUlz6z/lhlMpZXPzKPPQe/3O0xRBCxIoiGAPgUQDfMPM6ImoOIPg5hirII4NbY9qYCzCkfQOMvvAcd3m3prWDPud9n/9mqV5ZRQX26HpkpeX+V5BvmbQEZ0rMlU27J6d7lRkpGC09n/vJ73XDyTM/rLec58GI13VJgYT4Y+rqfZiz6WC0xYh7rPgRzGXmy5n5BWXR+BAz3xcB2eICdQSQmZaMt2/shqzqlcHZru7aOOjzzrb45T5b6p3AYPyPlWGq9xw7g9GfrfBaZJ67uRDjpno7iFdUMErLzZMi7Dl2BkdPlXiU7Tp8GodPepu2RoqKCsbWg97TY9Hknbnb8Lf/ro22GFFhW+FJnDhrvwEDAIz+bAVu/WBpRK5VlbFiNfQZEdUgogwAawGsJ6KH7RctPtHGCrISN8j0PBYPLTewH520YId7++nv1mHq6n24bIL39Mdni3fhh9V78e3KPRjzhWsEMnbKarR83DglNQPoM/5ndBk3E4c0DX/fl2ZHZHqgooKxevcxr/KJv2zHxa/Mxbcr9+C+z3/zqcgixfgfN+KTRTujLUZUGPDyXFz7TnQdG4XAsDI11FZJMXkFXEnrm8FlOSQY0LBmeljOU2bRQSBUs85tB0/h/i9W4r8r9wIAvly229Jxeqe04jL7G99JC3bg8gkL8Ou2Qx7ly3ceBeBaK/lu1V6fi79nSsp9TokJ4aEqGhZEg5PFZUGt1wWKFUWQrPgNXAHgO2YuheRYN+2xt2lYw2+dcPLgl6t87g+nCNr7CeS8zOweuaz8/VhAVjrMjCPKVJTauOw+6lv5+ZKtw1PT0ebv07zKtxyIfMNVXBY/Cunh/6yK2xzZp4rLUBGnGZ/aPzkdHZ7yXq8LN1YUwb8AFADIADCPiJoCkGAyPmiWleHe/uj2HoZ1sqqnhOVaRiGptcxc7zuHUCBx7Co0mmDd3uP4yGLe5D999htaPPY/AMAVby4wtdLZd/yMu3ev8vHCneg6bia2aU1kdb/pQBSu2Uhr4KvzrJ8kDBQcOoVzn5iGr5ZbG4FZ5WxpOf4xdX3Ye5H/CbOckaK4rBztnpyOZ37wXA/7ftVeHIywlVuwREKHWVksfoOZc5h5KLvYCeBC+0WLXx4a1AoA0LZRDSSatrSxEUlU24j66/E9O3WDe/vFaZvw5HfrLF1j6hrvNNTlFYwn/rvGQ5H1e3EOrn77V496qkXIzsO+FZ6WSIzEJs3fgTdnb7VUt7CoGHljp2L+lsoprc3KCGTa2vCm6P500U68+8sOL38Tp6IaU3ytUWQni8vw589/w03vRzf7XyxhZbG4JhG9QkTLlL+X4RodCCZc2rERfvvbQHRtUhvnNsg0rBOBCBWWeHXWZv+VgqCigpH/7Ex8uex3w/0b9p3Ap4t24d7JK9xlJT4WeT/6VbPw6vXsIj/sf+aH9Xhp+iZLddWYUR9oFvGD4VRxGR77Zo1PnxB1xBMLC+YxgfLVKCouw7bCk3hz9lb3NKU4wlViZWpoEoAiuJLTXAvXtNAHdgoVD/ibUqmd4Zr6yc5MRcH4Yeic6xmlNDFWNEGIbDaZWy+tqMChkyV44hvfJpT+noKag2Hu5kAS78TWfLCRNNrcEodOFuOFaRsNLcC0fLBgBz5bvAsT5/rv7cfWEzDmgwU7kDd2qt/7DgWt5d6Ify3CS9M3uafN4uEZRQoriqAFMz/JzNuVv6cBNPd3EBFNIqKDRGTYEhBRbSL6hohWE9ESImofqPDRJNDph2opns7YCVUkycwgk7n1MgOnNu0PXn1+Zo/hyrcWIG/sVGNFEWe/YNURz+xeH/9mDd6esw3ztx4yrqCgPj6754zfnrPN79pTOFD9XSI1eilWvPXV91ARiTnEOMGKIjhDROerH4ioDwArY6oPAQz2sf8xACuZuSOAmwG8buGccctrIzp7fK4iesAUI89k7Q9edThat/cE8sZO9QoV8dsub38B7e/2YFHgC33bDGIyRRbjl67GlvLnua0e/fPGg1hgojTUOhv2mdtzrNh11DC3BeCKePvCtI0RSXAUiWbYyNKNlKckeqASK4rgbgBvElEBERUAmADgLn8HMfM8AL6ypLQF8LNSdyOAPCKqb0GeuKSeLv1kcqJDUkFo2r5LXqscPejDYpiFitAqTO0wX+uwpP9Bm/3AB7w814+wofN7gIH9mCsVpDpK3Hf8DMZ88ZtpvKn1+07ghvcW+1Rsv2wxH11c9davuMwsvpLy7LS+FuGyrvly6e/o/Xxkw5D4autDcfi0myOnSiJqXmzFamgVM3cC0BFAR2buAuCiMFx7FVwRTUFEPQA0BWAYk4GIRqmL1YWFVSNJ++Q7e4YlUmmso+0D7zxc2Ug+8tXqgM/FlV06FBwOLZKqEYVFxch/dpbfVJt6hzaV2RsP4oIXZ+NHAyspPdrnomaVUxXBM9+vx39X7sVz/9vgoQz0o0gjxWZ1pLnfYuNeVl6BHmGKIfXI16uxV5u4KALtsK9Rlh1TbJ8t3oVftx7Cil1H8Y6FtRwzuo6bibs+WR5GyXxjuVvKzCcUD2MAeDAM1x4PoBYRrQTwZwC/ATBUgcw8kZnzmTk/Ozs7DJeOPo1qpePTO3pGW4yYZMoKrc26d8umX3/Q/56D/X3/tOEADp0sxqT5LuueorOlXguZr8/aguvfXex17LHTJbjtQ1fMm9U6hznfvdJKEnS/xo8X7sTT31sz0Q0Ha/cc9zL11ftd+Ju+soLVczBzQGbDKqXlFShTRllGV3KPBCyI8cmincgbOxWnS6z5ZTz2zRpc/95iXPXWrx4xv/zx266jXs9lzqbKTu//WbRQC5Zg5ydCnuFWFMttzNwZrjWCbADbQz1vpKGqPtkfIsGEntB6S8/aUOkQp/5QHvtmjUf9nzd6BuhbWnAE5RWMs6XluO2DJdh6MDCvYQKhpKwCHZ6agad0vhJm5rbf/LZHI6fJecl4W0UdEWiP/3xJpfmt0XfthvcW4aXp1hscPXd9sgw9n5sFALj0n/Pdz5YMZAEC60Xf9ckyw3L9OXcePo2pq10KaOK8bbh5ksu+//35O9DvpTl4dWZgJs4tH/8RFymjJY81AtZ3IPzfzL+UXv2hohI/NYNnxrr9uPKtX/HFUmNTawCYYNFnJViCVQQhdwuIqBYRqe61dwKYpxlxOIKGNdP8VxLcqPGQ/PHitE14c/ZWLCs4itmbCvHkd+vw8ozAelSqT8PXK6x51GrbGH0vVt1ntH6gbZyCsSRbsPUw3pxdOQVBuj5a3tipXspMy/R1B3DghHnkWH1jGYilzfR1xl7t7P7v2hry+jyM/szlT/Lc/zZinmIq/K3yvrXhws+UlGPl796GBHp2HTmN12dt8dnYW1Fqqpl3qBZGFRWVoVL0qFOm0Yyga6oIiKiIiE4Y/BUBaOTvxET0OYCFAM4lot1EdAcR3U1EdytV2gBYS0SbAAwBcH8Y7iditM+pid4t6uKZ4daTtW1+dojH56TEBBSMHxZu0QQAmw4UuX+8CUT458+B9ahUNw9mV8KdY6d99wi1zYQ+s5waRmLj/iIwM16dudlwjv69X7bj2OmSoBcxD5w4axhC40OLoUCMsLoQH9g59crF+rF3fbocV7y5wNJUzauzNuOYQT5v9fJWpqjUyABG6WAD4dVZm9F13ExDay2r+v+WSfZ5QieZ7WBmY5dYizDzdX72LwTQKpRrRJO05ER89sdeAR2TkpSAuQ/3R3qKJHizHa7sxVmdvtP+1NWedQUzej73E1KSErwUucexPhoK7fTWur0nPHq4szXzwDPWH0DilDVBNbZFZ0tDSgykDwSoPjK9KMH2jG//sDJnwMz1BzCkQ8OgzqOOFqw6ofkS18oZEpQeQbBObxUVjIQEwgxldHT4VDGyM1MN6/p7tIE5VQaGQ2wYY4emdTNQL1OmhOxm2c4j7kxt5RWBrVOUM6NI8XNQf5slQYbZXlrgaUH9o5/YQieLyzBDFyjwdEkZRn+2wtT2HzBei3nkq8q1Fn8pPc1MUUNdHL7nU5fli3YdR40iG8qp9YfuPnracOpFq7hY/9/K1BCFNjUU6kgiUpiOCAQhnjlwohh//Ni1WLlgq7W8xo8rC6VfLd8dcFRQs9/7l7oFQO18vlW+X7XXvZhqhlFwQ21uieveXYRZD/Zz59bWk5Fi3BToO8L+GsQynZewfpoMsL7A2LJedawxCVmuF+P8F2YDAG7tneeznqusslDtsZsR6oigvIKRbHECgOEK137v5MiZjarIiCAG0YejECKD4W/d4u/fbF5/gx+fBCvoF4CNsNJjvfiVuegybobhPr3pqnpF/YjgijcX4IAPH4Q3AlyL0TJFtzDfs3kdAMAfuhm4F5ncrn49xOi5aIu2G4TSWLHrqFuBq36fZ0vLvZScFcor2G8aV21U333Hz5gustuJjAhigHdvzsfq3cdQPTUJmw4U4ZFLWqNXhD0wncZ/TKKi6tE28Jf+8xfzeiYN09o9gSkCo/UMK4281Q6rUY5rwFzZ6C+9+cBJjJ68Al/d09uwviW7f+WkepH1SZZ8KcBweQVrH/fx06Xo9uxM94L7td1z3ZZcV7+9EB1yauL7P59vdBpT5m0uxD2aCLuxOlMkiiAGGNi2Pga2rbLRNWKShy16NpdqnNcCbdSDwajp+9c8/+41doVLMFJCahhso0xzRg3dou2eU3MMl4WU3+kWkwVrs+sYoZVfjem0ZIdx5Jvp6/Z7WV1pTXrNpql8oU+05ItoKglRBIIQAjsOnUJe3WphC1dgZOBkJRJoqI2ImY270X0dP+NaSD9scIyRGCMnesaRCtSU13CeP4RjrfqGANbCxa/ZfdyVhMqgrv59Lth6CK3qZxrW/fDXAr9mynYhikAQQuDC/5sT1vPtPx5cgLdQHZ70eaCPni5FhyenI7eOdzysfYqM5tn3woN7ncKg2WdmLN7u3wjA6LmokW/9sazgiNc95o2dis//2AvntagLwKUELpswH/cPaIkHBvq3hn926gaUlFfg3v7nGO636jQZbmSxWBBiCNW0MlDsCKBWVFyG9T7CWWu59l/hD1utrpdMWbEHN7znOao4WFSMERONI9Zq+XjhTq8yfayqtXuOG2YrO1lchkOnvBd63/ulcqpOXThfu+e4z9DfWgoikOshUEQRxBlLH7/YvV2/hrFjiuA8+oz/OeBjKjTaI5jOvfYYdd49HEHp3OfXbOtNgIe8br5wr2W7hRwUl/5zPnobPL+zpRXYXujdaGvvW7W2qmC2RRlGClEEcUKjmmkoGD/Mwyvx9j7NoiiREO+8FmK+arvDLeqV04x13j4J/jBy6NJOF508ax6q4u5Pzez5KwUjt8OZp2JVefeXHRYljS6iCOIEX/2sxrWNnYQEwRdzNclrftpg3Xbd3UAbaIJwzlDp2/BRQcTnN3Iq17bXf/cRkM8MjxFBEJ7Hx8+UYvXuY2j792n41U960kghiiDGGdbRf0yWAa3rRVy/YWwAABtMSURBVEASoaqxShPFc9Vu66aRZm3eF0t2+fWADoRwKBVjh7LKsqMayyerJrhqRjlmxttzXBZQvjLC6Zm+7gAun7AAp0vKcf173rktfGFX1jJRBDGO2tvXfp/fvqErPrq9h/trm5qciOeu7BB54QTHcuRUiVeCnrFT1pjUDpyfNhwIKc+CitHiu7a536UJDW61Uz9nUyG+X7UXS3YccWeXixRvzwk+65kvxHw0xklXApVoeytq5EbVSoEQ3GKfIARLKKGtrXDHR8ZJbcKBWYNv5mhmxKeLdqJvK89siZHwByvysaYRCjIiiHFGdm8CAKiRluyznugBIZK8oQmlHW+YWTZN0WSZ88fiHUfwki59ZKi+HFZISrTnly4jghinQc00PHVZWwxs18BnPRkRCII17GquzeI4hZNkfXTAMCEjgjjg1j7NTMMHq1iJUKln2pgLsOSxAWhi4D1qxN39WgR8DUGINYINKR0LWAl5EQyiCKoY13RrbDn9ZesGNVCvRprhkPaufs29yqy65gtCLKMPpxFPJNmkCGRqKEb57+g+gWXWCuH7oXeEOf+cLMM1ibMl9piuCYJgjaREmRpyFJ1za6Fb0zqW61cmEvEsvza/MS48t9K6obcSLEuLUcJzozWHa/INEoQIghAxDvlJchMstikCIppERAeJaK3J/ppE9D0RrSKidUR0m12yOIEezVxK48ouOR7lL17TCR/c1sP92SjvwZ8u8o6EaLTmoEZi7JFnXUEJghA+Plu8y5bz2jki+BDAYB/7RwNYz8ydAPQH8DIRpdgoT5VD2/tvWjcDBeOH4fyWWT6PMZpBuvm8PK8yo6lINX9rJMzkBEHwJiUpzqaGmHkeAF8eGgwgk1xRm6orde3xlqjqBLA+4CtRt8rd/Vp4ZGZyH6sUiSIQhOgQd4rAAhMAtAGwF8AaAPczs+HqKBGNIqJlRLSssLAwkjLGNRcZxCBSm/fhnRsZHqOOKozWCNRIi+UMXNxG4hsJQqRJS656iuASACsBNALQGcAEIqphVJGZJzJzPjPnZ2dnG1URdGx/bijevyXfq1xtzDNSPQ3GXry6o6lyUFHXCJgZLepV99j33s3e1xIEIbwkVUGHstsATGEXWwHsANA6ivJUKRISyN3oa7mkXQM0rJmG2/vkeZRf2z0Xr4/s4v5sdKyvkLs10n2HwIgk44a3M933l0H+0wkKQqxiVwCBaCqCXQAGAAAR1QdwLoDtPo8QPGiWlQEAOCe7up+alWRnpmLhowNwTr1Mn/XU9YCR3XPdZbl10pWyJh5++oPa1jdUDu0aGQ7wwkYweRjaNappgySCEBnsCiVjp/no5wAWAjiXiHYT0R1EdDcR3a1UGQegNxGtAfATgL8yc2xkaYgTBrdvgG9H98E13cJv3z+iey6u6pqDvw6uHKTVqpaCgvHDcGOvpu6y2tWS8frILgEtIA/v3AjPX9UBr4/s7C4bfWGEwldITCYhjjEaqYcD2zyLmfk6P/v3Ahhk1/WdQqfcWract1pKEl65trPpfrXZv6tfC6SnJHo5so25uCVmrDPOenXzeU3RrWkdzN54EADQr1U27u1/Dt6c7T/WevXUJJwsdhmXGVk2+UP0gBDP2PX9lRATDmHB2IvC+iVSQ/mq59SPCMZc3MpUEbirKgczrA95tfXU7U/v6InWDTOR/+ws70qCUIUIpvNj6by2nFWIOXJqpaORnwimwUBu34IgjlX+M7OHJ3O/VuaWYWSwnVM7HVnVUyt3iJ+DUEVpa9O6mygCISj0bW0gawSq8tD2brQdHV9hgo3mSPWJRnxJYnT8b38b6OMIQYgdnr/KnpS0ogiEkFB78voIpgDw6NDWqJeZ6lWuttuVown2UARlPqKuek4NKX4NXjKZU13xn2iuWFwBQO2MFLx0TUe8cV0Xs8O8+Pqe3pbrAkCv5hKfSQidNCV1bbgRRSAEhb7x1eqB63q40mte0DIbSx6/2PQcqhJh9gxy52tEcOx0KUbk5+KJYW2CWvPo1rQ2XhvRGROu7+pR/of8XFzeybdDnZZqKYH9IO/p7x3YT8WfI1+0aFXfullysJxb37cZs1CJXUlpAFEEQoioPXR17vKtG7p6DV/zm9bGs1e0R8OaaR7l6vea2TPInVYR/MHANPaFazrizgsqE+cEuiRwRZccZKSG1rNKIApoTdpX1T7n+A4UGC3e1ClLO3hQHPws07KefYpZFIEQFPrGN6dWOgrGD8PQDg296n51T2/c2Kup92K1x9SQ/xHBlV1yMP+vF3odb18WWnP0+UGM8jzoad3AuPfrzzHOKNCYkYJUiTWjqXQf0xkX+ImWGwlqV4sdr3gj1BSxdgZ7FEUgBAUH0fjqF3VrKmEpmtat5tFj/uuQSic2BvDyHzrh36N64dURndG4dmV+ZTvaO6upALOqp3oowwvP9R+E79s/9TEsr5eZivY5xtYgG54ZjD4GSiY/r7bpdbSL8I8MPte0Xo206FuP22UOaZUru+Sgad0M/xWjiNrpsNMYThSBEBKheDq2a1QTk27Nx9OXt/foxfZukYUXr+no/nx1t8bo2dy8xx3MD0Rdk9BPV9WvkWZU3Yta1QJLnUEEpCaZ9YzJHdBPT3pKomF6QqPEQSpaXZaR4mrs69fwXrR/6nLzmEyRwuzr06imtfcQKt3DnGSJCHh1RCdTxa4nw8JaU4KJUUQ4EUUgBEVQja/Br/6i1vWRnpJoqlB8XSccP5CkRM/rfjGql8/6TetWw1KDBXC9+AGHC/ahUFPCkKf2knYNvMqSbcp/q8dXX8FMAfY7NzJRhvudmx3WqTRm4MoujdGxsTWPfysdKff3XKaGhGjSuHY6/nhBM8N90RzYq78hs9+HNmCeVXLrVPO5v0GNNGQbmMTq+e1vgUVPSfORcKRGusEUjo8Hr23gjRoPVf4mdarhsaHhCfj76ohOQR1nNjXkq82rnhr4lJY+ltXFbeqhYPww5NRKD+t3uHm2a5rJapttxUFMtRaSqSEhqsz/60V4fFjbaIthitl6hR3Tz1bO2aBGGtJ1Q341p7SWkd1zcfN5TdE8KwOvjDCP6/To0DYByehv3v2uvs3xv/suQKfcWhjV17OB7Ni4MjprIM+vR7O6uLd/4IEDzTLm+Wr0MoNY29DbHwTjCa/FyD8G8D+i1PPgQP9WU25FENCZA0MUgRASgTQWgQxtrZz2JiUKasManlY3auPiy+5aldt83j4wtEP8RY8N8NjXObeW4XWaZWXgmeHtkZBAyKmVjq/v6e3RmKrTSzXSkjH34f4Yq1lED6ZVGKxMDyUQGfZEbz6vKabc09sdbtwKDZQ1lQSq9B/Ro38L/7bQWDbL9r2A+8Ft3d3b1+Y3xgMX+25Q9RY32u9iMOtcZofUywxsbcOKcYJMDQkxi51fSo/r+GjxbjovDwXjh6Gmzvzvqi6Ncct5TfHwIPNpj8a10zHm4paYdEt30zpa1J7b5Z1yDPf3NOjxq5gpJH2vtFvT2u4e+aC29bFx3BD3vqZ1M9xmhFYwamAaKAuwvtq9pMSEgNYO1AaWYN2vwtfCP+CautIv4msheFppvXhNJ9yqS7TkJafuYWs/acU26+kD3lONvkOn+/599GuVjdVPDbL0zNz+Nv6rBo0oAiEodAFELWFXLHU9qUkJeHp4ew8FMaB1PZyjccghIoy5uBWa1PW9JgAA39zbG/cNaIkt/xiC63pUNgbPXdkBdTJcORra55gnvDFTBEZKTtWv/qZ3fClIgmtUMvfh/j7PEQ7c3wOy9n5b+OnpA0B29VQMbFsfF7Wuhy9G9UJKUoLPBhoAkhN9X9vX1NCdmvUvo1HN2CGtcXe/FvjHlR0w68F+7vKHLzHvaPjrJ13XowlqpCXDyi8oEmsE0TckFuKaSDXuofL+rdZ6/kZ0aeKy2df3lK/v2QTX96xsOBITyNAZzmz4bxSfSS3y91h9Ngrk8nPQRmQl+B/Fee/2/261caPMrJu035GfHurv95yAKx/GJOWdbX52CK5/dxEOFhWb1veXy9fX1NDg9pVOkEYKRTsSC2ah2ogBbfz7naj4ShEbLmREIARFMN/JJ4a1QesGmT7TRaqNSSgKJlTdNOOBvvjx/gsCPu6XRy7EN/d6B6PTjggWa9YPjBYs1Z6+X0XgY58dqnniTd2M5dBMDWVnpgYUuE9PD8Wm38gqy9/3zd+IwOr39Q/5uWGJ6ePrejef19TdqbDyXW1cOx2Xd2qEt28wfgfhQBSBEBKBNLpdmtTGtDF9vSxqVN6+oStmPtjXszAKqQVa1c9Em4aBx31vVCvdPXrQoh0RaB3WjBoLdw87TM15juKJnZflf0rG17vsbJIJTzs1BCCgwH16vrz7PIy/qgNeM7CgUhVkVvVUvH2jd4NIRKYmzoB32BKz3nWdjBR8dmdPv7IaHZ5jkO+jic4ceVDb+nhmeHv3Z/0jz6ru7aiYlEh447ou6NDYvnzbogiEoAgmxIQ/hnRo6Hb3j48JJ2voLYb+fqnLFHdYR++4TO6nGsTUkGplpPVEHti2Pr4Y1Qu3nJfnLtOfOr9pbc9zGpy7Xo00TLjes7ev7TiH632N7NEEtTO8G0NVtjeu62yanlUf5uMyjVLSO/iZKdpEIsve5Va4u18L/PLIhXjnRuMAfurIt3a1ZLxwdQdkpnnHPYrE9KsoAiEkYrHBDtcPZ/qYvgHbhRuhD+Vw+/nNUDB+mMfitYo+BaieyXf2xD+ubO9VPuvBvm4LqDq6hrRX87qm9voAcLlJGGz9Y7y0o6veXf1ckV8TEwiT7+yJO89v5nVNj/OY7rFOpXGC59m0I5VSXa+/Vnoy1j59Cd6+oSta1qsM+JdTK90jhImWhARCXlYGfnqon+F+Ff9Tdy5ZEsjTSVF/nPqxce1qGNG9iWFY7kj8xmxbLCaiSQAuBXCQmb2+uUT0MIAbNHK0AZDNzEfskkkIH5HKBhnqZe4f0DLo9H7nmkQLDZQGQcTNMVNmfc7JQp9zsjB58U53WesGmTinXqbfd+LvWaqN1/NXdcBzP25Ebu1qaNuwBtbvO+GuUzB+GM6WluNfc7cjPTkRbRrWwBOX+nc2fOHqDmhQM4RUqQaL6OuevsRjAb97Xm10zq2Fto1q4LPFu1DBjOqpSRjSoSH+vXSXu97rIzv7TdvaIrs6Lm5T38t5rZZiiXZ7H+9pqItaV45I9MmXzN6Ne79ygy9f2wk3726KtXuO47n/bQQQmcB8dloNfQhgAoCPjXYy80sAXgIAIroMwAOiBOKHQe0aYPLiXcgPc9AulXB99x+w4LkZS1Saj/qu18qg52hVaeqVjP5SPZvXxbejXZFSvxndG6XlnmdOS07EQwNbYXB77/hFZozobuxsZhX3IrqmLENnwVMtJQn/Hd0Hny5yKUntAEF9rlnVUwy/s1/f0xtzNh30KHvvlnyvemnJiSgYP8yrfNGjAwzn9/UjGLPPqnwZqUno3SILa3Yfr6wTgSGBbYqAmecRUZ7F6tcB+NwuWYTw069VtuEPIlyoMfjNFparKhV+poZUuufVwcSbumHUJ8vdZc2zMnB5p0Z+Hc8CcQZMTUqEkcXknwe0ND3m3ZvzcbqkDPd/sdLSNT66vQdy/eRkqOxh+28