Gabriel_Sansigolo_Trabalho_de_DL_4.ipynb

{"cells":[{"cell_type":"markdown","metadata":{"id":"p0x3A8_iBMP1"},"source":["# **Atividade -Exercício sobre o material Transfer Learning**"]},{"cell_type":"markdown","metadata":{"id":"n3dmsuRlBMP3"},"source":["## **Disciplina:** CAP-421-3 Aprendizado Profundo\n","\n","**Discintes:**\n","\n","Gabriel Sansigolo"]},{"cell_type":"markdown","metadata":{"id":"OegWcnpsLCKM"},"source":["Create 2 different classifications, now exploring the EuroSAT database with Red, Green and Near-Infrared bands (RGNIR):\n","* instantiate a well-known architecture (e.g. ResNet50) with random weights, train with EuroSAT (RGNIR)\n","* instantiate the same well-known architecture, but pretrained with ImageNet, and fine-tune with EuroSAT (RGNIR)\n","\n","Show plots of accuracy and loss along the epochs, for both cases."]},{"cell_type":"markdown","metadata":{"id":"7FsJyUhsBMP4"},"source":["### **1 - Importando bibliotecas**"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"fmY-qk1ZBMP4"},"outputs":[],"source":["import os\n","import torch\n","from tqdm  import tqdm\n","import rasterio\n","from pathlib import Path\n","import torch.nn as nn\n","from torchvision import datasets, models, transforms\n","from PIL import Image\n","import matplotlib.pyplot as plt\n","import torch.optim as optim"]},{"cell_type":"markdown","metadata":{"id":"dHk8VZHmBMP5"},"source":["### **2 - RGNIR**"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"5y-j-7ZnBMP5"},"outputs":[],"source":["folders = [\"AnnualCrop\",\"Forest\",\"HerbaceousVegetation\",\"Highway\",\"Industrial\",\"Pasture\",\"PermanentCrop\",\"Residential\",\"River\",\"SeaLake\"]\n","\n","os.mkdir('RGNIR/')\n","os.mkdir('RGNIR/train/')\n","os.mkdir('RGNIR/validation/')\n","\n","for folder in folders:\n","    os.mkdir('RGNIR/train/'+folder)\n","    os.mkdir('RGNIR/validation/'+folder)"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"n6LkifhaBMP6"},"outputs":[],"source":["path_dir = os.path.dirname(\"\")\n","\n","def create_rgnir(path):\n","\n","    filenames = os.listdir(path)\n","    filename_len = len(filenames)\n","\n","    for i, filename in enumerate(tqdm(filenames)):\n","\n","        if i < int(filename_len*0.8):\n","            mode = 'train'\n","        else:\n","            mode = 'validation'\n","\n","        if filename == '_DS_Store':\n","            continue\n","\n","        image_src = rasterio.open(os.path.join(path, filename))\n","        new_image = image_src.read((3,4,8))\n","\n","        profile = image_src.profile\n","        profile.update(count=3, driver=\"PNG\")\n","\n","        with rasterio.open(\n","            os.path.join('RGNIR/',mode, path.split('/')[1], Path(filename).stem+\".png\"),\n","            mode=\"w\",\n","            **profile,\n","        ) as new_dataset:\n","            new_dataset.write(new_image)"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"U4tuz6rkBMP6","outputId":"bf903671-dc26-4935-e479-ab35556330db"},"outputs":[{"name":"stderr","output_type":"stream","text":["100%|██████████| 3000/3000 [01:31<00:00, 32.91it/s]\n","100%|██████████| 2500/2500 [01:14<00:00, 33.44it/s]\n","100%|██████████| 2000/2000 [00:48<00:00, 41.60it/s]\n","100%|██████████| 2500/2500 [01:21<00:00, 30.79it/s]\n","100%|██████████| 3000/3000 [01:39<00:00, 30.15it/s]\n","100%|██████████| 2500/2500 [01:13<00:00, 34.02it/s]\n","100%|██████████| 2500/2500 [01:22<00:00, 30.31it/s]\n","100%|██████████| 3597/3597 [02:01<00:00, 29.70it/s]\n","100%|██████████| 3000/3000 [01:37<00:00, 30.73it/s]\n","100%|██████████| 3000/3000 [01:36<00:00, 30.97it/s]\n"]}],"source":["base_path = os.path.dirname(\"\")\n","\n","path = os.path.join(base_path,'EuroSATallBands/')\n","\n","annual_crop = create_rgnir(os.path.join(path,\"AnnualCrop\"))\n","highway = create_rgnir(os.path.join(path,\"Highway\"))\n","pasture = create_rgnir(os.path.join(path,\"Pasture\"))\n","river = create_rgnir(os.path.join(path,\"River\"))\n","forest = create_rgnir(os.path.join(path,\"Forest\"))\n","industrial = create_rgnir(os.path.join(path,\"Industrial\"))\n","permanent_crop = create_rgnir(os.path.join(path,\"PermanentCrop\"))\n","sea_lake = create_rgnir(os.path.join(path,\"SeaLake\"))\n","herbaceous_vegetation = create_rgnir(os.path.join(path,\"HerbaceousVegetation\"))\n","residential = create_rgnir(os.path.join(path,\"Residential\"))"]},{"cell_type":"markdown","metadata":{"id":"6GCF-aFxBMP7"},"source":["### **3 - Função trasformers**"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"l7WTp7sfBMP7"},"outputs":[],"source":["normalize = transforms.Normalize(mean=[0.485, 0.456, 0.406],\n","                                 std=[0.229, 0.224, 0.225])\n","\n","data_transforms = {\n","    'train':\n","    transforms.Compose([\n","        transforms.Resize((64,64)),\n","        transforms.ToTensor(),\n","        normalize\n","    ]),\n","    'validation':\n","    transforms.Compose([\n","        transforms.Resize((64,64)),\n","        transforms.ToTensor(),\n","        normalize\n","    ]),\n","}"]},{"cell_type":"markdown","metadata":{"id":"-e6iMzHnBMP9"},"source":["### **4 - Datasets e dataloaders**"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"W8xMhoYtBMP9"},"outputs":[],"source":["input_path = \"RGNIR/\""]},{"cell_type":"code","execution_count":null,"metadata":{"id":"iYxXlBoiBMP9"},"outputs":[],"source":["def check_Image(path):\n","  try:\n","    im = Image.open(path)\n","    return True\n","  except:\n","    return False\n","\n","image_datasets = {\n","    'train':\n","    datasets.ImageFolder(input_path + 'train', data_transforms['train'], is_valid_file=check_Image),\n","    'validation':\n","    datasets.ImageFolder(input_path + 'validation', data_transforms['validation'], is_valid_file=check_Image)\n","}\n","\n","dataloaders = {\n","    'train':\n","    torch.utils.data.DataLoader(image_datasets['train'],\n","                                batch_size=32,\n","                                shuffle=True,\n","                                num_workers=0),\n","    'validation':\n","    torch.utils.data.DataLoader(image_datasets['validation'],\n","                                batch_size=32,\n","                                shuffle=False,\n","                                num_workers=0)\n","}"]},{"cell_type":"markdown","metadata":{"id":"IdMsEjyiBMP9"},"source":["### **4 - Utiliza GPU (cuda), caso contrário, utiliza CPU**"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"xwCBsqgxBMP9"},"outputs":[],"source":["device = 'cpu'"]},{"cell_type":"markdown","metadata":{"id":"6tWvCALuBMP9"},"source":["### **5 - resnet50 - normal**"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"gNOnc3BYBMP9","outputId":"684a8ef2-d7e7-4213-8fa5-d0e13d93982b"},"outputs":[{"name":"stderr","output_type":"stream","text":["c:\\Users\\Sansigolo\\AppData\\Local\\Programs\\Python\\Python312\\Lib\\site-packages\\torchvision\\models\\_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.\n","  warnings.warn(\n","c:\\Users\\Sansigolo\\AppData\\Local\\Programs\\Python\\Python312\\Lib\\site-packages\\torchvision\\models\\_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=None`.\n","  warnings.warn(msg)\n"]}],"source":["model = models.resnet50(pretrained=False).to(device)\n","\n","for param in model.parameters():\n","    param.requires_grad = False\n","\n","model.fc = nn.Sequential(\n","               nn.Linear(2048, 128),\n","               nn.ReLU(inplace=True),\n","               nn.Linear(128, 10)).to(device)"]},{"cell_type":"markdown","metadata":{"id":"GxTV78ejBMP9"},"source":["### **5 - Criterion e optimizer**"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"RXoXGXlDBMP-"},"outputs":[],"source":["criterion = nn.CrossEntropyLoss()\n","optimizer = optim.Adam(model.fc.parameters())"]},{"cell_type":"markdown","metadata":{"id":"pi8lgYCoBMP-"},"source":["### **6 - Treinamento**"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"7AqmB9QSBMP-"},"outputs":[],"source":["train_loss = []\n","train_acc = []\n","val_loss = []\n","val_acc = []\n","\n","def train_model(model, criterion, optimizer, num_epochs=3):\n","    for epoch in range(num_epochs):\n","        print('Epoch {}/{}'.format(epoch+1, num_epochs))\n","        print('-' * 10)\n","\n","        for phase in ['train', 'validation']:\n","            if phase == 'train':\n","                model.train()\n","            else:\n","                model.eval()\n","\n","            running_loss = 0.0\n","            running_corrects = 0\n","\n","            for inputs, labels in dataloaders[phase]:\n","                inputs = inputs.to(device)\n","                labels = labels.to(device)\n","\n","                outputs = model(inputs)\n","                loss = criterion(outputs, labels)\n","\n","                if phase == 'train':\n","                    optimizer.zero_grad()\n","                    loss.backward()\n","                    optimizer.step()\n","\n","                _, preds = torch.max(outputs, 1)\n","                running_loss += loss.item() * inputs.size(0)\n","                running_corrects += torch.sum(preds == labels.data)\n","\n","            epoch_loss = running_loss / len(image_datasets[phase])\n","            epoch_acc = running_corrects.double() / len(image_datasets[phase])\n","\n","            if phase == 'train':\n","                train_loss.append(epoch_loss)\n","                train_acc.append(epoch_acc)\n","            else:\n","                val_loss.append(epoch_loss)\n","                val_acc.append(epoch_acc)\n","\n","            print('{} loss: {:.4f}, acc: {:.4f}'.format(phase,\n","                                                        epoch_loss,\n","                                                        epoch_acc))\n","    return model"]},{"cell_type":"markdown","metadata":{"id":"UbP0ZEbyBMP-"},"source":["### **7 - Execução do treinamento - normal**"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"IXF9Eg62BMP_","outputId":"ec1fc2f9-ce2e-48eb-b0c5-58db120271f4"},"outputs":[{"name":"stdout","output_type":"stream","text":["Epoch 1/20\n","----------\n","train loss: 1.6290, acc: 0.4162\n","validation loss: 1.3053, acc: 0.5368\n","Epoch 2/20\n","----------\n","train loss: 1.3534, acc: 0.5116\n","validation loss: 1.2777, acc: 0.5408\n","Epoch 3/20\n","----------\n","train loss: 1.2700, acc: 0.5417\n","validation loss: 1.1788, acc: 0.5696\n","Epoch 4/20\n","----------\n","train loss: 1.2476, acc: 0.5495\n","validation loss: 1.2042, acc: 0.5560\n","Epoch 5/20\n","----------\n","train loss: 1.2190, acc: 0.5574\n","validation loss: 1.1404, acc: 0.5871\n","Epoch 6/20\n","----------\n","train loss: 1.1975, acc: 0.5706\n","validation loss: 1.2693, acc: 0.5444\n","Epoch 7/20\n","----------\n","train loss: 1.1696, acc: 0.5781\n","validation loss: 1.1675, acc: 0.5832\n","Epoch 8/20\n","----------\n","train loss: 1.1635, acc: 0.5819\n","validation loss: 1.1360, acc: 0.5933\n","Epoch 