VAEs_Conv/VAE_Conv_Big.ipynb

{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.10.13","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"nvidiaTeslaT4","dataSources":[{"sourceId":7617761,"sourceType":"datasetVersion","datasetId":4436752}],"dockerImageVersionId":30648,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import numpy as np\nimport matplotlib.pyplot as plt\nimport os\nimport time\nfrom tqdm import tqdm\nimport torch\nimport torchvision\nfrom torch import nn\nfrom torch import optim\nimport torch.nn.functional as F\nfrom torch.utils.data import DataLoader\nimport torch.nn as nn\nimport torch.nn.functional as F\nimport torch.optim as optim\nfrom torchvision import datasets, transforms\nimport torchvision.transforms as transforms\nfrom torchvision.utils import save_image\nfrom torchsummary import summary","metadata":{"execution":{"iopub.status.busy":"2024-02-13T13:03:32.217614Z","iopub.execute_input":"2024-02-13T13:03:32.218532Z","iopub.status.idle":"2024-02-13T13:03:36.842615Z","shell.execute_reply.started":"2024-02-13T13:03:32.218495Z","shell.execute_reply":"2024-02-13T13:03:36.841649Z"},"trusted":true},"execution_count":5,"outputs":[]},{"cell_type":"code","source":"device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\nprint(f\"Using device: {device}\")","metadata":{"execution":{"iopub.status.busy":"2024-02-13T13:11:30.993313Z","iopub.execute_input":"2024-02-13T13:11:30.993738Z","iopub.status.idle":"2024-02-13T13:11:30.999562Z","shell.execute_reply.started":"2024-02-13T13:11:30.993705Z","shell.execute_reply":"2024-02-13T13:11:30.998560Z"},"trusted":true},"execution_count":13,"outputs":[{"name":"stdout","text":"Using device: cuda\n","output_type":"stream"}]},{"cell_type":"code","source":"import sys\nsys.path.append('/kaggle/input/models-vae')","metadata":{"execution":{"iopub.status.busy":"2024-02-13T13:03:50.759914Z","iopub.execute_input":"2024-02-13T13:03:50.760803Z","iopub.status.idle":"2024-02-13T13:03:50.765483Z","shell.execute_reply.started":"2024-02-13T13:03:50.760758Z","shell.execute_reply":"2024-02-13T13:03:50.764478Z"},"trusted":true},"execution_count":6,"outputs":[]},{"cell_type":"markdown","source":"Importing the model","metadata":{}},{"cell_type":"code","source":"from vae_conv2 import VAEAutoEncoder","metadata":{"execution":{"iopub.status.busy":"2024-02-13T13:04:31.829491Z","iopub.execute_input":"2024-02-13T13:04:31.830350Z","iopub.status.idle":"2024-02-13T13:04:31.851372Z","shell.execute_reply.started":"2024-02-13T13:04:31.830308Z","shell.execute_reply":"2024-02-13T13:04:31.850285Z"},"trusted":true},"execution_count":7,"outputs":[]},{"cell_type":"code","source":"model = VAEAutoEncoder()\n\nif torch.cuda.device_count() > 1:\n    print(f\"using {torch.cuda.device_count()} GPUs\")\n    model = nn.DataParallel(model)\n\ndevice = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\nmodel.to(device)","metadata":{"execution":{"iopub.status.busy":"2024-02-13T13:07:02.172094Z","iopub.execute_input":"2024-02-13T13:07:02.172472Z","iopub.status.idle":"2024-02-13T13:07:02.191931Z","shell.execute_reply.started":"2024-02-13T13:07:02.172442Z","shell.execute_reply":"2024-02-13T13:07:02.190987Z"},"trusted":true},"execution_count":11,"outputs":[{"name":"stdout","text":"using 2 GPUs\n","output_type":"stream"},{"execution_count":11,"output_type":"execute_result","data":{"text/plain":"DataParallel(\n  (module): VAEAutoEncoder(\n    (encoder): VAEEncoder(\n      (conv1): Conv2d(1, 32, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))\n      (conv2): Conv2d(32, 64, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))\n      (conv3): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n      (flatten): Flatten(start_dim=1, end_dim=-1)\n      (z_mean): Linear(in_features=3136, out_features=50, bias=True)\n      (z_log): Linear(in_features=3136, out_features=50, bias=True)\n      (relu): ReLU()\n    )\n    (decoder): VAEDecoder(\n      (fc1): Linear(in_features=50, out_features=3136, bias=True)\n      (reshape): Unflatten(dim=1, unflattened_size=(64, 7, 7))\n      (conv1): ConvTranspose2d(64, 64, kernel_size=(3, 3), stride=(2, 2))\n      (conv2): ConvTranspose2d(64, 32, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))\n      (conv3): ConvTranspose2d(32, 1, kernel_size=(2, 2), stride=(1, 1), padding=(1, 1))\n      (relu): ReLU()\n      (sigmoid): Sigmoid()\n    )\n  )\n)"},"metadata":{}}]},{"cell_type":"code","source":"# model latent dimension is 50\ntrain_size = 0.9\nlr = 1e-3\neps = 1e-8\nbatch_size = 512\nepochs = 20","metadata":{"execution":{"iopub.status.busy":"2024-02-13T13:30:40.147645Z","iopub.execute_input":"2024-02-13T13:30:40.148518Z","iopub.status.idle":"2024-02-13T13:30:40.153443Z","shell.execute_reply.started":"2024-02-13T13:30:40.148480Z","shell.execute_reply":"2024-02-13T13:30:40.152395Z"},"trusted":true},"execution_count":23,"outputs":[]},{"cell_type":"markdown","source":"Dataset","metadata":{}},{"cell_type":"code","source":"def mnist(train_part, transform=None):\n    dataset = torchvision.datasets.MNIST('.', download=True, transform=transform)\n    train_part = int(train_part * len(dataset))\n    train_dataset, test_dataset = torch.utils.data.random_split(dataset, [train_part, len(dataset) - train_part])\n    return train_dataset, test_dataset","metadata":{"execution":{"iopub.status.busy":"2024-02-13T13:12:48.795979Z","iopub.execute_input":"2024-02-13T13:12:48.797004Z","iopub.status.idle":"2024-02-13T13:12:48.802389Z","shell.execute_reply.started":"2024-02-13T13:12:48.796969Z","shell.execute_reply":"2024-02-13T13:12:48.801408Z"},"trusted":true},"execution_count":15,"outputs":[]},{"cell_type":"code","source":"transform = transforms.Compose([transforms.ToTensor()])\n\ntrain_dataset, test_dataset = mnist(train_size, transform)\n\ntrain_dataloader = torch.utils.data.DataLoader(train_dataset, drop_last=True, batch_size=batch_size, shuffle=True)\ntest_dataloader = torch.utils.data.DataLoader(test_dataset, batch_size=1, shuffle=False)\ndataloaders = (train_dataloader, test_dataloader)","metadata":{"execution":{"iopub.status.busy":"2024-02-13T13:12:49.099783Z","iopub.execute_input":"2024-02-13T13:12:49.100819Z","iopub.status.idle":"2024-02-13T13:12:50.226482Z","shell.execute_reply.started":"2024-02-13T13:12:49.100782Z","shell.execute_reply":"2024-02-13T13:12:50.225399Z"},"trusted":true},"execution_count":16,"outputs":[{"name":"stdout","text":"Downloading http://yann.lecun.com/exdb/mnist/train-images-idx3-ubyte.gz\nDownloading http://yann.lecun.com/exdb/mnist/train-images-idx3-ubyte.gz to ./MNIST/raw/train-images-idx3-ubyte.gz\n","output_type":"stream"},{"name":"stderr","text":"100%|██████████| 9912422/9912422 [00:00<00:00, 113194945.85it/s]\n","output_type":"stream"},{"name":"stdout","text":"Extracting ./MNIST/raw/train-images-idx3-ubyte.gz to ./MNIST/raw\n\nDownloading http://yann.lecun.com/exdb/mnist/train-labels-idx1-ubyte.gz\nDownloading http://yann.lecun.com/exdb/mnist/train-labels-idx1-ubyte.gz to ./MNIST/raw/train-labels-idx1-ubyte.gz\n","output_type":"stream"},{"name":"stderr","text":"100%|██████████| 28881/28881 [00:00<00:00, 41713393.19it/s]","output_type":"stream"},{"name":"stdout","text":"Extracting ./MNIST/raw/train-labels-idx1-ubyte.gz to ./MNIST/raw\n\nDownloading http://yann.lecun.com/exdb/mnist/t10k-images-idx3-ubyte.gz\nDownloading http://yann.lecun.com/exdb/mnist/t10k-images-idx3-ubyte.gz to ./MNIST/raw/t10k-images-idx3-ubyte.gz\n","output_type":"stream"},{"name":"stderr","text":"\n100%|██████████| 1648877/1648877 [00:00<00:00, 27470066.44it/s]\n","output_type":"stream"},{"name":"stdout","text":"Extracting ./MNIST/raw/t10k-images-idx3-ubyte.gz to ./MNIST/raw\n\nDownloading