VK6tU3qe6NaB1fcNjujWtjW5NzUN7+0M/6rOulJX6ugO0H+N9RGAJIqoGYDCAP0VbFiF2GNK+IR4ceAq3+fEYrapYafBylMZTrZuUmBCS+Wa4pvsGtnU1ttnVU3H9e4v91u/Xyn+kUb11ki/cdvea1Nf6qRq76di4Jr5avtttreV/Ws4T7ZRUJESOhcXiywAs8DUtRESjiGgZES0rLCyMoGhCtEhMINw3oKWhFUVVptJ81D4i5QTYTvW2Nrhcs6wMDGpr3Ds3wt8iupb2ip9KH032M6v+GeHipl5NMfOBvl7BBr0Wi90jAk9VMFIzlRaJ9xX1EQGAkfAzLcTMEwFMBID8/PwoWJYLQuD8ZVArNAxwgfTiNvXRpmEN/Oki82T3dhFJJ/HZf+kfUP1ARgQdGtfEqicHuTPgAUB/xbTULDBeuCEitNSs45guFpuZsSYQeuTVwZKCI/G9RmAFIqoJoB+AG6MphyDYwZ8uMp9HN6NmtWTLXs2BTuX4tSoKdxcrjOerlM1aq6hVAkBlTu1oYz4i8K57Xou6WFJwxGcAvnBhp/no5wD6A8giot0AngSQDADM/I5S7UoAM5j5lF1yCM5hQOt6+GnjQf8VqxiBdhi9epg2dTlJmXiu68PHwCqBjAhikWByZtw3oCWu6dbYb7KkcGCn1dB1Fup8CJeZqSCEzJs3dMXxM6XRFiNuCfeAoEZaMp6/qgP6WlgM9ksAawSxjNlUkJGiSEygiCgBIDbWCAQhLKQlJyIt2VnmpoD1XvJtffLw88aDXvb/djau4ZqTrxwRxKcq8LdGECkHTTNEEQhCnGO1EWmeXR0Lxl5krzA2E59qoDKXca/mxlZE0baAEUUgCAKA6PdKfRFpP4Bw065RTSx89CJ3ak+VdGUE29iPQ53dxIIfgSAIIRCvjWMgqN7SVsJpxyoNa6Z7TW3l1qmGd27shtdHBu8IGA5kRCAIcUq4e/CxrFCGdWyIYR2jb/5pB4HEbLILGREIQpwTrgY8lqeGBHuREYEgxCnhSg6kxvTvf24YzDwdxrjh7XBOvfCEK48moggEweG0z6mJjeMGO9L0NlRu0mR+i2dkakgQBFECDkcUgSDEOeFKdC84F1EEghCnyOKuEC5EEQhCnBPLZp9CfCCKQBDilEQlsXFqkvyMhdAQqyFBiFPaNaqB+wa0xPURSrYiVF1EEQhCnEJEeHBgq2iLIVQBZEwpCILgcEQRCIIgOBxRBIIgCA5HFIEgCILDEUUgCILgcEQRCIIgOBxRBIIgCA5HFIEgCILDIY6zyFVEVAhgZ5CHZwE4FEZx4gG5Z2cg9+wMQrnnpsxsmH0o7hRBKBDRMmbOj7YckUTu2RnIPTsDu+5ZpoYEQRAcjigCQRAEh+M0RTAx2gJEAblnZyD37AxsuWdHrREIgiAI3jhtRCAIgiDoEEUgCILgcByjCIhoMBFtIqKtRDQ22vKEEyIqIKI1RLSSiJYpZXWIaCYRbVH+11bKiYjeUJ7DaiLqGl3prUFEk4joIBGt1ZQFfI9EdItSfwsR3RKNe7GKyT0/RUR7lHe9koiGavY9qtzzJiK6RFMeF999IsolotlEtJ6I1hHR/Up5lX3PPu45su+Zmav8H4BEANsANAeQAmAVgLbRliuM91cAIEtX9iKAscr2WAAvKNtDAfwIgAD0ArA42vJbvMe+ALoCWBvsPQKoA2C78r+2sl072vcW4D0/BeAvBnXbKt/rVADNlO97Yjx99wE0BNBV2c4EsFm5ryr7nn3cc0Tfs1NGBD0AbGXm7cxcAuALAMOjLJPdDAfwkbL9EYArNOUfs4tFAGoRUcNoCBgIzDwPwBFdcaD3eAmAmcx8hJmPApgJYLD90geHyT2bMRzAF8xczMw7AGyF63sfN999Zt7HzCuU7SIAGwDkoAq/Zx/3bIYt79kpiiAHwO+az7vh+2HHGwxgBhEtJ6JRSll9Zt6nbO8HUF/ZrkrPItB7rCr3/idlKmSSOk2CKnbPRJQHoAuAxXDIe9bdMxDB9+wURVDVOZ+ZuwIYAmA0EfXV7mTXmLJK2wk74R4V3gbQAkBnAPsAvBxdccIPEVUH8DWAMcx8Qruvqr5ng3uO6Ht2iiLYAyBX87mxUlYlYOY9yv+DAL6Ba5h4QJ3yUf4fVKpXpWcR6D3G/b0z8wFmLmfmCgDvwvWugSpyz0SUDFeDOJmZpyjFVfo9G91zpN+zUxTBUgAtiagZEaUAGAnguyjLFBaIKIOIMtVtAIMArIXr/lRriVsAfKtsfwfgZsXioheA45phd7wR6D1OBzCIiGorQ+1BSlncoFvPuRKudw247nkkEaUSUTMALQEsQRx994mIALwPYAMzv6LZVWXfs9k9R/w9R3vVPFJ/cFkYbIZrZf3xaMsTxvtqDpeFwCoA69R7A1AXwE8AtgCYBaCOUk4A3lSewxoA+dG+B4v3+TlcQ+RSuOY/7wjmHgHcDtcC21YAt0X7voK450+Ue1qt/NAbauo/rtzzJgBDNOVx8d0HcD5c0z6rAaxU/oZW5ffs454j+p4lxIQgCILDccrUkCAIgmCCKAJBEASHI4pAEATB4YgiEARBcDiiCARBEByOKAIhLiCik8r/PCK6Psznfkz3+ddwnj/cENGtRDQh2nIIVQdRBEK8kQcgIEVAREl+qngoAmbuHaBMcQURJUZbBiG2EEUgxBvjAVygxGh/gIgSieglIlqqBOi6CwCIqD8R/UJE3wFYr5T9VwnMt04NzkdE4wGkK+ebrJSpow9Szr2WXPkeRmjOPYeIviKijUQ0WfEQ9UCp8wIRLSGizUR0gVLu0aMnoh+IqL96beWa64hoFhH1UM6znYgu15w+VynfQkRPas51o3K9lUT0L7XRV877MhGtAnBeuF6GUEWItmed/MmflT8AJ5X//QH8oCkfBeAJZTsVwDK44rT3B3AKQDNNXdUjNR0ul/262nMbXOtquEIYJ8IV8XIXXPHj+wM4Dlc8lwQAC+EK/KeXeQ6Al5XtoQBmKdu3ApigqfcDgP7KNkPxFoUrbtQMAMkAOgFYqTl+H1wet+q95ANoA+B7AMlKvbcA3Kw577XRfo/yF5t//obMghDrDALQkYiuUT7XhCv+SgmAJeyK2a5yHxFdqWznKvUO+zj3+QA+Z+ZyuAKfzQXQHcAJ5dy7AYCIVsI1ZTXf4Bxq4LTlSh1/lACYpmyvAVDMzKVEtEZ3/ExmPqxcf4oiaxmAbgCWKgOUdFQGaCuHK7CZIHghikCIdwjAn5nZI6iYMtVySvf5YgDnMfNpIpoDIC2E6xZrtsth/lsqNqhTBs9pWa0cpcysxn2pUI9n5grdWoc+NgzD9Sw+YuZHDeQ4qyg0QfBC1giEeKMIrpR+KtMB3KOE8gURtVKisOqpCeCoogRaw5XaUKVUPV7HLwBGKOsQ2XCljlwShnsoANCZiBKIKBeVIYYDYSC5cvmmw5WxawFcgdmuIaJ6gDvXb9MwyCtUcWREIMQbqwGUK4ueHwJ4Ha4pkxXKgm0hKlMZapkG4G4i2gBX1MZFmn0TAawmohXMfIOm/Bu4FlZXwdXjfoSZ9yuKJBQWANgB1yL2BgArgjjHErimehoD+JSZlwEAET0BV7a6BLiilo4GsDNEeYUqjkQfFQRBcDgyNSQIguBwRBEIgiA4HFEEgiAIDkcUgSAIgsMRRSAIguBwRBEIgiA4HFEEgiAIDuf/AXaMgEMDqZsSAAAAAElFTkSuQmCC\n"},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"KPMf0jfF2zo6"},"source":["### 4.5 Quantitative Analysis\n","\n","We can analyze the accuracy of the predictions as follows. You should see around 38\\% accuracies. We can finetune the hyperparameters to obtain better results. But we will skip that and go for a bigger model.\n","\n","*Disclaimer: This code piece is taken from PyTorch examples.*"]},{"cell_type":"code","metadata":{"id":"5PAV1QQw218J","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1650222308032,"user_tz":-180,"elapsed":4167,"user":{"displayName":"Arda Karaman","userId":"09649945184885111594"}},"outputId":"499c41eb-2f39-4f01-cfbb-f95f4d282c1e"},"source":["correct = 0\n","total = 0\n","model = model.to(\"cpu\")\n","with torch.no_grad():\n","    for data in testloader:\n","        images, labels = data\n","        outputs = model(images)\n","        _, predicted = torch.max(outputs.data, 1)\n","        total += labels.size(0)\n","        correct += (predicted == labels).sum().item()\n","\n","print('Accuracy of the network on the 10000 test images: %d %%' % (\n","    100 * correct / total))"],"execution_count":10,"outputs":[{"output_type":"stream","name":"stdout","text":["Accuracy of the network on the 10000 test images: 38 %\n"]}]},{"cell_type":"markdown","metadata":{"id":"M4AJXtea2427"},"source":["## 5 Try a Larger Model\n","\n"]},{"cell_type":"markdown","metadata":{"id":"Dr7NvUrWW9ax"},"source":["### 5.1 Model Definition"]},{"cell_type":"code","metadata":{"id":"qmBmxHGoVE_F","executionInfo":{"status":"ok","timestamp":1650222312162,"user_tz":-180,"elapsed":426,"user":{"displayName":"Arda Karaman","userId":"09649945184885111594"}}},"source":["import torch.nn as nn\n","import torch.nn.functional as F\n","\n","class ThreeLayerNet(nn.Module):\n","    def __init__(self, D, H1, H2, C):\n","        super(ThreeLayerNet, self).