{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"## Image Classification with Pytorch","metadata":{}},{"cell_type":"markdown","source":"### there are 8 processes while doing this job.","metadata":{}},{"cell_type":"markdown","source":"1. import Library\n2. hyper parameter setting\n3. set seed and random value \n4. data load\n5. make our model\n6. set our loss function and optimizer\n7. train\n8. check our model performance or using test datasets in trained model","metadata":{}},{"cell_type":"markdown","source":"\n### 1. Import Library","metadata":{}},{"cell_type":"code","source":"#about torch...\nimport torch\nimport torch.nn as nn\nimport torch\nimport torch.optim as optim\nimport torch.nn.functional as F\n\nfrom torchvision import datasets, models, transforms\nfrom torch.utils.data import DataLoader, Dataset\n\n#using numpy\nimport numpy as np\n\n#for data load or save\nimport pandas as pd\n\n#visualize some datasets\nimport matplotlib.pyplot as plt\n\n#check our work directory\nimport os\n\n#to unzip datasets\nimport zipfile","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","execution":{"iopub.status.busy":"2022-02-28T20:08:28.192606Z","iopub.execute_input":"2022-02-28T20:08:28.193129Z","iopub.status.idle":"2022-02-28T20:08:28.201938Z","shell.execute_reply.started":"2022-02-28T20:08:28.193076Z","shell.execute_reply":"2022-02-28T20:08:28.200845Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### 2. Hyper parameters Setting","metadata":{}},{"cell_type":"code","source":"lr = 0.001 # learning_rate\nbatch_size = 100 # we will use mini-batch method\nepochs = 10 # How much to train a model","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:28.204252Z","iopub.execute_input":"2022-02-28T20:08:28.205053Z","iopub.status.idle":"2022-02-28T20:08:28.210955Z","shell.execute_reply.started":"2022-02-28T20:08:28.205013Z","shell.execute_reply":"2022-02-28T20:08:28.20988Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"3. Set seed and Random Value","metadata":{}},{"cell_type":"code","source":"device = 'cuda' if torch.cuda.is_available() else 'cpu'\n\ntorch.manual_seed(1234)\nif device =='cuda':\n    torch.cuda.manual_seed_all(1234)","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:28.212709Z","iopub.execute_input":"2022-02-28T20:08:28.213531Z","iopub.status.idle":"2022-02-28T20:08:28.220463Z","shell.execute_reply.started":"2022-02-28T20:08:28.21349Z","shell.execute_reply":"2022-02-28T20:08:28.219498Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### 4. Load train,  test data","metadata":{}},{"cell_type":"code","source":"os.listdir('../input/dogs-vs-cats-redux-kernels-edition')","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:28.222324Z","iopub.execute_input":"2022-02-28T20:08:28.223014Z","iopub.status.idle":"2022-02-28T20:08:28.232306Z","shell.execute_reply.started":"2022-02-28T20:08:28.222915Z","shell.execute_reply":"2022-02-28T20:08:28.23125Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### make a directory for our datasets after unzipping","metadata":{}},{"cell_type":"code","source":"os.makedirs('../data', exist_ok=True)","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:28.235862Z","iopub.execute_input":"2022-02-28T20:08:28.236447Z","iopub.status.idle":"2022-02-28T20:08:28.241376Z","shell.execute_reply.started":"2022-02-28T20:08:28.23641Z","shell.execute_reply":"2022-02-28T20:08:28.240317Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"base_dir = '../input/dogs-vs-cats-redux-kernels-edition'\ntrain_dir = '../data/train'\ntest_dir = '../data/test'","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:28.245149Z","iopub.execute_input":"2022-02-28T20:08:28.246021Z","iopub.status.idle":"2022-02-28T20:08:28.250232Z","shell.execute_reply.started":"2022-02-28T20:08:28.245983Z","shell.execute_reply":"2022-02-28T20:08:28.249467Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with zipfile.ZipFile(os.path.join(base_dir, 'train.zip')) as