http://yann.lecun.com/exdb/mnist/t10k-labels-idx1-ubyte.gz\nDownloading http://yann.lecun.com/exdb/mnist/t10k-labels-idx1-ubyte.gz to ./MNIST/raw/t10k-labels-idx1-ubyte.gz\n","output_type":"stream"},{"name":"stderr","text":"100%|██████████| 4542/4542 [00:00<00:00, 7772553.56it/s]\n","output_type":"stream"},{"name":"stdout","text":"Extracting ./MNIST/raw/t10k-labels-idx1-ubyte.gz to ./MNIST/raw\n\n","output_type":"stream"}]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Loss function -> reconstruction loss + KL divergence","metadata":{}},{"cell_type":"code","source":"def vae_loss(preds, targets, z_mean, z_log, z_val):\n    mse = nn.MSELoss()\n    reconstruction_loss = mse(preds.view(targets.shape[0], -1), targets.view(targets.shape[0], -1)) * 784.0\n    temp = 1.0 + z_log - torch.square(z_mean) - torch.exp(z_log)\n    kl_loss = -0.5 * torch.sum(temp, axis=-1)\n    return torch.mean(reconstruction_loss + kl_loss)","metadata":{"execution":{"iopub.status.busy":"2024-02-13T13:13:18.540492Z","iopub.execute_input":"2024-02-13T13:13:18.540887Z","iopub.status.idle":"2024-02-13T13:13:18.549337Z","shell.execute_reply.started":"2024-02-13T13:13:18.540856Z","shell.execute_reply":"2024-02-13T13:13:18.548226Z"},"trusted":true},"execution_count":17,"outputs":[]},{"cell_type":"code","source":"optimizer = optim.RMSprop(model.parameters(), lr=lr, eps=eps)","metadata":{"execution":{"iopub.status.busy":"2024-02-13T13:13:27.886311Z","iopub.execute_input":"2024-02-13T13:13:27.886709Z","iopub.status.idle":"2024-02-13T13:13:27.892473Z","shell.execute_reply.started":"2024-02-13T13:13:27.886677Z","shell.execute_reply":"2024-02-13T13:13:27.891146Z"},"trusted":true},"execution_count":18,"outputs":[]},{"cell_type":"code","source":"def train_vae(dataloaders, model, optimizer, epochs, device):\n    tqdm_iter = tqdm(range(epochs))\n    train_dataloader, test_dataloader = dataloaders[0], dataloaders[1]\n    epoch_train_losses = []\n\n    for epoch in tqdm_iter:\n        model.train()\n        train_loss = 0.0\n        test_loss = 0.0\n\n        for batch in train_dataloader:\n            imgs, _ = batch  \n            imgs = imgs.to(device)\n\n            optimizer.zero_grad()\n\n            preds,z_mean, z_log, z_val = model(imgs)\n            #print(preds.shape)\n            #print(z_val.shape)\n            loss = vae_loss(preds, imgs, z_mean, z_log, z_val)\n\n            loss.backward()\n            optimizer.step()\n\n            train_loss += loss.item()\n\n        model.eval()\n        with torch.no_grad():\n            for batch in test_dataloader:\n                imgs, _ = batch\n                imgs = imgs.to(device)\n\n                preds,z_mean, z_log, z_val = model(imgs)\n                \n                loss = vae_loss(preds, imgs, z_mean, z_log, z_val)\n\n                test_loss += loss.item()\n\n        train_loss /= len(train_dataloader)\n        test_loss /= len(test_dataloader)\n        \n        epoch_train_losses.append(train_loss)\n\n        tqdm_dct = {'train loss': train_loss, 'test loss': test_loss}\n        tqdm_iter.set_postfix(tqdm_dct, refresh=True)\n        tqdm_iter.refresh()\n        \n    plt.figure(figsize=(5, 3))\n    plt.plot(range(1, epochs + 1), epoch_train_losses, label='Train Loss')\n    plt.title('Loss vs. Epochs')\n    plt.xlabel('Epochs')\n    plt.ylabel('Loss')\n    plt.legend()\n    plt.show()","metadata":{"execution":{"iopub.status.busy":"2024-02-13T13:13:45.544314Z","iopub.execute_input":"2024-02-13T13:13:45.544680Z","iopub.status.idle":"2024-02-13T13:13:45.556849Z","shell.execute_reply.started":"2024-02-13T13:13:45.544651Z","shell.execute_reply":"2024-02-13T13:13:45.555720Z"},"trusted":true},"execution_count":19,"outputs":[]},{"cell_type":"code","source":"train_vae(dataloaders, model, optimizer, epochs, device)","metadata":{"execution":{"iopub.status.busy":"2024-02-13T13:15:03.919157Z","iopub.execute_input":"2024-02-13T13:15:03.919549Z","iopub.status.idle":"2024-02-13T13:25:19.563328Z","shell.execute_reply.started":"2024-02-13T13:15:03.919520Z","shell.execute_reply":"2024-02-13T13:25:19.562361Z"},"trusted":true},"execution_count":20,"outputs":[{"name":"stderr","text":"100%|██████████| 