11_tensorflow_exam_skill_checklist.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"11_tensorflow_exam_skill_checklist.ipynb","provenance":[{"file_id":"1dDmKr6ZoPsuKqTUV66CmOKXpbJMGk-pi","timestamp":1630056709163},{"file_id":"1GEb2rUHkUYB3sh-J-ItxbIL-yqyyRr6Q","timestamp":1630036176000}],"collapsed_sections":[],"toc_visible":true},"kernelspec":{"name":"python3","display_name":"Python 3"},"language_info":{"name":"python"}},"cells":[{"cell_type":"code","metadata":{"id":"6rMdpdCp9Dy8","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1630050696811,"user_tz":-420,"elapsed":1302,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"909fd035-7237-4abd-e4b3-037726595d04"},"source":["!nvidia-smi"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Fri Aug 27 07:51:35 2021       \n","+-----------------------------------------------------------------------------+\n","| NVIDIA-SMI 470.57.02    Driver Version: 460.32.03    CUDA Version: 11.2     |\n","|-------------------------------+----------------------+----------------------+\n","| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\n","| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\n","|                               |                      |               MIG M. |\n","|===============================+======================+======================|\n","|   0  Tesla K80           Off  | 00000000:00:04.0 Off |                    0 |\n","| N/A   73C    P0    75W / 149W |   2269MiB / 11441MiB |      0%      Default |\n","|                               |                      |                  N/A |\n","+-------------------------------+----------------------+----------------------+\n","                                                                               \n","+-----------------------------------------------------------------------------+\n","| Processes:                                                                  |\n","|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |\n","|        ID   ID                                                   Usage      |\n","|=============================================================================|\n","|  No running processes found                                                 |\n","+-----------------------------------------------------------------------------+\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"KFKZSEeUOdSE"},"source":["## 2. Building and training neural network using TensorFlow 2.x"]},{"cell_type":"markdown","metadata":{"id":"u0zf0mpWOshF"},"source":["#### Use TensorFlow 2.x"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":35},"id":"MR98O2gRO0_I","executionInfo":{"status":"ok","timestamp":1630035778242,"user_tz":-420,"elapsed":8,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"60f08d90-2bb8-4834-ffa3-420d3cebd460"},"source":["import tensorflow as tf\n","\n","tf.__version__"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"application/vnd.google.colaboratory.intrinsic+json":{"type":"string"},"text/plain":["'2.6.0'"]},"metadata":{},"execution_count":164}]},{"cell_type":"markdown","metadata":{"id":"usGtsZjWO6py"},"source":["### Preprocess data, build, predict\n","\n","* Build, compile and train machine learning (ML) models using TensorFlow.\n","* Preprocess data to get it ready for use in a model.\n","* Use models to predict results.\n","* Build sequential models with multiple layers.\n"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"H439XRP9PQnd","executionInfo":{"status":"ok","timestamp":1630035779047,"user_tz":-420,"elapsed":811,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"a188b0fa-3b1f-4fea-f991-8834efa418b0"},"source":["from tensorflow.keras import datasets\n","import tensorflow_datasets as tfds\n","\n","(X_train, y_train), (X_test, y_test) = datasets.mnist.load_data()\n","\n","X_train.shape, X_test.shape, y_train.shape, y_test.shape"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["((60000, 28, 28), (10000, 28, 28), (60000,), (10000,))"]},"metadata":{},"execution_count":165}]},{"cell_type":"markdown","metadata":{"id":"fAhW6sQaPcrj"},"source":["#### Preprocess data"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":265},"id":"ztHh9glIRZSU","executionInfo":{"status":"ok","timestamp":1630035779048,"user_tz":-420,"elapsed":18,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"5419e914-c954-4372-93b0-ec747d908c22"},"source":["import