train_zip:\n    train_zip.extractall('../data')\n    \nwith zipfile.ZipFile(os.path.join(base_dir, 'test.zip')) as test_zip:\n    test_zip.extractall('../data')","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:28.252085Z","iopub.execute_input":"2022-02-28T20:08:28.252674Z","iopub.status.idle":"2022-02-28T20:08:45.182204Z","shell.execute_reply.started":"2022-02-28T20:08:28.252634Z","shell.execute_reply":"2022-02-28T20:08:45.181361Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"os.listdir(train_dir)[:5]","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:45.183753Z","iopub.execute_input":"2022-02-28T20:08:45.184178Z","iopub.status.idle":"2022-02-28T20:08:45.210787Z","shell.execute_reply.started":"2022-02-28T20:08:45.184135Z","shell.execute_reply":"2022-02-28T20:08:45.209804Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import glob\n\ntrain_list = glob.glob(os.path.join(train_dir,'*.jpg'))\ntest_list = glob.glob(os.path.join(test_dir, '*.jpg'))","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:45.212738Z","iopub.execute_input":"2022-02-28T20:08:45.213141Z","iopub.status.idle":"2022-02-28T20:08:45.607945Z","shell.execute_reply.started":"2022-02-28T20:08:45.213093Z","shell.execute_reply":"2022-02-28T20:08:45.606988Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"len(train_list)","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:45.609564Z","iopub.execute_input":"2022-02-28T20:08:45.609924Z","iopub.status.idle":"2022-02-28T20:08:45.61588Z","shell.execute_reply.started":"2022-02-28T20:08:45.609888Z","shell.execute_reply":"2022-02-28T20:08:45.614944Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### check our datasets","metadata":{}},{"cell_type":"code","source":"from PIL import Image\nrandom_idx = np.random.randint(1,25000,size=10)\n\nfig = plt.figure()\ni=1\nfor idx in random_idx:\n    ax = fig.add_subplot(2,5,i)\n    img = Image.open(train_list[idx])\n    plt.imshow(img)\n    i+=1\n\nplt.axis('off')\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:45.617508Z","iopub.execute_input":"2022-02-28T20:08:45.618181Z","iopub.status.idle":"2022-02-28T20:08:46.765056Z","shell.execute_reply.started":"2022-02-28T20:08:45.618141Z","shell.execute_reply":"2022-02-28T20:08:46.764184Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_list[0].split('/')[-1].split('.')[0]","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:46.766299Z","iopub.execute_input":"2022-02-28T20:08:46.766755Z","iopub.status.idle":"2022-02-28T20:08:46.774913Z","shell.execute_reply.started":"2022-02-28T20:08:46.766718Z","shell.execute_reply":"2022-02-28T20:08:46.77385Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"int(test_list[0].split('/')[-1].split('.')[0])","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:46.776864Z","iopub.execute_input":"2022-02-28T20:08:46.777692Z","iopub.status.idle":"2022-02-28T20:08:46.786451Z","shell.execute_reply.started":"2022-02-28T20:08:46.777649Z","shell.execute_reply":"2022-02-28T20:08:46.7856Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(len(train_list), len(test_list))","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:46.7892Z","iopub.execute_input":"2022-02-28T20:08:46.789872Z","iopub.status.idle":"2022-02-28T20:08:46.795806Z","shell.execute_reply.started":"2022-02-28T20:08:46.789832Z","shell.execute_reply":"2022-02-28T20:08:46.794813Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from sklearn.model_selection import train_test_split\ntrain_list, val_list = train_test_split(train_list, test_size=0.2)","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:46.797567Z","iopub.execute_input":"2022-02-28T20:08:46.798336Z","iopub.status.idle":"2022-02-28T20:08:46.81541Z","shell.execute_reply.started":"2022-02-28T20:08:46.798138Z","shell.execute_reply":"2022-02-28T20:08:46.814751Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Image Augumentation\n- we have to do this job before making our model. cuz it