20/20 [10:15<00:00, 30.77s/it, train loss=29.6, test loss=30.4]\n","output_type":"stream"},{"output_type":"display_data","data":{"text/plain":"<Figure size 500x300 with 1 Axes>","image/png":"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"},"metadata":{}}]},{"cell_type":"code","source":"def plotn(n, data):\n    fig, ax = plt.subplots(1, n)\n    for i, z in enumerate(data):\n        if i == n:\n            break\n        preprocess = z[0].reshape(1, 28, 28)\n        ax[i].imshow(preprocess[0],cmap ='gray')\n        ax[i].axis('off')\n    plt.show()","metadata":{"execution":{"iopub.status.busy":"2024-02-13T13:29:13.067131Z","iopub.execute_input":"2024-02-13T13:29:13.068286Z","iopub.status.idle":"2024-02-13T13:29:13.074438Z","shell.execute_reply.started":"2024-02-13T13:29:13.068250Z","shell.execute_reply":"2024-02-13T13:29:13.073579Z"},"trusted":true},"execution_count":21,"outputs":[]},{"cell_type":"code","source":"model.eval()\npredictions = []\nplots = 10\nfor i, data in enumerate(test_dataset):\n    if i == plots:\n        break\n        \n    output_imgs = model(data[0].to(device).unsqueeze(0))\n    #print(output_imgs[0].shape)\n    #print(output_imgs[3].shape)\n    op_img = output_imgs[0].view(1, 28, 28)\n    #print(op_img.shape)\n    predictions.append(op_img.detach().cpu())\nplotn(plots, test_dataset)\nplotn(plots, predictions)","metadata":{"execution":{"iopub.status.busy":"2024-02-13T13:29:14.615173Z","iopub.execute_input":"2024-02-13T13:29:14.615546Z","iopub.status.idle":"2024-02-13T13:29:15.263325Z","shell.execute_reply.started":"2024-02-13T13:29:14.615517Z","shell.execute_reply":"2024-02-13T13:29:15.262382Z"},"trusted":true},"execution_count":22,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 640x480 with 10 Axes>","image/png":"iVBORw0KGgoAAAANSUhEUgAAAgMAAAA+CAYAAAC2oBgNAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8WgzjOAAAACXBIWXMAAA9hAAAPYQGoP6dpAAAv/0lEQVR4nO2daWxb55W/H+6buEiiFmqjFmu3JMeSbVnxktqK08STxE2bNGhSNGnapOgMZjATTD9MgX4o2g8doGj7L6YtpsgADQZJm2niuk2TNHZT77a8KJItWfu+UyQlijspkf8Pxr21bCdxYomk7fsAhm2Kkt6X9973Pe85v3OOLB6Px5GQkJCQkJC4Z5EnewASEhISEhISyUUyBiQkJCQkJO5xJGNAQkJCQkLiHkcyBiQkJCQkJO5xJGNAQkJCQkLiHkcyBiQkJCQkJO5xJGNAQkJCQkLiHkcyBiQkJCQkJO5xJGNAQkJCQkLiHkd5q2+UyWTrOY415eOKKt4t84C7Zy53yzzg7pnL3TIPuHvmcrfMA+6eudwt8wDJMyAhISEhIXHPIxkDEhISEhIS9ziSMSAhISEhIXGPc8uagc+KTCYT/1z72rXxi5vFMuLx+CfGOCTWl5tdu1gsJl0XiXXh2vtNus8kJBLLuhkDer2ezMxMduzYQX19PTabDa1Wi8ViQa/X4/P5iMfjLC8vc+XKFaLRqPi9k5OTXLx4kZ6eHgKBwHoNUeIT2LZtG48++igVFRUAeL1efvGLXzAyMoLL5Ury6CTuJmQyGV/60peora2lubmZH/zgB3R1dbG4uJjsoUlI3BOsuTGg1WrJy8tjw4YN5Ofns337dsrLy7FarWg0GtLS0tDpdAQCAWKxGLFYDJPJxPLysvgzZmdnyczMBGBqaorZ2dm1HuaaoFAoSE9PJy0tDYPB8JHvy87ORq/Xo9VqxdeEk4/f7+fSpUt4PB5CoVAihv2JyGQyzGYzJSUlNDU1UVpaikwmw+/3k5OTg8PhkIyBBKLVatHpdOh0OqLRKKFQSDSm7waUSiV6vZ6NGzdSV1dHcXExBoMBuVyKYq43BoOBtLQ0jEYjWq1WXKM0Gg1GoxGdTsfKygqBQICenh4WFxfx+/1JHvXdiUaj4YEHHhCvwYcffsji4iJLS0sJ+f1rbgxkZmayb98+nnrqKaqqqsjNzb3p+8xms/jvwsLCVV8TPAZpaWmcPHmSd999d62HedsoFAq0Wi01NTVUVFSIGyasDnvIZDL27dtHXl6e+FnE43FWVlaIRCKMjo7yb//2b1y+fJnp6emkzOV65HI5ZWVlNDQ00NzcjF6vF123eXl5TExMMD4+nuxh3jNkZGSQl5dHYWEhS0tLzMzMMDg4SCQSSfbQ1gS9Xk9BQQEtLS0UFxcTDAZZWVkhFosle2h3PTabjfLycmpra8nOziYnJwe4eoCpqKggPz+fSCTC2NgY3//+97l48SLDw8NJHvXdidls5uc//zk2mw25XM63vvUt2tvb6e7uTsjvX3NjwGg00tjYSEZGBgqFQnzd5/Px9ttvMzMzg8fjuen3ymQyamtrSU9Px2Qy4Xa7U+a0fD0NDQ3cd999/NM//RMmkwm1Wi2eZK5fxCwWCyqVikgkQiAQYGJigtdff52Ojg6mpqYYGxtLqXBIPB7H5/OJi3I8HmdxcZGxsTEuXbrE6Ohosod4V6JUKjGZTDz88MNkZGSQnp7Oli1bMJlMGAwGDAYDwWAQl8vFd7/7XUZGRpifn0/2sG9ApVKRnp7OgQMH6O3tZXh4mKmpqZt6MvLz82lpaeHrX/86DQ0NBINBOjs7cTgceL3eJIz+7sdoNJKVlcXXvvY16uvrqaqqQqfTEQwG8fl8RCIRHA4Hb775JhaLhfz8fJqbm7FarZhMpmQP/67EbDaTm5uLTqdDoVCgUCh46qmnUCqV9PT0JMQwXnNjIBQKMTExQX5+Pmq1mqysLCKRCG63m9OnTzM1NfWRcUCFQoHb7cZsNmM2mxkeHsbtdq/1EG8bhUKB3W6nsbGR6upq1Gr1DSK7cDgsLn6RSITFxUWmp6eZm5tjdHSU06dP09XVlZLudrlcTk5ODhaLBYVCgUwmIxqN4vF48Hg8CXUTyuVyVCoVNpsNgPn5eVGAGo1GyczMRK/XA1c3U5VKhdVqJRKJEIlESEtLE6+NMP65uTkxTJUKyGQy1Go1xcXF2O12duzYQXp6OhaLhc2bN6PT6VCpVCiVSiKRCF6vl9zcXBwOR0oZA3K5HK1WS1FRESUlJezYsUM0LKenp2/wmKlUKjZt2sSWLVtoaGhArVYzPT1NZ2cnHo+HlZWVJM5mbVEoFKSlpWE2m1GpVPj9fjweD8FgMGFjkMlkaLVaKioqqKqqorm5mZycHLRaLU6nE6fTicPhIBQKMT09TXd3N3q9nsrKSmpra5HL5SiV6645XzfkcvkNgujrEULXiUav15Oeno5arUahUCCXy7Hb7eTk5IgHyfUOC675lR0dHeWHP/whX/3qV9m5cydVVVXMz8/T19fHb37zGwKBwMc+5EeOHFnrIa0pMpkMnU7H1q1befTRR1GpVKu+LoQA5ubmRFHk3NwcPT09/N///R8XLlxgYWEhGUO/ZZRKJQ8//DCNjY2iFkIIa0Sj0YQu0hqNBqvVynPPPUc8HufNN99EqVSyvLyMy+Xi85//PJWVlcBVD0xWVhaPPvooc3NzzM7OUl9fj1qtJhaLcfr0aU6dOsXrr79Of39/Qhfij0OpVJKVlcULL7zA3r17qa+vvyH7RkCtVmM2myksLGR6ejqlvDRqtZqioiK+8Y1vsGvXLu677z5kMhk+n4/Lly+vWmTVajUWi4XvfOc7VFVVkZmZyejoKBcuXOCXv/zlXScc1Ov11NXV0dLSQkZGBj09PZw6dYrBwcGEjUGpVJKfn89zzz3HF7/4RYxGIxcuXODPf/4z7777LtPT00xPT686yABs2rQJu92esNj1eqHValEqlas81tcTCAQSsvFeT2ZmJkVFRej1epRKJfF4nKysLDIyMjAajbjd7jvPGICrG0d2djbZ2dlEIhGOHj3KiRMnCIVCd7S1L3g6hMXOarWu0gnEYjF6eno4c+YMBw8eFN2ckUgEv9/P7Oxsyrs+y8rKqK2t5fOf/zwFBQXi66FQiJmZmVVZH4kgFosRiUR4++23WV5eZnR0lI0bN1JZWckzzzxDTk4OaWlpwN9T0+bm5uju7qa/v1+0uNPS0qiqqsLtdlNYWMjIyEhKGAM5OTkUFhbyhS98gd27d6/SnnwUCoWCp59+mj179jA/Py+Gno4fP05fXx8+ny9Bo1+NwWCgubmZhoYGiouLmZqaYnR0lPHx8RtOW1arlZqaGvLy8kTX8+XLl+nq6kp5r4BcLhdFdxaLhezsbPLz8wFYWFhgbGyMyspKTCYTRqORsrIyrFYrRUVFZGZmIpPJqKqqYnJyMqHGgEwmQ6PRiGGns2fP8qc//Yl33nkHl8tFKBS66Ubodrs5efLkR4Z3k4Ewl/T0dHQ6HePj42JIExDF6o2NjRQVFbFhwwbsdjsGg0H0JN6M8fFxZmdnGRoaoq2tja6urkRNaRXxeJxAIIDP58Pj8dyZYQK4ulhlZmaSmZnJ4uIivb29dHZ2rsoYuJMQbryCggLKysrYuXMnRUVFaDSaG94bjUbx+XzMzMwwOzsrZgncKcpvm83G5s2bsdvtGI1G8fVAIMDk5GRSjIFwOMzo6CjLy8sEg0FsNhsbN25k9+7dKBQKMSvD7XazsLDAxMQEV65cYWhoCJPJRGZmJlarlYyMDILBYEpcC8GlXllZSU1NDdu3b6eoqOgjY7JCaEQwFCorKykuLiYcDuP3+xkdHSUSiRAMBpmenmZpaSmh7k7BBV1SUkJOTo54mvF6vSwuLt7wmSsUCjQaDVqtFpVKRTweJxgMEgwGE36P3QpyuVx0k2dnZ5OXl4fNZiMjI4Pc3FyKiooAcLlc2Gw2amtrsVgsmEwmKioqxPVQoVAQjUaRy+VYLJaEz0MISSmVShwOB2NjY59okPj9fnp7ewmHwymjbVIqlRQXF4vPjNPpJB6Po1KpyMjIwGKxkJmZKYpSq6ursdvt6PV61Gq16MH1+/2ratoI96TX6/1YoyERCIcbIbxxx3kGBBFUfX09paWlnD59mmPHjtHW1rbWvyphaDQaysrKeOmll9i9ezcbN2684fQmk8lQKBTU19dTXl5OSUkJ586d491336Wvry8lTqG3wubNm3n22WcxGAyr5jg5Ockf//jHhJ8OotGo6DJWqVSYzWa+8pWv8MADD6BUKsUH+uzZs7z//vu0tbVx7tw58ZTwyiuvkJmZSW5uLqWlpSwtLTE0NJT06yF4Kv7jP/6Dbdu2idk1N3vghetwrUEgpN4Kr2/cuJH9+/fz05/+lDNnznD48GECgUDCDHCdTkd6ejq1tbWYzWbkcjlGo5FQKMT8/PwN83K5XHR1dYkZETKZjPLy8pTMUhEMHaPRiNVq5Zvf/CYtLS1s3rz5hnXgel3EzRZxtVrNhg0bkirG+6TY+bW4XC7+9re/rfOIPh3p6em8/PLLtLS0kJmZSWtrKxaLhbKyMp555hlKS0ux2+3iRioQDodxOp1EIhEWFhY4f/48wWBQfE7GxsaYm5tjamqKycnJZE0PuVyO1WolPz+f4uJihoeH191IXnNjQKPRkJeXh8FgIBaLMTY2ljS35VqgVqvJzc3lySefZMuWLRQVFX3sQySc+JqamrDb7WzdupVXX32Vnp4eent7Ezjyz4bBYCA7O1vMjIjH40xMTDA6OrpKB5Fo9Ho9JSUlfPnLX6ampkYMDQwNDdHX18fPfvYzpqencTqdq9yFQprq8vIyoVAIlUpFfn6+mKufDKNAqVRSUFDAV7/6VcrKytDpdKvGKixGS0tL1NXVYTQa0Wg0nD17lomJCSYnJ2lqaiIvL4+CggJ0Op14vR555BFKS0sBOHPmTMLSVUtLS2lsbKSpqYmMjAw8Hg+//vWvOXfuHB6P54YNMT8/n23btq2a+8WLF7l8+XJCxnurGAwGsrKyePrppyktLRVT8axW66o6CML8bqb1uN4gWFlZSUqmVCwWw+PxsLCwgN/vp7GxkRMnTohfF4zmgYGBlE5bLSgoEAWQNpsNlUrFv/zLv2AwGLBarWzcuBGVSoXH4yEcDuPxeJiZmaGrq4uZmRlGRkYIhUKEQiFcLtcq0aDP5yMcDhMMBpNaT0FYC8Lh8A0ajvViXTwDFotFTLWLx+MYDAZMJpOoGbiTSo1mZGRgt9vZvn07xcXFq+ojXItw8YR55efnk5ubS3l5Ob29vSwvL9Pf358yCvabIRgygmhQmIugWk9mLNdgMFBYWMiDDz4oLgDhcJj+/n7a2to4fvz4TceWlpaG1WrFZrORm5srioeETXdiYiKh85DJZOTl5VFZWUlLSwtWqxWFQkEwGMTtdrO0tMTY2BgjIyN4PB5KSkpE4+b8+fP09vaKi3VZWRmhUIjMzEwMBgNms5mKigqMRiNDQ0MMDQ0xPz+/7gacXC6nuLiYyspK8vPzkclkLCwscObMGSYnJwmHwzd8T0ZGBrW1tajVavG1sbExpqam