__init__()\n","        torch.manual_seed(501)        \n","        self.fc1 = nn.Linear(D, H1)\n","        self.fc2 = nn.Linear(H1, H2)\n","        self.fc3 = nn.Linear(H2, C)\n","        ###########################################################\n","        # @TODO: Create three linear layers                       #\n","        #  (1) One with size D x H1 and store in self.fc1         #\n","        #  (2) Second with size H1 x H2 and store in self.fc2     #\n","        #  (3) Third with size H2 x HC and store in self.fc3      #\n","        ###########################################################\n","        ###########################################################\n","        #                         END OF YOUR CODE                #\n","        ###########################################################\n","\n","    def forward(self, x):\n","      ###########################################################\n","      # @TODO: Feedforward x through the layers. Be careful     #\n","      # about the shape of x. It needs to be reshaped to        #\n","      # (batchsize, 3072) first. Needs to return x. Apply relu  #\n","      # nonlinearity after the first and second layers.         #\n","      ###########################################################\n","      x = x.view(200, -1)\n","      x = self.fc1(x)\n","      x = F.relu(x)\n","      x = self.fc2(x)\n","      x = F.relu(x)\n","      x = self.fc3(x)\n","      ###########################################################\n","      #                         END OF YOUR CODE                #\n","      ###########################################################\n","      return x"],"execution_count":11,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"OjayBqpCXBJG"},"source":["### 5.2 Define the Loss Function and the Optimizer"]},{"cell_type":"code","metadata":{"id":"MfE6ztl-Vsxr","executionInfo":{"status":"ok","timestamp":1650222315304,"user_tz":-180,"elapsed":516,"user":{"displayName":"Arda Karaman","userId":"09649945184885111594"}}},"source":["import torch.optim as optim\n","\n","D = 3072 # dimensionality\n","C = 10 # num of classes\n","H1 = 1000 # number of hidden neurons\n","H2 = 100 # number of hidden neurons\n","\n","model = ThreeLayerNet(D, H1, H2, C)\n","\n","criterion = nn.CrossEntropyLoss()\n","optimizer = optim.SGD(model.parameters(), lr=5e-2, weight_decay=1e-2)"],"execution_count":12,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"AQw2uB8kXFu-"},"source":["### 5.3 Train the Model"]},{"cell_type":"code","metadata":{"id":"M_fNyLKWV5kg","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1650222431850,"user_tz":-180,"elapsed":114029,"user":{"displayName":"Arda