will prevent overfitting Possibility","metadata":{}},{"cell_type":"code","source":"#data Augumentation\ntrain_transforms =  transforms.Compose([\n        transforms.Resize((224, 224)),\n        transforms.RandomResizedCrop(224),\n        transforms.RandomHorizontalFlip(),\n        transforms.ToTensor(),\n    ])\n\nval_transforms = transforms.Compose([\n        transforms.Resize((224, 224)),\n        transforms.RandomResizedCrop(224),\n        transforms.RandomHorizontalFlip(),\n        transforms.ToTensor(),\n    ])\n\n\ntest_transforms = transforms.Compose([   \n    transforms.Resize((224, 224)),\n     transforms.RandomResizedCrop(224),\n    transforms.RandomHorizontalFlip(),\n    transforms.ToTensor()\n    ])\n","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:46.816709Z","iopub.execute_input":"2022-02-28T20:08:46.817111Z","iopub.status.idle":"2022-02-28T20:08:46.827946Z","shell.execute_reply.started":"2022-02-28T20:08:46.817082Z","shell.execute_reply":"2022-02-28T20:08:46.827005Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Load datasets\n- this code is for load our image sets. you can find this code easily in pytorch.org , main page","metadata":{}},{"cell_type":"code","source":"class dataset(torch.utils.data.Dataset):\n    def __init__(self,file_list,transform=None):\n        self.file_list = file_list\n        self.transform = transform\n        \n        \n    #dataset length\n    def __len__(self):\n        self.filelength = len(self.file_list)\n        return self.filelength\n    \n    #load an one of images\n    def __getitem__(self,idx):\n        img_path = self.file_list[idx]\n        img = Image.open(img_path)\n        img_transformed = self.transform(img)\n        \n        label = img_path.split('/')[-1].split('.')[0]\n        if label == 'dog':\n            label=1\n        elif label == 'cat':\n            label=0\n            \n        return img_transformed,label\n        ","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:46.829289Z","iopub.execute_input":"2022-02-28T20:08:46.83012Z","iopub.status.idle":"2022-02-28T20:08:46.840249Z","shell.execute_reply.started":"2022-02-28T20:08:46.830081Z","shell.execute_reply":"2022-02-28T20:08:46.839146Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### use torchvision.datasets","metadata":{}},{"cell_type":"code","source":"train_data = dataset(train_list, transform=train_transforms)\ntest_data = dataset(test_list, transform=test_transforms)\nval_data = dataset(val_list, transform=test_transforms)","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:46.842135Z","iopub.execute_input":"2022-02-28T20:08:46.842659Z","iopub.status.idle":"2022-02-28T20:08:46.849077Z","shell.execute_reply.started":"2022-02-28T20:08:46.842622Z","shell.execute_reply":"2022-02-28T20:08:46.848086Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_loader = torch.utils.data.DataLoader(dataset = train_data, batch_size=batch_size, shuffle=True )\ntest_loader = torch.utils.data.DataLoader(dataset = test_data, batch_size=batch_size, shuffle=True)\nval_loader\n= torch.utils.data.DataLoader(dataset = val_data, batch_size=batch_size, shuffle=True)","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:46.85061Z","iopub.execute_input":"2022-02-28T20:08:46.851057Z","iopub.status.idle":"2022-02-28T20:08:46.859411Z","shell.execute_reply.started":"2022-02-28T20:08:46.851021Z","shell.execute_reply":"2022-02-28T20:08:46.858336Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(len(train_data), len(train_loader))","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:46.861113Z","iopub.execute_input":"2022-02-28T20:08:46.861986Z","iopub.status.idle":"2022-02-28T20:08:46.869744Z","shell.execute_reply.started":"2022-02-28T20:08:46.861946Z","shell.execute_reply":"2022-02-28T20:08:46.868781Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(len(val_data), len(val_loader))","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:46.87143Z","iopub.execute_input":"2022-02-28T20:08:46.872148Z","iopub.status.idle":"2022-02-28T20:08:46.87841Z","shell.execute_reply.started":"2022-02-28T20:08:46.872108Z","shell.execute_reply":"2022-02-28T20:08:46.877374Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"#check