matplotlib.pyplot as plt\n","\n","plt.figure()\n","plt.imshow(X_train[0])\n","plt.colorbar()\n","plt.grid(False)\n","plt.show()"],"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 2 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"-Xff-nunRdGq","executionInfo":{"status":"ok","timestamp":1630035779048,"user_tz":-420,"elapsed":17,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"7b24fa93-8fbb-4417-dcc7-2491ae5da388"},"source":["img = X_train[0]\n","tf.reduce_max(img), tf.reduce_min(img)"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(<tf.Tensor: shape=(), dtype=uint8, numpy=255>,\n"," <tf.Tensor: shape=(), dtype=uint8, numpy=0>)"]},"metadata":{},"execution_count":167}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"xOeXg_c9RuxE","executionInfo":{"status":"ok","timestamp":1630035779049,"user_tz":-420,"elapsed":12,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"b79e0df5-ec20-454a-d148-543b02a8de2c"},"source":["label = y_train[0]\n","label"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["5"]},"metadata":{},"execution_count":168}]},{"cell_type":"markdown","metadata":{"id":"tZGaf6LpR3cs"},"source":["##### Normalize the data\n"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"zfGqH7xjSdKx","executionInfo":{"status":"ok","timestamp":1630035779389,"user_tz":-420,"elapsed":345,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"08a301ff-6e45-4de8-9b6f-071eb2af502c"},"source":["X_train, X_test = X_train/255., X_test/255.\n","\n","tf.reduce_max(X_train), tf.reduce_min(X_train)"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(<tf.Tensor: shape=(), dtype=float64, numpy=1.0>,\n"," <tf.Tensor: shape=(), dtype=float64, numpy=0.0>)"]},"metadata":{},"execution_count":169}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"jg00ZPqXT7S_","executionInfo":{"status":"ok","timestamp":1630035779389,"user_tz":-420,"elapsed":11,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"b19bf5d0-2490-4dfe-a10b-debab02966f2"},"source":["X_train.shape, X_test.shape"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["((60000, 28, 28), (10000, 28, 28))"]},"metadata":{},"execution_count":170}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"ziFt1rpKTebs","executionInfo":{"status":"ok","timestamp":1630035779742,"user_tz":-420,"elapsed":355,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"f5a51070-73a6-49b1-db76-ebc20ae767af"},"source":["y_train.shape, y_test.shape"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["((60000,), (10000,))"]},"metadata":{},"execution_count":171}]},{"cell_type":"markdown","metadata":{"id":"r9rdww-3SlJU"},"source":["#### Build the model"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"A9nBoJnxSrsX","executionInfo":{"status":"ok","timestamp":1630035819984,"user_tz":-420,"elapsed":40245,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"82dcba99-7062-45fa-d681-2d62b87e21d1"},"source":["import tensorflow as tf\n","from tensorflow.keras import layers\n","\n","# Create a model\n","model_0 = tf.keras.Sequential([\n","  layers.Flatten(),\n","  layers.Dense(100, activation=\"relu\"),\n","  layers.Dense(10, activation=\"softmax\")\n","])\n","# Compile the model\n","model_0.compile(\n","    loss=tf.keras.losses.SparseCategoricalCrossentropy(),\n","    optimizer=tf.keras.optimizers.Adam(),\n","    metrics=[\"accuracy\"]\n",")\n","# Fit the model\n","model_0_history = model_0.fit(x=X_train, \n","                              y=y_train,\n","                              validation_data=(X_test, y_test),\n","                              epochs=10)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Epoch 1/10\n","1875/1875 [==============================] - 4s 2ms/step - loss: 0.2672 - accuracy: 0.9240 - val_loss: 0.1448 - val_accuracy: 0.9554\n","Epoch 2/10\n","1875/1875 [==============================] - 4s 2ms/step - loss: 0.1184 - accuracy: 0.9654 - val_loss: 0.1179 - val_accuracy: 0.9631\n","Epoch 3/10\n","1875/1875 [==============================] - 4s 2ms/step - loss: 0.0829 - accuracy: 0.9750 - val_loss: 0.0821 - val_accuracy: 0.9738\n","Epoch 4/10\n","1875/1875 [==============================] - 4s 2ms/step - loss: 0.0629 - accuracy: 0.9808 - val_loss: 0.0799 - val_accuracy: 0.9760\n","Epoch 