9/20\n","----------\n","train loss: 1.1535, acc: 0.5844\n","validation loss: 1.1928, acc: 0.5761\n","Epoch 10/20\n","----------\n","train loss: 1.1303, acc: 0.5949\n","validation loss: 1.0734, acc: 0.6161\n","Epoch 11/20\n","----------\n","train loss: 1.1325, acc: 0.5898\n","validation loss: 1.0625, acc: 0.6123\n","Epoch 12/20\n","----------\n","train loss: 1.1196, acc: 0.6003\n","validation loss: 1.0459, acc: 0.6333\n","Epoch 13/20\n","----------\n","train loss: 1.1209, acc: 0.5965\n","validation loss: 1.0493, acc: 0.6263\n","Epoch 14/20\n","----------\n","train loss: 1.1202, acc: 0.5988\n","validation loss: 1.1047, acc: 0.6009\n","Epoch 15/20\n","----------\n","train loss: 1.0922, acc: 0.6091\n","validation loss: 1.1157, acc: 0.6067\n","Epoch 16/20\n","----------\n","train loss: 1.1065, acc: 0.6018\n","validation loss: 1.0592, acc: 0.6159\n","Epoch 17/20\n","----------\n","train loss: 1.0965, acc: 0.6064\n","validation loss: 1.0510, acc: 0.6208\n","Epoch 18/20\n","----------\n","train loss: 1.0847, acc: 0.6131\n","validation loss: 1.0269, acc: 0.6332\n","Epoch 19/20\n","----------\n","train loss: 1.0767, acc: 0.6119\n","validation loss: 1.0697, acc: 0.6083\n","Epoch 20/20\n","----------\n","train loss: 1.0752, acc: 0.6154\n","validation loss: 1.0884, acc: 0.6179\n"]}],"source":["model_trained = train_model(model, criterion, optimizer, num_epochs=20)"]},{"cell_type":"markdown","metadata":{"id":"bXqIO9NABMP_"},"source":["### **8 - Visualização das curvas de treinamento e teste - normal**"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"naZ7mfkiBMP_"},"outputs":[],"source":["num_epochs = 20"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"WsGFy8IcBMP_","outputId":"83343abe-6875-4b4c-caf7-96bc078cb764"},"outputs":[{"data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAmIAAAHJCAYAAAAmZVVWAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy80BEi2AAAACXBIWXMAAA9hAAAPYQGoP6dpAABkHklEQVR4nO3dd3wT9f8H8NclbdOd7gWlpbRAKVALSBkigoVSoIKoIPD7Ag4ERUURB/qVpYKyvjgQRBlOZAiCArKHDEFGoWxKJ9Bdundyvz/ShoYO2pL20vT1fDzukeRW3pdruBefu3xOEEVRBBERERE1OpnUBRARERE1VwxiRERERBJhECMiIiKSCIMYERERkUQYxIiIiIgkwiBGREREJBEGMSIiIiKJMIgRERERSYRBjIiIiEgiDGJE1CDWrl0LQRAQGxsrdSm19thjj+Gxxx6TugwiakYYxIga2Y0bNzBp0iT4+PjA3Nwctra26N27Nz7//HMUFBRIXZ5B8vb2hiAI9x3Wrl0rdalN0o4dOzB79mypyyBqlkykLoCoOdm+fTueeeYZKBQKjBs3Dh07dkRxcTGOHDmCt99+GxcvXsTKlSulLtPgLF26FLm5udrXO3bswLp16/C///0PTk5O2vG9evV6oPfZvXv3Ay3fVO3YsQPLli1jGCOSAIMYUSOJiYnBs88+Cy8vL+zfvx/u7u7aaVOmTEFUVBS2b9+ul/fKy8uDlZWVXtZlCIYPH67zOikpCevWrcPw4cPh7e1d7XJ1/RzMzMzqWSERUf3w1CRRI1mwYAFyc3OxatUqnRBWztfXF1OnTgUAxMbGVnuqTRAEnZaL2bNnQxAEXLp0CWPGjIG9vT0eeeQRLFq0CIIgIC4urtI6ZsyYATMzM9y5cwcA8Pfff+OZZ55Bq1atoFAo4OnpiTfffLPWp0ovXryI/v37w8LCAi1btsTHH38MtVpd5bw7d+5Enz59YGVlBRsbGwwZMgQXL16s1fvUZMKECbC2tsaNGzcwePBg2NjYYOzYsQAAtVqNpUuXIiAgAObm5nB1dcWkSZO021/u3mvEDh48CEEQsGHDBnzyySdo2bIlzM3N8fjjjyMqKkpn2dp+huV1xsfHY+jQobC2tkaLFi2wbNkyAEBkZCT69+8PKysreHl54Zdffqm0rZmZmXjjjTfg6ekJhUIBX19ffPbZZzqfefnf0KJFi7By5Uq0adMGCoUCDz/8MP7991+desrfu+Jp3nJ5eXl46623tO/Vrl07LFq0CKIo1mX3EFE12CJG1Ej++OMP+Pj4PPDps+o888wz8PPzw7x58yCKIoYOHYp33nkHGzZswNtvv60z74YNGzBw4EDY29sDADZu3Ij8/Hy8/PLLcHR0xMmTJ/Hll1/i5s2b2LhxY43vm5SUhH79+qG0tBTvvfcerKyssHLlSlhYWFSa98cff8T48eMRGhqKzz77DPn5+Vi+fDkeeeQRnD17tsbWrdooLS1FaGioNohaWloCACZNmoS1a9fiueeew+uvv46YmBh89dVXOHv2LI4ePQpTU9Ma1/vpp59CJpNh+vTpyMrKwoIFCzB27FicOHFCO09dPkOVSoWwsDA8+uijWLBgAX7++We8+uqrsLKywgcffICxY8dixIgRWLFiBcaNG4eePXuidevWAID8/Hz07dsXt27dwqRJk9CqVSscO3YMM2bMQGJiIpYuXarzXr/88gtycnIwadIkCIKABQsWYMSIEYiOjoapqSkmTZqE27dvY8+ePfjxxx91lhVFEU888QQOHDiAF154AQ899BB27dqFt99+G7du3cL//ve/+u4qIionElGDy8rKEgGIw4YNq9X8MTExIgBxzZo1laYBEGfNmqV9PWvWLBGAOHr06Erz9uzZU+zatavOuJMnT4oAxB9++EE7Lj8/v9Ky8+fPFwVBEOPi4mqs9Y033hABiCdOnNCOS0lJEZVKpQhAjImJEUVRFHNyckQ7Oztx4sSJOssnJSWJSqWy0viaLFy4UGfdoiiK48ePFwGI7733ns68f//9twhA/Pnnn3XG//XXX5XG9+3bV+zbt6/29YEDB0QAor+/v1hUVKQd//nnn4sAxMjISO242n6G5XXOmzdPO+7OnTuihYWFKAiC+Ouvv2rHX7lypdL+/uijj0QrKyvx2rVrOu/13nvviXK5XIyPjxdF8e7fkKOjo5iRkaGdb+vWrSIA8Y8//tCOmzJliljV4eD3338XAYgff/yxzvinn35aFARBjIqKqrQMEdUNT00SNYLs7GwAgI2NTYO9x+TJkyuNGzVqFE6fPo0bN25ox61fvx4KhQLDhg3TjqvYepWXl4e0tDT06tULoiji7NmzNb7vjh070KNHD3Tv3l07ztnZWXtasNyePXuQmZmJ0aNHIy0tTTvI5XIEBwfjwIEDdd7mqrz88ss6rzdu3AilUokBAwbovG/Xrl1hbW1dq/d97rnndK4f69OnDwAgOjpaO66un+GLL76ofW5nZ4d27drBysoKI0eO1I5v164d7OzsdN5n48aN6NOnD+zt7XW2JyQkBCqVCocPH9Z5n1GjRmlbPqurvTo7duyAXC7H66+/rjP+rbfegiiK2Llz533XQUQ146lJokZga2sLAMjJyWmw9yg/dVXRM888g2nTpmH9+vV4//33IYoiNm7ciLCwMG1NABAfH4+ZM2di27Ztla6bysrKqvF94+LiEBwcXGl8u3btdF5fv34dANC/f/8q11OxnvoyMTFBy5YtK71vVlYWXFxcqlwmJSXlvutt1aqVzuvyYFPxs6rLZ2hubg5nZ2edcUqlEi1bttS5Pqt8fMX1Xb9+HefPn6+0fHXbU5vaqxMXFwcPD49K/4Hw9/fXTieiB8MgRtQIbG1t4eHhgQsXLtRq/nsPxuVUKlW1y1R1TZaHhwf69OmDDRs24P3338c///yD+Ph4fPbZZzrrHDBgADIyMvDuu++iffv2sLKywq1btzBhwoRqL7qvq/L1/Pjjj3Bzc6s03cTkwf85UigUkMl0G/rVajVcXFzw888/V7lMdYGmIrlcXuV4seyC9bp+htWt737vU749AwYMwDvvvFPlvG3btq3zOolIOgxiRI1k6NChWLlyJY4fP46ePXvWOG95q0VmZqbO+Pq0QIwaNQqvvPIKrl69ivXr18PS0hLh4eHa6ZGRkbh27Rq+//57jBs3Tjt+z549tVq/l5eXtrWroqtXr+q8btOmDQDAxcUFISEhdd6O+mrTpg327t2L3r17VxlW9eFBP8O6aNOmDXJzc/X6GVYX/L28vLB3717k5OTotIpduXJFO510iaKI0tLSGv/TRMZNLpfDxMSk2u/VvRjEiBrJO++8g59//hkvvvgi9u/fD1dXV53pN27cwJ9//ompU6fC1tYWTk5OOHz4MN544w3tPF9//XWd3/epp57Ca6+9hnXr1mHjxo0YOnSoTt9a5S0mFVtIRFHE559/Xqv1Dx48GEuXLsXJkye114mlpqZWaoEKDQ2Fra0t5s2bh379+lX6pWJqamqtWqfqauTIkfj666/x0UcfYd68eTrTSktLkZubCzs7uwd6jwf9DOti5MiRmD17Nnbt2oXQ0FCdaZmZmbC2tq5z62L530NmZqbOZzF48GCsXLkSX331FWbMmKEd/7///Q+CICAsLKz+G2KEiouLkZiYiPz8fKlLIYlZWlrC3d29Vn0TMogRNZI2bdrgl19+wahRo+Dv76/Ts/6xY8ewceNGTJgwQTv/iy++iE8//RQvvvgiunXrhsOHD+PatWt1fl8XFxf069cPS5YsQU5ODkaNGqUzvX379mjTpg2mT5+OW7duwdbWFr/99lutriECNAHzxx9/xKBBgzB16lRt9xVeXl44f/68dj5bW1ssX74c//nPf9ClSxc8++yzcHZ2Rnx8PLZv347evXvjq6++qvP23U/fvn0xadIkzJ8/HxERERg4cCBMTU1x/fp1bNy4EZ9//jmefvrpB3qPB/0M6+Ltt9/Gtm3bMHToUEyYMAFdu3ZFXl4eIiMjsWnTJsTGxurcbaA2unbtCgB4/fXXERoaCrlcjmeffRbh4eHo168fPvjgA8TGxiIwMBC7d+/G1q1b8cYbb2hbOUlzyjgmJgZyuRweHh4wMzOrdYsIGQ9RFFFcXIzU1FTExMTAz8+v0uUSVS1ERI3o2rVr4sSJE0Vvb2/RzMxMtLGxEXv37i1++eWXYmFhoXa+/Px88YUXXhCVSqVoY2Mjjhw5UkxJSam2+4rU1NRq3/Pbb78VAYg2NjZiQUFBpemXLl0SQ0JCRGtra9HJyUmcOHGieO7cuWq70LjX+fPnxb59+4rm5uZiixYtxI8++khctWpVpS4mRFHTJURoaKioVCpFc3NzsU2bNuKECRPEU6dO3fd9ylXXfYWVlVW1y6xcuVLs2rWraGFhIdrY2IidOnUS33nnHfH27dvaearrvmLjxo0666qqe5HafobV1dm3b18xICCg0ngvLy9xyJAhOuNycnLEGTNmiL6+vqKZmZno5OQk9urVS1y0aJFYXFysU+PChQsrrfPev6HS0lLxtddeE52dnUVBEHS6ssjJyRHffPNN0cPDQzQ1NRX9/PzEhQsXimq1utJ6m7OCggLx0qVLYl5entSlkAHIy8sTL126VOW/t/cSRJFXbBIRET2IwsJCxMTEoHXr1jA3N5e6HJJYXf4e2I8YERERkUQYxIiIiEgvvL29K91mq77K7/V676/HjQ0v1iciImrGHnvsMTz00EN6CVD//vuvzq+y6f4YxIiIiKhaoihCpVLVqluUhuiCxtjx1CQREVEzNWHCBBw6dAiff/45BEGAIAhYu3YtBEHAzp070bVrVygUChw5cgQ3btzAsGHD4OrqCmtrazz88MPYu3evzvruPTUpCAK+++47PPnkk7C0tISfnx+2bdtW73p/++03BAQEQKFQwNvbG4sXL9aZ/vXXX8PPzw/m5uZwdXXV6Zpm06ZN6NSpEywsLODo6IiQkBDk5eVpp3/33Xfw9/eHubk52rdvr9NvY3FxMV599VW4u7vD3NwcXl5emD9/fr23oyK2iBERETUAURRRUCJND/sWpvJa9WP2+eef49q1a+jYsSPmzp0LALh48SIA4L333sOiRYvg4+MDe3t7JCQkYPDgwfjkk0+gUCjwww8/IDw8HFevXq10T9OK5syZgwULFmDhwoX48ssvMXbsWMTFxcHBwaFO23T69Glth8ajRo3CsWPH8Morr8DR0RETJkzAqVOn8Prrr+PHH39Er169kJGRgb///hsAkJiYiNGjR2PBggV48sknkZOTg7///lvbCfPPP/+MmTNn4quvvkJQUBDOnj2LiRMnwsrKCuPHj8cXX3yBbdu2YcOGDWjVqhUSEhKQkJBQp/qr0+yCmFqtxu3bt2FjY8PO9oiISC+Ki4uhVquhUqm0tzfKLy5Fpzl777Nkw4icFQJLs/sf4q2trWFqagoLC4tKpxVnz56N/v37a18rlUp07NhRZ/qWLVvw+++/Y8qUKdrx5Z9DuXHjxmHkyJEAgI8++ghffPEFjh8/jkGDBtVYW/k6yj/TRYsWoV+/fvjvf/8LQRDQtm1bXLp0CQsXLsSECRMQHx8PKysrDB06FDY2NvDy8kJQUBAATRArLS3FiBEjtLfm6tSpk/a9Zs2ahcWLF2PEiBEAgNatW+PSpUv45ptvMH78eMTHx8PPzw+PPPIIBEHQ6+29ml0Qu337Njw9PaUug4iIjIiXlxdWrFiBgoIC7bjCUum66Tx37jzMTWrX2JCbm4uUlBScPXsWABAVFQUAUCgU2nEAkJ+fj5UrV+Lo0aNIS0uDSqVCUVERTp06pZ2vuLgYN2/e1FnOxsZG57WVlRVOnTpV6TZv9yqv4/z589p19O3bFyUlJdpbB/Xu3RtLly6FSqXCgAED4OXlBR8fHwwaNAiDBg3SnhINDAzE448/jk6dOiE0NBQDBw7E008/DXt7e+Tl5eHGjRt44YUXMHHiRO37l5aWQqlUAtCcwh0wYADatWuHQYMGYejQoRg4cGCtPt/7aXZBrPzGtQkJCbC1tZW4GiIiMgbFxcVITk6Gt7e3tgNPURQRGWjYpyYBTauYi4uLtvUoKysLABAcHKxz79FXXnkFx48fx6JFi9CmTRtYWFhg1KhRsLe31y5rZmaGli1bal8DQNu2bXVem5iYwNPTU2dcVcrr6Ny5M+zs7GBhYQEA1d4yyMbGBmfOnMHBgwexe/duzJw5E7Nnz8a///4LOzs77NmzB8eOHcPu3bvx5Zdf4oMPPsCJEydgaWkJAPj2228RHByss87y+8h26dIFMTEx2LlzJ/bu3YuRI0ciJCQEmzZtqvnDrYVmF8TK/zBtbW0ZxIiISC8KCwuRmpoKuVyuPXgDgE0db8AuBYVCAbVara274mPFbTl27BgmTJiAp556CoCmJS02NhaPPfaYznwymazG19WNu9e9dfj7++PcuXM6AfPo0aNo27atdl4TExOEhIQgJCQEs2bNgp2dHfbv348RI0ZAEAT07t0bvXv3xsyZM+Hl5YUtW7Zg2rRp8PDwQHR0NMaOHVttPba2thg1ahRGjRqFp59+GoMGDUJGRkadr3W7l+H/hRAREVGD8fb2xokTJxAbGwtra2uo1eoq5/Pz88PmzZsRHh4OQRDw4YcfVjtvQ3jzzTfRo0cPfPzxxxg9ejSOHz+Or776Svvrxj///BPR0dF49NFHYW9vjx07dkCtVqNdu3Y4ceIE9u3bh4EDB8LFxQUnTpxAamoq/P39AWh+UPD6669DqVRi0KBB2lOud+7cwbRp07BkyRK4u7sjKCgIMpkMGzduhJubm06LYX2x+woiIqJmbPr06ZDL5ejQoQOcnZ0RHx9f5XxLliyBvb09evXqhfDwcISGhqJLly6NVmeXLl0wf/58bNiwAR07dsTMmTMxd+5cTJgwAQBgZ2eHzZs3o3///vD398eKFSuwbt06BAQEwNbWFocPH8bgwYPRtm1b/Pe//8XixYsRFhYGAHjxxRfx3XffYc2aNejUqRP69u2LtWvXonXr1gA0pz0XLFiAbt264eGHH0ZsbCx27NhR7WnSumh2N/3Ozs6GUqlEVlYWT00SEZFe8KbfDU+lUuHs2bMICgq672lNqfGm30RERERNAIMYERERNbrJkyfD2tq6ymHy5MlSl9doeLE+ERERNbq5c+di+vTpVU5rTpcOMYgRERFRo3NxcYGLi4vUZUiOpyaJiIiIJMIgRkRERCQRBjEiIiIiiTCIEREREUmEQYyIiIhIIgxiepRdWILzNzOlLoOIiIiaCAYxPTl/MxNdP9qDF74/BbW6Wd01ioiICIDmBuJLly6Vuowmhf2I6Ul7N1uYm8iRmlOEiJuZ6NLKXuqSiIiI7uuxxx7DQw89pJcA9e+//8LKyurBi2pG2CKmJ2YmMvRt5wwA2HMpWeJqiIiI9EMURZSWltZqXmdnZ1haWjZwRcaFQUyPBga4AQB2X0ySuBIiIpKcKALFedIMYu0ukZkwYQIOHTqEzz//HIIgQBAErF27FoIgYOfOnejatSsUCgWOHDmCGzduYNiwYXB1dYW1tTUefvhh7N27V2d9956aFAQB3333HZ588klYWlrCz88P27Ztq1VtKpUKL7zwAlq3bg0LCwt06NAB69atqzTf6tWrERAQAIVCAXd3d7z66qvaaZmZmZg0aRJcXV1hbm6Ojh074s8//wQAxMXFITw8HPb29rCyskJAQAB27NhRq9r0iacm9eixds4wlQu4kZqHG6m5aONsLXVJREQklZJ8YJ6HNO/9/m3A7P6nCD///HNcu3YNHTt2xNy5cwEAFy9eBAC89957WLRoEXx8fGBvb4+EhAQMHjwYn3zyCRQKBX744QeEh4fj6tWraNWqVbXvMWfOHCxYsAALFy7El19+ibFjxyIuLg4ODg411qZWq9GyZUts3LgRjo6OOHLkCF566SU8/PDDePbZZwEAy5cvx7Rp0/Dpp58iLCwMWVlZOHr0qHb5sLAw5OTk4KeffkKbNm1w6dIlyOVyAMCUKVNQXFyMw4cPw8rKCpcuXYK1deMftyUNYocPH8bChQtx+vRpJCYmYsuWLRg+fHiNyxQVFWHu3Ln46aefkJSUBHd3d8ycORPPP/984xRdA1tzU/TwccTf19Ow51Iy2vRlECMiIsOlVCphZmYGS0tLuLlpzupcuXIFgOam3AMGDNDO6+DggMDAQO3rjz76CFu2bMG2bdt0WqHuNWHCBIwePRoAMG/ePHzxxRc4efIkBg0aVGNtpqammDNnjvZ1q1at8Mcff2Djxo3aIPbxxx/jrbfewtSpU7XzPfzwwwCAvXv34uTJk7h8+TLatm0LAPDx8dHOFx8fj6eeegqdOnWqNK0xSRrE8vLyEBgYiOeffx4jRoyo1TIjR45EcnIyVq1aBV9fXyQmJkKtVjdwpbU3MMBNG8Qm920jdTlERCQVU0tNy5RU7/2AunXrpvM6NzcXs2fPxvbt25GYmIjS0lIUFBQgPj6+xvV07txZ+9zKygq2trZISUmpVQ3Lli3D6tWrER8fj4KCAhQVFeGhhx4CAKSkpOD27dt4/PHHq1w2IiICLVu21Iawe73++ut4+eWXsXv3boSEhOCpp57SqbWxSBrEwsLCEBYWVuv5//rrLxw6dAjR0dHaJk1vb+8Gqq5+Bvi74sPfL+BM/B2k5BTCxcZc6pKIiEgKglCr04OG6t5fP06fPh179uzBokWL4OvrCwsLCzz99NMoLi6ucT2mpqY6rwVBqFUDyq+//orp06dj8eLF6NmzJywtLTFjxgxER0cDACwsLGpc/n7TX3zxRYSGhmL79u3YvXs35s+fj8WLF+O11167b2361KQu1t+2bRu6deuGBQsWoEWLFmjbti2mT5+OgoKCapcpKipCdna2ztCQ3JTmCGyphCgC+y7XLvETERFJxczMDCqV6r7zHT16FBMmTMCTTz6JTp06wc3NDbGxsQ1W19GjR9GrVy+88sorCAoKgq+vL27evKmdbmNjA29vb+zbt6/K5Tt37oybN2/i2rVr1b6Hp6cnJk+ejM2bN+Ott97Ct99+q/ftuJ8mFcSio6Nx5MgRXLhwAVu2bMHSpUuxadMmvPLKK9UuM3/+fCiVSu3g6enZ4HUO6