our images shape\ntrain_data[0][0].shape","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:46.880021Z","iopub.execute_input":"2022-02-28T20:08:46.880683Z","iopub.status.idle":"2022-02-28T20:08:46.892223Z","shell.execute_reply.started":"2022-02-28T20:08:46.880609Z","shell.execute_reply":"2022-02-28T20:08:46.89157Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### build Model\n- 3 Convolution layer and 2 fully connected layer\n- batchNormalization for limit overfitting","metadata":{}},{"cell_type":"code","source":"class Cnn(nn.Module):\n    def __init__(self):\n        super(Cnn,self).__init__()\n        \n        self.layer1 = nn.Sequential(\n            nn.Conv2d(3,16,kernel_size=3, padding=0,stride=2),\n            nn.BatchNorm2d(16),\n            nn.ReLU(),\n            nn.MaxPool2d(2)\n        )\n        \n        self.layer2 = nn.Sequential(\n            nn.Conv2d(16,32, kernel_size=3, padding=0, stride=2),\n            nn.BatchNorm2d(32),\n            nn.ReLU(),\n            nn.MaxPool2d(2)\n            )\n        \n        self.layer3 = nn.Sequential(\n            nn.Conv2d(32,64, kernel_size=3, padding=0, stride=2),\n            nn.BatchNorm2d(64),\n            nn.ReLU(),\n            nn.MaxPool2d(2)\n        )\n        \n        \n        self.fc1 = nn.Linear(3*3*64,10)\n        self.dropout = nn.Dropout(0.5)\n        self.fc2 = nn.Linear(10,2)\n        self.relu = nn.ReLU()\n        \n        \n    def forward(self,x):\n        out = self.layer1(x)\n        out = self.layer2(out)\n        out = self.layer3(out)\n        out = out.view(out.size(0),-1)\n        out = self.relu(self.fc1(out))\n        out = self.fc2(out)\n        return out","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:46.895085Z","iopub.execute_input":"2022-02-28T20:08:46.895324Z","iopub.status.idle":"2022-02-28T20:08:46.909896Z","shell.execute_reply.started":"2022-02-28T20:08:46.895301Z","shell.execute_reply":"2022-02-28T20:08:46.909123Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"model = Cnn().to(device)\nmodel.train()","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:46.913399Z","iopub.execute_input":"2022-02-28T20:08:46.913636Z","iopub.status.idle":"2022-02-28T20:08:46.923993Z","shell.execute_reply.started":"2022-02-28T20:08:46.913612Z","shell.execute_reply":"2022-02-28T20:08:46.923067Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### set Loss function and optimizer","metadata":{}},{"cell_type":"code","source":"optimizer = optim.Adam(params = model.parameters(),lr=0.001)\ncriterion = nn.CrossEntropyLoss()","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:46.925458Z","iopub.execute_input":"2022-02-28T20:08:46.925998Z","iopub.status.idle":"2022-02-28T20:08:46.931225Z","shell.execute_reply.started":"2022-02-28T20:08:46.925962Z","shell.execute_reply":"2022-02-28T20:08:46.930182Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Train our Network","metadata":{}},{"cell_type":"code","source":"epochs = 10\n\nfor epoch in range(epochs):\n    epoch_loss = 0\n    epoch_accuracy = 0\n    \n    for data, label in train_loader:\n        data = data.to(device)\n        label = label.to(device)\n        \n        output = model(data)\n        loss = criterion(output, label)\n        \n        optimizer.zero_grad()\n        loss.backward()\n        optimizer.step()\n        \n        acc = ((output.argmax(dim=1) == label).float().mean())\n        epoch_accuracy += acc/len(train_loader)\n        epoch_loss += loss/len(train_loader)\n        \n    print('Epoch : {}, train accuracy : {}, train loss : {}'.format(epoch+1, epoch_accuracy,epoch_loss))\n    \n    \n    with torch.no_grad():\n        