5/10\n","1875/1875 [==============================] - 4s 2ms/step - loss: 0.0501 - accuracy: 0.9849 - val_loss: 0.0854 - val_accuracy: 0.9725\n","Epoch 6/10\n","1875/1875 [==============================] - 4s 2ms/step - loss: 0.0406 - accuracy: 0.9872 - val_loss: 0.0757 - val_accuracy: 0.9769\n","Epoch 7/10\n","1875/1875 [==============================] - 4s 2ms/step - loss: 0.0329 - accuracy: 0.9898 - val_loss: 0.0779 - val_accuracy: 0.9757\n","Epoch 8/10\n","1875/1875 [==============================] - 4s 2ms/step - loss: 0.0271 - accuracy: 0.9916 - val_loss: 0.0781 - val_accuracy: 0.9771\n","Epoch 9/10\n","1875/1875 [==============================] - 4s 2ms/step - loss: 0.0224 - accuracy: 0.9931 - val_loss: 0.0875 - val_accuracy: 0.9769\n","Epoch 10/10\n","1875/1875 [==============================] - 4s 2ms/step - loss: 0.0189 - accuracy: 0.9940 - val_loss: 0.0761 - val_accuracy: 0.9793\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"Sy-6nCwgTXxV","executionInfo":{"status":"ok","timestamp":1630035820525,"user_tz":-420,"elapsed":566,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"5751c604-6bc2-4601-9d5c-bf38c7d1a121"},"source":["model_0.evaluate(X_test, y_test)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["313/313 [==============================] - 1s 2ms/step - loss: 0.0761 - accuracy: 0.9793\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["[0.07612026482820511, 0.9793000221252441]"]},"metadata":{},"execution_count":173}]},{"cell_type":"markdown","metadata":{"id":"gCBAa9DqXONy"},"source":["#### Plot the history and evaluate the model"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"EJug8hlDUjjC","executionInfo":{"status":"ok","timestamp":1630035820864,"user_tz":-420,"elapsed":344,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"468461bd-10c6-4ae1-9c4f-467fef820068"},"source":["from sklearn.metrics import accuracy_score\n","\n","model_0_pred_prob = model_0.predict(X_test)\n","model_0_pred = tf.argmax(model_0_pred_prob, axis=1)\n","\n","model_0_accuracy = accuracy_score(y_true=y_test, y_pred=model_0_pred)\n","model_0_accuracy"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["0.9793"]},"metadata":{},"execution_count":174}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"-VqRg4DjVMFj","executionInfo":{"status":"ok","timestamp":1630035820864,"user_tz":-420,"elapsed":12,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"af2f54bd-a98c-4dd9-cd0d-cc0419bc5742"},"source":["from sklearn.metrics import precision_recall_fscore_support\n","\n","model_0_precision, model_0_recall, model_0_f1, _ = precision_recall_fscore_support(y_true=y_test,\n","                                                                                y_pred=model_0_pred,\n","                                                                                average=\"weighted\")\n","model_0_precision, model_0_recall, model_0_f1"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(0.979319533176659, 0.9793, 0.9792944002446339)"]},"metadata":{},"execution_count":175}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"pZinuPfbWjlC","executionInfo":{"status":"ok","timestamp":1630035820864,"user_tz":-420,"elapsed":9,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"a376ab1f-0a2a-4d60-d16f-c42a1f45189c"},"source":["model_0_results = {\n","    \"accuracy\": model_0_accuracy,\n","    \"precision\": model_0_precision,\n","    \"recall\": model_0_precision,\n","    \"f1\": model_0_f1\n","}\n","\n","model_0_results"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["{'accuracy': 0.9793,\n"," 'f1': 0.9792944002446339,\n"," 'precision': 0.979319533176659,\n"," 'recall': 0.979319533176659}"]},"metadata":{},"execution_count":176}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":514},"id":"8iv6UG54XYSr","executionInfo":{"status":"ok","timestamp":1630035821322,"user_tz":-420,"elapsed":464,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"a67d2d70-f6b1-4e9a-aaf4-ff039d45cd4a"},"source":["import pandas as pd\n","\n","# plt.figure()\n","pd.DataFrame(model_0_history.history).plot(figsize=(10,8))\n","plt.xlabel(\"Epochs\")\n","plt.ylabel(\"Accuracy\")\n","plt.legend()"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.legend.Legend at 