OAKgN1YEBGR4fP29saJEycQGxuLtLS0alur/Pz8sHnzZkRERODcuXMYM2ZMg14a5Ofnh1OnTmHXrl24du0aZs6ciUuXLunMM3v2bCxevBhffPEFrl+/jjNnzuDLL78EAPTt2xePPvoonnrqKezZswcxMTHYuXMn/vrrLwDAG2+8gV27diEmJgZnzpzBgQMH4O/v32DbU50mFcTUajUEQcDPP/+M7t27Y/DgwViyZAm+//77alvFZsyYgaysLO2QkJDQ4HWWd2NxJCoNeUW163uFiIhICtOnT4dcLkeHDh3g7Oxc7TVfS5Ysgb29PXr16oXw8HCEhoaiS5cuDVbXpEmTMGLECIwaNQrBwcHIyMjA008/rTPP+PHjsXTpUnz99dcICAjA0KFDcf36de303377DQ8//DBGjx6NDh064J133tG2/qlUKkyZMgX+/v4YNGgQ2rZti6+//rrBtqc6gijWsrORBiYIwn1/NTl+/HgcPXoUUVFR2nGXL19Ghw4dcO3aNfj5+d33fbKzs6FUKpGVlQVbW1t9lF6JKIrot+ggYtPzsXxsF4R1cm+Q9yEiIsNQWFiImJgYtG7dGubmvDa4IahUKpw9exZBQUHaLigMVV3+HppUi1jv3r1x+/Zt5Obmasddu3YNMpkMLVu2lLAyXYIgaE9P7ubpSSIiIqqGpEEsNzcXERERiIiIAADExMQgIiJC2yw6Y8YMjBs3Tjv/mDFj4OjoiOeeew6XLl3C4cOH8fbbb+P555+/768jGlv56cn9V1JQojKc7jWIiIgMweTJk2FtbV3lMHnyZKnLazSSdl9x6tQp9OvXT/t62rRpADSnINeuXYvExESdc9XW1tbYs2cPXnvtNXTr1g2Ojo4YOXIkPv7440av/X66tLKHo5UZ0vOK8W9MBnr5OkldEhERkcGYO3cupk+fXuW0hrp0yBBJGsQee+wx1HSJ2tq1ayuNa9++Pfbs2dOAVemHXCbgcX8XbDh1E7svJTOIERERVeDi4gIXFxepy5Bck7pGrKkZ0EFzenLPpeQaAycRERkH/ltPQN3+DhjEGlAfPydYmMpxK7MAF283bEeyREQknfLe4/Pz8yWuhAxB+d/BvXcVqIqkpyaNnbmpHI+2dcKui8nYcykZHVsopS6JiIgagFwuh52dnfYeipaWlhAEQeKqjEt5/1+FhYUG232FKIrIz89HSkoK7OzsalUng1gDG9DBDbsuJmP3pWS8OaDqG48SEVHT5+amuRyltje0prpRq9VIS0tDbGwsZDLDPqFnZ2en/Xu4HwaxBvZ4exfIBOByYjYSMvLh6WApdUlERNQABEGAu7s7XFxcUFJSInU5Ric3NxdDhgzBqVOnYG1tLXU51TI1Na1Tix2DWAOztzLDw94OOBGTgT2XkvH8I62lLomIiBqQXC432FNnTVlxcTHi4uJgZmZmVHcvMOy2PSNR3rnr7ktJEldCREREhoRBrBEMLLvd0b+xd3Anr1jiaoiIiMhQMIg1Ak8HS7R3s4FKLWL/FV7ESURERBoMYo2k/PTkHt4EnIiIiMowiDWS8tOTh66lorBEJXE1REREZAgYxBpJgIctPJTmKChR4cj1NKnLISIiIgPAINZIBEHAgLJWMZ6eJCIiIoBBrFGVXye293IyVGreGJaIiKi5YxBrRN1bO8DW3ATpecU4G39H6nKIiIhIYgxijchULkP/9i4AgN08PUlERNTsMYg1Mm0v+xeTIIo8PUlERNScMYg1skfbOsNMLkNsej6iUnKlLoeIiIgkxCDWyKwVJujt6wiApyeJiIiaOwYxCdy9CTiDGBERUXPGICaBx/1dIAjAuYRMJGcXSl0OERERSYRBTAIuNuYI8rQDwM5diYiImjMGMYkM6MDTk0RERM0dg5hEBgZobnd0/EYasgtLJK6GiIiIpMAgJpE2ztbwcbZCiUrEoaupUpdDREREEmAQk9BAnp4kIiJq1hjEJFR+evLglRQUl6olroaIiIgaG4OYhB5qaQdnGwVyikrxT3S61OUQERFRI2MQk5BMJiDEX9Mqxm4siIiImh8GMYkN7HA3iKnVvAk4ERFRc8IgJrGebRxhZSZHUnYhIm9lSV0OERERNSIGMYmZm8rxWDsXADw9SURE1NwwiBmAAWWnJ3dfSpK4EiIiImpMDGIGoF87F5jIBFxLzkVsWp7U5RAREVEjYRAzAEpLUwT7OADg6UkiIqLmhEHMQJT3ss8gRkRE1HwwiBmIkLLrxE7FZSAtt0jiaoiIiKgxMIgZiBZ2FujYwhZqEdh/OUXqcoiIiKgRMIgZEN4EnIiIqHlhEDMg5d1Y/H09FfnFpRJXQ0RERA2NQcyAtHezgaeDBYpK1fj7eprU5RAREVEDYxAzIIIgYIB/2enJizw9SUREZOwYxAzMwADN6cl9V5JRqlJLXA0RERE1JAYxA9PNyx72lqbIzC/Bqbg7UpdDREREDUjSIHb48GGEh4fDw8MDgiDg999/r/WyR48ehYmJCR566KEGq08KJnIZ+rcvu/ckT08SEREZNUmDWF5eHgIDA7Fs2bI6LZeZmYlx48bh8ccfb6DKpFV+enLP5SSIoihxNURERNRQTKR887CwMISFhdV5ucmTJ2PMmDGQy+V1akVrKvr4OUFhIkNCRgGuJOXA391W6pKIiIioATS5a8TWrFmD6OhozJo1q1bzFxUVITs7W2cwdJZmJujj5wyA954kIiIyZk0qiF2/fh3vvfcefvrpJ5iY1K4xb/78+VAqldrB09OzgavUj4FlnbvuvpQkcSVERETUUJpMEFOpVBgzZgzmzJmDtm3b1nq5GTNmICsrSzskJCQ0YJX687i/C2QCcOFWNm5lFkhdDhERETUASa8Rq4ucnBycOnUKZ8+exauvvgoAUKvVEEURJiYm2L17N/r3719pOYVCAYVC0djlPjBHawW6eTngZGwG9l5Kxvhe3lKXRERERHrWZIKYra0tIiMjdcZ9/fXX2L9/PzZt2oTWrVtLVFnDGdDBFSdjM7D7UhKDGBERkRGSNIjl5uYiKipK+zomJgYRERFwcHBAq1atMGPGDNy6dQs//PADZDIZOnbsqLO8i4sLzM3NK403FgM6uOKTHZdxIjoDWfklUFqaSl0SERER6ZGk14idOnUKQUFBCAoKAgBMmzYNQUFBmDlzJgAgMTER8fHxUpYoKW8nK7R1tUapWsSBqylSl0NERER6JojNrMfQ7OxsKJVKZGVlwdbW8PvnWrTrKr46EIUhndyxbGwXqcshIiKSRFM7ftdWk/nVZHNV3sv+waspKCxRSVwNERER6RODmIHr1EIJN1tz5BWrcPxGutTlEBERkR4xiBk4QRAwQNu5K3vZJyIiMiYMYk1AeRDbezkZanWzuqSPiIjIqDGINQE9fBxhozBBak4RIm5mSl0OERER6QmDWBNgZiLDY+1dAAB/nLstcTVERESkLwxiTcTQzu4AgDVHY/Hj8VhpiyEiIiK9YBBrIgZ2cMWLj2hu4/Th1otYdSRG4oqIiIjoQTGINRGCIOCDIf545bE2AICP/ryE5QdvSFwVERERPQgGsSZEEAS8HdoOb4T4AQA+++sKvth3XeKqiIiIqL4YxJoYQRDwRkhbvB3aDgCwZM81LNp1Fc3sTlVERERGgUGsiZrSzxf/HeIPAPjqQBTm77zCMEZERNTEMIg1YS/28cGcJwIAACsPR2POH5cYxoiIiJoQBrEmbnwvb8x7shMEAVh7LBYf/H6Bve8TERE1EQxiRmBMcCsseKozBAH45UQ83v3tPFQMY0RERAaPQcxIPNPNE0tHPQS5TMDG0zfx1oYIlKrUUpdFRERENWAQMyLDHmqBL0cHwUQm4PeI25j6awRKGMaIiIgMFoOYkRncyR1fj+0CU7mA7ZGJmPLzGRSVqqQui4iIiKrAIGaEBga4YeV/usHMRIbdl5Ix+cfTKCxhGCMiIjI0DGJGql97F6wa3w3mpjIcuJqKiT+cQkExwxgREZEhYRAzYn38nLFmQndYmsnx9/U0PLf2JPKKSqUui4iIiMowiBm5nm0c8cPz3WGtMME/0RkYv/okcgpLpC6LiIiIwCDWLHTzdsCPL3SHjbkJTsXdwX9WnURWAcMYERGR1BjEmomgVvZYN7EH7CxNEZGQibHf/YM7ecVSl0VERNSsMYg1Ix1bKLFuYg84Wpnhwq1sjP72H6TlFkldFhERUbPFINbM+Lvb4teXesDZRoErSTkYvfIfpGQXSl0WERFRs8Qg1gz5udpg/Us94GZrjuspuXh25T9IymIYIyIiamwMYs2Uj7M11k/qgRZ2FohOy8PIb47j5p18qcsiIiJqVhjEmjEvRyusn9QDrRwsEZ+Rj1Hf/IPvj8UiKiUHoihKXR4REZHRE8RmdsTNzs6GUqlEVlYWbG1tpS7HICRmFWDstycQnZanHedqq0CvNk7o1cYRvX2d4GFnIWGFRETU3Bnr8ZtBjAAAGXnFWHcyHkeup+F0/B0Ul6p1prd2stKGsp4+jrC3MpOoUiIiao6M9fjNIEaVFJaocDruDo5GpeHojXRE3syEusJfiSAA/m626O3riF6+Tuju7QArhYl0BRMRkdEz1uM3gxjdV1ZBCU5Ep+PYjXQcjUrD9ZRcnekmMgFBrezQq40Tevs64SFPO5iZ8PJDIiLSH2M9fjOIUZ2l5BTieFkoOxqVjluZBTrTLc3keNjbQdNi1sYJHdxtIZMJElVLRETGwFiP3wxi9EBEUUR8Rj6ORqXj6I00HL+Rjox7bp1kZ2mKnj6O6OPnjIEBrnCyVkhULRERNVXGevxmECO9UqtFXE3OwdGoNBy7kY4T0enIK1Zpp8sEoGcbRwzu5I5BAW5wZCgjIqJaMNbjN4MYNagSlRrnb2biaFQ69lxKRuStLO00uUxADx8HDOnkgdAAV4YyIiKqlrEevxnEqFHFp+dje2QitkfexoVb2drxcpmAnj6OGNLZHaEBbnBg9xhERFSBsR6/GcRIMnHpedgemYgdkYmVQlmvNo4Y0kkTythnGRERGevxm0GMDEJsWh52XEjE9vOJuHiboYyIiHQZ6/GbQYwMTmyapqVs+/lEXErUDWW9fZ0wpJMbBnZgKCMiak6M9fjNIEYGLSYtDzsiE/Hn+URcrhDKTGQCevk6YWgndwwMcIWdJUMZEZExM9bjN4MYNRnRqbnYEZmI7ZFJlUJZb18nhHRwRU8fR7RxtoIgsANZIiJjYqzHbwYxapJupOZix/lEbI9MxJWkHJ1pTtYK9PBxQM82jujh4wgfJwYzIqKmzliP35IGscOHD2PhwoU4ffo0EhMTsWXLFgwfPrza+Tdv3ozly5cjIiICRUVFCAgIwOzZsxEaGlrr9zTWHdmc3UjNxc7IRByNSsfp+DsoLlXrTHe2UaCHjyN6+jiih48DWjOYERE1OcZ6/DaR8s3z8vIQGBiI559/HiNGjLjv/IcPH8aAAQMwb9482NnZYc2aNQgPD8eJEycQFBTUCBWTIWrjbI1X+/vh1f5+KCxRISIhE/9Ep+Of6HScic9Eak4R/jh3G3+cuw0AcCkLZj18HNGzjSO8HS0ZzIiISBIGc2pSEIT7tohVJSAgAKNGjcLMmTNrNb+xJmqqWmGJCmfj7wazs/GZKFbptpi52lYIZj6O8GIwIyIyOMZ6/Ja0RexBqdVq5OTkwMHBQepSyECZm8rRs42m5QvQBLMz8XfwT3QG/olOR0R8JpKzi7A14ja2RmhazNxszdHDx0EbzhjMiIioodQ7iOXl5eHQoUOIj49HcXGxzrTXX3/9gQurjUWLFiE3NxcjR46sdp6ioiIUFRVpX2dnZ1c7Lxk/c1M5erVxQq82TgAqBLMb6fgnOgNnE+4gKbsQv0fcxu9lwcxdaY4ePo7o194F/du7wFrRpP//QkREBqReR5SzZ89i8ODByM/PR15eHhwcHJCWlgZLS0u4uLg0ShD75ZdfMGfOHGzduhUuLi7Vzjd//nzMmTOnweuhpuneYFZQrMLZ+Dv4Jzodx6PTEZGQicSsQmw5ewtbzt6CmYkMj/o5I6yjG0I6uEJpYSrxFhARUVNWr2vEHnvsMbRt2xYrVqyAUqnEuXPnYGpqiv/7v//D1KlTa3XhfaVC6nCN2K+//ornn38eGzduxJAhQ2qct6oWMU9PT6M7x0wNo6BY02L29/U0/HUhEbHp+dpppnIBvdo4IayjGwbyRuVERA3KWK8Rq1cQs7Ozw4kTJ9CuXTvY2dnh+PHj8Pf3x4kTJzB+/HhcuXKl7oXUMoitW7cOzz//PH799VcMGzaszu9jrDuSGp4oiriSlIOdF5Lw14VEXEvO1U6TywQEt3ZAWEc3hAa4wcXWXMJKiYiMj7Eev+t1atLU1BQymQwA4OLigvj4ePj7+0OpVCIhIaHW68nNzUVUVJT2dUxMDCIiIuDg4IBWrVphxowZuHXrFn744QcAmtOR48ePx+eff47g4GAkJSUBACwsLKBUKuuzKUS1JggC/N1t4e9ui2kD2iIqJRd/XUjEzgtJuHg7G8dupOPYjXTM3HYR3bzsMaijOwZ1dEMLOwupSyciIgNVrxaxgQMHYsKECRgzZgwmTpyI8+fP4/XXX8ePP/6IO3fu4MSJE7Vaz8GDB9GvX79K48ePH4+1a9diwoQJiI2NxcGDBwFoTokeOnSo2vlrw1gTNUkrPj0fO8tCWURCps60QE87hHV0Q1hHN3g5WklTIBFRE2esx+96BbFTp04hJycH/fr1Q0pKCsaNG4djx47Bz88Pq1evRmBgYEPUqhfGuiPJcNzOLMCui0nYGZmEf+MyUPEb1sHdVhPKOrnB18VGuiKJiJoYYz1+G0yHro3FWHckGaaUnELsvpiMnRcS8U90BlTqu183PxdrhHV0w6CO7vB3t2FfZURENTDW4zeDGFEjycgrxt5LydhxIRFHo9JQorr71fNQmsNNaQ47SzPYWZhqHi1NYWdpCqWFKezLX1uYQWlpCltzEwY3ImpWjPX4Xesg1qVLF+zbtw/29vYICgqq8SBw5swZvRWob8a6I6lpySoowf4rydgRmYTD11JRdM+Nyu9HLhOgtDCFnYUplJaaR3tLs7Lnd0NcebBztDaDq605TOWyBtoiIqKGZazH71r/anLYsGFQKBQAUOf7QRKRLqWFKZ4Maokng1oir6gUF25lIbOgBFn5JbiTX4zMghJk5pcgM79Y81hQgqz8YtzJL0FBiQoqtYiMvGJk5BXf/83KyATA1dYcHnYWZYM5WthZwENpAfey50oLU7a0ERE1Ip6aJGpiCktUyC7QhLM7ecXaAJdZoAlqmfklyCrQBLg7ZWEuPbe40s3Oq2JpJtcGtRZ25vBQWugENzelORQm8kbYSiIiXcZ6/K5XP2L//vsv1Go1goODdcafOHECcrkc3bp100txRFSZuakc5qbyOnUaq1aLSMsrwu3MQtzOLMDtzALcKntMzNKMS8stRn6xClEpuYhKya12Xc42ikpBrYOHLbq0soeZCU99EhHVRb2C2JQpU/DOO+9UCmK3bt3CZ599Vut+xIiocchkAlxszOFiY46HPO2qnKewRKUNZeUhrTyolb8uLFEjNacIqTlFOHdP382WZnL08HHEI75OeLStE9o4W/M0JxHRfdTr1KS1tTXOnz8PHx8fnfExMTHo3LkzcnJy9Fagvhlr0yZRQxNFEXfySyoFtYSMAvwbm4H0e65Xc7M1Rx8/Jzzi54RHfJ3gaK2QqHIiMgbGevyuV4uYQqFAcnJypSCWmJgIE5N6rZKIDJwgCHCwMoODlRk6ttC9pZhaLeJyUjb+vp6GI9fTcDI2A0nZhdh4+iY2nr4JAAjwsMUjfk7o4+uMbt72MDfltWZERPVqERs9ejQSExOxdetW7T0eMzMzMXz4cLi4uGDDhg16L1RfjDVRExmSwhIV/o3NwN/X0/D39TRcTszWma4wkaF7awf08XNCHz9ntHdjh7ZEVDNjPX7XK4jdunULjz76KNLT0xEUFAQAiIiIgKurK/bs2QNPT0+9F6ovxrojiQxZak4RjkallQWzVKTkFOlMd7JW4BFfR/Txc8Yjfk5wrcMPEYioeTDW43e9u6/Iy8vDzz//jHPnzsHCwgKdO3fG6NGjYWpqqu8a9cpYdyRRUyGKIq6n5JadxkzFP9EZKChR6czT1tUaj/g6o09bJwS3doCFqRylahHFpWoUl6pRolKjqFSNYpXmefn44nue607TXb58XrlMgLejJVo7WcPH2QputuaQydg6R2RojPX4zX7EiEhSRaUqnInLxJGoVPx9PQ2Rt7J0bpRefsaysf6lsjCVw9vJCj7OVvDRPlqjtbMVbM0N+z+aRMbMWI/fDxTELl26hPj4eBQX6/5a6oknnnjgwhqKse5IImNxJ68Yx26k4+/rmmB2K7Og0jwyATAzkcFMLoOZiRxmckHz2kQGU7mswrQKj2XPTcseFWXzFpWqEJOWj+i0XMSn56NUXf0/iU7WCp2AVt6K1srBkrePImpgxnr8rlcQi46OxpNPPonIyEgIgoDyVZRfbKtSqWpaXFLGuiOJjJEoikjN1VxPppDLtYFK3kCnDktVaiTcKUBMWi6iU/NwIzVP+/ze69oqkssEtHKwhI+TFVo7WcHH2Vob2ByszFCqFqFSixUe1VCVPdcZr9KdXqoWodZZToRKrdaZXxAAmSBAEDR1yAQBMkHz77FMECCX3X0uK5tX+1x2d1552fjy9clkgLmJHF6OlvwhBRkEYz1+16uvialTp6J169bYt28fWrdujZMnTyI9PR1vvfUWFi1apO8aiaiZEgRNR7SNxUQuQ+uyMNW/ve60nMISxJa1nEWn5iE6LQ/RqbmISctDfrEKMWl5iEnLa7RaG0sLOwuEdXRDWCc3BHna8/o5Ij2rV4uYk5MT9u/fj86dO0OpVOLkyZNo164d9u/fj7feegtnz55tiFr1wlgTNRFJQxRFJGcXITo1tyyclbWipeUhISMfVZ3plAmAiUwGmUzzKJcJMJEJdx/lgma6UGG6XLhnPpk2FKnVItRi+aCpSaW++1wtAuqycWLZc7Woea4qX059d16VKEIURWQXlqK49O49Sl1tFRgU4IawTu542NuhwVomiapirMfverWIqVQq2NjYANCEstu3b6Ndu3bw8vLC1atX9VogEZEhEwQBbkrNDdF7+TrpTCsuVSO/uBQmcpk2QMkFocm0KhWWqHDoWir+upCEvZeSkZxdhO+Px+H743FwsjbDwAA3hHV0Qw8fR14jR1RP9QpiHTt2xLlz59C6dWsEBwdjwYIFMDMzw8qVKyv1tk9E1Fxprmkzk7qMejM3lSM0wA2hAW4oKlXhWFQ6dkQmYvelZKTlFuOXE/H45UQ87CxNMbCDK8I6uqO3r5OkN39XqUXEpedpb1yvuUG9BewsTXmtGxmkep2a3LVrF/Ly8jBixAhERUVh6NChuHbtGhwdHbF+/Xr079+/IWrVC2Nt2iQiaiwlKjX+iU7Hjsgk7L6YpHOfURtzE4T4uyKsoxsebevcYLeyKg9c15JzEZWSg2vJubiWnIPotDyd06nlzE1l2lDmobSAh50F3O3MNa/tLOCuNOdttwycsR6/9daPWEZGBuzt7Q3+fxzGuiOJiKSgUos4GZOBnRcS8deFJJ1fl1qZydGvvQsGd3LHY+2cYWlW95MwdQ1cgCZ0tXG2holMwK3MQqTlVv+L14qcrM20oUwb2soHpTmcrBVN5rSyMTLW43edg1hJSQksLCwQERGBjh07NlRdDcZYdyQRkdTUahFn4u9g54Uk7IxMxO2sQu00c1MZHmvrgrBObujf3gU293SOW9/A5etijbYuNvB11Ty2dbVBC3sLnR8SFJWqkJRViFuZBbidWYjbmQWaIUvz/Nadgkp3d6iKqVyAu9ICHnaaoOZpbwkvR0t4OVrBy9ESjlZmBt8Y0ZQZ6/G7Xi1iPj4+2LJlCwIDAxuipgZlrDuSiMiQiKKIczezsPNCInZGJiE+I187zUwuQx8/J3RsoURMWp7eAteD1JpVUKINaolZBZVCW3J2YZW/gK3IykyuDWWtHC3h7WgFLwfNc3elfmptzoz1+F2vILZq1Sps3rwZP/74IxwcHBqirgZjrDuSiMhQiaKIS4nZ2BmZhB0XEhGdWnV/a9UFrpb2FpKfEixVqZGcU6QNZjfvFCAhIx9x6fmIS89DYnZhjbfhMpPL0NLBAl4Od1vQvBwt0crBCp4OFlCY8Pq0+zHW43e9glhQUBCioqJQUlICLy8vWFlZ6Uw/c+aM3grUN2PdkURETUH5Td93Ribh5p18tHa2MqjAVV+FJSrcvFOAuPQ8xKXnIz4jH7HpeYhPz0fCnXyUqKo/1AoC4KG0QCsHS21rWisHS7jamsPZWgFnGwWsFPXq5MCoGOvxu157dvjw4Xoug4iImgNBENDWVRO8jIm5qRy+LtbwdbGuNE2lFpGYVVDWepaPuIw8xKXlIy4jH/HpecgrVuFWpuZ06PHo9CrXb2kmh7ONQhvMKj0vGxytFJJ2H0J1p7dfTTYVxpqoiYio6RFFEel5xdqWtPJTnTfvFCA1twgp2UW1+iFBRXaWpvcNbEoLU5jIZDAtu2ODaYVOhw31BwfGevxmWycREZFEBEGAk7UCTtYKdPWq+prrvKJSpOYUITW3SPNYcSgbl1b2WKoWkZlfgsz8Elwv69S2rsoDmalcVvZ497ZaFYOb5tZbmgBncs/89pZmWPhM0/tBnxTqFcRkMlmNiVmlqlt6JyIioqpZKUxgpTCBt5NVjfOp1Zpff1YKbFW8zi4oQWk1PwMtVYsoVYsoquZXrLXhaquo97LNTb2C2JYtW3Rel5SU4OzZs/j+++8xZ84cvRRGREREtSeTCbC3MoO9lVmtrsErvzl8efAqVanLHkWUqtVlj/c