Karaman","userId":"09649945184885111594"}},"outputId":"afdac071-ad92-4605-ca6c-0669c298cd8a"},"source":["model = model.to(device)\n","epochs = 10\n","\n","loss_history = train(model, criterion, optimizer, epochs, trainloader)"],"execution_count":13,"outputs":[{"output_type":"stream","name":"stdout","text":["Epoch 0 / 10: avg. loss of last 5 iterations 1.7390442609786987\n","Epoch 1 / 10: avg. loss of last 5 iterations 1.6098030090332032\n","Epoch 2 / 10: avg. loss of last 5 iterations 1.5395172357559204\n","Epoch 3 / 10: avg. loss of last 5 iterations 1.5421529054641723\n","Epoch 4 / 10: avg. loss of last 5 iterations 1.4511483192443848\n","Epoch 5 / 10: avg. loss of last 5 iterations 1.445723867416382\n","Epoch 6 / 10: avg. loss of last 5 iterations 1.4436224937438964\n","Epoch 7 / 10: avg. loss of last 5 iterations 1.3744275093078613\n","Epoch 8 / 10: avg. loss of last 5 iterations 1.3570579290390015\n","Epoch 9 / 10: avg. loss of last 5 iterations 1.3051367044448852\n"]}]},{"cell_type":"markdown","metadata":{"id":"2KD1t5l6WpzE"},"source":["### 5.4 Loss Curve\n","\n","Let us visualize the loss curve."]},{"cell_type":"code","metadata":{"id":"XOALtgdcWsHj","colab":{"base_uri":"https://localhost:8080/","height":279},"executionInfo":{"status":"ok","timestamp":1650222436039,"user_tz":-180,"elapsed":442,"user":{"displayName":"Arda Karaman","userId":"09649945184885111594"}},"outputId":"10bcbb1a-6de7-47b0-c083-2bf07b8bca7d"},"source":["plt.plot(loss_history)\n","plt.xlabel('Iteration number')\n","plt.ylabel('Loss value')\n","plt.show()"],"execution_count":14,"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"}}]},{"cell_type":"markdown","metadata":{"id":"XRPl6GSAW1mz"},"source":["### 5.5 Quantitative Analysis\n","\n","You should see around 52\\% accuracy here. By finetuning the model, you should be able to obtain better results.\n","\n","*Disclaimer: This code piece is taken from PyTorch examples.*"]},{"cell_type":"code","metadata":{"id":"QM1Fin7mW2B0","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1650222443406,"user_tz":-180,"elapsed":3691,"user":{"displayName":"Arda Karaman","userId":"09649945184885111594"}},"outputId":"69952e19-e3fe-4eae-d15b-3b357fc1e965"},"source":["correct = 0\n","total = 0\n","model = model.to(\"cpu\")\n","with torch.no_grad():\n","    for data in testloader:\n","        images, labels = data\n","        outputs = model(images)\n","        _, predicted = torch.max(outputs.data, 1)\n","        total += labels.size(0)\n","        correct += (predicted == labels).sum().item()\n","\n","print('Accuracy of the network on the 10000 test images: %d %%' % (\n","    100 * correct / total))"],"execution_count":15,"outputs":[{"output_type":"stream","name":"stdout","text":["Accuracy of the network on the 10000 test images: 51 %\n"]}]}]}