0x7f781c0d9750>"]},"metadata":{},"execution_count":177},{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAmEAAAHgCAYAAADt8bqrAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeXydZZ3///d17rNlX5ruLbSVbtBSSkNZFEQRRQXq8ii1X0SpLG7AiIoyiMogzoI6zjjDIHVBQBAQhvkxiqBInYKsoRYKbSmltDRladqkWZqc7T7X74/7nJOT5CRN25zeSfp6Ph7nce77uq77Pp80bfLudW/GWisAAAAcWgG/CwAAADgcEcIAAAB8QAgDAADwASEMAADAB4QwAAAAHxDCAAAAfBD0u4D9VVdXZ6dNm+Z3GQAAAPv0/PPP77LWji3UN+JC2LRp09TQ0OB3GQAAAPtkjNnWXx+HIwEAAHxACAMAAPABIQwAAMAHhDAAAAAfEMIAAAB8QAgDAADwASEMAADAB4QwAAAAHxDCAAAAfEAIAwAA8AEhDAAAwAeEMAAAAB8ULYQZY35pjNlpjHmpn35jjPmJMWazMeZFY8zxxaoFAABguCnmTNivJJ01QP+HJc3MvC6VdHMRawEAABhWihbCrLWrJTUPMGSJpNut52lJ1caYicWqBwAAYDjx85ywyZK25603Ztr6MMZcaoxpMMY0NDU1HZLiAAAAiinodwGDYa1dKWmlJNXX11ufywEAAMOVtd5LVrLpzCuz3LstGJVCUd9K9TOE7ZA0NW99SqYNAAD/WSulU1LalazbvZxbz3/P/GLv0+ZK6XSBtoG26W8/B7JN5r1Pv+35NeT60z3b+gQYW6Atb71Hmwq0FWlcfrjaHx/4B+k9Xxm6vzP7yc8Q9qCky4wxd0s6UVKrtfYtH+sBgMObtZmQkfQCh9vrPbecbU95ywX780NLqvsXfr+BJtUzJOxzu1R3YMiNS3mho8d22XG9a0n33a73/jUSD7wYKeBIJiAZJ7PsSIFA3zYTyLQ7vbYJdPfnXsZ7l8nbLtizrdA4Ywq0FRoXkIwGub/Me5+2Axh3xEn+fJsyihbCjDG/kXS6pDpjTKOk70oKSZK19qeSHpL0EUmbJXVKWlGsWgBgyGRnR/KDR/7LzQ8gye5f6j3Gu/2HFzfVvW2f/vzPTXaHoIMKTL326adA0HsZJ7McyFvODxTBvOXsK9i9Hgz32i4TFnqMy+w/f7vcuGCvsYFe2znd47L7zQUZUyDoOD3H9dimv3DU631QgSrz+RgxihbCrLXL99FvJX25WJ8PYBix1vtF78alVPYVk9yE955K9O0bMOgUaNuvEDTQ/vK2L7Sv/T3cMRRywSAkOZn3QFByer3nljPr4dLMNqFMqAjl9Tvdy04wb/+995W//319fq/+QH5o2VeY4t7hOPyMiBPzARwga3sGnR7BJxN43Pg++uK91gvtq7/1vG2LcWgn94s8mBckgt0ho0eYcPL6gt4JufmBI//lFGgbaF+5QNPfNr1ry9+mdyAqFKgIKMBoRAgDDjVrvYCS6JSSe6Vkl5TYKyU7M22ZVyLTl1vuzBvb1R2WCgalvOWhEAhJwYjkhL3wEsy859YjUmmt5ES85eyr4Hqh7bPr+WNDfUNL/qwLsycABsmm07KplJRKyaZSsq4rm0wqUFomp7zMt7oIYUBv1noBJheG8sPSvoLTYPo7td+zQsGoFCqRQmXeIaZQSXf4iVTkBZ1CwSYvKB1oMCLsHDTrurKJhGw8rnQ8IZtMygRM5nyjgEwge7Kz8ZYdR8YU6A8EvHbsF2tt5mT8tGz23c2e2N+rzaYl1+3TZt3MVYWu67Xly/+e5C/2aC+w3ON7uR/76G+/Q7GP7LK1XnBxXS+4pFzJTeWWbSqZ19dzXHdf3nIyJeum8pYzfanMdm4qbzmvr8+4gft6BK1M8PJuWdHXuKu+rjEXXVSw71AghGHkS8Wlrj1SV4sUy7639Q1OeWHIxr11G++UYp2yib2y8Zhs3Btv0zZzxbPJu91MZjnzbq2RsldIm5AUiMgab7bImrBsICyZkKypkDW1kkKyJihrgpIcWQVlFZCsI6tAZjng7dcaWZv9PMm66e4fUK4rua4UCsqEQjKhsEw45C2HwzKhkAKZd+Utm9x7UCYUkAkbmZCVCbky4aRMSDJhZdZdmVBCJpzIfEbe9iM0kFlrZZNJ2Xg890rHE7KJzHIs3r0cj8vm98UTmW1iectxpROZcbn1hGws1r2c167kEJ/0PkBQGzC8BQKSE5DJXuEWMDIBp3vZ9LOd4/To73c7J9B99Vw6LZt2pbR31aVNZwKMTUuZANS3LROIXNcLTq6bCUA92wYcnx+qMn39/RKGT0IhmWBQxnFkgkGpx7IjEwwV7DPRiAKZvr7jHBkn2HcfoaD3b8UJdi8HQzJBRyULFvj6x0AIw/CQTkvxtp5BKhusulpku1qUbt0ld/duuS0tcve0yW3vkNvepVRnSm4iIDfe/Uone4WYTKjJhqnCE1GOpPLM62AkMq99CGV+QDhO9w+ZfpZzP1yyyyYgG08o3bHXm11JJgu/JwZRx/4KBnuFulCfoBbIBcO8gJjf1mP7cN+gFw5LNq10LJYLQ7lgFI/3Wo/16BsoGB0sEw7LRCIykYgXbjPLJhJWIByRU1kpM3asAvntkYhMuMB6KOjNNGRnWdLpTFjJCy492t3u5Wx7btvMLE82rKTTXkjpvb8e7W7P7fLDUKH9uSkpab3DOpmA4+0vO8OUt7/smGxQy4a+bJDLhLf8GT/jBKVQ374+47NBL9Crr782kwmd2T7Hu5Kwb1/vNm9/2TZv//kB1emeaerx8yRvxRZettnlHv0qOLbg/grtqxj7kGSCmYCTF1zyl72+YM9xocIBSkEvII3U/8gVAyEMQysZ6zdIZdtsZ4vSrbvltjTLbW2T29out71LbtwoFQ/0CVRups2mCx2CKZGMkVMelVNZJqeuUqGaGjkVVVI4IhOKyISjmcAT7P4BklvO/oBwcj9I8v/HZJwC7bkfQj37+yzn+rvH5n4QHaIfQrlZg/6CWt5yus+YpGwyf7mfoJcs3Jfu7JJNtu1zu/0SDBYMP7lgVFoqp6bGW46EC4ef3Hq0QF84L0BFZMJ56+EwvzwADClCGPpKu1KsdYAgtSfXZjublW7fI7dlj9y2drmdyVxwKhioEkG5cdPrKv9I5iUvUFWUyqmqlDOhWqGaWkXHjFVwTJ2c6ho51dVyamrk1FTLqa5WsKZGgcpKfjn2wxiT+9/ncGStlTJhLJ0f/hIJGSfQHa6yIWiYfh0AcCD4iXa42fOGtPnPUuv2PuHKC1Stcls75CZMj/DUI1AlQnKTXphyY1bWlbx5+cqenxUwcirKveA0uVahmjGK1tYoWFNDoIKkTEgMh70ZpzL/rlACAD8QwkYZm0zK7ehQeu9epTs6lG5rk/vGWqVfe1rprS8o3fy23KSRm3DkpiJyk6FM0JLcTlfWLZVU2nfHgYCcygo5NbVyxtUqVF2taE11JlBVe6GKQAUAwKARwoaJHuGpvV3pjg5vvWOv0ns75La3e8sdHV7f3kxfdmwmdNlYbB+fVOkFqqoqLzRNqOk/UFVXy6khUAEAUAyEsINkE4lcAMoFpPYOpfd29AxS7e1emOroNTbzPqgrtxxHgfJyOeXlCmReTt0YhY88UoFQWoHETjl7tymwd6sCTkqBslI50xcpMOs9Csw+XYExE+WUl8uUlHCfIQAAfEYI6yXR2KiO1at7zDp5M1EdBWehBnULAMfpEZwCFeUK1tUpMG1apq1MTkWFAmXl3evl5QqUV+Qtl8tEo93hyU1K25+RNj0svfKwtPtVqUTSUXOlWRdJs86SppzgPS4FAAAMO/yG7iW+aZPeuf573kow2DM8lZcpOG6cwuUzeoSj3mHJC1OZYFVe7l3hNRQzT53N0uZHveC1+VHvCkYnLE17j7T4EmnmB6Xa6Qf/OQAAoOgIYb2UnXKKZj7xuAIVFd4l8X4etrNWanrFC12bHpG2P+09TqNsrDTnHGn2WdKM073H1gAAgBGFENZLIBpVIBr1r4BUXNr2V+8Q46aHpT3bvPYJx0qnft07zDhpIc/yAwBghCOEDQcdO6VX/+iFrtdWSYkO78HJM06X3vMVaeaHpKrJflcJAACGECHMD9ZKb6/zDjFuelja8bwkK1VOlo49z5vtmnaqFC5wvy4AADAqEMIOlWSX9Ppq6ZU/eLNebTskGWnyIul935JmfUiaMF/i1hEAABwWCGHF1LpDevURb8Zry/9JqS4pXC69633S+67xrmYsH+d3lQAAwAeEsKGUTktv/i1zNeMfvEOOklR9pLTos95s15HvloIRf+