8r2EelVpEiUqESq3mdWp1oNdrxH755ResX78eW7du1dcq9c5YzzETEREZM2M9fus1svbo0QP79u3T5yqJiIiIjJbeglhBQQG++OILtGjRQl+rJCIiIjJq9bpG7N6be4uiiJycHFhaWuKnn37SW3FERERExqxeQex///ufThCTyWRwdnZGcHAw7O3t9VYcERERkTGrVxCbMGGCnssgIiIian7qdY3YmjVrsHHjxkrjN27ciO+///6BiyIiIiJqDuoVxObPnw8nJ6dK411cXDBv3rwHLoqIiIioOahXEIuPj0fr1q0rjffy8kJ8fPwDF0VERETUHNQriLm4uOD8+fOVxp87dw6Ojo4PXBQRERFRc1CvIDZ69Gi8/vrrOHDgAFQqFVQqFfbv34+pU6fi2Wef1XeNREREREapXr+a/OijjxAbG4vHH38cJiaaVajVaowbNw6ffPKJXgskIiIiMlYPdK/J69evIyIiAhYWFujUqRO8vLz0WVuDMNZ7VRERERkzYz1+16tFrJyfnx/8/PwAaD6g5cuXY9WqVTh16pReiiMiIiIyZg8UxADgwIEDWL16NTZv3gylUoknn3xSH3URERERGb16Xax/69YtfPLJJ/D19cUzzzyDX375BatXr8atW7ewbNmyWq/n8OHDCA8Ph4eHBwRBwO+//37fZQ4ePIguXbpAoVDA19cXa9eurc8mEBEREUmuTkHst99+w+DBg9GuXTtERERg8eLFuH37NmQyGTp16qRz/8nayMvLQ2BgYK3DW0xMDIYMGYJ+/fohIiICb7zxBl588UXs2rWrTu9LREREZAjqdGpy1KhRePfdd7F+/XrY2Ng88JuHhYUhLCys1vOvWLECrVu3xuLFiwEA/v7+OHLkCP73v/8hNDT0geshIiIiakx1ahF74YUXsGzZMgwaNAgrVqzAnTt3GqquKh0/fhwhISE640JDQ3H8+PFqlykqKkJ2drbOQERERGQI6hTEvvnmGyQmJuKll17CunXr4O7ujmHDhkEURajV6oaqUSspKQmurq4641xdXZGdnY2CgoIql5k/fz6USqV28PT0bPA6iYiIiGqjzhfrW1hYYPz48Th06BAiIyMREBAAV1dX9O7dG2PGjMHmzZsbos56mzFjBrKysrRDQkKC1CURERERAajnrybL+fn5Yd68eUhISMBPP/2E/Px8jB49Wl+1VeLm5obk5GSdccnJybC1tYWFhUWVyygUCtja2uoMRERERIbggfsRAwCZTIbw8HCEh4cjJSVFH6usUs+ePbFjxw6dcXv27EHPnj0b7D2JiIiIGsoDtYhVxcXFpdbz5ubmIiIiAhEREQA03VNEREQgPj4egOa04rhx47TzT548GdHR0XjnnXdw5coVfP3119iwYQPefPNNvW4DERERUWPQexCri1OnTiEoKAhBQUEAgGnTpiEoKAgzZ84EACQmJmpDGQC0bt0a27dvx549exAYGIjFixfju+++Y9cVRERE1CQ90E2/myJjvWkoERGRMTPW43edWsSio6Mbqg4iIiKiZqdOQaxz587o2LEj3n//fZw4caKhaiIiIiJqFuoUxNLS0jB//nykpKRg2LBhcHd3x8SJE/HHH3+gsLCwoWokIiIiMkr1vkZMFEUcP34c27Ztw7Zt2xAfH4+QkBA88cQTCA8Ph7Ozs75r1QtjPcdMRERkzIz1+F3vX00KgoBevXrh008/xaVLl3D27Fn06dMHa9euRcuWLbFs2TJ91klERERkdBrkV5Pp6enIyMiAn5+fvlf9wIw1URMRERkzYz1+66Vn/Xs5OjrC0dGxIVZNREREZDQk7dCViIiIqDljECMiIiKSCIMYERERkUTqFcQSEhJw8+ZN7euTJ0/ijTfewMqVK/VWGBEREZGxq1cQGzNmDA4cOAAASEpKwoABA3Dy5El88MEHmDt3rl4LJCIiIjJW9QpiFy5cQPfu3QEAGzZsQMeOHXHs2DH8/PPPWLt2rT7rIyIiIjJa9QpiJSUlUCgUAIC9e/fiiSeeAAC0b98eiYmJ+quOiIiIyIjVK4gFBARgxYoV+Pvvv7Fnzx4MGjQIAHD79m32H0ZERERUS/UKYp999hm++eYbPPbYYxg9ejQCAwMBANu2bdOesiQiIiKimtX7FkcqlQrZ2dmwt7fXjouNjYWlpSVcXFz0VqC+GestEoiIiIyZsR6/69UiVlBQgKKiIm0Ii4uLw9KlS3H16lWDDmFEREREhqReQWzYsGH44YcfAACZmZkIDg7G4sWLMXz4cCxfvlyvBRIREREZq3oFsTNnzqBPnz4AgE2bNsHV1RVxcXH44Ycf8MUXX+i1QCIiIiJjVa8glp+fDxsbGwDA7t27MWLECMhkMvTo0QNxcXF6LZCIiIjIWNUriPn6+uL3339HQkICdu3ahYEDBwIAUlJSjOoCOiIiIqKGVK8gNnPmTEyfPh3e3t7o3r07evbsCUDTOhYUFKTXAomIiIiMVb27r0hKSkJiYiICAwMhk2ny3MmTJ2Fra4v27dvrtUh9MtafvxIRERkzYz1+m9R3QTc3N7i5ueHmzZsAgJYtW7IzVyIiIqI6qNepSbVajblz50KpVMLLywteXl6ws7PDRx99BLVare8aiYiIiIxSvVrEPvjgA6xatQqffvopevfuDQA4cuQIZs+ejcLCQnzyySd6LZKIiIjIGNXrGjEPDw+sWLECTzzxhM74rVu34pVXXsGtW7f0VqC+Ges5ZiIiImNmrMfvep2azMjIqPKC/Pbt2yMjI+OBiyIiIiJqDuoVxAIDA/HVV19VGv/VV18hMDDwgYsiIiIiag7qdY3YggULMGTIEOzdu1fbh9jx48eRkJCAHTt26LVAIiIiImNVrxaxvn374tq1a3jyySeRmZmJzMxMjBgxAlevXtXeg5KIiIiIalbvDl2rcvPmTcydOxcrV67U1yr1zlgv9iMiIjJmxnr8rleLWHXS09OxatUqfa6SiIiIyGjpNYgRERERUe0xiBERERFJhEGMiIiISCJ16r5ixIgRNU7PzMx8kFqIiIiImpU6BTGlUnnf6ePGjXuggoiIiIiaizoFsTVr1jRUHURERETNDq8RIyIiIpIIgxgRERGRRBjEiIiIiCTCIEZEREQkEYMIYsuWLYO3tzfMzc0RHByMkydP1jj/0qVL0a5dO1hYWMDT0xNvvvkmCgsLG6laIiIiIv2QPIitX78e06ZNw6xZs3DmzBkEBgYiNDQUKSkpVc7/yy+/4L333sOsWbNw+fJlrFq1CuvXr8f777/fyJUTERERPRjJg9iSJUswceJEPPfcc+jQoQNWrFgBS0tLrF69usr5jx07ht69e2PMmDHw9vbGwIEDMXr06Pu2ohEREREZGkmDWHFxMU6fPo2QkBDtOJlMhpCQEBw/frzKZXr16oXTp09rg1d0dDR27NiBwYMHN0rNRERERPpSpw5d9S0tLQ0qlQqurq46411dXXHlypUqlxkzZgzS0tLwyCOPQBRFlJaWYvLkydWemiwqKkJRUZH2dXZ2tv42gIiIiOgBSH5qsq4OHjyIefPm4euvv8aZM2ewefNmbN++HR999FGV88+fPx9KpVI7eHp6NnLFRERERFUTRFEUpXrz4uJiWFpaYtOmTRg+fLh2/Pjx45GZmYmtW7dWWqZPnz7o0aMHFi5cqB33008/4aWXXkJubi5kMt1sWVWLmKenJ7KysmBra6v/jSIiIiK9y87OhlKpNLrjt6QtYmZmZujatSv27dunHadWq7Fv3z707NmzymXy8/MrhS25XA4AqCpTKhQK2Nra6gxEREREhkDSa8QAYNq0aRg/fjy6deuG7t27Y+nSpcjLy8Nzzz0HABg3bhxatGiB+fPnAwDCw8OxZMkSBAUFITg4GFFRUfjwww8RHh6uDWRERERETYHkQWzUqFFITU3FzJkzkZSUhIceegh//fWX9gL++Ph4nRaw//73vxAEAf/9739x69YtODs7Izw8HJ988olUm0BERERUL5JeIyYFYz3HTEREZMyM9fjd5H41SURERGQsGMSIiIiIJMIgRkRERCQRBjEiIiIiiTCIEREREUmEQYyIiIhIIgxiRERERBJhECMiIiKSCIMYERERkUQYxIiIiIgkwiBGREREJBEGMSIiIiKJMIgRERERSYRBjIiIiEgiDGJEREREEmEQIyIiIpIIgxgRERGRRBjEiIiIiCTCIEZEREQkEQYxIiIiIokwiBERERFJhEGMiIiISCIMYkREREQSYRAjIiIikgiDGBEREZFEGMSIiIiIJMIgRkRERCQRBjEiIiIiiTCIEREREUmEQYyIiIhIIgxiRERERBJhECMiIiKSCIMYERERkUQYxIiIiIgkwiBGREREJBEGMSIiIiKJMIgRERERSYRBjIiIiEgiDGJEREREEmEQIyIiIpIIgxgRERGRRBjEiIiIiCTCIEZEREQkEROpCyAiIqImRlUClOQDJQVAcZ7msaSgbFw+IMiAtqFSV9kkMIgRERE1JlEECrOA/HQgP6PssWwoKHtdmKUJM4IckJkAMrnmdflzmUnZNFmF5+XjK8ynXf6edYnqCsGpmjClnVbheUkBUJIHqEtr3kYbD+Cty43zeTZxBhHEli1bhoULFyIpKQmBgYH48ssv0b1792rnz8zMxAcffIDNmzcjIyMDXl5eWLp0KQYPHtyIVRMRUa2VFgPZt4CsBCDrJpCZABTnAJZOgJVz2eB099HUQuqKa0cUgeLcCmHqnmClMz7jbti6X5BpKgQZYGpZNlhoHs0sASsXqStrMiQPYuvXr8e0adOwYsUKBAcHY+nSpQgNDcXVq1fh4lJ5RxYXF2PAgAFwcXHBpk2b0KJFC8TFxcHOzq7xiyciIo2iHE24ykoAMuM1YSsr4e64nCQAYu3XZ2ZTIZiVh7SqQpszYOEAyOtxOFOrNCGqKKfCkH3P69wqxpXNVx6s1CV1f28AMLMGLB0AS0fNNlg63h3MlQBETWBTqzSPogpQqys8Lx+vvjufWDZOra7wvOJ8Zc8FoSw0Wd0NUNrHsufaaRXHl08rey4306yL6k0QRbEO3wz9Cw4OxsMPP4yvvvoKAKBWq+Hp6YnXXnsN7733XqX5V6xYgYULF+LKlSswNTWt8/tlZ2dDqVQiKysLtra2D1w/EVGVRBEouAOUFpUdtKzqFxYMgSgCuSll4Sq+LFxVDFrxmlNp92NiDihblg2egMJWE2TyUsuGNM1jnYONoAk094Y0QVZNuCobinPr9XFUvW0WZSHqnkClHXfPeAsHwNRcf+/fDBjr8VvSfxWKi4tx+vRpzJgxQztOJpMhJCQEx48fr3KZbdu2oWfPnpgyZQq2bt0KZ2dnjBkzBu+++y7kcnml+YuKilBUVKR9nZ2drf8NIaLmRxQ1oSEzAciMuxtKMuPvtgrde6CXm5W1MljdbVEws7p7Oufe8dppVc1T9iiqNWFPVawZSosBVVHZuJKy52XjdKaXzV9x+r3jSvKArFua0KUqqvpzqMjcDrDzBJStNGHLzlMTuMofrZzv33oiiprgVB7K7g1p9z7PzwAg3j0FmHql7vtSZgqY2wIKm7Kh4vMKg9k9ryu2ZplZ1v19iSBxEEtLS4NKpYKrq6vOeFdXV1y5UvWXKTo6Gvv378fYsWOxY8cOREVF4ZVXXkFJSQlmzZpVaf758+djzpw5DVI/EelZSUHZ9UNxd0NNbrKmJcVcWXawtC17XjZoX9tqwoq+TpOo1UBukqaG8laf8uflp95KC+6/HkGuOUUEaEJOQbGmpazJEQAb93vCVUtN6Cp/rrDRw9sId/etY5v7z68q1XyelUJbiqbmqgLVvYHLRPHgdRPVU5NrJ1er1XBxccHKlSshl8vRtWtX3Lp1CwsXLqwyiM2YMQPTpk3Tvs7Ozoanp2djlkxE5Ypy77YWlQ/a1wllB88HIDMpC2a294S0KkJb+XNBVhb+4nXDVtbNWpwiKw8nZWHErtXdkGLnpQknJuaaAFacp/nlWXG+pqWs/HlJ3t1fpWnnydN9XuW8Za8FGSBXACZmmke5qSZYVDnO7O5jxedVjZObaa4FKg9fti006zE0chPA2lkzEDVBkgYxJycnyOVyJCcn64xPTk6Gm5tblcu4u7vD1NRU5zSkv78/kpKSUFxcDDMzM535FQoFFAr+b4eoURRm3dNqlKDbulWQcf91mFlXCDStABs3TZApzAIKszWnrQqz7g7lr8svWC7IqN371IYgB5Qtylp9qghbti01Yed+TBRlrS4O+qmLiIyGpEHMzMwMXbt2xb59+zB8+HAAmhavffv24dVXX61ymd69e+OXX36BWq2GTKa5McC1a9fg7u5eKYQRETSn2EoLNYOquOx58T2vi2qeVlp097qjSq8LgZxkTdAqqsUF2+bKaoJN2TgL+7qfXhRFTcuRNphlV3ieec/rewKduvTuxeN2Xro12bg33QvsiahJkPxfmGnTpmH8+PHo1q0bunfvjqVLlyIvLw/PPfccAGDcuHFo0aIF5s+fDwB4+eWX8dVXX2Hq1Kl47bXXcP36dcybNw+vv/66lJtB9GBE8Z7TV3mVT0/d93luhdNWFZ6X5Dfutlg66gYrnaDlWfazfD0TBEBhrRnQQv/rJyJqIJIHsVGjRiE1NRUzZ85EUlISHnroIfz111/aC/jj4+O1LV8A4OnpiV27duHNN99E586d0aJFC0ydOhXvvvuuVJtAVLXi/Cp+7XXvUDYtP71xOngUZJprlsqvBdJeS2SuOcWmnVbVa0U10xSazhvtyn4pp7Bu+O0gIjISkvcj1tiMtR8SagRqlebXWbkpVfyMPqXyz+vr20fRvV0c1Ol5xa4NKjw3tSgLTpL/34uIqF6M9fjNf5WJKlKrNBeXp10vG65pHjNuaAKWqK7b+uQKTd9J1s73dDbpotvxpKWjpiWpKXf6SUREdcZ/8al5KswG0q8DaVFlYatC4FIV17ysRYUevCsFLOeykFX2XGHD238QEVG1GMTIeKnVQPbNspBVIXClRwE5idUvJ1cAjr6Akx/g1Fbz6OgL2HpoWq4MsS8lIiJqkhjEqGkrLSq7ZisFyIjRPZ2YHlVzz+fWrpqg5ehbFrjaAk6+ml/4ySrfLouIiEjfGMTI8JQW370APrd8SNaMK3+dVzbufjcalplqbpPi5Ac4+ukGroboRoGIiKgOGMSocahKNL8k1Aaq5Lu/Pix/Xh6w6nofPpkpYO2iaclyblshcPlpOujkxe9ERGSgeISi+qsUru5puarYolXXW87ITDQXvVs7a04h6jwve7R20QzmdrwgnoiImiQGMdKlKqkQpPQcrgT53V8aasNV2VDxubWrJlxV6MiXiIjIGDGINXelxcCN/UDkBiD6EJCfVrflK4arqoJVxdYrCweGKyIiogoYxJojtRqIPw5EbgQu/V75mqxK4cq1+qDFcEVERFRvDGLNhSgCyReA8xuAC78B2bfuTrN2BTo+BXQYrvmFIcMVERFRo2AQM3YZMcCFTUDkJiD1yt3xClugwxNAp2cA7z7sN4uIiEgCDGLGKDcVuLhZc+rx5r93x8sVQNtQTfjyGwiYmktXIxERETGIGY3CbODKdk34ij4IiCrNeEEGtO6rCV/+Q9mJKRERkQFhEGvKSouAqL2a676u/QWUFt6d1qIr0GkkEPAkYOMqXY1ERERULQaxpkatAuKOlv3icavuLX4c/YDOIzUX3ju2ka5GIiIiqhUGsaaiOA84tAA4vx7ISbw73sZdE7w6jwTcOrOHeSIioiaEQawpUJUAG8YDUXs0r82Vmq4mOj0DePXiLx6JiIiaKAYxQyeKwLbXNSHMxAIYvgxoPxQwUUhdGRERET0gBjFDt28ucO4XTW/3z6wF2g2SuiIiIiLSE3afbshOfAMcWaJ5Hv45QxgREZGRYRAzVBe3ADvf1Tzv/1+gy3+krYeIiIj0jkHMEMX8DWx+CYAIPPwi0Ge61BURERFRA2AQMzRJF4BfxwCqYsA/HAhbwC4piIiIjBSDmCHJjAd+fhooyga8egMjvmPXFEREREaMQcxQ5GcAPz2l6azV2R949mfelJuIiMjIMYgZguJ84JdRQNo1wLYF8H+/ARb2UldFREREDYxBTGqqUmDT88DNk4C5HfB/mwFlC6mrIiIiokbAICYlUQS2vwlc2wmYmANj1gMu7aWuioiIiBoJg5iUDs4HzvwACDLgqVVAqx5SV0RERESNiEFMKv+uAg59pnk+ZDHgP1TaeoiIiKjRMYhJ4fIfwI6yTlr7vgd0e17aeoiIiEgSDGKNLe44sOkFQFQDXcYDj70ndUVEREQkEQaxxpRyGVg3ClAVAe0GA0OWsNd8IiKiZoxBrLFk3dJ02FqYBbTsrrk4X24idVVEREQkIQaxxlBwRxPCsm8BTm013VSYWUpdFREREUmMQayhlRQA68YAqZcBG3dNr/mWDlJXRURERAaAQawhqVXAby8C8ccAhVITwuxaSV0VERERGQgGsYYiipouKq78CcjNgNG/AK4BUldFREREBoRBrKEcXgScWg1AAEZ8C3g/InVFREREZGAYxBrC6e+BAx9rnoctAAKGS1oOERERGSYGMX27uhP48w3N80emAcEvSVoOERERGS4GMX1KOAlsfE7Ta37gGODxmVJXRERERAaMQUxfUq8Bv4wESgsA3wHAE1+w13wiIiKqkUEEsWXLlsHb2xvm5uYIDg7GyZMna7Xcr7/+CkEQMHz48IYtsDZMFIClI+DRBRj5PSA3lboiIiIiMnCSB7H169dj2rRpmDVrFs6cOYPAwECEhoYiJSWlxuViY2Mxffp09OnTp5EqvQ97L+D53cDYjYCZldTVEBERURMgeRBbsmQJJk6ciOeeew4dOnTAihUrYGlpidWrV1e7jEqlwtixYzFnzhz4+Pg0YrX3YeUIWDlJXQURERE1EZIGseLiYpw+fRohISHacTKZDCEhITh+/Hi1y82dOxcuLi544YUX7vseRUVFyM7O1hmIiIiIDIGkQSwtLQ0qlQqurq46411dXZGUlFTlMkeOHMGqVavw7bff1uo95s+fD6VSqR08PT0fuG4iIiIifZD81GRd5OTk4D//+Q++/fZbODnV7hTgjBkzkJWVpR0SEhIauEoiIiKi2jGR8s2dnJwgl8uRnJysMz45ORlubm6V5r9x4wZiY2MRHh6uHadWqwEAJiYmuHr1Ktq0aaOzjEKhgEKhaIDqiYiIiB6MpC1iZmZm6Nq1K/bt26cdp1arsW/fPvTs2bPS/O3bt0dkZCQiIiK0wxNPPIF+/fohIiKCpx2JiIioSZG0RQwApk2bhvHjx6Nbt27o3r07li5diry8PDz33HMAgHHjxqFFixaYP38+zM3N0bFjR53l7ezsAKDSeCIiIiJDJ3kQGzVqFFJTUzFz5kwkJSXhoYcewl9//aW9gD8+Ph4yWZO6lI2IiIioVgRRFEWpi2hM2dnZUCqVyMrKgq2trdTlEBERUS0Y6/GbTU1EREREEmEQIyIiIpIIgxgRERGRRBjEiIiIiCTCIEZEREQkEcm7r2hs5T8S5c2/iYiImo7y47axdfbQ7IJYTk4OALAXfiIioiYoJycHSqVS6jL0ptn1I6ZWq3H79m3Y2NhAEAS9rjs7Oxuenp5ISEgwqj5OqtKcthVoXtvLbTVezWl7ua3GRxRF5OTkwMPDw6g6em92LWIymQwtW7Zs0PewtbU16i9DRc1pW4Hmtb3cVuPVnLaX22pcjKklrJzxREoiIiKiJoZBjIiIiEgiDGJ6pFAoMGvWLCgUCqlLaXDNaVuB5rW93Fbj1Zy2l9tKTUWzu1ifiIiIyFCwRYyIiIhIIgxiRERERBJhECMiIiKSCIMYERERkUQYxOpo2bJl8Pb2hrm5OYKDg3Hy5Mka59+4cSPat28Pc3NzdOrUCTt27GikSutv/vz5ePjhh2FjYwMXFxcMHz4cV69erXGZtWvXQhAEncHc3LyRKn4ws2fPrlR7+/bta1ymKe5XAPD29q60rYIgYMqUKVXO35T26+HDhxEeHg4PDw8IgoDff/9dZ7ooipg5cybc3d1hYWGBkJAQXL9+/b7rret3vrHUtL0lJSV499130alTJ1hZWcHDwwPjxo3D7du3a1xnfb4LjeF++3bChAmV6h40aNB912uI+/Z+21rV91cQBCxcuLDadRrqfiUNBrE6WL9+PaZNm4ZZs2bhzJkzCAwMRGhoKFJSUqqc/9ixYxg9ejReeOEFnD17FsOHD8fw4cNx4cKFRq68bg4dOoQpU6bgn3/+wZ49e1BSUoKBAwciLy+vxuVsbW2RmJioHeLi4hqp4gcXEBCgU/uRI0eqnbep7lcA+Pfff3W2c8+ePQCAZ555ptplmsp+zcvLQ2BgIJYtW1bl9AULFuCLL77AihUrcOLECVhZWSE0NBSFhYXVrrOu3/nGVNP25ufn48yZM/jwww9x5swZbN68GVevXsUTTzxx3/XW5bvQWO63bwFg0KBBOnWvW7euxnUa6r6937ZW3MbExESsXr0agiDgqaeeqnG9hrhfqYxItda9e3dxypQp2tcqlUr08PAQ58+fX+X8I0eOFIcMGaIzLjg4WJw0aVKD1qlvKSkpIgDx0KFD1c6zZs0aUalUNl5RejRr1iwxMDCw1vMby34VRVGcOnWq2KZNG1GtVlc5vanuVwDili1btK/VarXo5uYmLly4UDsuMzNTVCgU4rp166pdT12/81K5d3urcvLkSRGAGBcXV+08df0uSKGqbR0/frw4bNiwOq2nKezb2uzXYcOGif37969xnqawX5sztojVUnFxMU6fPo2QkBDtOJlMhpCQEBw/frzKZY4fP64zPwCEhoZWO7+hysrKAgA4ODjUOF9ubi68vLzg6emJYcOG4eLFi41Rnl5cv34dHh4e8