1nWst4pCocBkMpGfn09ZWRn79u2jvLxc1AYIXB++uR5BQCks7NFolHA4zNjYWMKFxLFYDJ/Ph8vlwu12U1BQIBaDW1lZISsri5qaGsbGxlLWGFAqldjtdmpra7Hb7eLYGxsbUalUaDQawuEwLpdLLJA0Pz8vZnBNTEwwNDREOBxO6bUYrnpFI5FIwsLM65onkpGRwTPPPINGo+HcuXO0tbWJC16q1g+4ngMHDrBr1y727t37sRXRYrEYw8PDYtyqqKgIpVKJwWDgS1/6ElarldOnT+PxeFJSOyGTyTAYDKsWZ7hqEExOTopx6GRRWlpKXV0dW7ZsQSaTEQgEGBoa4r//+7957733bmoIqFQqHnzwQVpbW3nsscfEmG08Hufw4cMcPXqUH//4xwmbg1wuR61W86Mf/Yht27ZRWFgozqW7u5uf/vSnHD9+nGAwiF6vJy8vj6effpq2tjb++te/8rvf/Q6v18vy8jInTpwQq8R97nOfo6WlhRdffBGFQkFubi7f+ta3xPoV/f396zYnhUKBXq9n37597N69W9wUV1ZWRFHa9SiVSmw2G42Njat0N0tLSykjsM3MzOSZZ57hK1/5Cg0NDWKK7bV81Mn/+q+fP3+eoaEhsX6Cw+Hg9OnTCV8HYrGY6Bo3m81885vfxGazUV1dzcLCAg888ADPPPMMp0+fTklvrkqlIicnhy9/+cvs27cPg8GA1+tlYWGB7u5uMTXy2LFjOJ1OXC4XPp9PNKav/ZPqCLVdnE5nwirwrkudgdHRUYaHh7HZbGRlZbFr1y4qKyvZt28fc3NzuN1uUV0/MTFBW1tbyomGBO1DQ0MDDQ0NNxgCk5OTTE1NceLECdHin5mZQaFQoNPpeOmll7DZbBiNRrKzs9m0aRPf+MY3OHjwIOPj4ylnDKnVavbs2SO6mAVisRgnTpzg0qVLSRrZVa4V0sDVPgnvvPMOExMTNzUEKisr2bhxI08++SQVFRWit0r4fpPJJIYaEoFKpcJut7N3716qq6vFTJSenh4GBgZ47bXXaG9vZ2FhQdwkvF4vp06dEj0fwglTED8Jnrbz58/j9/sxm83s2bOHnJwcNBoNtbW1zM/PMzAwsG4LoMFgoKamhg0bNpCXlwdAV1cXnZ2dzMzM3OBqVSqV3H///TQ3N1NRUYFarcbpdDIwMMCVK1dSwjNQU1NDbW0tBw4cwG63iwLH6xE213A4TCgUwuPxMDs7S39/P1lZWcBV4W17ezvT09MMDQ0RCATE9LVkMT4+zqVLl1heXmbTpk187WtfIxKJoNfr6evrS7m1GP6edfPFL36R5uZmMjIyGBgY4OTJk3R0dNDX10cgECAYDDI5OUkwGCQcDqesh2Otkcvlotfzs16/NTcGIpEIs7Oz9PX1kZWVhU6nIz8/XzwpO51OUWHf19fHlStXmJ6exuv1ii605eXlpN+QGo1GjA+WlJQQi8WIRqNibfj+/n66urp47bXX8Pl8RKNRlpaWUCqV4mlNpVJhNBrF9KLHH3+cc+fOMT8/n5LGQFNT06p0QkHJf+nSJUZGRpI4OlbVTYersb0TJ07cUHRH8HBUV1fT2trKrl27MJvNq74X/p4KmiiEokL/8A//QEFBAVqtlmAwyJUrVzh79ixvvvnmqg0nGo3i9/s5f/48Fy5cuGlJUuH6CILItLQ06uvrsVqtqFQqCgoKxFTF9TIG9Ho9VVVV5OXlkZ6eDsDAwAAXLlwQhVrXolAo2LZtG/X19avS8To6OpiYmGBpaQmz2SzWhlhcXEx4WNFut1NfX8+2bdtWGQLXj2FlZYWRkRG8Xi8+n080BE6ePElJSQlyuZzFxUWGhoZYWFhgbm4uYXP4OBwOB4ODg0SjUUpLS0lPT0cul9Pd3c2pU6dScgPNycmhpqaGJ554guzsbAC6u7tFD5/D4RCLjl17+jcajavW7jvBK/BZEDJdhAJxn4U1NwaE2Pl//ud/8qtf/YqamhqKi4vJycmhtraWsrIy8vPzqaqqoqGhgZWVFf793/+doaEhOjo6+OCDDxgYGBBTEZN18Ww2G88//zylpaUoFAocDgcXLlygp6eHP/7xj0xOTuJ0OsV4zrXjdLvdfO973+PAgQN873vfA64WxNmyZQvFxcVMTEykVOEhpVJJWloaTU1Nq2Kibreb0dFR5ufnk96c5JFHHmHHjh3i/30+H2fOnFk1LplMRlpaGt/+9rdpbW1lx44dqFSqmyq+Dx06xJkzZxI2/pycHKqqqnjwwQdRKpXMzc1x+PBhfvnLX9Ld3X3Dfe73+wkEAvzqV7+6JZ3GzMwMv//973n66afF02wiMJvNbN++XTQEAHp6ejh37txNc9Y1Gg3f/va3xYqScHVzOnnyJAqFgqamJp588kl27tzJ7Ows3/rWt3A4HAlNaVOpVGL63ceJhQOBAM8//zzT09PiOrCyssLKygoXL14E/m50ptIm5HQ6USqVDA4OYrfbKS8vx+/3c/ToUf70pz8l/Vm/Gdu2bWPXrl2UlZUxMzPDwMAA3/3ud8X07Xg8Tn19PY2NjQSDQZaWlvB4POzevRu3283Fixfp7OxMybndLkJmyO0+8+umGRBcmP39/TgcDtLS0rhy5Qq5ublkZWWRl5dHTk4OOTk5bN68Wczbz8rKYmJigu7ubs6dO8fs7GxSGvgsLS3R1tZGdnY2NpuN2dlZTp06RW9vL0NDQywtLX1kzvrKygpTU1MMDQ3R29tLSUkJGo1GTD1cXFxMaLGRT6KqqopNmzZRWlqKxWIRT5Kzs7O0tbWxuLiY9NOCUKJTQAgbCP9Wq9WUlZVRUVFBa2sr5eXlN+gfrkWoppgIZDIZW7dupa6uTjy5LC0t0dHRIRYNuhnxePyWP3ehPPPg4CB5eXls3rxZ7Ha4noTDYSYnJwkEAsTjceRyOffddx8rKyuEw2FmZmZWGb7X1xYAKCoq4oknnmDHjh2YTCY2btyIzWbD6/UmpXOhUBteQAgHCP0STCYTKpWKUCiEQqEgFovdcA1TuWiSUMhrZmaGzMxMMjIyOH/+PN3d3TgcjpQcu8ViISMjQyz4lJ+fz/PPP4/f7xefEbvdTlFRkRhCC4VClJSUEAwGue+++3jrrbcYHBxMyayupaUlHA4H4XAYjUZzy2mf8PeqtzcT6n4a1s0YEBayqampVXFAvV6PXq8nNzeXmpoa6urqKCoqwmw2U15eTnV1NYuLi4yOjqJUKrl06RIul4toNLrubl2hVoBcLsfv93P8+HEMBgM5OTnMzc1x9OjRWy7/6nQ6GR8fp7e3F5vNJrqpGxoa8Pl8HDp0KGXUrLW1tbS2tmK328W4eiQSYWJigrNnz6aEMSA83AIKhQKLxSLeZ2azmc2bN9PS0iJ6BITP9/qc6lgsJqYbJgK5XM62bduoq6sDroYAFhYW6OzsFN3ga8Xk5CQTExOiMfBxrbXXAkGg6PF4iEajaDQatm7dis1mw+Fw0NnZydjYmPh+o9EoivGuNQaEoj2CS9fn8yWtf4RGo0Gj0Yj3VigUYnx8nCtXrjA7O0tBQYGoQcnKymJxcfGOKp8sfPYLCwtiLYoPP/yQ/v7+lJ2HTqcTiwAZDAaKiop46aWXVnW3VKlU4uk4FouxsrIiNv6Bq+EolUqVssbA3NycKB7+tKd8IWR4OyS860QwGCQUComn43feeYf/9//+H42NjbS2tvK5z32O7OxsGhoa+MEPfkBPTw+/+c1vOHLkyE0FSWuJkKNaVlaGy+Xi0KFD/O53v0Mul4vW9K1yrbV2rffg/vvvR6vV8sorryTc/Xk9Qp+FvLw8SkpKxEUiEonwhz/8gXfffZdDhw4lvUAPwB//+EcWFhbEOK7dbueVV14RxZj19fXk5eWRkZGBWq0WF+jl5WXMZvMqQZfL5WJwcDBhHieZTMauXbuora0lHo/zwQcfcPLkSU6dOrXm3onm5ma2b98OQFNTE1qt9lOdMj4tTqeTQ4cOUVpaysLCAg8++CAZGRmYzWZ++MMfsry8vOqkKZPJSE9Pv+kmH41GGR8f57333uPQoUMMDQ19pEB0PRDEv3v27GH79u1MTk5y5MgR2tvbxecgFouRnp7OP//zP/PSSy/x/e9/nzfeeINf/OIXSTeYb5WioiJqa2vZv38/ZrMZmUzGtm3bGBsbE/t+pBodHR2kpaWxa9culEolSqUStVrNwMCA6Jman5/H4XCIXtmBgQG+973vUV9fT11dneiVTkUE8alQAClRYb5rSbgxIMTXhdMZXI3/9vT0iHXli4qKqK6uFgv3PPTQQywuLnLlypV1VbUbDAaKi4vZsmULk5OTHDx48LY2QqfTSXt7u9iFDq4KyTQazSfGI9eD0tJSdDqdWLZ3ZWWFwsJCioqKKCgoWFVoKBAI4Pf7U6Z878TEBB0dHfzhD38gJycHnU4nem2i0Sg6nQ6fz8fS0hKjo6NMT08zOztLKBSiubmZ/fv3I5PJWFlZIRAIJLz74rUNUi5fvszly5fXbPMQmjnV1tZSWFhIWloa8Xic+fl5ZmZm1uR3fBSxWIxQKCTGyGtqasjIyECn0920wt6197ywFgSDQRYXF+nq6qK/v58jR47Q19eHy+VKqMtarVaTk5NDdnY2BoOB/v5+Lly4wIULF1a5z4W0SY/Hg9VqFSsuJqJk7O0guNjr6+vZsWMHsVgMt9tNOBwmLy9PNKYT/bnfCv39/aysrNwQ+pubmxMzOrxeL0tLS7jdbhYXF/H5fKSlpYnVVIWusamIcHhMJinTj3JsbIyxsTGOHj0qFvkpLCzEbrfz+OOP43A4UKlUdHd3r5sgJy0tjcrKSrZv305fX99tP9xzc3N88MEHfP3rX0+JRaKurg6r1UpPTw/RaJRgMEhlZaXYxAP+vkBHIpGkCjivZ3p6Gr/fTygUEkNLW7ZsEV2DQpERp9PJn//8ZwYHBxkfHycYDPLtb3+bhx9+GIVCwfLyMoFAAK/Xm1CvjKBRUKvVdHR0rJlRK5PJMJlMVFVV8bWvfY3i4mJRJzA8PMyVK1cScg3PnDmDy+Vi9+7d1NTUoFKpbigFe7MCPZFIBJfLxcDAAK+//jrd3d20t7ev+3hvhtB/RDDcP/zwQ86ePUtnZ+eq9wkt2WdmZsjOzkar1X5sJ7xUQaFQkJWVxf3338/+/ftXFebZsWMHhYWFZGdns7i4mPSN6Xp6e3vp7e3lz3/+8y2932KxkJeXR2FhIVarlXg8zsDAQNKzolKZlDEGBEKhEENDQ0xPT7OyssKOHTt47rnneOqpp9i4caPY79zpdK75787KyuKBBx4gNzeXxcVFGhsb6e/vv+NbdwpEIhG8Xi+9vb1Eo1FycnJ49tlnqaqqEt+zuLjI9PQ0//u//8vQ0FASR3sjXq+XCxcucOnSJbHa2LUIccJAICC63z/3uc+JvdiTRSwW48iRI/h8Pnbu3MkjjzyC0Wjk17/+9W39XIvFgt1u59lnn6Wuro7m5uZVgkGhglkiELx7L7/8MjU1NZSWlrJv3z5MJpOol7FarWJ9BSHs9uqrr9LW1sb777+P1+tNqqtdiLu+8cYbGAwG+vr6blhnZDIZer2egoICampqUCgUmM1mMWslVYznm6HVarn//vspLy9HoVDwjW98Q0z9bmpqSvbw1gyFQsHOnTt54YUXqKioQKvVipUIU1UTIZDM++e2jQEh/9toNDI3N3fbD7Ogio5Go3R2dqJQKKiurqauro78/Hy2bt3KxMTEuhgDWq2W/Px8dDodOp2OrKys2+oXL5TQvda15fF4xLrkiba+p6enUalU+P1+9Ho9FouF8vLyVd3TBgYGOHXqFOPj4ynVshT+7pK+1RoNCoUCn88npn0lOixzLTMzM+I9W1VVRSAQ4OLFizidTnw+H263+2O/XyaTiUJUrVYrnuLsdjtbtmwRwwPwd4X/hx9+yKVLlxKywMTjcUKhEGNjY2Kp50gkgsFgQKfTYbPZaGhoEBssCXn5wsk7GRlD1xONRpmfn6ejowOVSiVeGwGhe2l9fT0FBQWoVCqWlpaSJnT8NGi1WjIzM9m2bRtqtVosDCe0947FYmIcPpnPye0gVPhsbGxk69at1NTUoNPpcDqdXL58mZmZmZSsrHgzBD1XWloaJpMJr9eb+l0LBXFWeXk5x48fx+12r9mgOzs7cblcAHznO98hPz+fxx57jFOnTq1LiVXB4oerG3lubu5N2xTfKpmZmTQ2Nq5qhTkxMcHIyAiLi4sJL6x0rbszMzOTrKwsqqqqVolVTp48yU9+8hNmZ2dTfoH7JGKxGH19fUxPTyfV4hbSNAVjYNOmTWRmZordFoeHh1e1+b3ZYqxSqWhoaCA3N5fs7GwOHDiAzWYjJydn1alUKGN67Ngx3njjDTo7OxM2d6H2fV9fH319fRw/fhy46n5vbW1Fq9WKGRUOh4P29nbee++9lInjBoNBsafDzVAoFKSlpXHgwAGqq6uJx+OMjY0xPz9/05oKqUR6ejolJSU88cQTdHR0cPbsWVwu1ypPklKpRK/X37HGgFqtJj09nRdeeIHNmzezYcMG4OoB55VXXqG3t/eO8AwIBxeLxUJWVhY2my0hrb1v2xgIhUKkpaWxbds2GhoaGB0d5be//S2RSGRNNhOPx0N7ezs+n09sJZqI+JwgJvq0qk6hEtTmzZvZtWsXTzzxxKrmRocPH+bkyZNJP02YTCYyMjJuePB9Ph8Oh+OONwTg6uK2a9cuqqurkxrTjcfjvP/++ywuLpKXl8f27dux2Ww888wzfOELXyAYDOJ0Ouns7MTj8bBnzx7UarU4ZiE1Mi0tTRQiWiwWVCqV+P9QKITT6eT999+ns7OTgwcP4nK5UkIZrtFo2Lt3L2VlZWK4xufzMTk5mRLjuxUUCgWtra3s3LmTJ598EovFQigU4mc/+xnt7e0J6yz3WTGbzeTm5mIymVhYWGBwcFAMqwlrtcFgID8/n46OjmQP91Mhk8kwGo18/vOfZ//+/ezbtw+z2UwsFuOtt97ib3/7mxiGSmXC4TB//etfaWlpoaGhAZ1Oh9FoFAWq681tGwPRaBSv18vU1BSlpaWUlZXR2tqK0+nE6/XidrvF8sOfZYMRVJZC+EBIv1hv9Ho91dXVYtGKWzm96PV68QLu3LmTxsbGVR32nE4n/f39DA8PJ3WzlcvlFBQUUFZWJhoDsViMubk5MR3vbkBIY1vvXPtbYXFxkeHhYY4ePYrRaKSkpGRVbn1hYSFarRav18t9990nCvDg78bAtfeMUBbb6/XicDhYWFhgenqaEydO0N/fn/BujB/HtU2whPstEomwtLSU8kanSqVCq9VSVlZGU1MTW7duJScnh8XFRSYmJujq6kq65+lWyMjIICcnRyygJIxZSLUNhUKiHiKVPANarVbUcF0fIhQaF1mtVux2O/fffz/33XcfFosFr9fL9PQ0p06d4vLlyynvEYCre53L5RKFzULxq0Rlnt22MRAIBOjr62N8fJyXXnqJpqYmXnzxRTo6OhgaGuLo0aO0t7czOTn5merxq9VqMjMzxf7UH3zwwbroBa4nOzubJ598EqfTicVi4eDBgx/7frlcjs1mIz8/n9LSUl5++WWxhjbA7OwsR48epa2tbV27yH0SQtyzubmZhx56SNxwwuEwx48fvy2NRCoipHGmAoODg/zkJz9hfn5eFDjB1XvHYDDQ1NS0alO52b+Fv5eWlpienqanp4ff//73DA0NMTo6itfrTTkleDwex+fzrdIThUIh3G53yo31eoRqd//4j/9IS0uLKLa9dOkSf/nLX+jp6bkjStwWFRVRUVFBMBhkZGSE7u5uVlZWcDgcYvvs5eVljEZjUsW212O1Wnn88cdpa2tjampKNHLlcjkmk4nW1lYeeOAB9u7di8ViwWAw4Pf7aW9v5w9/+AOvvfbaHSMAF9Jsk9VDYU1WSUHB/dvf/pbjx49z7NgxHn30UXbu3MnOnTsZGRlhbGyMd955R8xxF9xUNyM7O5uKigry8vKoqKjgwIEDlJWV4XQ61/WDmp2d5d1336W5uVmsMHbgwAFaWlrYv38/brcbt9tNW1ubmKfu9/vJyMggLy+PF198kdzcXNLT01fVap+YmODMmTP8+Mc/Znh4+LYrRd0OmZmZPPnkk+zdu5fa2lrR4lxeXqa/vz8hhlYyEeqUJ7PY05EjR+jq6uLs2bO0trZSVlZGeXm5GAJbWVkhGo2KG2U4HCYQCHDmzBlCoRCRSITDhw8zPz+Py+USe2QEg8GU3Fzlcjm5ubkYjUbxNZPJRElJSVKKq1xLc3MzL7/8MkeOHBHr3O/YsYOCggKxVK/JZCI/P1/sUTIyMsLbb7/Nm2++mRIFuW4FoWGaUJjH6XQSj8fFNTY7O5vx8XFmZmZS6h6yWq089thjNDU14XQ66erqEhvA7dy5E7vdTnZ2Nunp6Xg8HkZGRnj11Ve5fPkyly5duiMMNQFBhLu8vJwU78yaGAOCK39yclIsXVtVVSV2jzMajRQUFOD1esX0DqvV+pE3XW5uLtXV1eTl5VFWVsamTZtYWFhgZmaGmZmZdev453K5OHv2LJmZmRQVFZGTk0NBQQFZWVlkZmbidrtxuVyo1WrcbrfYfz07O5uCggK2bNkiFlyBq67cQCBAb28v3d3d9PX1EQ6Hk+oa1el01NTUYLPZxKIwPp+Pubk5+vv7RcHm3Yrf72dycjKpXTHn5uZYWlrC5/Oh1WqZnZ1lfn4ei8WCTqcjHA7j9/tFo1H4/+nTp0Vj4MyZMywuLibVsPw0XN9bQq1WY7FYkn4KtVqt7N69m1AoJBoDO3fuFPPTTSYTGo2GlZUVJiYmmJ6epr29nZ6enpQKxXwSSqVS7I8ihDLdbjc2m426ujo0Gg2RSEQsRpYqyOVyVCoVVVVVRKNR8ZBmNBppaWkRjcnp6WnGxsbEtsaCYXMnEYvFktoUbs39pz6fj46ODn70ox+xYcMGXnrpJe6//342bdrEpk2biMViYv3xj0KhUIgKacFCOn78OKdPn+bQoUPrlos8PDzMf/3XfzE6OkpLSwsvvviiWNRGUKYCPPzww+KiHIlExGp4wlgFz4Xb7aa7u5v/+Z//obu7OyVOERqNhoqKilWntN7eXs6dO8fBgwdTrrXyWuP1eunv70/qJirEagcHB1c1rKqqqsJqtYqG7yelG94pCBkO197/Op2O9PT0pBfrEao3fuUrX7np1wWtRiAQoK2tjRMnTvDee+/dkR40pVKJ3W5n27ZtYhnppqYmnn/+eRQKhdilNJU0Qy6Xi3feeYfnn3+e8vJympubV319cnKSsbEx3nrrLc6fP09nZyc+ny/lNRw3IxwOc+zYsRvChYli3YKpbrebK1eu8POf/5zDhw+Tn5+P3W6ntLRUjKt/1EIgGAFCWlhfXx9/+ctfGBwcXLMshZshiCGPHTvG1NQU0Wh0lcvwWoQWp4JrVzAEBHXumTNnuHLlCm1tbbS1taXEiTs9PZ3c3Fxyc3NXpRQ5HI6kixoThdFopLy8nGPHjiV7KDcwOTmJw+EQxYF3C+FwmKNHj2K1WqmqqmJkZIRz585x5MiRpLtxA4EAMzMz6PV6Mc8+FosRDAYZHR1FoVDg9/t55513uHDhAr29vTgcjjumD4HAwMAAWVlZPPzww2zfvp28vDx27txJQ0MDhYWFYjGvmZmZlDIGHA4Hb731FrOzs5SUlFBdXS22956bm2N4eFhszrWwsJAy5dM/C4K2ZmpqisHBQUpKShL6+9e1hXE4HMbj8TAzMyPWIpibm6O0tJRIJPKJLsKOjg5GR0e5dOkS7e3tOJ3Odd2whDK8QvMbk8mEXq/H5/NRWVm56r1CYSKlUil2WRNiux6Ph7Nnz3LlyhUuXLjAzMxMSrhzBSGd0H9dMAj6+voYHR29J4wBhULxqVuEJoo7pSDKp2V5eVkMldntdnp6emhvb+fSpUtJfy48Hg8dHR2kp6eL7ZVXVlbwer10dXWhVCrxer188MEHDA0NpUxNhE/LzMwM/f39TE5OYjKZaGhooLi4GIPBwMrKCp2dnQwPDyfdOLueYDBIf38/kUiEgYEBMUtNSE0dGxtL2bbLnwWh7Xdvb69Y2CpRok5Z/BbNqLVYPK91+9/Kz7tWQf1prL2Pe++nmYdMJhN1A0888cSqr+3YsYO6ujrMZjMTExOMjo5y/vx5BgcH6enp4fz587dtpX7S936WayKTyW64sYTPdz0t6rW6Jp8GtVrN888/z0MPPcTjjz8OwKlTp/jFL37BX/7yFxYWFj71z1yPa5IsEn1Nrm/ms1bG5+1cE+F5yM3NxWAwoNfrxf4VY2Njn3kN+qys5zVJS0ujtbWVZ599lj179mAymbh8+TLnz5/nBz/4AfPz82sWylyvtevT/p61INHPSX5+PhUVFbz22muMjY1x+PBhfvKTn9x22PCTPquE5lwl6oFaK4RYZzQa5fe///2qC3/ixAkx5dHv94viSK/XKwq7UnGuqdAdK1EI6Z7XlluGO+8+vFtIRc+T8DwIgmClUikW47nbnpNQKER7ezsej4eDBw+KQuj5+XmcTmfKhz7ulWd2YWGBnp4e/vVf/xW/3y82aVtvUiMBO4URCl2kQsxf4tMhVO0TSkqHw2HC4XDKL3oSiScVxL3rzfLyMuPj44yPjyd7KBIfQyAQEFP1E0nqVJeQkFhjYrEYExMTYr+Mixcv0t7ezsjISNJj1RISEhKphOQZkLhrWV5e5sSJE4yPj3P06FGmpqaYnJwUM0UkJCQkJK6SUAFhokiGWG09kMRqqYd0TVIP6ZqkHtI1ST0+8ZrcqjEgISEhISEhcXciaQYkJCQkJCTucSRjQEJCQkJC4h5HMgYkJCQkJCTucSRjQEJCQkJC4h5HMgYkJCQkJCTucSRjQEJCQkJC4h5HMgYkJCQkJCTucSRjQEJCQkJC4h5HMgYkJCQkJCTucf4/1EmA5Vgatm8AAAAASUVORK5CYII="},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"<Figure size 640x480 with 10 Axes>","image/png":"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"},"metadata":{}}]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Since the latent dimension of the model is 50 , obiviously can't plot the latent space. But the reconstruction of the images is much better due to increased size of latent dimension to 50 as compared to previous example where latent dimension was of size 2.","metadata":{}},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]}]}