epoch_val_accuracy=0\n        epoch_val_loss =0\n        for data, label in val_loader:\n            data = data.to(device)\n            label = label.to(device)\n            \n            val_output = model(data)\n            val_loss = criterion(val_output,label)\n            \n            \n            acc = ((val_output.argmax(dim=1) == label).float().mean())\n            epoch_val_accuracy += acc/ len(val_loader)\n            epoch_val_loss += val_loss/ len(val_loader)\n            \n        print('Epoch : {}, val_accuracy : {}, val_loss : {}'.format(epoch+1, epoch_val_accuracy,epoch_val_loss))","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:08:46.932696Z","iopub.execute_input":"2022-02-28T20:08:46.933139Z","iopub.status.idle":"2022-02-28T20:29:38.50512Z","shell.execute_reply.started":"2022-02-28T20:08:46.933098Z","shell.execute_reply":"2022-02-28T20:29:38.504165Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"","metadata":{}},{"cell_type":"code","source":"dog_probs = []\nmodel.eval()\nwith torch.no_grad():\n    for data, fileid in test_loader:\n        data = data.to(device)\n        preds = model(data)\n        preds_list = F.softmax(preds, dim=1)[:, 1].tolist()\n        dog_probs += list(zip(list(fileid), preds_list))","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:29:38.506467Z","iopub.execute_input":"2022-02-28T20:29:38.507035Z","iopub.status.idle":"2022-02-28T20:30:41.795683Z","shell.execute_reply.started":"2022-02-28T20:29:38.50697Z","shell.execute_reply":"2022-02-28T20:30:41.794925Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"- In this Problem, we used softmax except sigmoid. \n- why? isn't it binary classification?","metadata":{}},{"cell_type":"code","source":"dog_probs.sort(key = lambda x : int(x[0]))\ndog_probs","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:30:41.796967Z","iopub.execute_input":"2022-02-28T20:30:41.797291Z","iopub.status.idle":"2022-02-28T20:30:41.858308Z","shell.execute_reply.started":"2022-02-28T20:30:41.797258Z","shell.execute_reply":"2022-02-28T20:30:41.857383Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"idx = list(map(lambda x: x[0],dog_probs))\nprob = list(map(lambda x: x[1],dog_probs))","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:30:41.860126Z","iopub.execute_input":"2022-02-28T20:30:41.860822Z","iopub.status.idle":"2022-02-28T20:30:41.872924Z","shell.execute_reply.started":"2022-02-28T20:30:41.860758Z","shell.execute_reply":"2022-02-28T20:30:41.872121Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"submission = pd.DataFrame({'id':idx,'label':prob})\nsubmission","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:30:41.878142Z","iopub.execute_input":"2022-02-28T20:30:41.878423Z","iopub.status.idle":"2022-02-28T20:30:41.908552Z","shell.execute_reply.started":"2022-02-28T20:30:41.878398Z","shell.execute_reply":"2022-02-28T20:30:41.907751Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### check our model performance and visualize some data!","metadata":{}},{"cell_type":"code","source":"import random\n\nid_list = []\nclass_ = {0: 'cat', 1: 'dog'}\n\nfig, axes = plt.subplots(2, 5, figsize=(20, 12), facecolor='w')\n\nfor ax in axes.ravel():\n    \n    i = random.choice(submission['id'].values)\n    \n    label = submission.loc[submission['id'] == i, 'label'].values[0]\n    if label > 0.5:\n        label = 1\n    else:\n        label = 0\n        \n    img_path = os.path.join(test_dir, '{}.jpg'.format(i))\n    img = Image.open(img_path)\n    \n    ax.set_title(class_[label])\n    ax.imshow(img)","metadata":{"execution":{"iopub.status.busy":"2022-02-28T20:30:41.910038Z","iopub.execute_input":"2022-02-28T20:30:41.910452Z","iopub.status.idle":"2022-02-28T20:30:43.302882Z","shell.execute_reply.started":"2022-02-28T20:30:41.910417Z","shell.execute_reply":"2022-02-28T20:30:43.302052Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{"trusted":true},"execution_count":null,"outputs":[]}]}