PHxwdixYxEfH1/tvMayX4uLi/HTTz/h+eefhyAI1c7XlPdruZiYGCQlJensN6VSieDg4Gr3W32+84YsKysLgiDAzs6uxvnq8l0wJAcPHoSLiwvatWuHl19+Genp6dXOayz7Njk5Gdu3b8cLL7xw33mb6n5tDhjEaiktLQ0qlQqurq46411dXZGUlFTlMklJSXWa3xCp1Wq88cYb6N27Nzp27FjtfO3atcPq1auxdetW/PTTT1Cr1ejVqxdu3rzZiNXWT3BwMNauXYu//voLy5cvR0xMDPr06YOcnJwq5zeG/QoAv//+OzIzMzFhwoRq52nK+7Wi8n1Tl/1Wn++8oSosLMS7776L0aNH13hT6Lp+FwzFoEGD8MMPP2Dfvn347LPPcOjQIYSFhUGlUlU5v7Hs2++//x42NjYYMWJEjfM11f3aXJhIXQAZtilTpuDChQv3vZ6gZ8+e6Nmzp/Z1r1694O/vj2+++QYfffRRQ5f5QMLCwrTPO3fujODgYHh5eWHDhg21+p9mU7Vq1SqEhYXBw8Oj2nma8n4ljZKSEowcORKiKGL58uU1zttUvwvPPvus9nmnTp3QuXNntGnTBgcPHsTjjz8uYWUNa/Xq1Rg7dux9f0DTVPdrc8EWsVpycnKCXC5HcnKyzvjk5GS4ublVuYybm1ud5jc0r776Kv78808cOHAALVu2rNOypqamCAoKQlRUVANV13Ds7OzQtm3bamtv6vsVAOLi4rB37168+OKLdVquqe7X8n1Tl/1Wn++8oSkPYXFxcdizZ0+NrWFVud93wVD5+PjAycmp2rqNYd/+/fffuHr1ap2/w0DT3a/GikGslszMzNC1a1fs27dPO06tVmPfvn06LQYV9ezZU2d+ANizZ0+18xsKURTx6quvYsuWLdi/fz9at25d53WoVCpERkbC3d29ASpsWLm5ubhx40a1tTfV/VrRmjVr4OLigiFDhtRpuaa6X1u3bg03Nzed/ZadnY0TJ05Uu9/q8503JOUh7Pr169i7dy8cHR3rvI77fRcM1c2bN5Genl5t3U193wKaFu2uXbsiMDCwzss21f1qtKT+tUBT8uuvv4oKhUJcu3ateOnSJfGll14S7ezsxKSkJFEURfE///mP+N5772nnP3r0qGhiYiIuWrRIvHz5sjhr1izR1NRUjIyMlGoTauXll18WlUqlePDgQTExMVE75Ofna+e5d1vnzJkj7tq1S7xx44Z4+vRp8dlnnxXNzc3FixcvSrEJdfLWW2+JBw8eFGNiYsSjR4+KISEhopOTk5iSkiKKovHs13IqlUps1aqV+O6771aa1pT3a05Ojnj27Fnx7NmzIgBxyZIl4tmzZ7W/Evz0009FOzs7cevWreL58+fFYcOGia1btxYLCgq06+jfv7/45Zdfal/f7zsvpZq2t7i4WHziiSfEli1bihERETrf46KiIu067t3e+30XpFLTtubk5IjTp08Xjx8/LsbExIh79+4Vu3TpIvr5+YmFhYXadTSVfXu/v2NRFMWsrCzR0tJSXL58eZXraCr7lTQYxOroyy+/FFu1aiWamZmJ3bt3F//55x/ttL59+4rjx4/XmX/Dhg1i27ZtRTMzMzEgIEDcvn17I1dcdwCqHNasWaOd595tfeONN7Sfi6urqzh48GDxzJkzjV98PYwaNUp0d3cXzczMxBYtWoijRo0So6KitNONZb+W27VrlwhAvHr1aqVpTXm/HjhwoMq/2/LtUavV4ocffii6urqKCoVCfPzxxyt9Bl5eXuKsWbN0xtX0nZdSTdsbExNT7ff4wIED2nXcu733+y5IpaZtzc/PFwcOHCg6OzuLpqamopeXlzhx4sRKgaqp7Nv7/R2Loih+8803ooWFhZiZmVnlOprKfiUNQRRFsUGb3IiIiIioSrxGjIiIiEgiDGJEREREEmEQIyIiIpIIgxgRERGRRBjEiIiIiCTCIEZEREQkEQYxIiIiIokwiBGRQZk6dSpeeuklqNVqqUshImpwDGJEZDASEhLQrl07fPPNN5DJ+M8TERk/9qxPREREJBH+l5OIJDdhwgQIglBpGDRokNSlERE1KBOpCyAiAoBBgwZhzZo1OuMUCoVE1RARNQ62iBGRQVAoFHBzc9MZ7O3tAQCCIGD58uUICwuDhYUFfHx8sGnTJp3lIyMj0b9/f1hYWMDR0REvvfQScnNzdeZZvXo1AgICoFAo4O7ujldffVU7bcmSJejUqROsrKzg6emJV155pdLyRET6xiBGRE3Chx9+iKeeegrnzp3D2LFj8eyzz+Ly5csAgLy8PISGhsLe3h7//vsvNm7ciL179+oEreXLl2PKlCl46aWXEBkZiW3btsHX11c7XSaT4YsvvsDFixfx/fffY//+/XjnnXcafTuJqHnhxfpEJLkJEybgp59+grm5uc74999/H++//z4EQcDkyZOxfPly7bQePXqgS5cu+Prrr/Htt9/i3XffRUJCAqysrAAAO3bsQHh4OG7fvg1XV1e0aNECzz33HD7++ONa1bRp0yZMnjwZaWlp+ttQIqJ78BoxIjII/fr10wlaAODg4KB93rNnT51pPXv2REREBADg8uXLCAwM1IYwAOjduzfUajWuXr0KQRBw+/ZtPP7449W+/969ezF//nxcuXIF2dnZKC0tRWFhIfLz82FpaamHLSQiqoynJonIIFhZWcHX11dnqBjEHoSFhUWN02NjYzF06FB07twZv/32G06fPo1ly5YBAIqLi/VSAxFRVRjEiKhJ+Oeffyq99vf3BwD4+/vj3LlzyMvL004/evQoZDIZ2rVrBxsbG3h7e2Pfvn1Vrvv06dNQq9VYvHgxevTogbZt2+L27dsNtzFERGV4apKIDEJRURGSkpJ0xpmYmMDJyQkAsHHjRnTr1g2PPPIIfv75Z5w8eRKrVq0CAIwdOxazZs3C+PHjMXv2bKSmpuK1117Df/7zH7i6ugIAZs+ejcmTJ8PFxQVhYWHIycnB0aNH8dprr8HX1xclJSX48ssvER4ejqNHj2LFihWN+wEQUfMkEhFJbPz48SKASkO7du1EURRFAOKyZcvEAQMGiAqFQvT29hbXr1+vs47z58+L/fr1E83NzUUHBwdx4sSJYk5Ojs48K1asENu1ayeampqK7u7u4muvvaadtmTJEtHd3V20sLAQQ0NDxR9++EEEIN65c6fBt5+Imi/+apKIDJ4gCNiyZQuGDx8udSlERHrFa8SIiIiIJMIgRkRERCQRXqxPRAaPV1AQkbFiixgRERGRRBjEiIiIiCTCIEZEREQkEQYxIiIiIokwiBERERFJhEGMiIiISCIMYkREREQSYRAjIiIikgiDGBEREZFE/h9bO2epRnPuSgAAAABJRU5ErkJggg==","text/plain":["<Figure size 640x480 with 1 Axes>"]},"metadata":{},"output_type":"display_data"}],"source":["plt.plot(range(num_epochs), train_loss[:20] , label='train_losses')\n","plt.plot(range(num_epochs), train_acc[:20], label='train_accs')\n","plt.title('Curva de Treinamento')\n","plt.xlabel('Época')\n","plt.ylabel('Loss / Acuracia')\n","plt.legend(bbox_to_anchor=(1.1, 1.05))\n","plt.show()"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"ABeroE9XBMQA","outputId":"3a1274a9-1c85-4080-8a5b-c24d34717ff6"},"outputs":[{"data":{"image/png":"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","text/plain":["<Figure size 640x480 with 1 Axes>"]},"metadata":{},"output_type":"display_data"}],"source":["plt.plot(range(num_epochs), val_loss[:20] , label='train_losses')\n","plt.plot(range(num_epochs), val_acc[:20], label='train_accs')\n","plt.title('Curva de Validação')\n","plt.xlabel('Época')\n","plt.ylabel('Loss / Acuracia')\n","plt.legend(bbox_to_anchor=(1.1, 1.05))\n","plt.show()"]},{"cell_type":"markdown","metadata":{"id":"bgycuHKJBMQB"},"source":["### **8.5 - Salva pesos - normal**"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"PDDuhfnwBMQB"},"outputs":[],"source":["torch.save(model_trained.state_dict(), 'weights.h5')"]},{"cell_type":"markdown","metadata":{"id":"KG75GpnABMQB"},"source":["### **9 - Validação - normal**"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"CwliBaGSBMQB"},"outputs":[],"source":["validation_img_paths = [\"train/River/River_1.png\",\n","                        \"train/Residential/Residential_1.png\",\n","                        \"train/SeaLake/Bombay_000001.png\"]\n","img_list = [Image.open(input_path + img_path) for img_path in validation_img_paths]\n","\n","validation_batch = torch.stack([data_transforms['validation'](img).to(device)\n","                                for img in img_list])"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"VYnjJ3lsBMQB","outputId":"6dd42162-ca5a-4901-e861-4489f891a63f"},"outputs":[{"data":{"text/plain":["tensor([[-0.3379, -3.1141, -2.1886,  1.0018,  2.1967, -0.1580, -0.3453, -0.0842,\n","          1.3839, -6.1068],\n","        [-0.3985, -2.9574, -1.4564,  0.8135,  2.2156, -0.5779, -1.4851,  2.0806,\n","         -0.4474, -3.9132],\n","        [-1.0177, -2.2023,  0.5217, -1.7249, -0.1710, -2.5843, -4.2879, -0.6905,\n","          2.3281, 10.1210]], grad_fn=<AddmmBackward0>)"]},"execution_count":18,"metadata":{},"output_type":"execute_result"}],"source":["pred_logits_tensor = model(validation_batch)\n","pred_logits_tensor"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"D_RtqILTBMQB","outputId":"05cdee0c-4a91-4d1e-e8e4-1e0f4daf7f33"},"outputs":[{"data":{"text/plain":["Softmax(\n","  dim=tensor([[-0.3379, -3.1141, -2.1886,  1.0018,  2.1967, -0.1580, -0.3453, -0.0842,\n","            1.3839, -6.1068],\n","          [-0.3985, -2.9574, -1.4564,  0.8135,  2.2156, -0.5779, -1.4851,  2.0806,\n","           -0.4474, -3.9132],\n","          [-1.0177, -2.2023,  0.5217, -1.7249, -0.1710, -2.5843, -4.2879, -0.6905,\n","            2.3281, 10.1210]], grad_fn=<AddmmBackward0>)\n",")"]},"execution_count":19,"metadata":{},"output_type":"execute_result"}],"source":["pred_probs = nn.Softmax(pred_logits_tensor).cpu()\n","pred_probs"]},{"cell_type":"markdown","metadata":{"id":"jEyzuUysBMQC"},"source":["### **10 - resnet50 - pré-treinada**"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"_pJhn_hWBMQC"},"outputs":[],"source":["pretmodel = models.resnet50(pretrained=\"IMAGENET1K_V1\").to(device)\n","\n","for param in model.parameters():\n","    param.requires_grad = False\n","\n","model.fc = nn.Sequential(\n","               nn.Linear(2048, 128),\n","               nn.ReLU(inplace=True),\n","               nn.Linear(128, 10)).to(device)"]},{"cell_type":"markdown","metadata":{"id":"wLn5j4qKBMQC"},"source":["### **11 - Execução do treinamento - pré-treinada**"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"1ozhGWxeBMQC","outputId":"de83825e-47c0-491f-f84e-970897c5509e"},"outputs":[{"name":"stdout","output_type":"stream","text":["Epoch 1/3\n","----------\n","train loss: 12.7544, acc: 0.0007\n","validation loss: 14.1321, acc: 0.0009\n","Epoch 2/3\n","----------\n","train loss: 12.7731, acc: 0.0006\n","validation loss: 14.3292, acc: 0.0005\n","Epoch 3/3\n","----------\n","train loss: 12.7489, acc: 0.0005\n","validation loss: 14.3927, acc: 0.0004\n"]}],"source":["pmodel_trained = train_model(pretmodel, criterion, optimizer, num_epochs=3)"]},{"cell_type":"markdown","metadata":{"id":"j8LhuxXYBMQC"},"source":["### **12 - Visualização das curvas de treinamento e teste - pré-treinada**"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"CRQ0zKSUBMQD"},"outputs":[],"source":["num_epochs = 3"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"-tIngienBMQD","outputId":"5f421685-8cf1-40e5-92e7-eb0e8cad135e"},"outputs":[{"data":{"image/png":"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","text/plain":["<Figure size 640x480 with 1 Axes>"]},"metadata":{},"output_type":"display_data"}],"source":["plt.plot(range(num_epochs), train_loss[:3] , label='train_losses')\n","plt.plot(range(num_epochs), train_acc[:3], label='train_accs')\n","plt.title('Curva de Treinamento')\n","plt.xlabel('Época')\n","plt.ylabel('Loss / Acuracia')\n","plt.legend(bbox_to_anchor=(1.1, 1.05))\n","plt.show()"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"9U0OIsayBMQD","outputId":"f040bb31-9b88-44e2-fde8-220143522328"},"outputs":[{"data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAmIAAAHJCAYAAAAmZVVWAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy80BEi2AAAACXBIWXMAAA9hAAAPYQGoP6dpAABcd0lEQVR4nO3deVwU9f8H8NfuArvch9yK4H2LV5KaqYXiRZodaIdgqVlqGvktMW9LLO/UNMsz07ytX5p5pZaRlkqpeKSioHKZwsJy787vD2R03QU5FmZZX8/HYx+ws5+deX9YcF7OfOYzMkEQBBARERFRtZNLXQARERHR44pBjIiIiEgiDGJEREREEmEQIyIiIpIIgxgRERGRRBjEiIiIiCTCIEZEREQkEQYxIiIiIokwiBERERFJhEGMiMzC2rVrIZPJcO3aNalLKZFMJsP06dPF5+WpOSAgABEREVVW2+3btxEYGAgvLy+sW7cOx44dQ5s2bapse0RkGgxiRGbuypUreOutt1C/fn2oVCo4OTmhS5cuWLx4MXJycqQuzyy9++67kMlkuHz5coltPvroI8hkMvzzzz/VWFnV2bJlC+zt7fH2229j/Pjx6Nq1K958802pyyKiR2AQIzJju3fvRqtWrbBlyxaEhoZiyZIliI6ORt26dfG///0P48aNk7pEs/Tqq68CADZu3Fhim02bNqFVq1Zo3bp1hbfz+uuvIycnB/7+/hVeh6kMGTIEP/74I6ZPn45bt24hJSUFY8eOlbosInoEK6kLICLj4uPjMXjwYPj7++PQoUPw8fERXxs9ejQuX76M3bt3m2RbGo0G9vb2JlmXOQgKCkLDhg2xadMmTJ061eD1mJgYxMfHY86cOZXajkKhgEKhqNQ6TMXV1VX83tbWFra2thJWQ0RlxSNiRGbqs88+Q1ZWFlatWqUXwoo1bNhQPCJ27do1yGQyrF271qDdw+Oapk+fDplMhri4OLzyyitwdXXFU089hXnz5kEmk+H69esG64iKioKNjQ3u3r0LAPj111/x0ksvoW7dulAqlfDz88N7771X5lOl586dwzPPPANbW1vUqVMHH3/8MXQ6ndG2P/30E7p27Qp7e3s4OjqiX79+OHfu3CO38eqrr+LChQs4deqUwWsbN26ETCbDkCFDkJ+fj6lTp6J9+/ZwdnaGvb09unbtil9++eWR2zA2RkwQBHz88ceoU6cO7Ozs0KNHD6P13rlzBxMmTECrVq3g4OAAJycn9OnTB3///bdB29zcXEyfPh2NGzeGSqWCj48PBg0ahCtXrohtPv30U3Tu3Bm1atWCra0t2rdvj23bthmsq7CwELNmzUKDBg2gVCoREBCASZMmIS8v75H9JSLTYxAjMlP/93//h/r166Nz585Vsv6XXnoJ2dnZmD17NkaMGIGXX34ZMpkMW7ZsMWi7ZcsW9OrVSzzqsnXrVmRnZ+Ptt9/GkiVLEBISgiVLlmDo0KGP3G5ycjJ69OiB2NhYTJw4EePHj8f69euxePFig7bffPMN+vXrBwcHB3z66aeYMmUK4uLi8NRTTz1ygHxJpye1Wi22bNmCrl27om7dulCr1fj666/RvXt3fPrpp5g+fTrS0tIQEhKC2NjYR/bnYVOnTsWUKVMQGBiIuXPnon79+ujVqxc0Go1eu6tXr2LXrl3o378/FixYgP/97384c+YMunXrhlu3bunV279/f8yYMQPt27fH/PnzMW7cOGRkZODs2bNiu0WLFqFt27aYOXMmZs+eDSsrK7z00ksGR02HDx+OqVOnol27dli4cCG6deuG6OhoDB48uNx9JSITEIjI7GRkZAgAhAEDBpSpfXx8vABAWLNmjcFrAIRp06aJz6dNmyYAEIYMGWLQtlOnTkL79u31lp04cUIAIKxfv15clp2dbfDe6OhoQSaTCdevXy+11vHjxwsAhOPHj4vLUlNTBWdnZwGAEB8fLwiCIGRmZgouLi7CiBEj9N6fnJwsODs7Gyw35oknnhDq1KkjaLVacdnevXsFAMKXX34pCIIgFBYWCnl5eXrvu3v3ruDl5SW88cYbessf/lmuWbNGr+bU1FTBxsZG6Nevn6DT6cR2kyZNEgAI4eHh4rLc3Fy9ugSh6HNUKpXCzJkzxWWrV68WAAgLFiww6N+D29BoNHqv5efnCy1bthSeeeYZcVlsbKwAQBg+fLhe2wkTJggAhEOHDhlsg4iqFo+IEZkhtVoNAHB0dKyybYwaNcpgWVhYGE6ePKl3ymvz5s1QKpUYMGCAuOzB8UcajQa3b99G586dIQgCTp8+Xep29+zZgyeffBIdO3YUl3l4eIhHsIrt378f6enpGDJkCG7fvi0+FAoFgoKCynTq8LXXXsONGzdw9OhRcdnGjRthY2ODl156CUDROC8bGxsAgE6nw507d1BYWIgOHToYPa1ZmgMHDiA/Px9jx46FTCYTl48fP96grVKphFxe9E+wVqvFf//9BwcHBzRp0kRvu9u3b4e7u7vRgfcPbsPOzk78/u7du8jIyEDXrl311rVnzx4AQGRkpN563n//fQAw2ZhDIio7BjEiM+Tk5AQAyMzMrLJt1KtXz2DZSy+9BLlcjs2bNwMoGu+0detW9OnTR6wJABISEhAREQE3Nzc4ODjAw8MD3bp1AwBkZGSUut3r16+jUaNGBsubNGmi9/zff/8FADzzzDPw8PDQe+zbtw+pqamP7OPgwYOhUCjE05O5ubnYuXMn+vTpoze4fd26dWjdujVUKhVq1aoFDw8P7N69+5F9MdY3AAb98/Dw0NseUBT6Fi5ciEaNGkGpVMLd3R0eHh74559/9LZ75coVNGnSBFZWpV9b9eOPP+LJJ5+ESqWCm5sbPDw8sHz5cr11Xb9+HXK5HA0bNtR7r7e3N1xcXIyODySiqsWrJonMkJOTE3x9ffXGAJXmwSMjD9JqtSW+x9hVdb6+vujatSu2bNmCSZMm4Y8//kBCQgI+/fRTvXX27NkTd+7cwYcffoimTZvC3t4eN2/eRERERImD7sureD3ffPMNvL29DV5/VDABAE9PT/Ts2RPbt2/HsmXL8H//93/IzMzUO/q2YcMGREREYODAgfjf//4HT09PKBQKREdH6x0ZNLXZs2djypQpeOONNzBr1iy4ublBLpdj/Pjx5f4Z/vrrr3juuefw9NNP44svvoCPjw+sra2xZs0ao1N4lPT7QkTVj0GMyEz1798fK1euRExMDDp16lRq2+KjLenp6XrLK3KEIywsDO+88w4uXryIzZs3w87ODqGhoeLrZ86cwaVLl7Bu3Tq9wfn79+8v0/r9/f3Fo10Punjxot7zBg0aACgKU8HBweXuR7FXX30Ve/fuxU8//YSNGzfCyclJrz/btm1D/fr1sWPHDr2AMm3atHJvq3g+sX///Rf169cXl6elpYlXnD643R49emDVqlV6y9PT0+Hu7i4+b9CgAY4fP46CggJYW1sb3e727duhUqnw888/Q6lUisvXrFljUJ9Op8O///6LZs2aictTUlKQnp5uFvOh1XSCIKCwsLDU/wSRZVMoFLCysirzf3gYxIjM1AcffIBvv/0Ww4cPx6FDh+Dl5aX3+pUrV/Djjz9i3LhxcHJygru7O44ePao3HumLL74o93ZfeOEFjB07Fps2bcLWrVvRv39/vTnGiufNEgRBXCYIgtGrHo3p27cvFi1ahBMnTojjxNLS0vDtt9/qtQsJCYGTkxNmz56NHj16GISQtLQ0eHh4PHJ7AwcOhJ2dHb744gscPnwYQ4YMgUqlMtqf4n84jx8/jpiYGNStW7dMfSoWHBwMa2trLFmyBL169RLXt2jRIoO2CoVC72cIFF2NevPmTb1Thy+88AJ2796NpUuX4r333tNrX1yzQqGATCbT2/lfu3YNu3bt0mvft29fTJo0CYsWLcKXX34pLl+wYAEAoF+/fuXqL+nLz89HUlISsrOzpS6FJGZnZwcfHx9x/GlpGMSIzFSDBg2wceNGhIWFoVmzZhg6dChatmyJ/Px8/P7779i6davevQuHDx+OOXPmYPjw4ejQoQOOHj2KS5culXu7np6e6NGjBxYsWIDMzEyEhYXpvd60aVM0aNAAEyZMwM2bN+Hk5ITt27cbHPEpyQcffIBvvvkGvXv3xrhx42Bvb4+VK1fC399f73ZDTk5OWL58OV5//XW0a9cOgwcPhoeHBxISErB792506dIFS5cufeT2HBwcMHDgQPEU3cMXBfTv3x87duzA888/j379+iE+Ph4rVqxA8+bNkZWVVaY+FfPw8MCECRMQHR2N/v37o2/fvjh9+jR++uknvaNcxdudOXMmhg0bhs6dO+PMmTP49ttv9Y6kAcDQoUOxfv16REZG4sSJE+jatSs0Gg0OHDiAd955BwMGDEC/fv2wYMEC9O7dG6+88gpSU1OxbNkyNGzYUO9nGhgYiPDwcKxcuRLp6eno1q0bTpw4gXXr1mHgwIHo0aNHufpL9+l0OsTHx0OhUMDX1xc2NjY8BfwYEgQB+fn5SEtLQ3x8PBo1aiRelFPam4jIjF26dEkYMWKEEBAQINjY2AiOjo5Cly5dhCVLlgi5ubliu+zsbOHNN98UnJ2dBUdHR+Hll18WUlNTS5y+Ii0trcRtfvXVVwIAwdHRUcjJyTF4PS4uTggODhYcHBwEd3d3YcSIEcLff/9d4hQaD/vnn3+Ebt26CSqVSqhdu7Ywa9YsYdWqVXpTQRT75ZdfhJCQEMHZ2VlQqVRCgwYNhIiICOGvv/565HaK7d69WwAg+Pj4GEwZodPphNmzZwv+/v6CUqkU2rZtK/z4449CeHi44O/vr9f24Z/lw9NXCIIgaLVaYcaMGYKPj49ga2srdO/eXTh79qzg7+9vMH3F+++/L7br0qWLEBMTI3Tr1k3o1q2b3nazs7OFjz76SKhXr54AQLCyshJefPFF4cqVK2KbVatWCY0aNRKUSqXQtGlTYc2aNeJn/aCCggJhxowZQr169QRra2vBz89PiIqK0vtdovLLyckR4uLiDKYRoceTRqMR4uLijP77+TCZIDx0bJyIiMzWhg0bsGfPnlLvo0nVLzc3F/Hx8ahXr57eqW96PJXn94HTVxAR1SChoaHYtm0bb0lEZCEYxIiIaoDz589j8eLFmD17NgoKCpCbmyt1SUQGAgICjF6cUhGHDx+GTCYzuBrc0nCwPhFRDZCbm4uPP/4Yubm5mDRpEpydnaUuiSxE9+7d0aZNG5MEqD///FPvKmt6NAYxIqIaoG3btkhLS5O6DHoMCYIArVZbpkmUyzKlDOnjqUkiIqLHVEREBI4cOYLFixdDJpNBJpNh7dq1kMlk+Omnn9C+fXsolUr89ttvuHLlCgYMGAAvLy84ODjgiSeewIEDB/TW9/CpSZlMhq+//hrPP/887Ozs0KhRI/zwww8Vrnf79u1o0aIFlEolAgICMH/+fL3Xv/jiCzRq1AgqlQpeXl548cUXxde2bduGVq1awdbWFrVq1UJwcDA0Go34+tdff41mzZpBpVKhadOmevMw5ufnY8yYMfDx8YFKpYK/vz+io6Mr3I8H8YgYERFRFRAEATkF0sywb2utKNM8ZosXL8alS5fQsmVLzJw5EwBw7tw5AMDEiRMxb9481K9fH66urkhMTETfvn3xySefQKlUYv369QgNDcXFixdLnfx4xowZ+OyzzzB37lwsWbIEr776Kq5fvw43N7dy9enkyZN4+eWXMX36dISFheH333/HO++8g1q1aiEiIgJ//fUX3n33XXzzzTfo3Lkz7ty5g19//RUAkJSUhCFDhuCzzz7D888/j8zMTPz666/ipMrffvstpk6diqVLl6Jt27Y4ffo0RowYAXt7e4SHh+Pzzz/HDz/8gC1btqBu3bpITExEYmJiueovyWMXxHQ6HW7dugVHR0dOtkdERCaRn58PnU4HrVYr3uEgO78QrWYceMQ7q8aZacGws3n0Lt7BwQHW1tawtbU1OK04ffp0PPPMM+JzZ2dntGzZUu/1nTt3YteuXRg9erS4vPjnUGzo0KF4+eWXAQCzZs3C559/jpiYGPTu3bvU2orXUfwznTdvHnr06IHJkydDJpOhcePGiIuLw9y5cxEREYGEhATY29ujf//+cHR0hL+/P9q2bQugKIgVFhZi0KBB4q28WrVqJW5r2rRpmD9/PgYNGgQAqFevHuLi4vDll18iPDwcCQkJaNSoEZ566inIZDKT3g7ssQtit27dgp+fn9RlEBGRBfH398eKFSuQk5MjLsstlG6azr///gcqq7IdbMjKykJqaipOnz4NALh8+TIAQKlUissAIDs7GytXrsSxY8dw+/ZtaLVa5OXl4a+//hLb5efn48aNG3rvc3R01Htub2+Pv/76y+C2bQ8rruOff/4R19GtWzcUFBSItw7q0qULFi1aBK1Wi549e8Lf3x/169dH79690bt3b/GUaGBgIJ599lm0atUKISEh6NWrF1588UW4urpCo9HgypUrePPNNzFixAhx+4WFheJFMREREejZsyeaNGmC3r17o3///ujVq1eZfr6P8tgFMUdHRwBAYmIinJycJK6GiIgsQX5+PlJSUhAQECBO4CkIAs4EmvepSaDoqJinp6d49CgjIwMAEBQUBBcXF7HdO++8g5iYGMybNw8NGjSAra0twsLC4OrqKr7XxsYGderUEZ8DQOPGjfWeW1lZwc/PT2+ZMcV1tG7dGi4uLrC1tQWAEm8Z5OjoiFOnTuHw4cPYt28fpk6diunTp+PPP/+Ei4sL9u/fj99//x379u3DkiVL8NFHH+H48eOws7MDAHz11VcICgrSW2fxvWjbtWuH+Ph4/PTTTzhw4ABefvllBAcHY9u2baX/cMvgsQtixb+YTk5ODGJERGQSubm5SEtLg0KhEHfeAOBYhisNpaZUKqHT6cS6H/z6YF9+//13RERE4IUXXgBQdCTt2rVr6N69u147uVxe6vOSlj3s4TqaNWuGv//+Wy9gHjt2DI0bNxbbWllZITg4GMHBwZg2bRpcXFxw6NAhDBo0CDKZDF26dEGXLl0wdepU+Pv7Y+fOnYiMjISvry+uXr1qcC/aBzk5OSEsLAxhYWF48cUX0bt3b9y5c6fcY90eZv6/IURERFRlAgICcPz4cVy7dg0ODg7Q6XRG2zVq1Ag7duxAaGgoZDIZpkyZUmLbqvDee+/hySefxMcff4whQ4YgJiYGS5cuFa9u/PHHH3H16lU8/fTTcHV1xZ49e6DT6dCkSRMcP34cBw8eRK9eveDp6Ynjx48jLS0NzZo1A1B0QcG7774LZ2dn9O7dWzzlevfuXURGRmLBggXw8fFB27ZtIZfLsXXrVnh7e+sdMawoTl9BRET0GJswYQIUCgWaN28ODw8PJCQkGG23YMECuLq6onPnzggNDUVISAjatWtXbXW2a9cO0dHR2LJlC1q2bImpU6di5syZiIiIAAC4uLhgx44deOaZZ9CsWTOsWLECmzZtQosWLeDk5ISjR4+ib9++aNy4MSZPnoz58+ejT58+AIDhw4fj66+/xpo1a9CqVSt069YNa9euRb169QAUnfb87LPP0KFDBzzxxBO4du0a9uzZU+Jp0vJ47G76rVar4ezsjIyMDJ6aJCIik+BNv6ueVqvF6dOn0bZt20ee1pQab/pNREREVAMwiBEREVG1GzVqFBwcHIw+Ro0aJXV51YaD9YmIiKjazZw5ExMmTDD62uM0dIhBjIiIiKqdp6cnPD09pS5Dcjw1SURERCQRBjEiIiIiiTCIEREREUmEQYyIiIhIIgxiRERERBKRNIgdPXoUoaGh8PX1hUwmw65du0pt/9tvv6FLly6oVasWbG1t0bRpUyxcuLB6in0EnU5A/G0NUjNzkZOvxWN2wwIiIiKqAEmnr9BoNAgMDMQbb7yBQYMGPbK9vb09xowZg9atW8Pe3h6//fYb3nrrLdjb22PkyJHVUHHJMnIK0GPeYfG5Qi6DvY0CjiprOCitYK9UwEFlDUel1b3nVnBQWcHR2PdKKziq7n9vY8UDl0REZP4CAgIwfvx4jB8/XupSagxJg1ifPn3EG26WRdu2bdG2bVvxeUBAAHbs2IFff/1V8iCWXaCFo9IKWfmFEARAqxOgzi2EOrew0uu2sZLD4V4oEx+qh74aec3+XqBzeCDgKeQyE/SWiIgsRffu3dGmTRssWrSo0uv6888/YW9vX/miHiM1ekLX06dP4/fff8fHH39cYpu8vDzk5eWJz9VqdZXUUtvFFmdmhEAQBGTna5GVV1j0yC00/v2955q8QmQWf59f9LX4eU6BFgCQX6jDncJ83NHkV7pOOxtFUUC7F9jsbQyPzIlH5Iy8Vvy9nY0CMhlDHRGRpRMEAVqtFlZWj44MHh4e1VCRZamRQaxOnTpIS0tDYWEhpk+fjuHDh5fYNjo6GjNmzKi22mQyGezvhRWvSq6rUKuD5l6o0+QVIvNekNPk3Q9smnuhLjP3/vcGgS+vEPmFOgBAdr4W2flapGXmPWLrpZPLIB5lM3Zk7uGjcY6qkl9TWskZ6ojI8ggCUJAtzbat7YAy/LsaERGBI0eO4MiRI1i8eDEAYM2aNRg2bBj27NmDyZMn48yZM9i3bx/8/PwQGRmJP/74AxqNBs2aNUN0dDSCg4PF9T18alImk+Grr77C7t278fPPP6N27dqYP38+nnvuuUfWptVqMXLkSBw6dAjJycmoW7cu+vfvr3dmDABWr16N+fPn4/Lly3Bzc8MLL7yApUuXAgDS09Px4YcfYteuXcjIyEDDhg0xZ84c9O/fH9evX8eYMWPw22+/IT8/HwEBAZg7dy769u1b1p+ySdTIIPbrr78iKysLf/zxByZOnIiGDRtiyJAhRttGRUUhMjJSfK5Wq+Hn51ddpVaKlUIOZ1s5nG2tK72uvEItNHnah0JaAbLEZfrfa/K0947MFX1fFPYKoMnXQqsToBOAzNyiAFhZ1gqZ3unTB8fHPXhk7uHQp9dOaQ17pQJWCo6nIyIzUZANzPaVZtuTbgE2jz5FuHjxYly6dAktW7bEzJkzAQDnzp0DAEycOBHz5s1D/fr14erqisTERPTt2xeffPIJlEol1q9fj9DQUFy8eBF169YtcRszZszAZ599hrlz52LJkiV49dVXcf36dbi5uZVam06nQ506dbB161bUqlULv/32G0aOHIknnngCgwcPBgAsX74ckZGRmDNnDvr06YOMjAwcO3ZMfH+fPn2QmZmJDRs2oEGDBoiLi4NCoQAAjB49Gvn5+Th69Cjs7e0RFxcHBweHR/9sTaxGBrF69eoBAFq1aoWUlBRMnz69xCCmVCqhVCqrszyzpLRSQGmlgJu9TaXWIwgCcgt0yLwX1oqOzBUYnFo1dtTu4dO0mvyiU68FWgF3swtwN7ug0v1UWcvhoLSGg1JhOH7uXnBzNDgyVxTiir93UFnBzloBOcfTEZGFc3Z2ho2NDezs7ODt7Q0AuHDhAoCim3L37NlTbOvm5obAwEDx+axZs7Bz50788MMPGDNmTInbiIiIEPfRs2fPxueff44TJ06gd+/epdZmbW2td0arbt26+L//+z9s3bpVDGIff/wx3n//fYwbN05s98QTTwAADhw4gBMnTuD8+fNo3LgxAKB+/fpiu4SEBLzwwgto1aqVwWvVqUYGsQfpdDq9MWBUtWQyGWxtFLC1UQCOlVuXTicUhTcjp16Lv8/KLURW/v0QZ3CK9t7zvHunXnMLdMgtyMPtrMr39f6p1PtXvNorFXBQWt87GqcQg5tD8fdGTtWqrHnqleixZG1XdGRKqm1XUocOHfSeZ2VlYfr06di9ezeSkpJQWFiInJwcJCQklLqe1q1bi9/b29vDyckJqampZaph2bJlWL16NRISEpCTk4O8vDy0adMGAJCamopbt27h2WefNfre2NhY1KlTRwxhD3v33Xfx9ttvY9++fQgODsYLL7ygV2t1kTSIZWVl4fLly+Lz+Ph4xMbGws3NDXXr1kVUVBRu3ryJ9evXAyj6QOrWrYumTZsCKJqHbN68eXj33XclqZ8qRy6XwVFlDUdV5U+9Fmh1YkgzdmTu4QsmjL5273mhrmgOuOJllaWQy4xf8WrsSN3D39voj6fjVCZENYhMVqbTg+bq4asfJ0yYgP3792PevHlo2LAhbG1t8eKLLyI/v/QLyayt9f+Nl8lk0Ol0j9z+d999hwkTJmD+/Pno1KkT7OzsEBUVhatXrwIAbG1tS33/o14fPnw4QkJCsHv3buzbtw/R0dGYP38+xo4d+8jaTEnSIPbXX3+hR48e4vPisVzh4eFYu3YtkpKS9JK2TqdDVFQU4uPjYWVlhQYNGuDTTz/FW2+9Ve21k3mxVsjhYmcDF7vKn3rNK9SVcCrV+FG7B4/MPXgxxYNTmWTkFCAjp/KnXm2s5HrzzZU2F52xcXUODwQ8TmVCRABgY2MDrVb7yHbHjh1DREQEnn/+eQBFB1OuXbtWZXUdO3YMnTt3xjvvvAOgaPD+jRs3xNcdHR0REBCAgwcP6mWJYq1bt8aNGzdw6dKlEo+K+fn5YdSoURg1ahSioqLw1VdfPV5BrHv37qXOQL927Vq952PHjq32HxA9XmQyGVTWCqisFXB3qNzYQp1OQE6BVgxpRq9wLWWakwenNnlwKpP/CvPxn4mmMintKJ3Rq1+NHLWzteZUJkQ1WUBAAI4fP45r167BwcGhxKNVjRo1wo4dOxAaGgqZTIYpU6aU6chWRTVq1Ajr16/Hzz//jHr16mHdunWIi4tDw4YNxTbTp0/HqFGj4OnpKQ7MP3bsGMaOHYtu3brh6aefxgsvvIAFCxagYcOGuHDhAmQyGXr37o3x48ejT58+aNy4Me7evYtffvkFzZo1q7L+lKTGjxEjMldy+QNTmThVbl0PTmVicKTuobnoSp/mpAAF2qL//BRPZZJqoqlMHp6LzqGEo3YlnqZVWUFppajcD4qIym3ChAkIDw9H8+bNkZOTgzVr1hhtt2DBArzxxhvo3Lkz3N3d8eGHH1bZ3JwA8NZbb+H06dMICwuDTCbD4MGD8eKLLyI2NlZsEx4ejtzcXCxcuBATJkyAu7s7XnzxRfH17du3Y8KECRgyZAg0Go04fQVQdIRt9OjRuHHjBpycnNC7d29JbpsoEx6zmyKq1Wo4OzsjIyMDTk6V3DsS1UAPTmUiXv2aV3DvSF3R90Vh7973efe+zy24F/C0yLz3vc7E/3oUT2VibHzcw3eJMHbFq71SAWdb04w7JCqP3NxcxMfHo169elCpVFKXY5G0Wi1Onz6Ntm3bilNQmKvy/D7wiBjRY8aUU5kUn3p91KnVku4goRGXmXYqE19nFVrUdkYLXye08C366uOs4ilUIjI7DGJEVCEymQx2Nlaws7GCZyWnMtHqBGTn3w9yRueiM3KF64PPiy+gyCvU4VZGLm5l5GJ/XIq4DVc7azGUNb8X0Oq52/OiBSKJjBo1Chs2bDD62muvvYYVK1ZUc0XSYBAjIskpHpzKxLly61LnFuD8LTXOiY8MXE7Nwt3sAvx2+TZ+u3xbbGtno0BTb0e0fODoWSMvB45VI6oGM2fOxIQJE4y+9jgNHWIQIyKL4qSyRlD9WgiqX0tcllugxaWUTDGYnbulxvkkNbLztTiVkI5TCeliWyu5DI28HO8Fs6Jw1tzXCQ5K/nNJZEqenp7w9PSUugzJ8V8WIrJ4KmsFWtdxQes6LuIyrU5A/O0svSNnZ2+qkZFTgPNJRUFt28n766jnbn/vlOb9cWeVneKELM9jdv0blaA8vwcMYkT0WFLIZWjo6YiGno4Y0KY2gKJ/PG+m54jhLO7e0bOkjFzE39Yg/rYGu/9JEtfh5aQUQ1lxQKvjasuLAh5DxbPHZ2dnP3JGd7J82dnZAAzvKmAMgxgR0T0ymQx1XO1Qx9UOIS28xeX/ZeUhLkl/3Fn8bQ1S1HlIUafi0IX7981zUlndD2e1i8JZfXd7WCl4eypLplAo4OLiIt5D0c7OjoHcxIpn/8/NzTXb6SsEQUB2djZSU1Ph4uJSpjo5jxgRUQVo8gpxPkmtN+7sUkqmOGHug5RWcjT1cdI7ctbU2xEqa/PcmVDFCIKA5ORkpKenS12KRdLpdEhMTISfnx/kcvP+j42Liwu8vb3LFMYZxIiITCS/UId/UzPvndZU4+zNDJxPUovzpD1IIZehoYeD3nQazX2d4GzLyWhrOq1Wi4KCyt9blvRlZWWhQ4cO+Ouvv+Dg4CB1OSWytrYu1xE7BjEioiqk0wm49p9G77Rm3C11ifcL9XOzRQuf+6c2W/o6w9OJM7UTWer+m0GMiKiaCYKAFHWeeErz7M2irzfTc4y2d3dQ6p3WbOHrhLpudpBzMlp6jFjq/ptBjIjITKRn5yPulv64sytpWUbv6emgtEJzHye9KTUaeTnAmhcFkIWy1P03gxgRkRnLydfiQrJ+OLuQnIn8Qp1BWxuFHE28709G29zXGc18HGFnwwvkqeaz1P03gxgRUQ1ToNXhSloWzt3UH3eWmVdo0FYuK5qMtviUZvHtnFzsKnfTd6LqZqn7bwYxIiILIAgCEu/k4OytDPHI2blbaqRl5hltX9vF1uBOAT7OKs59RWbLUvffDGJERBYsNTNXnE6jOKBd/y/baFtXO2sxlDW/d/SsXi17XhRAZsFS998MYkREjxl1bgHO31Lj7AOnNf9NzYLWyFUBdjYKNHtoMtpGXg5QWnEyWqpelrr/ZhAjIiLkFmhxKSVT76KA80lq5BYYXhRgrSi6T2cLXye09HVCi9rOaObjBAclLwqgqmOp+28GMSIiMkqrE3A1LUsvnJ27pUZGjuGs8TIZEFDL3mDcmbuDUoLKyRJZ6v6bQYyIiMpMEATcTM8RQ1ncvYCWlJFrtL2Xk/L+TdDvfa3jasuLAqjcLHX/zSBGRESV9l9WnsFtnOL/08DYHsbZ1hrNi8ed1S4KaPXd7WHFyWipFJa6/2YQIyKiKpGVV4gLSfqT0V5KyUSB1nC3o7KWo4n3vTFn946cNfF2hMqaFwVQEUvdfzOIERFRtckv1OFSSqbedBrnk9TQ5GsN2irkMjT0cBCn02jh64zmvk5wtrWWoHKSmqXuvxnEiIhIUjqdgGv/afRObZ67pcYdTb7R9n5utmjh44yWte8fPfN0UlVz1VTdLHX/zSBGRERmRxAEJKtz9W7jdO6WGjfTc4y2d3dQ6s111sLXCXXd7DgZrQWx1P03gxgREdUY6dn5905rFoWzs7fUuJqWBSNz0cJRaYVmD4Wzhp4OsOZFATWSpe6/GcSIiKhGy8nX4nyy/nQaF5IzkV9oOBmtjZUcTbwcxaNnzX2d0czHEXY2nIzW3Fnq/ptBjIiILE6BVocraVk4d1N970boapy/pUZmXqFBW7kMqOduj5a19ec7c7GzkaByKoml7r8ZxIiI6LGg0wlIvJttcKeAtMw8o+1ru9ga3CnAx1nFyWglYqn7bwYxIiJ6rKVm5haFspv3w1nCnWyjbd3sbYomo33gis16tex5UUA1sNT9N4MYERHRQ9S5BXoXBcTdUuPf1CxojVwVYGejQLPiOwXcO3rWyMsBSitORmtKlrr/ljSIHT16FHPnzsXJkyeRlJSEnTt3YuDAgSW237FjB5YvX47Y2Fjk5eWhRYsWmD59OkJCQsq8TUv9IImIqGrlFmhxKSUTZ2/eP7V5IVmN3ALDiwKsFTI08rx/UUCL2s5o5uMEByUvCqgoS91/S/obodFoEBgYiDfeeAODBg16ZPujR4+iZ8+emD17NlxcXLBmzRqEhobi+PHjaNu2bTVUTEREjyuVtQKt67igdR0XcVmhVof42xqDcWcZOQWIS1IjLkmNrSeL2spkQEAte4NxZ+4OSmk6RGbBbE5NymSyRx4RM6ZFixYICwvD1KlTy9TeUhM1ERGZB0EQcONujt50GuduqZGszjXa3stJiZb3Qlnze1/ruNryooCHWOr+u0YfI9XpdMjMzISbm1uJbfLy8pCXd/+KGLVaXR2lERHRY0omk8HPzQ5+bnbo3dJbXP5fVp7ebZzibqlx9bYGKeo8pKhTcfBCqtjW2da66KIA3/sXBtR3t4cVJ6O1ODU6iM2bNw9ZWVl4+eWXS2wTHR2NGTNmVGNVREREhmo5KPF0Yw883dhDXJaVV4jzSfpXbP6bmomMnALEXP0PMVf/E9uqrOVo6q1/WrOJtyNU1rwooCarsacmN27ciBEjRuD7779HcHBwie2MHRHz8/OzuEObRERkGfILdbiUknnvqs2igBaXpEZ2vtagrUIuQ0MPh3unNYsCWnNfJzjbWktQedXiqUkz8t1332H48OHYunVrqSEMAJRKJZRKDoQkIqKawcZKjpa1ndGytjMAPwCAVifg2n8avdOa526pcUeTj4spmbiYkokdp2+K66jrZmdwE3RPJ5VEPaLS1LggtmnTJrzxxhv47rvv0K9fP6nLISIiqnIKuQwNPBzQwMMBzwX6Aii6KCBZnYtzN9V6V23eTM9Bwp1sJNzJxk9nk8V1uDso9cJZy9pO8HO142S0EpM0iGVlZeHy5cvi8/j4eMTGxsLNzQ1169ZFVFQUbt68ifXr1wMoOh0ZHh6OxYsXIygoCMnJRb9gtra2cHZ2lqQPREREUpDJZPBxtoWPsy2Cm3uJy+9q8hGXpD+dxtW0LNzOysORS2k4cilNbOuotEKzh46cNfR0gDUvCqg2ko4RO3z4MHr06GGwPDw8HGvXrkVERASuXbuGw4cPAwC6d++OI0eOlNi+LCz1HDMREVFJsvMLcSE5U+9WTheTM5GvNZyM1sZKjiZe9yejbe7rjGY+jrCzkfYkmqXuv81msH51sdQPkoiIqDwKtDpcTs3SO615/pYamXmFBm3lMqD+vYsCHjx65mJnU231Wur+m0GMiIiIAAA6nYDEu9k4d0uNsw9MqXE7K89o+9outnp3CmhZ2wneTqoqmYzWUvffDGJERERUqlR1rsFtnBLuZBtt62Zvgxa+Tlgd8YRJx5pZ6v67xl01SURERNXL00kFTycVejT1FJdl5BTgfFLRkbPi6TQup2XhjiYf1//L5oD/MmIQIyIionJztrXGk/Vr4cn6tcRluQVaXEzOhDq3QMLKahYGMSIiIjIJlbUCgX4uUpdRo/C4IREREZFEGMSIiIiIJMIgRkRERCQRBjEiIiIiiTCIEREREUmEQYyIiIhIIgxiRERERBJhECMiIiKSCIMYERERkUQYxIiIiIgkwiBGREREJBEGMSIiIiKJMIgRERERSYRBjIiIiEgiDGJEREREEmEQIyIiIpIIgxgRERGRRBjEiIiIiCTCIEZEREQkEQYxIiIiIokwiBERERFJhEGMiIiISCIMYkREREQSYRAjIiIikgiDGBEREZFEGMSIiIiIJMIgRkRERCQRSYPY0aNHERoaCl9fX8hkMuzatavU9klJSXjllVfQuHFjyOVyjB8/vlrqJCIiIqoKkgYxjUaDwMBALFu2rEzt8/Ly4OHhgcmTJyMwMLCKqyMiIiKqWlZSbrxPnz7o06dPmdsHBARg8eLFAIDVq1dXVVlERERE1YJjxIiIiIgkIukRseqQl5eHvLw88blarZawGiIiIqL7LP6IWHR0NJydncWHn5+f1CURERERAXgMglhUVBQyMjLER2JiotQlEREREQF4DE5NKpVKKJVKqcsgIiIiMiBpEMvKysLly5fF5/Hx8YiNjYWbmxvq1q2LqKgo3Lx5E+vXrxfbxMbGiu9NS0tDbGwsbGxs0Lx58+oun4iIiKhSZIIgCFJt/PDhw+jRo4fB8vDwcKxduxYRERG4du0aDh8+LL4mk8kM2vv7++PatWtl2qZarYazszMyMjLg5ORU0dKJiIioGlnq/lvSICYFS/0giYiILJml7r8tfrA+ERERkbliECMiIiKSCIMYERERkUQYxIiIiIgkwiBGREREJBEGMSIiIiKJMIgRERERSYRBjIiIiEgiDGJEREREEmEQIyIiIpIIgxgRERGRRBjEiIiIiCTCIEZEREQkEQYxIiIiIokwiBERERFJhEGMiIiISCIMYkREREQSYRAjIiIikgiDGBEREZFEGMSIiIiIJMIgRkRERCQRBjEiIiIiiVhV9I0ajQZHjhxBQkIC8vPz9V579913K10YERERkaWrUBA7ffo0+vbti+zsbGg0Gri5ueH27duws7ODp6cngxgRERFRGVTo1OR7772H0NBQ3L17F7a2tvjjjz9w/fp1tG/fHvPmzTN1jUREREQWqUJBLDY2Fu+//z7kcjkUCgXy8vLg5+eHzz77DJMmTTJ1jUREREQWqUJBzNraGnJ50Vs9PT2RkJAAAHB2dkZiYqLpqiMiIiKyYBUaI9a2bVv8+eefaNSoEbp164apU6fi9u3b+Oabb9CyZUtT10hERERkkSp0RGz27Nnw8fEBAHzyySdwdXXF22+/jbS0NKxcudKkBRIRERFZKpkgCILURVQntVoNZ2dnZGRkwMnJSepyiIiIqAwsdf/NCV2JiIiIJFLmMWLt2rXDwYMH4erqirZt20Imk5XY9tSpUyYpjoiIiMiSlTmIDRgwAEqlEgAwcODAqqqHiIiI6LEh6Rixo0ePYu7cuTh58iSSkpKwc+fOR4a8w4cPIzIyEufOnYOfnx8mT56MiIiIMm/TUs8xExERWTJL3X9XaIzYn3/+iePHjxssP378OP76668yr0ej0SAwMBDLli0rU/v4+Hj069cPPXr0QGxsLMaPH4/hw4fj559/LvM2iYiIiMxFhYLY6NGjjU7cevPmTYwePbrM6+nTpw8+/vhjPP/882Vqv2LFCtSrVw/z589Hs2bNMGbMGLz44otYuHBhmbdJREREZC4qFMTi4uLQrl07g+Vt27ZFXFxcpYsqSUxMDIKDg/WWhYSEICYmpsT35OXlQa1W6z2IiIiIzEGFgphSqURKSorB8qSkJFhZVWiy/jJJTk6Gl5eX3jIvLy+o1Wrk5OQYfU90dDScnZ3Fh5+fX5XVR0RERFQeFQpivXr1QlRUFDIyMsRl6enpmDRpEnr27Gmy4kyhuM7iB++FSUREROaiQoev5s2bh6effhr+/v5o27YtACA2NhZeXl745ptvTFrgg7y9vQ2OxKWkpMDJyQm2trZG36NUKsVpN4iIiIjMSYWCWO3atfHPP//g22+/xd9//w1bW1sMGzYMQ4YMgbW1talrFHXq1Al79uzRW7Z//3506tSpyrZJREREVFUqPKDL3t4eI0eOrNTGs7KycPnyZfF5fHw8YmNj4ebmhrp16yIqKgo3b97E+vXrAQCjRo3C0qVL8cEHH+CNN97AoUOHsGXLFuzevbtSdRARERFJoVIj6+Pi4pCQkID8/Hy95c8991yZ3v/XX3+hR48e4vPIyEgAQHh4ONauXYukpCQkJCSIr9erVw+7d+/Ge++9h8WLF6NOnTr4+uuvERISUpluEBEREUmiQjPrX716Fc8//zzOnDkDmUyG4lUU339Sq9WatkoTstSZeYmIiCyZpe6/K3TV5Lhx41CvXj2kpqbCzs4O586dw9GjR9GhQwccPnzYxCUSERERWaYKnZqMiYnBoUOH4O7uDrlcDrlcjqeeegrR0dF49913cfr0aVPXSURERGRxKnRETKvVwtHREQDg7u6OW7duAQD8/f1x8eJF01VHREREZMEqdESsZcuW+Pvvv1GvXj0EBQXhs88+g42NDVauXIn69eubukYiIiIii1ShIDZ58mRoNBoAwMyZM9G/f3907doVtWrVwubNm01aIBEREZGlqtBVk8bcuXMHrq6u4pWT5spSr7ogIiKyZJa6/y73GLGCggJYWVnh7Nmzesvd3NzMPoQRERERmZNyBzFra2vUrVvXrOcKIyIiIqoJKnTV5EcffYRJkybhzp07pq6HiIiI6LFRocH6S5cuxeXLl+Hr6wt/f3/Y29vrvX7q1CmTFEdERERkySoUxAYOHGjiMoiIiIgePya7arKmsNSrLoiIiCyZpe6/KzRGjIiIiIgqr0KnJuVyealTVfCKSiIiIqJHq1AQ27lzp97zgoICnD59GuvWrcOMGTNMUhgRERGRpTPpGLGNGzdi8+bN+P777021SpOz1HPMRERElsxS998mHSP25JNP4uDBg6ZcJREREZHFMlkQy8nJweeff47atWubapVEREREFq1CY8Qevrm3IAjIzMyEnZ0dNmzYYLLiiIiIiCxZhYLYwoUL9YKYXC6Hh4cHgoKC4OrqarLiiIiIiCxZhYJYRESEicsgIiIievxUaIzYmjVrsHXrVoPlW7duxbp16ypdFBEREdHjoEJBLDo6Gu7u7gbLPT09MXv27EoXRURERPQ4qFAQS0hIQL169QyW+/v7IyEhodJFERERET0OKhTEPD098c8//xgs//vvv1GrVq1KF0VERET0OKhQEBsyZAjeffdd/PLLL9BqtdBqtTh06BDGjRuHwYMHm7pGIiIiIotUoasmZ82ahWvXruHZZ5+FlVXRKnQ6HYYOHYpPPvnEpAUSERERWapK3Wvy33//RWxsLGxtbdGqVSv4+/ubsrYqYan3qiIiIrJklrr/rtARsWKNGjVCo0aNABT9gJYvX45Vq1bhr7/+MklxRERERJasUkEMAH755ResXr0aO3bsgLOzM55//nlT1EVERERk8SoUxG7evIm1a9dizZo1SE9Px927d7Fx40a8/PLLerc+IiIiIqKSleuqye3bt6Nv375o0qQJYmNjMX/+fNy6dQtyuRytWrViCCMiIiIqh3IdEQsLC8OHH36IzZs3w9HRsapqIiIiInoslOuI2Jtvvolly5ahd+/eWLFiBe7evWuSIpYtW4aAgACoVCoEBQXhxIkTJbYtKCjAzJkz0aBBA6hUKgQGBmLv3r0mqYOIiIioOpUriH355ZdISkrCyJEjsWnTJvj4+GDAgAEQBAE6na5CBWzevBmRkZGYNm0aTp06hcDAQISEhCA1NdVo+8mTJ+PLL7/EkiVLEBcXh1GjRuH555/H6dOnK7R9IiIiIqlUeh6xNWvWYN26dcjKykK/fv3w4osvYtCgQWVeR1BQEJ544gksXboUQNHEsH5+fhg7diwmTpxo0N7X1xcfffQRRo8eLS574YUXYGtriw0bNjxye5Y6DwkREZEls9T9d4VucVSsUaNGmD17NhITE7FhwwZkZ2djyJAhZX5/fn4+Tp48ieDg4PsFyeUIDg5GTEyM0ffk5eVBpVLpLbO1tcVvv/1WYnu1Wq33ICIiIjIHlQpi4krkcoSGhmLXrl1ITEws8/tu374NrVYLLy8vveVeXl5ITk42+p6QkBAsWLAA//77L3Q6Hfbv348dO3YgKSnJaPvo6Gg4OzuLDz8/v7J3jIiIiKgKmSSIPcjT09PUq9SzePFiNGrUCE2bNoWNjQ3GjBmDYcOGQS433pWoqChkZGSIj/IERSIiIqKqZPIgVh7u7u5QKBRISUnRW56SkgJvb2+j7/Hw8MCuXbug0Whw/fp1XLhwAQ4ODqhfv77R9kqlEk5OTnoPIiIiInMgaRCzsbFB+/btcfDgQXGZTqfDwYMH0alTp1Lfq1KpULt2bRQWFmL79u0YMGBAVZdLREREZFLlCmJXr141eQGRkZH46quvsG7dOpw/fx5vv/02NBoNhg0bBgAYOnQooqKixPbHjx/Hjh07cPXqVfz666/o3bs3dDodPvjgA5PXRkRERFSVyjWzfuvWrREQEIDnnnsOAwYMQFBQUKULCAsLQ1paGqZOnYrk5GS0adMGe/fuFQfwJyQk6I3/ys3NxeTJk3H16lU4ODigb9+++Oabb+Di4lLpWoiIiIiqU7nmEcvNzcX+/fvx/fff48cff4RMJkP//v3x3HPPoWfPngbTSpgjS52HhIiIyJJZ6v67whO6CoKAmJgY/PDDD/jhhx+QkJCA4OBgPPfccwgNDYWHh4epazUJS/0giYiILJml7r8rPFhfJpOhc+fOmDNnDuLi4nD69Gl07doVa9euRZ06dbBs2TJT1klERERkcSp1i6OS/Pfff7hz5w4aNWpk6lVXmqUmaiIiIktmqfvvcg3WL6tatWqhVq1aVbFqIiIiIosh6TxiRERERI8zBjEiIiIiiTCIEREREUmkQkEsMTERN27cEJ+fOHEC48ePx8qVK01WGBEREZGlq1AQe+WVV/DLL78AAJKTk9GzZ0+cOHECH330EWbOnGnSAomIiIgsVYWC2NmzZ9GxY0cAwJYtW9CyZUv8/vvv+Pbbb7F27VpT1kdERERksSoUxAoKCqBUKgEABw4cwHPPPQcAaNq0KZKSkkxXHREREZEFq1AQa9GiBVasWIFff/0V+/fvR+/evQEAt27d4vxhRERERGVUoSD26aef4ssvv0T37t0xZMgQBAYGAgB++OEH8ZQlEREREZWuwrc40mq1UKvVcHV1FZddu3YNdnZ28PT0NFmBpmapt0ggIiKyZJa6/67QEbGcnBzk5eWJIez69etYtGgRLl68aNYhjIiIiMicVCiIDRgwAOvXrwcApKenIygoCPPnz8fAgQOxfPlykxZIREREZKkqFMROnTqFrl27AgC2bdsGLy8vXL9+HevXr8fnn39u0gKJiIiILFWFglh2djYcHR0BAPv27cOgQYMgl8vx5JNP4vr16yYtkIiIiMhSVSiINWzYELt27UJiYiJ+/vln9OrVCwCQmppqUQPoiIiIiKpShYLY1KlTMWHCBAQEBKBjx47o1KkTgKKjY23btjVpgURERESWqsLTVyQnJyMpKQmBgYGQy4vy3IkTJ+Dk5ISmTZuatEhTstTLX4mIiCyZpe6/rSr6Rm9vb3h7e+PGjRsAgDp16nAyVyIiIqJyqNCpSZ1Oh5kzZ8LZ2Rn+/v7w9/eHi4sLZs2aBZ1OZ+oaiYiIiCxShY6IffTRR1i1ahXmzJmDLl26AAB+++03TJ8+Hbm5ufjkk09MWiQRERGRJarQGDFfX1+sWLECzz33nN7y77//Hu+88w5u3rxpsgJNzVLPMRMREVkyS91/V+jU5J07d4wOyG/atCnu3LlT6aKIiIiIHgcVCmKBgYFYunSpwfKlS5ciMDCw0kURERERPQ4qNEbss88+Q79+/XDgwAFxDrGYmBgkJiZiz549Ji2QiIiIyFJV6IhYt27dcOnSJTz//PNIT09Heno6Bg0ahIsXL4r3oCQiIiKi0lV4Qldjbty4gZkzZ2LlypWmWqXJWepgPyIiIktmqfvvCh0RK8l///2HVatWmXKVRERERBbLpEGMiIiIiMqOQYyIiIhIImYRxJYtW4aAgACoVCoEBQXhxIkTpbZftGgRmjRpAltbW/j5+eG9995Dbm5uNVVLREREZBrlmr5i0KBBpb6enp5e7gI2b96MyMhIrFixAkFBQVi0aBFCQkJw8eJFeHp6GrTfuHEjJk6ciNWrV6Nz5864dOkSIiIiIJPJsGDBgnJvn4iIiEgq5Qpizs7Oj3x96NCh5SpgwYIFGDFiBIYNGwYAWLFiBXbv3o3Vq1dj4sSJBu1///13dOnSBa+88goAICAgAEOGDMHx48fLtV0iIiIiqZUriK1Zs8akG8/Pz8fJkycRFRUlLpPL5QgODkZMTIzR93Tu3BkbNmzAiRMn0LFjR1y9ehV79uzB66+/brR9Xl4e8vLyxOdqtdqkfSAiIiKqqArNrG8qt2/fhlarhZeXl95yLy8vXLhwweh7XnnlFdy+fRtPPfUUBEFAYWEhRo0ahUmTJhltHx0djRkzZpi8diIiIqLKMovB+uVx+PBhzJ49G1988QVOnTqFHTt2YPfu3Zg1a5bR9lFRUcjIyBAfiYmJ1VwxERERkXGSHhFzd3eHQqFASkqK3vKUlBR4e3sbfc+UKVPw+uuvY/jw4QCAVq1aQaPRYOTIkfjoo48gl+tnS6VSCaVSWTUdICIiIqoESY+I2djYoH379jh48KC4TKfT4eDBg+LNxB+WnZ1tELYUCgUAwIR3ayIiIiKqcpIeEQOAyMhIhIeHo0OHDujYsSMWLVoEjUYjXkU5dOhQ1K5dG9HR0QCA0NBQLFiwAG3btkVQUBAuX76MKVOmIDQ0VAxkRERERDWB5EEsLCwMaWlpmDp1KpKTk9GmTRvs3btXHMCfkJCgdwRs8uTJkMlkmDx5Mm7evAkPDw+Ehobik08+kaoLRERERBUiEx6z83mWevd2IiIiS2ap++8ad9UkERERkaVgECMiIiKSCIMYERERkUQYxIiIiIgkwiBGREREJBEGMSIiIiKJMIgRERERSYRBjIiIiEgiDGJEREREEmEQIyIiIpIIgxgRERGRRBjEiIiIiCTCIEZEREQkEQYxIiIiIokwiBERERFJhEGMiIiISCIMYkREREQSYRAjIiIikgiDGBEREZFEGMSIiIiIJMIgRkRERCQRBjEiIiIiiTCIEREREUmEQYyIiIhIIgxiRERERBJhECMiIiKSCIMYERERkUQYxIiIiIgkwiBGREREJBEGMSIiIiKJMIgRERERSYRBjIiIiEgiZhHEli1bhoCAAKhUKgQFBeHEiRMltu3evTtkMpnBo1+/ftVYMREREVHlSR7ENm/ejMjISEybNg2nTp1CYGAgQkJCkJqaarT9jh07kJSUJD7Onj0LhUKBl156qZorJyIiIqocyYPYggULMGLECAwbNgzNmzfHihUrYGdnh9WrVxtt7+bmBm9vb/Gxf/9+2NnZMYgRERFRjSNpEMvPz8fJkycRHBwsLpPL5QgODkZMTEyZ1rFq1SoMHjwY9vb2VVUmERERUZWwknLjt2/fhlarhZeXl95yLy8vXLhw4ZHvP3HiBM6ePYtVq1aV2CYvLw95eXnic7VaXfGCiYiIiExI8lOTlbFq1Sq0atUKHTt2LLFNdHQ0nJ2dxYefn181VkhERERUMkmDmLu7OxQKBVJSUvSWp6SkwNvbu9T3ajQafPfdd3jzzTdLbRcVFYWMjAzxkZiYWOm6iYiIiExB0iBmY2OD9u3b4+DBg+IynU6HgwcPolOnTqW+d+vWrcjLy8Nrr71WajulUgknJye9BxEREZE5kHSMGABERkYiPDwcHTp0QMeOHbFo0SJoNBoMGzYMADB06FDUrl0b0dHReu9btWoVBg4ciFq1aklRNhEREVGlSR7EwsLCkJaWhqlTpyI5ORlt2rTB3r17xQH8CQkJkMv1D9xdvHgRv/32G/bt2ydFyUREREQmIRMEQZC6iOqkVqvh7OyMjIwMnqYkIiKqISx1/12jr5okIiIiqskYxIiIiIgkwiBGREREJBEGMSIiIiKJMIgRERERSYRBjIiIiEgiDGJEREREEmEQIyIiIpIIgxgRERGRRBjEiIiIiCTCIEZEREQkEQYxIiIiIokwiBERERFJhEGMiIiISCIMYkREREQSYRAjIiIikgiDGBEREZFEGMSIiIiIJMIgRkRERCQRBjEiIiIiiTCIEREREUmEQYyIiIhIIgxiRERERBJhECMiIiKSCIMYERERkUQYxIiIiIgkwiBGREREJBEGMSIiIiKJMIgRERERSYRBjIiIiEgiDGJEREREEmEQIyIiIpIIgxgRERGRRMwiiC1btgwBAQFQqVQICgrCiRMnSm2fnp6O0aNHw8fHB0qlEo0bN8aePXuqqVoiIiIi07CSuoDNmzcjMjISK1asQFBQEBYtWoSQkBBcvHgRnp6eBu3z8/PRs2dPeHp6Ytu2bahduzauX78OFxeX6i+eiIiIqBJkgiAIUhYQFBSEJ554AkuXLgUA6HQ6+Pn5YezYsZg4caJB+xUrVmDu3Lm4cOECrK2ty709tVoNZ2dnZGRkwMnJqdL1ExERUdWz1P23pKcm8/PzcfLkSQQHB4vL5HI5goODERMTY/Q9P/zwAzp16oTRo0fDy8sLLVu2xOzZs6HVao22z8vLg1qt1nsQERERmQNJg9jt27eh1Wrh5eWlt9zLywvJyclG33P16lVs27YNWq0We/bswZQpUzB//nx8/PHHRttHR0fD2dlZfPj5+Zm8H0REREQVYRaD9ctDp9PB09MTK1euRPv27REWFoaPPvoIK1asMNo+KioKGRkZ4iMxMbGaKyYiIiIyTtLB+u7u7lAoFEhJSdFbnpKSAm9vb6Pv8fHxgbW1NRQKhbisWbNmSE5ORn5+PmxsbPTaK5VKKJVK0xdPREREVEmSHhGzsbFB+/btcfDgQXGZTqfDwYMH0alTJ6Pv6dKlCy5fvgydTicuu3TpEnx8fAxCGBEREZE5k/zUZGRkJL766iusW7cO58+fx9tvvw2NRoNhw4YBAIYOHYqoqCix/dtvv407d+5g3LhxuHTpEnbv3o3Zs2dj9OjRUnWBiIiIqEIkn0csLCwMaWlpmDp1KpKTk9GmTRvs3btXHMCfkJAAufx+XvTz88PPP/+M9957D61bt0bt2rUxbtw4fPjhh1J1gYiIiKhCJJ9HrLpZ6jwkRERElsxS99+Sn5okIiIielwxiBERERFJhEGMiIiISCIMYkREREQSYRAjIiIikojk01cQERFRDSIIQL4GyFMDeZlFj9yM+9/nZQIKa6DjCKkrrREYxIiIiB4HggAUZN8LTsUhSq0fqAxC1YNh6973+ZmAoCt9W44+DGJlxCBGRERkzgQBKMh5IBw9dPSppFCVqzYMVILWdHXJFIDSEVA6ASqne9/fe27vYbrtWDgGMSIioqogCEBh3gPhSP1QOHooVBkEpwee6wpNV5dMfj8w6X2991A5GX9Nb7kjYG0HyGSmq+sxxSBGRET0sMI842OfSgxVav2vxa/pCkxYlKyU4FS8/OHgZCRw2dgzQJkRBjEiIrIchfkVHPv0UKjS5pu2LpuHw5GjYXgyeM1ZP1RZ2wNyTnZgaRjEiIhIetqCkgeHlzj26eGvmUBhrmnrsnEwEpxKGBf14NcHX7NxZICiEjGIERFRxWkLi66iMxj79EB4Km3sU/FrhTmmrcva7hFjnB4+vedsJFQ5AnKFaesiegiDGBHR40inrdzYp+JlBdmmrcvK9tFjnx41oNzGEVBw90Y1A39TiYhqEp0OyM8qZYzTQ0HJaKjKLFqHKSmURk7VGRs4bmTsU3F7GwfAysa0dRGZOQYxIqLqoNMBBZqKj316cDJNU5JbGz+6VNL4J6MDyh0AK6Vp6yJ6TDCIERGVRrydSwXHPj0YqCCYri651UOhyVhwenjsk5FQxQBFJCkGMSKyTOJs5KVNpllaqHpwNvJH3M6lPGSKkud3euS4qAdClZWKc0ERWQAGMaKHCcK9h+4RD1O0KeF1PLy8guuprnr1Xq+u7ZTwmq7wfqAy6e1c5MYDUkljn0oKVda2DFBEJGIQMxWdFlDfLOOORCjjzsacdrC6on5KuYOt8DrKGWpMefqIzICslODkpP+1tAHlvJ0LEVUBBjFT0aQBi1pJXQVJQlZ0tKTEx6NeRyXfX5Y2lX39oTYl9tnE26lIG7lC/6gUZyMnIjPGIGYqMvm9MRvVveOqyh2frJq2Y8L+iD/n6qyXR0mIiKhiGMRMxcETmJwidRVERERUg/B4PREREZFEGMSIiIiIJMIgRkRERCQRBjEiIiIiiTCIEREREUmEQYyIiIhIIgxiRERERBJhECMiIiKSCIMYERERkUQYxIiIiIgkYhZBbNmyZQgICIBKpUJQUBBOnDhRYtu1a9dCJpPpPVQqVTVWS0RERGQakgexzZs3IzIyEtOmTcOpU6cQGBiIkJAQpKamlvgeJycnJCUliY/r169XY8VEREREpiF5EFuwYAFGjBiBYcOGoXnz5lixYgXs7OywevXqEt8jk8ng7e0tPry8vKqxYiIiIiLTkDSI5efn4+TJkwgODhaXyeVyBAcHIyYmpsT3ZWVlwd/fH35+fhgwYADOnTtXYtu8vDyo1Wq9BxEREZE5sJJy47dv34ZWqzU4ouXl5YULFy4YfU+TJk2wevVqtG7dGhkZGZg3bx46d+6Mc+fOoU6dOgbto6OjMWPGDIPlDGREREQ1R/F+WxAEiSsxLUmDWEV06tQJnTp1Ep937twZzZo1w5dffolZs2YZtI+KikJkZKT4/ObNm2jevDn8/PyqpV4iIiIynczMTDg7O0tdhslIGsTc3d2hUCiQkpKitzwlJQXe3t5lWoe1tTXatm2Ly5cvG31dqVRCqVSKzx0cHJCYmAhHR0fIZLKKF2+EWq2Gn58fEhMT4eTkZNJ1mwNL7x9g+X1k/2o+S+8j+1fzVVUfBUFAZmYmfH19TbZOcyBpELOxsUH79u1x8OBBDBw4EACg0+lw8OBBjBkzpkzr0Gq1OHPmDPr27Vum9nK53OgpTFNycnKy2D8wwPL7B1h+H9m/ms/S+8j+1XxV0UdLOhJWTPJTk5GRkQgPD0eHDh3QsWNHLFq0CBqNBsOGDQMADB06FLVr10Z0dDQAYObMmXjyySfRsGFDpKenY+7cubh+/TqGDx8uZTeIiIiIyk3yIBYWFoa0tDRMnToVycnJaNOmDfbu3SsO4E9ISIBcfv/izrt372LEiBFITk6Gq6sr2rdvj99//x3NmzeXqgtEREREFSJ5EAOAMWPGlHgq8vDhw3rPFy5ciIULF1ZDVeWnVCoxbdo0vTFplsTS+wdYfh/Zv5rP0vvI/tV8j0MfTUkmWNp1oEREREQ1hOQz6xMRERE9rhjEiIiIiCTCIEZEREQkEQYxIiIiIokwiD3CsmXLEBAQAJVKhaCgIJw4caLU9lu3bkXTpk2hUqnQqlUr7NmzR+91QRAwdepU+Pj4wNbWFsHBwfj333+rsgulKk//vvrqK3Tt2hWurq5wdXVFcHCwQfuIiAjIZDK9R+/evau6GyUqT//Wrl1rULtKpdJrY26fH1C+Pnbv3t2gjzKZDP369RPbmNNnePToUYSGhsLX1xcymQy7du165HsOHz6Mdu3aQalUomHDhli7dq1Bm/L+XVeV8vZvx44d6NmzJzw8PODk5IROnTrh559/1mszffp0g8+vadOmVdiLkpW3f4cPHzb6+5mcnKzXzlw+P6D8fTT29yWTydCiRQuxjTl9htHR0XjiiSfg6OgIT09PDBw4EBcvXnzk+2ravlBKDGKl2Lx5MyIjIzFt2jScOnUKgYGBCAkJQWpqqtH2v//+O4YMGYI333wTp0+fxsCBAzFw4ECcPXtWbPPZZ5/h888/x4oVK3D8+HHY29sjJCQEubm51dUtUXn7d/jwYQwZMgS//PILYmJi4Ofnh169euHmzZt67Xr37o2kpCTxsWnTpurojoHy9g8omgn6wdqvX7+u97o5fX5A+fu4Y8cOvf6dPXsWCoUCL730kl47c/kMNRoNAgMDsWzZsjK1j4+PR79+/dCjRw/ExsZi/PjxGD58uF5YqcjvRVUpb/+OHj2Knj17Ys+ePTh58iR69OiB0NBQnD59Wq9dixYt9D6/3377rSrKf6Ty9q/YxYsX9er39PQUXzOnzw8ofx8XL16s17fExES4ubkZ/A2ay2d45MgRjB49Gn/88Qf279+PgoIC9OrVCxqNpsT31LR9oeQEKlHHjh2F0aNHi8+1Wq3g6+srREdHG23/8ssvC/369dNbFhQUJLz11luCIAiCTqcTvL29hblz54qvp6enC0qlUti0aVMV9KB05e3fwwoLCwVHR0dh3bp14rLw8HBhwIABpi61QsrbvzVr1gjOzs4lrs/cPj9BqPxnuHDhQsHR0VHIysoSl5nTZ/ggAMLOnTtLbfPBBx8ILVq00FsWFhYmhISEiM8r+zOrKmXpnzHNmzcXZsyYIT6fNm2aEBgYaLrCTKQs/fvll18EAMLdu3dLbGOun58gVOwz3LlzpyCTyYRr166Jy8z1MxQEQUhNTRUACEeOHCmxTU3bF0qNR8RKkJ+fj5MnTyI4OFhcJpfLERwcjJiYGKPviYmJ0WsPACEhIWL7+Ph4JCcn67VxdnZGUFBQieusKhXp38Oys7NRUFAANzc3veWHDx+Gp6cnmjRpgrfffhv//fefSWsvi4r2LysrC/7+/vDz88OAAQNw7tw58TVz+vwA03yGq1atwuDBg2Fvb6+33Bw+w4p41N+gKX5m5kSn0yEzM9Pgb/Dff/+Fr68v6tevj1dffRUJCQkSVVgxbdq0gY+PD3r27Iljx46Jyy3t8wOK/gaDg4Ph7++vt9xcP8OMjAwAMPide1BN2heaAwaxEty+fRtarVa81VIxLy8vg/EKxZKTk0ttX/y1POusKhXp38M+/PBD+Pr66v0x9e7dG+vXr8fBgwfx6aef4siRI+jTpw+0Wq1J63+UivSvSZMmWL16Nb7//nts2LABOp0OnTt3xo0bNwCY1+cHVP4zPHHiBM6ePWtwn1Zz+QwroqS/QbVajZycHJP83puTefPmISsrCy+//LK4LCgoCGvXrsXevXuxfPlyxMfHo2vXrsjMzJSw0rLx8fHBihUrsH37dmzfvh1+fn7o3r07Tp06BcA0/26Zk1u3buGnn34y+Bs0189Qp9Nh/Pjx6NKlC1q2bFliu5q0LzQHZnGLI6p55syZg++++w6HDx/WG9A+ePBg8ftWrVqhdevWaNCgAQ4fPoxnn31WilLLrFOnTujUqZP4vHPnzmjWrBm+/PJLzJo1S8LKqsaqVavQqlUrdOzYUW95Tf4MHycbN27EjBkz8P333+uNoerTp4/4fevWrREUFAR/f39s2bIFb775phSlllmTJk3QpEkT8Xnnzp1x5coVLFy4EN98842ElVWNdevWwcXFBQMHDtRbbq6f4ejRo3H27FnJxqtZKh4RK4G7uzsUCgVSUlL0lqekpMDb29voe7y9vUttX/y1POusKhXpX7F58+Zhzpw52LdvH1q3bl1q2/r168Pd3R2XL1+udM3lUZn+FbO2tkbbtm3F2s3p8wMq10eNRoPvvvuuTP+oS/UZVkRJf4NOTk6wtbU1ye+FOfjuu+8wfPhwbNmyxeAU0MNcXFzQuHHjGvH5GdOxY0exdkv5/ICiqwZXr16N119/HTY2NqW2NYfPcMyYMfjxxx/xyy+/oE6dOqW2rUn7QnPAIFYCGxsbtG/fHgcPHhSX6XQ6HDx4UO+oyYM6deqk1x4A9u/fL7avV68evL299dqo1WocP368xHVWlYr0Dyi60mXWrFnYu3cvOnTo8Mjt3LhxA//99x98fHxMUndZVbR/D9JqtThz5oxYuzl9fkDl+rh161bk5eXhtddee+R2pPoMK+JRf4Om+L2Q2qZNmzBs2DBs2rRJb9qRkmRlZeHKlSs14vMzJjY2VqzdEj6/YkeOHMHly5fL9J8hKT9DQRAwZswY7Ny5E4cOHUK9evUe+Z6atC80C1JfLWDOvvvuO0GpVApr164V4uLihJEjRwouLi5CcnKyIAiC8PrrrwsTJ04U2x87dkywsrIS5s2bJ5w/f16YNm2aYG1tLZw5c0ZsM2fOHMHFxUX4/vvvhX/++UcYMGCAUK9ePSEnJ8fs+zdnzhzBxsZG2LZtm5CUlCQ+MjMzBUEQhMzMTGHChAlCTEyMEB8fLxw4cEBo166d0KhRIyE3N9fs+zdjxgzh559/Fq5cuSKcPHlSGDx4sKBSqYRz586Jbczp8xOE8vex2FNPPSWEhYUZLDe3zzAzM1M4ffq0cPr0aQGAsGDBAuH06dPC9evXBUEQhIkTJwqvv/662P7q1auCnZ2d8L///U84f/68sGzZMkGhUAh79+4V2zzqZ2bO/fv2228FKysrYdmyZXp/g+np6WKb999/Xzh8+LAQHx8vHDt2TAgODhbc3d2F1NRUs+/fwoULhV27dgn//vuvcObMGWHcuHGCXC4XDhw4ILYxp89PEMrfx2KvvfaaEBQUZHSd5vQZvv3224Kzs7Nw+PBhvd+57OxssU1N3xdKjUHsEZYsWSLUrVtXsLGxETp27Cj88ccf4mvdunUTwsPD9dpv2bJFaNy4sWBjYyO0aNFC2L17t97rOp1OmDJliuDl5SUolUrh2WefFS5evFgdXTGqPP3z9/cXABg8pk2bJgiCIGRnZwu9evUSPDw8BGtra8Hf318YMWKEZP9ACkL5+jd+/HixrZeXl9C3b1/h1KlTeuszt89PEMr/O3rhwgUBgLBv3z6DdZnbZ1g8ncHDj+I+hYeHC926dTN4T5s2bQQbGxuhfv36wpo1awzWW9rPrDqVt3/dunUrtb0gFE3X4ePjI9jY2Ai1a9cWwsLChMuXL1dvx+4pb/8+/fRToUGDBoJKpRLc3NyE7t27C4cOHTJYr7l8foJQsd/R9PR0wdbWVli5cqXRdZrTZ2isbwD0/q4sYV8oJZkgCEKVHW4jIiIiohJxjBgRERGRRBjEiIiIiCTCIEZEREQkEQYxIiIiIokwiBERERFJhEGMiIiISCIMYkREREQSYRAjIrMybtw4jBw5EjqdTupSiIiqHIMYEZmNxMRENGnSBF9++SXkcv7zRESWjzPrExEREUmE/+UkIslFRERAJpMZPHr37i11aUREVcpK6gKIiACgd+/eWLNmjd4ypVIpUTVERNWDR8SIyCwolUp4e3vrPVxdXQEAMpkMy5cvR58+fWBra4v69etj27Zteu8/c+YMnnnmGdja2qJWrVoYOXIksrKy9NqsXr0aLVq0gFKphI+PD8aMGSO+tmDBArRq1Qr29vbw8/PDO++8Y/B+IiJTYxAjohphypQpeOGFF/D333/j1VdfxeDBg3H+/HkAgEajQUhICFxdXfHnn39i69atOHDggF7QWr58OUaPHo2RI0fizJkz+OGHH9CwYUPxdblcjs8//xznzp3DunXrcOjQIXzwwQfV3k8ierxwsD4RSS4iIgIbNmyASqXSWz5p0iRMmjQJMpkMo0aNwvLly8XXnnzySbRr1w5ffPEFvvrqK3z44YdITEyEvb09AGDPnj0IDQ3FrVu34OXlhdq1a2PYsGH4+OOPy1TTtm3bMGrUKNy+fdt0HSUiegjHiBGRWejRo4de0AIANzc38ftOnTrpvdapUyfExsYCAM6fP4/AwEAxhAFAly5doNPpcPHiRchkMty6dQvPPvtsids/cOAAoqOjceHCBajVahQWFiI3NxfZ2dmws7MzQQ+JiAzx1CQRmQV7e3s0bNhQ7/FgEKsMW1vbUl+/du0a+vfvj9atW2P79u04efIkli1bBgDIz883SQ1ERMYwiBFRjfDHH38YPG/WrBkAoFmzZvj777+h0WjE148dOwa5XI4mTZrA0dERAQEBOHjwoNF1nzx5EjqdDvPnz8eTTz6Jxo0b49atW1XXGSKie3hqkojMQl5eHpKTk/WWWVlZwd3dHQCwdetWdOjQAU899RS+/fZbnDhxAqtWrQIAvPrqq5g2bRrCw8Mxffp0pKWlYezYsXj99dfh5eUFAJg+fTpGjRoFT09P9OnTB5mZmTh27BjGjh2Lhg0boqCgAEuWLEFoaCiOHTuGFStWVO8PgIgeTwIRkcTCw8MFAAaPJk2aCIIgCACEZcuWCT179hSUSqUQEBAgbN68WW8d//zzj9CjRw9BpVIJbm5uwogRI4TMzEy9NitWrBCaNGkiWFtbCz4+PsLYsWPF1xYsWCD4+PgItra2QkhIiLB+/XoBgHD37t0q7z8RPb541SQRmT2ZTIadO3di4MCBUpdCRGRSHCNGREREJBEGMSIiIiKJcLA+EZk9jqAgIkvFI2JEREREEmEQIyIiIpIIgxgRERGRRBjEiIiIiCTCIEZEREQkEQYxIiIiIokwiBERERFJhEGMiIiISCIMYkREREQS+X+P27gE9y2HjwAAAABJRU5ErkJggg==","text/plain":["<Figure size 640x480 with 1 Axes>"]},"metadata":{},"output_type":"display_data"}],"source":["plt.plot(range(num_epochs), val_loss[:3] , label='train_losses')\n","plt.plot(range(num_epochs), val_acc[:3], label='train_accs')\n","plt.title('Curva de Validação')\n","plt.xlabel('Época')\n","plt.ylabel('Loss / Acuracia')\n","plt.legend(bbox_to_anchor=(1.1, 1.05))\n","plt.show()"]},{"cell_type":"markdown","metadata":{"id":"Dlq2b8ZEBMQE"},"source":["### **12.5 - Salva pesos - pré-treinada**"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"ULQsGZwLBMQE"},"outputs":[],"source":["torch.save(pmodel_trained.state_dict(), 'weights-pre-treined.h5')"]},{"cell_type":"markdown","metadata":{"id":"NhhUOk6YBMQE"},"source":["### **13 - Validação - pré-treinada**"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"ml0EYn6aBMQE"},"outputs":[],"source":["validation_img_paths = [\"train/River/River_1.png\",\n","                        \"train/Residential/Residential_1.png\",\n","                        \"train/SeaLake/Bombay_000001.png\"]\n","img_list = [Image.open(input_path + img_path) for img_path in validation_img_paths]\n","\n","validation_batch = torch.stack([data_transforms['validation'](img).to(device)\n","                                for img in img_list])"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"ioIMz03TBMQE","outputId":"afc7921f-3225-42a8-f6e3-d8f742145ecb"},"outputs":[{"data":{"text/plain":["tensor([[ 0.0272, -0.9226,  0.0950,  0.1259, -0.2709, -0.0060, -0.4868,  0.1690,\n","         -0.2195, -0.5499],\n","        [ 0.0157, -0.5184,  0.1189,  0.1778, -0.2566, -0.1138, -0.4979,  0.3074,\n","         -0.0127, -0.3980],\n","        [ 0.2275, -0.7395,  0.2603, -0.0031, -0.1404, -0.2608, -0.2846,  0.2013,\n","         -0.0738, -0.3537]], grad_fn=<AddmmBackward0>)"]},"execution_count":39,"metadata":{},"output_type":"execute_result"}],"source":["pred_logits_tensor = model(validation_batch)\n","pred_logits_tensor"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"FOrgQVudBMQF","outputId":"0eb4a713-e1ba-4cc6-c2db-59c396f64651"},"outputs":[{"data":{"text/plain":["Softmax(\n","  dim=tensor([[ 0.0272, -0.9226,  0.0950,  0.1259, -0.2709, -0.0060, -0.4868,  0.1690,\n","           -0.2195, -0.5499],\n","          [ 0.0157, -0.5184,  0.1189,  0.1778, -0.2566, -0.1138, -0.4979,  0.3074,\n","           -0.0127, -0.3980],\n","          [ 0.2275, -0.7395,  0.2603, -0.0031, -0.1404, -0.2608, -0.2846,  0.2013,\n","           -0.0738, -0.3537]], grad_fn=<AddmmBackward0>)\n",")"]},"execution_count":40,"metadata":{},"output_type":"execute_result"}],"source":["pred_probs = nn.Softmax(pred_logits_tensor).cpu()\n","pred_probs"]}],"metadata":{"colab":{"provenance":[],"toc_visible":true},"kernelspec":{"display_name":"Python 3","name":"python3"},"language_info":{"codemirror_mode":{"name":"ipython","version":3},"file_extension":".py","mimetype":"text/x-python","name":"python","nbconvert_exporter":"python","pygments_lexer":"ipython3","version":"3.12.6"}},"nbformat":4,"nbformat_minor":0}