sEAAC+I4QdrHi7dzL9pke8Wa+9TZIJSFNPkj7wD975XWNnc5gRAAD0QAg7EC1bu0+q3/qE5CakaJV01AekWR+WjjpDKq31u0oAADCMEcIGw01Jjc9lDjM+LDVt9NrrZkknft6b7Zp6ouSE/K0TAACMGISw/nS1SJv/7M14bf6Ttx4Ieud0HZ85v2vMu/yuEgAAjFCEsN7eeFp67AZp25OSdaXSMd5M16yzvKsao1V+VwgAAEYBQlhvgZA36/Wer3jBa/IiKeD4XRUAABhlCGG9TVkkffGvflcBAABGOZ4CDQAA4ANCGAAAgA8IYQAAAD4ghAEAAPiAEAYAAOADQhgAAIAPCGEAAAA+IIQBAAD4gBAGAADgA0IYAACADwhhAAAAPiCEAQAA+IAQBgAA4ANCGAAAgA8IYQAAAD4ghAEAAPiAEAYAAOADQhgAAIAPCGEAAAA+IIQBAAD4gBAGAADgA0IYAACADwhhAAAAPiCEAQAA+IAQBgAA4ANCGAAAgA8IYQAAAD4ghAEAAPiAEAYAAOADQhgAAIAPCGEAAAA+IIQBAAD4gBAGAADgA0IYAACADwhhAAAAPiCEAQAA+IAQBgAA4ANCGAAAgA8IYQAAAD4ghAEAAPiAEAYAAOADQhgAAIAPCGEAAAA+IIQBAAD4gBAGAADgg6KGMGPMWcaYV4wxm40xVxfoP8IYs8oY8zdjzIvGmI8Usx4AAIDhomghzBjjSLpJ0oclHS1puTHm6F7DrpV0r7V2oaRPSfqvYtUDAAAwnBRzJmyxpM3W2i3W2oSkuyUt6TXGSqrMLFdJerOI9QAAAAwbxQxhkyVtz1tvzLTlu07Sp40xjZIeknR5oR0ZYy41xjQYYxqampqKUSsAAMAh5feJ+csl/cpaO0XSRyTdYYzpU5O1dqW1tt5aWz927NhDXiQAAMBQK2YI2yFpat76lExbvosk3StJ1tqnJEUl1RWxJgAAgGGhmCHsOUkzjTHTjTFheSfeP9hrzBuSzpAkY8xceSGM440AAGDUK1oIs9amJF0m6RFJG+RdBfmyMeZ6Y8y5mWFfk3SJMeYFSb+RdKG11harJgAAgOEiWMydW2sfknfCfX7bd/KW10t6dzFrAAAAGI78PjEfAADgsEQIAwAA8AEhDAAAwAeEMAAAAB8QwgAAAHxACAMAAPABIQwAAMAHhDAAAAAfEMIAAAB8QAgDAADwASEMAADAB4QwAAAAHxDCAAAAfEAIAwAA8AEhDAAAwAeEMAAAAB8QwgAAAHxACAMAAPABIQwAAMAHhDAAAAAfEMIAAAB8QAgDAADwASEMAADAB4QwAAAAHxDCAAAAfEAIAwAA8AEhDAAAwAeEMAAAAB8QwgAAAHxACAMAAPABIQwAAMAHhDAAAAAfEMIAAAB8QAgDAADwASEMAADAB4QwAAAAHxDCAAAAfEAIAwAA8AEhDAAAwAeEMAAAAB8QwgAAAHxACAMAAPABIQwAAMAHhDAAAAAfEMIAAAB8QAgDAADwASEMAADAB4QwAAAAHxDCAAAAfEAIAwAA8AEhDAAAwAeEMAAAAB8QwgAAAHxACAMAAPABIQwAAMAHhDAAAAAfEMIAAAB8QAgDAADwASEMAADAB4QwAAAAHxDCAAAAfEAIAwAA8AEhDAAAwAeEMAAAAB8QwgAAAHxACAMAAPABIQwAAMAHhDAAAAAfEMIAAAB8QAgDAADwASEMAADAB4QwAAAAHxQ1hBljzjLGvGKM2WyMubqfMecZY9YbY142xtxVzHoAAACGi2CxdmyMcSTdJOlMSY2SnjPGPGitXZ83Zqakv5f0bmttizFmXLHqAQAAGE6KORO2WNJma+0Wa21C0t2SlvQac4mkm6y1LZJkrd1ZxHoAAACGjWKGsMmStuetN2ba8s2SNMsY81djzNPGmLMK7cgYc6kxpsEY09DU1FSkcgEAAA4dv0/MD0qaKel0Scsl/cwYU917kLV2pbW23lpbP3bs2ENcIgAAwNArZgjbIWlq3vqUTFu+RkkPWmuT1trXJW2SF8oAAABGtWKGsOckzTTGTDfGhCV9StKDvcb8j7xZMBlj6uQdntxSxJoAAACGhaKFMGttStJlkh6RtEHSvdbal40x1xtjzs0Me0TSbmPMekmrJF1lrd1drJoAAACGC2Ot9buG/VJfX28bGhr8LgMAAGCfjDHPW2vrC/UV7T5hAACgeJLJpBobGxWLxfwuBZKi0aimTJmiUCg06G0IYQAAjECNjY2qqKjQtGnTZIzxu5zDmrVWu3fvVmNjo6ZPnz7o7fy+RQUAADgAsVhMY8aMIYANA8YYjRkzZr9nJQlhAACMUASw4eNAvheEMAAAcEDKy8v9LmFEI4QBAAD4gBAGAAAOirVWV111lebNm6f58+frnnvukSS99dZbOu2003Tcccdp3rx5evzxx+W6ri688MLc2B//+Mc+V+8fro4EAGCE+4f/fVnr32wb0n0ePalS3z3nmEGN/e///m+tXbtWL7zwgnbt2qUTTjhBp512mu666y596EMf0re+9S25rqvOzk6tXbtWO3bs0EsvvSRJ2rNnz5DWPZIwEwYAAA7KE088oeXLl8txHI0fP17vfe979dxzz+mEE07Qrbfequuuu07r1q1TRUWFZsyYoS1btujyyy/Xww8/rMrKSr/L9w0zYQAAjHCDnbE61E477TStXr1av//973XhhRfqq1/9qj7zmc/ohRde0COPPKKf/vSnuvfee/XLX/7S71J9wUwYAAA4KKeeeqruueceua6rpqYmrV69WosXL9a2bds0fvx4XXLJJbr44ou1Zs0a7dq1S+l0Wp/85Cd1ww03aM2aNX6X7xtmwgAAwEH5+Mc/rqeeekoLFiyQMUY33nijJkyYoNtuu00/+MEPFAqFVF5erttvv107duzQihUrlE6nJUn/9E//5HP1/tnnA7yNMedI+r21Nn1oShoYD/AGAEDasGGD5s6d63cZyFPoezLQA7wHczhymaRXjTE3GmPmDEGNAAAAh719hjBr7aclLZT0mqRfGWOeMsZcaoypKHp1AAAAo9SgTsy31rZJuk/S3ZImSvq4pDXGmMuLWBsAAMCotc8QZow51xjzgKS/SApJWmyt/bCkBZK+VtzyAAAARqfBXB35SUk/ttauzm+01nYaYy4qTlkAAACj22BC2HWS3squGGNKJI231m611v65WIUBAACMZoM5J+y3kvJvT+Fm2gAAAHCABhPCgtbaRHYlsxwuXkkAAADdUqmU3yUUxWBCWJMx5tzsijFmiaRdxSsJAACMFB/72Me0aNEiHXPMMVq5cqUk6eGHH9bxxx+vBQsW6IwzzpAkdXR0aMWKFZo/f76OPfZY3X///ZKk8vLy3L7uu+8+XXjhhZKkCy+8UF/4whd04okn6hvf+IaeffZZnXzyyVq4cKFOOeUUvfLKK5Ik13X19a9/XfPmzdOxxx6r//iP/9Bjjz2mj33sY7n9/ulPf9LHP/7xQ/HHsV8Gc07YFyTdaYz5T0lG0nZJnylqVQAAYPD+cLX09rqh3eeE+dKH/3mfw375y1+qtrZWXV1dOuGEE7RkyRJdcsklWr16taZPn67m5mZJ0ve+9z1VVVVp3TqvzpaWln3uu7GxUU8++aQcx1FbW5sef/xxBYNBPfroo7rmmmt0//33a+XKldq6davWrl2rYDCo5uZm1dTU6Etf+pKampo0duxY3Xrrrfrc5z53cH8eRbDPEGatfU3SScaY8sx6R9GrAgAAI8JPfvITPfDAA5Kk7du3a+XKlTrttNM0ffp0SVJtba0k6dFHH9Xdd9+d266mpmaf+166dKkcx5Ektba26rOf/axeffVVGWOUTCZz+/3CF76gYDDY4/MuuOAC/frXv9aKFSv01FNP6fbbbx+ir3joDOoB3saYj0o6RlLUGCNJstZeX8S6AADAYA1ixqoY/vKXv+jRRx/VU089pdLSUp1++uk67rjjtHHjxkHvI5srJCkWi/XoKysryy1/+9vf1vve9z498MAD2rp1q04//fQB97tixQqdc845ikajWrp0aS6kDSeDuVnrT+U9P/JyeYcjl0o6ssh1AQCAYa61tVU1NTUqLS3Vxo0b9fTTTysWi2n16tV6/fXXJSl3OPLMM8/UTTfdlNs2ezhy/Pjx2rBhg9LpdG5Grb/Pmjx5siTpV7/6Va79zDPP1C233JI7eT/7eZMmTdKkSZN0ww03aMWKFUP3RQ+hwZyYf4q19jOSWqy1/yDpZEmzilsWAAAY7s466yylUinNnTtXV199tU466SSNHTtWK1eu1Cc+8QktWLBAy5YtkyRde+21amlp0bx587RgwQKtWrVKkvTP//zPOvvss3XKKado4sSJ/X7WN77xDf393/+9Fi5c2ONqyYsvvlhHHHGEjj32WC1YsEB33XVXru/888/X1KlTNXfu3CL9CRwcY60deIAxz1prFxtjnpb0CUm7Jb1srT3qUBTYW319vW1oaPDjowEAGDY2bNgwbMPFcHHZZZdp4cKFuuiiQ/OAn0LfE2PM89ba+kLjB3OA9H+NMdWSfiBpjSQr6WcHWygAAECxLFq0SGVlZfrRj37kdyn9GjCEGWMCkv5srd0j6X5jzO8kRa21rYekOgAAgAPw/PPP+13CPg14Tpi1Ni3pprz1OAEMAADg4A3mxPw/G2M+afKvIQUAAMBBGUwI+7y8B3bHjTFtxph2Y0xbkesCAAAY1QZzx/yKQ1EIAADA4WSfIcwYc1qhdmvt6qEvBwAA4PAwmFtUXJW3HJW0WNLzkt5flIoAAMCoU15ero6Owo+f3rp1q84++2y99NJLh7gqfw3mcOQ5+evGmKmS/q1oFQEAABwGDuRplo2SuEUvAADDxL88+y/a2Dz4h2YPxpzaOfrm4m/223/11Vdr6tSp+vKXvyxJuu666xQMBrVq1Sq1tLQomUzqhhtu0JIlS/brc2OxmL74xS+qoaFBwWBQ//qv/6r3ve99evnll7VixQolEgml02ndf//9mjRpks477zw1NjbKdV19+9vfzj0maSQYzDlh/yHvLvmSdzXlcfLunA8AAA5Ty5Yt01e+8pVcCLv33nv1yCOP6IorrlBlZaV27dqlk046Seeee6725y5XN910k4wxWrdunTZu3KgPfvCD2rRpk37605/q7/7u73T++ecrkUjIdV099NBDmjRpkn7/+99L8h7yPZIMZiYs/0GNKUm/sdb+tUj1AACA/TTQjFWxLFy4UDt37tSbb76ppqYm1dTUaMKECbryyiu1evVqBQIB7dixQ++8844mTJgw6P0+8cQTuvzyyyVJc+bM0ZFHHqlNmzbp5JNP1ve//301NjbqE5/4hGbOnKn58+fra1/7mr75zW/q7LPP1qmnnlqsL7coBnOfsPsk/dpae5u19k5JTxtjSotcFwAAGOaWLl2q++67T/fcc4+WLVumO++8U01NTXr++ee1du1ajR8/XrFYbEg+6//9v/+nBx98UCUlJfrIRz6ixx57TLNmzdKaNWs0f/58XXvttbr++uuH5LMOlUHdMV9SSd56iaRHi1MOAAAYKZYtW6a7775b9913n5YuXarW1laNGzdOoVBIq1at0rZt2/Z7n6eeeqruvPNOSdKmTZv0xhtvaPbs2dqyZYtmzJihK664QkuWLNGLL76oN998U6Wlpfr0pz+tq666SmvWjKyzpQZzODJqrc1dU2qt7WAmDAAAHHPMMWpvb9fkyZM1ceJEnX/++TrnnHM0f/581dfXa86cOfu9zy996Uv64he/qPnz5ysYDOpXv/qVIpGI7r33Xt1xxx0KhUKaMGGCrrnmGj333HO66qqrFAgEFAqFdPPNNxfhqyweY60deIAxf5V0ubV2TWZ9kaT/tNaefAjq66O+vt42NDTseyAAAKPYhg0bNHcuNysYTgp9T4wxz1tr6wuNH8xM2Fck/dYY86YkI2mCpJFz/ScAAMAwNJibtT5njJkjaXam6RVrbbK4ZQEAgNFm3bp1uuCCC3q0RSIRPfPMMz5V5K/B3Cfsy5LutNa+lFmvMcYst9b+V9GrAwAAo8b8+fO1du1av8sYNgZzdeQl1to92RVrbYukS4pXEgAAwOg3mBDmmLxb3RpjHEnh4pUEAAAw+g3mxPyHJd1jjLkls/55SX8oXkkAAACj32BC2DclXSrpC5n1F+VdIQkAAIADtM/DkdbatKRnJG2VtFjS+yVtKG5ZAABgNCkvL/e7hGGn35kwY8wsScszr12S7pEka+37Dk1pAAAAQyuVSikYHMyBwOIbqIqNkh6XdLa1drMkGWOuPCRVAQCAQXv7H/9R8Q0bh3SfkblzNOGaa/rtv/rqqzV16lR9+ctfliRdd911CgaDWrVqlVpaWpRMJnXDDTdoyZIl+/ysjo4OLVmypOB2t99+u374wx/KGKNjjz1Wd9xxh9555x194Qtf0JYtWyRJN998syZNmqSzzz5bL730kiTphz/8oTo6OnTdddfp9NNP13HHHacnnnhCy5cv16xZs3TDDTcokUhozJgxuvPOOzV+/Hh1dHTo8ssvV0NDg4wx+u53v6vW1la9+OKL+rd/+zdJ0s9+9jOtX79eP/7xjw/qz1caOIR9QtKnJK0yxjws6W55d8wHAACHuWXLlukrX/lKLoTde++9euSRR3TFFVeosrJSu3bt0kknnaRzzz1XeTdZKCgajeqBBx7os9369et1w96o3MAAACAASURBVA036Mknn1RdXZ2am5slSVdccYXe+9736oEHHpDruuro6FBLS8uAn5FIJJR97GFLS4uefvppGWP085//XDfeeKN+9KMf6Xvf+56qqqq0bt263LhQKKTvf//7+sEPfqBQKKRbb71Vt9xyy0AfNWj9hjBr7f9I+h9jTJmkJfIeXzTOGHOzpAestX8ckgoAAMBBGWjGqlgWLlyonTt36s0331RTU5Nqamo0YcIEXXnllVq9erUCgYB27Nihd955RxMmDHw9n7VW11xzTZ/tHnvsMS1dulR1dXWSpNraWknSY489pttvv12S5DiOqqqq9hnCli3rfuJiY2Ojli1bprfeekuJRELTp0+XJD366KO6++67c+NqamokSe9///v1u9/9TnPnzlUymdT8+fP380+rsME8tmivpLsk3WWMqZG0VN4Vk4QwAAAOY0uXLtV9992nt99+W8uWLdOdd96ppqYmPf/88wqFQpo2bZpisdg+93Og2+ULBoNKp9O59d7bl5WV5ZYvv/xyffWrX9W5556rv/zlL7ruuusG3PfFF1+sf/zHf9ScOXO0YsWK/aprIIO5WWuOtbbFWrvSWnvGkFUAAABGpGXLlunuu+/Wfffdp6VLl6q1tVXjxo1TKBTSqlWrtG3btkHtp7/t3v/+9+u3v/2tdu/eLUm5w5FnnHGGbr75ZkmS67pqbW3V+PHjtXPnTu3evVvxeFy/+93vBvy8yZMnS5Juu+22XPuZZ56pm266KbeenV078cQTtX37dt11111avnz5YP949mm/QhgAAEDWMccco/b2dk2ePFkTJ07U+eefr4aGBs2fP1+333675syZM6j99LfdMccco29961t673vfqwULFuirX/2qJOnf//3ftWrVKs2fP1+LFi3S+vXrFQqF9J3vfEeLFy/WmWeeOeBnX3fddVq6dKkWLVqUO9QpSddee61aWlo0b948LViwQKtWrcr1nXfeeXr3u9+dO0Q5FIy1dsh2dijU19fb7Il1AAAcrjZs2KC5c+f6XcZh4+yzz9aVV16pM87o/2Bgoe+JMeZ5a219ofHMhAEAAPRjz549mjVrlkpKSgYMYAdieNytDAAAjHrr1q3TBRdc0KMtEonomWee8amifauurtamTZuKsm9CGAAAOCTmz5+vtWvX+l3GsMHhSAAARqiRdl73aHYg3wtCGAAAI1A0GtXu3bsJYsOAtVa7d+9WNBrdr+04HAkAwAg0ZcoUNTY2qqmpye9SIC8UT5kyZb+2IYQBADAChUKh3ON2MDJxOBIAAMAHhDAAAAAfEMIAAAB8QAgDAADwASEMAADAB0UNYcaYs4wxrxhjNhtjrh5g3CeNMdYYU/ABlwAAAKNN0UKYMcaRdJOkD0s6WtJyY8zRBcZVSPo7ScP3wVEAAABDrJgzYYslbbbWbrHWJiTdLWlJgXHfk/QvkmJFrAUAAGBYKWYImyxpe956Y6YtxxhzvKSp1trfF7EOAACAYce3E/ONMQFJ/yrpa4MYe6kxpsEY08DjGQAAwGhQzBC2Q9LUvPUpmbasCknzJP3FGLNV0kmSHix0cr61dqW1tt5aWz927NgilgwAAHBoFDOEPSdppjFmujEmLOlTkh7MdlprW621ddbaadbaaZKelnSutbahiDUBAAAMC0ULYdbalKTLJD0iaYOke621LxtjrjfGnFuszwUAABgJgsXcubX2IUkP9Wr7Tj9jTy9mLQAAAMMJd8wHAADwASEMAADAB4QwAAAAHxDCAAAAfEAIAwAA8AEhDAAAwAeEMAAAAB8QwgAAAHxACAMAAPABIQwAAMAHhDAAAAAfEMIAAAB8QAgDAADwASEMAADAB4QwAAAAHxDCAAAAfEAIAwAA8AEhDAAAwAeEMAAAAB8QwgAAAHxACAMAAPABIQwAAMAHhDAAAAAfEMIAAAB8QAgDAADwASEMAADAB4QwAAAAHxDCAAAAfEAIAwAA8AEhDAAAwAeEMAAAAB8QwgAAAHxACAMAAPABIQwAAMAHhDAAAAAfEMIAAAB8QAgDAADwASEMAADAB4QwAAAAHxDCAAAAfEAIAwAA8AEhDAAAwAeEMAAAAB8QwgAAAHxACAMAAPABIQwAAMAHhDAAAAAfEMJ66Uq4uuX/XpObtn6XAgAARjFCWC9/XP+2/ukPG3Xt/6yTtQQxAABQHEG/Cxhulhw3Wa+83a7/+strqiwJ6e8/PNfvkgAAwChECCvgqg/NVlssqVv+b4uqSkL60ulH+V0SAAAYZQhhBRhjdP2589QeS+nGh19RZTSkT590pN9lAQCAUYQQ1o9AwOiHSxeoI5bSt/+/l1QRDWrJcZP9LgsAAIwSnJg/gJAT0E3nH6/F02r1tXtf0GMb3/G7JAAAMEoQwvYhGnL088/Wa+7ESn3x12v09JbdfpcEAABGAULYIFREQ7rtc4s1tbZUF9/WoHWNrX6XBAAARjhC2CDVloV1x0WLVVUS0mdvfVabd3b4XRIAABjBCGH7YWJVie68+EQFjNGnf/6Mtjd3+l0SAAAYoQhh+2laXZnuuGixOhMpXfCLZ7SzPeZ3SQAAYAQihB2AuRMrdeuKxXqnLa7P/OJZtXYm/S4JAACMMISwA7ToyBqt/MwibWnaqxW/eladiZTfJQEAgBGEEHYQTp05Vj9ZfpzWbt+jz9/xvOIp1++SAADACEEIO0hnzZuof/7ksXr81V36yt1rlXLTfpcEAABGAELYEDivfqqu/ehc/eGlt3XNA+tkrfW7JAAAMMzx7MghcvGpM9TWldRPHtusymhI3/roXBlj/C4LAAAMU4SwIXTlmbPUFkvp50+8rqqSkC4/Y6bfJQEAgGGKEDaEjDH6ztlHqy2W1I/+tEmVJSF99pRpfpcFAACGIULYEAsEjG785LFqj6X03QdfVmVJUB9fOMXvsgAAwDDDiflFEHQC+o/lC3XKu8bo6799UX98+W2/SwIAAMMMIaxIoiFHKz9Tr3mTq3TZb/6mJ1/b5XdJAABgGCGEFVF5JKhfXXiCpo0p1SW3NWjt9j1+lwQAAIYJQliR1ZSFdcdFJ2pMeUQX3vqsNr3T7ndJAABgGChqCDPGnGWMecUYs9kYc3WB/q8aY9YbY140xvzZGHNkMevxy/jKqH590YkKOwF9+ufP6I3dnX6XBAAAfFa0EGaMcSTdJOnDko6WtNwYc3SvYX+TVG+tPVbSfZJuLFY9fjtiTKnuuOhExVNpffoXz2hnW8zvkgAAgI+KORO2WNJma+0Wa21C0t2SluQPsNaustZmp4WeljSq7+Uwe0KFbvvcYu3uiOuCXzyrPZ0Jv0sCAAA+KWYImyxpe956Y6atPxdJ+kMR6xkWjptarZ99pl6v796rC299TnvjKb9LAgAAPhgWJ+YbYz4tqV7SD/rpv9QY02CMaWhqajq0xRXBKUfV6T+XL9S6Ha269I4GxZKu3yUBAIBDrJghbIekqXnrUzJtPRhjPiDpW5LOtdbGC+3IWrvSWltvra0fO3ZsUYo91D54zATd+Mlj9dfNu3XFb/6mlJv2uyQAAHAIFTOEPSdppjFmujEmLOlTkh7MH2CMWSjpFnkBbGcRaxmWPrloiq4752j9cf07+sb9Lyqdtn6XBAAADpGiPTvSWpsyxlwm6RFJjqRfWmtfNsZcL6nBWvugvMOP5ZJ+a4yRpDestecWq6bh6MJ3T1dbLKV//dMmVUZD+u45RyvzZwEAAEaxoj7A21r7kKSHerV9J2/5A8X8/JHi8vcfpdaupH7xxOuqKgnpyjNn+V0SAAAosqKGMAyOMUbXfnSu2rqS+vc/v6qqkpA+957pfpcFAACKiBA2TBhj9E+fmK/2WErX/269KqJBLa2fuu8NAQDAiDQsblEBT9AJ6N+XH6dTZ9bpm/e/qIdfesvvkgAAQJEQwoaZSNDRLRcs0nFTq3XFb9bqiVd3+V0SAAAoAkLYMFQaDurWCxdrxtgyXXpHg9a80eJ3SQAAYIgRwoapqtKQbr9oscZVRHThL5/Vxrfb/C4JAAAMIULYMDauIqo7LjpRpeGgLvjFs9q6a6/fJQEAgCFCCBvmptaW6tcXL1bKTev8nz+jt1tjfpcEAACGACFsBDhqXIVu+9xitXYl9elfPKPmvQm/SwIAAAeJEDZCHDulWj//bL22N3fqwlufVXss6XdJAADgIBDCRpCTZozRf51/vNa/2aZLbm9QLOn6XRIAADhAhLAR5oy54/Wj8xbomdebddlda5R0036XBAAADgAhbARactxkXb9knh7dsFNf/+0LSqet3yUBAID9xLMjR6gLTjpSbV1J/eCRV1QZDen6JcfIGON3WQAAYJAIYSPYl05/l9q6krpl9RZVlYT09Q/N9rskAAAwSISwEcwYo6s/PEdtsaT+c9VmVZWEdMlpM/wuCwAADAIhrJcXml7Qjc/dqLm1czW3dq7mjJmjmdUzFXbCfpdWkDFGN3xsvtpiKX3/oQ2qiAb1qcVH+F0WAADYB0JYL0k3qXAgrN9v+b3ueeUeSVLQBPWu6ndp7pi5mlM7R0ePOVqza2arNFTqc7UeJ2D04/OOU0cspb9/YJ0qoiF99NiJfpcFAAAGYKwdWVfW1dfX24aGhqJ/Ttqm1djeqA3NG7SxeaM27N6gDc0b1BxrliQZGR1ZeaQ3Y5YJZ3Nr56o6Wl302vrTlXD1mV8+o7Xb9+hnn6nX6bPH+VYLAACQjDHPW2vrC/YRwgbPWqudnTu1sXmj1jev18bdG7WheYPe2vtWbszEsom5w5hH1x6tObVzNK503CG7crG1K6nlK5/Wll0d+vVFJ6p+Wu0h+VwAANAXIazI9sT29Jkx29a2TVben21ttLbPjNmUiikKmOLcpm1XR1zn/fQpNXXEdfelJ+mYSVVF+RwAADAwQpgPOpOdeqXlFa3fvT4Xzl7b85pSNiVJKg+Va3bt7Fw4m1s7V9OrpisYGJrT9Hbs6dLSm59Uwk3r3s+frBljy4dkvwAAYPAIYcNEwk3o1T2v5g5jbmjeoE3NmxRzY5KkiBPRrJpZ3mxZJpjNrJmpiBM5oM97ralD5/30KUWCAd33xVM0qbpkKL8cAACwD4SwYSyVTmlb27buGbPmDdq4e6Pak+2SJMc4mlE9I3fLjOwhzbJQ2aD2/9KOVi1f+bTGVkZ07+dPVl35gQU6AACw/whhI4y1Vo0djT3OMduwe4N2x3bnxhxZeWTu/LLshQC10cIn4T+3tVkX/OIZvWtsuX5z6UmqjIYO1ZcCAMBhjRA2SjR1NuUCWfZCgB0dO3L940vH5w5jZmfNxpeOlzFGq17ZqUtua9DxR9Tots8tVknY8fErAQDg8EAIG8Va4609Z8yaN2hr69bclZk1kZrcOWad7eP1i8eSOnX6bP3sgsUKB4tzdSYAAPAQwg4znclObWrZlJs129i8Ua/ueVWptHdlpnUjCiSmaVrZAr1nykn62DGLNXNs5SG7lxkAAIcLQhiUdJPavGezNjRv0B83N2ht0xrttd6hTOtG5SSO0rvKF+h9R56ic48+XkeO4ZYWAAAcLEIYCtrVuUsPbX5Cf3r9CW3c8zfFtFOSlE6VKZycpVmVx+kD09+tj86dr0nVw+M5mQAAjCSEMAzKjvYd+t2mx/XYtie1uf1vSmiPJCmdrFKJO1tzq4/XB2e8W2fNmauxFdzqAgCAfSGEYb9Za/V661Y9+Mpqrd7+lF7veEEp0yFJSsfrVG5na17tIn34qPfoA7PfperSsM8VAwAw/BDCcNDSNq2Nuzfpfzet1hONT+uNzpeUNl2SJDc2QdVmrhbU1eujs96t9x51hCq4FxkAAIQwDL1UOqUXd76s/920Wk+9+YzejG2QNQlZa5SOTdYY5xgdP+4EnTP7ZL17xmTuSwYAOCwRwlB0CTeh5976m/5302o99/az2pnYJJm0rHVku47QuNA8nTBhsc6dc5JOmDZWkSChDAAw+hHCcMh1Jjv11I4G/e7Vx7Vm53NqTm6RjJVNh2S7pmtSdJ5Onniizp5zghYeMUYhhxvHAgBGH0IYfNcab9Xj25/VQ68+rhd2N6jN3S7Ju0eZYjM0teRYvWfKSTp7zvGaN7laToAbxwIARj5CGIadXV279JdtT+vh1x7XS81rtDf9tiTvHmWB2FGaXr5A7516ij4yd55mj69UgFAGABiBCGEY9t7qeEuPvv5X/fH1v2rjnucVsy2SvHuUOfGZmlm5UO+fdorOmjNHM+rKeMQSAGBEIIRhRLHW6o32N/TIlif02OtPalPb35RUuyQpnRijUGKW5lYfrzNnnKIzZx+lqbXczR8A9lcynVRbvE1ticwr3qbWRGuuLe7GVRWuUk20RjXRGtVGa73lSI1KQ/zcHSxCGEa0tE3r1ZZX9fBrT+j/3nhSWzrWyVX3PcpK3NmaV3O8PnjUKXr/rCM1sarE54oPb9ZaJdNJdaW6FEvFFHNjiqVi3npmueC625VbjqViCjkhVUWqVBWu8t7zlivDlaqMVKoiXKGA4aIOHL7ctKuOZIda460Fg1Rboq27r1d/Z6pzwH07xpFr3YJ9USeaC2c10RrVRmp7hrVIz+BWHio/bI9gEMIwqqTSKa3fvV5/2Py4Hm98Wts71yut7nuUlds5mlk9RxMryzShslQTqko1qapc5eGwnIAjxzgKBoJyjCMn4ChoggXbC62P9F/4btpV3I3nAlBXMvNeKDANFKB6Babe+0jb9H7XFnEiigajijpRRYNRJdyE2hJt2pvc2+82RkaVkcrucBapVGW4smdwKxDeqsJVCjncUBjDQ9qmtTe5t0dYyg9VhYJU9r092T7gvqNONPf3vsd7uLLPv4n8vspIpYImqL3JvWqJtag53qyWWIu3HMssx/OWM+tdqa6CdYQCoVwwGyisZQNdZaRyxP+8zSKEYVRLuAm9sPMFPbT5CT355tN6q+sVWVP4f28HK2ACPcJZwAQGDG5Bs4++TPtAYbDHuExbwASUcBP7Dkj54SgVUyKd2O+v2TFOj3BUEizJLeevlwRL+rTl1p3uvj7rTlQRJyInUPjecUk36f0SyvzvvTXeqtZEa+6XUmu8NdeW398Wb5NV/z/fSoOlPUJaLsD1Cm7ZX1TZ95JgyWH7P3r0z1qrrlRXv7NO/c5IJdrUnmgf8D8uoUCo37CUH6R6rGfaIs6hfc5vV6qrZ1iL9wpuvQJdR7Kj4H4c46gqUtXjEGihsJZdro5UKxgIHtKvdbAIYTisdCY79WbHm+qIx7V9T4caW/ZqR+tevbVnr95q26u327rUGovLyJVMWlJaVaWO6sqDGlMR0pjyoGpKg6ouc1RV4sgE0nLTrlI2JTftyrVun/VUOlWwfaBxrt1HX69117p9flDnB50eQahASBpsIOoRppwSBQPBERk60jbdfZimV3jrEeAKhLtkOtnvfkOBUOHZtQKHTvN/aRbj0Gna9v272d/fxd5/n5LpZL9/3wbz9zKVTvX/dzazLHnhRFIuEOfee7V3v+1jXKF99vMZg91nv+MHGGdlFXNjPWalUjbV7/fKMU6/M1L7mpUazcE/4SYKzqrlB7j89dZ4a7/7qopUqSbSM6j1Xs+ffQs7h+aZx4QwoJf2WFLbdndq6+692rprr17f5S1v271Xuzp6zhZNrIpq2pgyTasr07QxpZpWV6bpdWU6orZU0dChvfN/2qZzv/TCTnjUTNcPJ9kZjfyZi/zZtvwQl53NyLYNdI5N70OnpcHSQYfwlC3cdyCHfYdS/uxs75ndQofvjbwgkQ0U/a3nxvce12v8YMbs7z5z43Nv/Y8LO+EeM08DzUqVhbiqeyik0intie8pOKuWf5g0u74nvqfffyfloXJdtvAynT/3/KLWPFAIG55zd0CRVURDmje5SvMmV/Xpa4sltW1Xp17PBLRsUHvk5bfVvLc7oBkjTayMeuGsrkzT84La1CIFtIAJKGACCgU4n6lYjDEqDZWqNFSqCWUT9mvbpJvMHWIqNPuWm3lLtKkr2SUn4CgUDPU5dD2Y8xSDgeDgD2P3s/8e+zoMz5HEyBMMBFVXUqe6krpBjU/btNribf2GtRlVM4pc8cCYCQP2Q2tXMi+YebNnr2fW93R2H8IyRppUVaLpdWU6ckyppteV5WbTjqgtVTjILy8AOBwwEwYMkaqSkBZMrdaCqdV9+vZ0JrR1d2fm8GYmqO3u1O9efEutXd0BLWCkSdUlPYLZ9LpSHTmmTFNrCGgAcLgghAFDpLo0rONKwzquQEBr2ZvIO7zZmZtN+5+1O9Qe6z6Z1wkYTa4u6T7/bIx3/tm0ujJNqSnhQecAMIoQwoBDoKYsrJqysI4/oqZHu7VWzXsTBQ9vrtnWoo54z4A2pabEmz0bU6pxlVGNrYhobHlEYysiqiuPaEx5mKAGACMEIQzwkTFGY8ojGlMe0aIja3v0WWu1e2+iz+HNrbu8gNYeL3w5fE1pyAtnmWA2tjyiuore72GNKYvI4cHoAOAbQhgwTBljVFfuBan6abV9+rsSrnZ1xLWzPa5dHXE15b1nl9e80aKm9rhiyb6XaAeMVFsWUV15uM+MWv772IqIqktCChDYAGBIEcKAEaok7Ghqbek+H2BurdXehKtd7XE1dcRz7z1CW0dCW5r2qqkjrkSqb2BzAkZ15eE+AS0X1MojGlsR1tjyqCpLRubNXQHgUCOEAaOcMUblkaDKI0FNqysbcKy1Vm2xVL8za957QhvfateujrhS6b63uAk7AS+wZQ995s2odYc3b/atPEJgA3D4IoQByDHGqKokpKqSkN41tnzAsem0VWtXMm82rft9V3tCTR1xvdUa04s7WrW7I64CeU2RYKDPoc/scm1pWFUlIVWXevVUlYZUQWgDMIoQwgAckEDA5K76nDm+YsCxbtqqpTPR78xaU3tc25s7tWZbi5o7E+rvHtJOwKgyGlR1aViVJSFV54W06pKQ15YX3qpLugNcJHhoHzEFAPtCCANQdN45Zd4s15x9PAko5abVvDehls6k9nQmtKcrqdaupFo7vfc9XQm1dqW8vs5E7mkFbbFkv+FNkqKhgKpLwrlQlg1o3TNt4VyY6w52YVVEg1yUAKAoCGEAhpWgE9C4yqjGVUb3a7t02qo9lsoLaknt6UxqT1dSbV1eoMu2tXYl9UZzZ269K+n2u19jpMpor8DWK6j1CHal3W3RUIDDpwD6RQgDMCoEAkZVpV4IOkIDXzHaWzzl9pxty826JdWaDW9d3X07Wrpy626hk90ywsFAbnatO7iFc0GuPBpURSTovUe9iye8d6+vNOQwCweMYoQwAIe9SNDRuApH4yr2b/bNWquOeCoX2lq7eoe4RGYWznu9uSemDW+1q7Ur2eNpCP0xRioPeyGtPNL9ng1s5ZGQKvICXHmBIJe9MpYb8wLDDyEMAA6QMUYV0ZAqoiFN3c9tU25ae+Ou2uNeIOuIpdSeee+93h7LjImn1B5L6a3WWPe4QYQ5SSoNO7mgVhEJqiIaKhzsege5Xn08FgsYOoQwAPBB0AmoqjSgqtLQQe0nnbbam+gZ3Npj2TCX9Jbzwl1+0Gtqj6s9lvTa4qkBL2zIioYCuRm47Cxb/mHV/PWScFClYUclYUclIUelYe8VDTkqzfRFgpw3h8MXIQwARrBAoHs2TlUHvh9rrToTbm62rTu49Qxy7T36vRm67c2dPWbqBjpPrjdjpJKQF9JKMiGteznYT3t2OdhPe/e2XByB4YwQBgCQMUZlkaDKIkGNrzzw/VhrFU+l1RZLqivhqivpqjPhKpbw3juT2eVU3rJbcHlne6x726SrroSreIHHag38dRUv5EVDAUWDXDyBA0cIAwAMGWOMoiHvkGMxuGmrrkwg60q46kymupczoa8rE/K6kml1JVK5IJgfCrsSrpra4964vPb9DXmS96iuSCigSNCbeYuGvMOs0cxMXDToKJJ779XXaz23jwJj8/fPuXmjAyEMADBiOIHuZ6EWg5u2imVn7zLvnZkg1zvodSVdxZNpxVLe2FgyrXgq05b0Al0s6d3+JNsXy/Yl00q4+x/4spyA6Q5owUCPwNZ/yOs7plBf/nbhYEAhx3tFMstcaTt0CGEAAGQ4ge7DssWWTttcUMu+xzJBLZ50Fcu2ZfrjeUEvG+Zi2dDXa+yezkTPsanuvsFcgDEQJ2AUcozCjhfSwk5AoUxAyy5HnIBCQW9MKH+c0x3svDbTYz27j55tJrd9oVCYPyYUCIyow8OEMAAAfBAIGO/K0fChe66ptVYJN50Lernw10+4S7rejF0i5b0nU1YJ11XStXltmffcOKtEZl/tsVRuXCLVPSbp2lzbUAs5JhfUeoY30yfoLTthqs4+dtKQ1zBYhDAAAA4TxhhFgo73QPuSg7s9ylCw1iqVtj0CWiIvqCXdtOI9wlvhMYWDXrbNFmhLqyOeUvIgDgkPBUIYAADwhTEmN3N1ODo8v2oAAACfEcIAAAB8UNQQZow5yxjzijFmszHm6gL9EWPMPZn+Z4wx04pZz//f3v2GXHnXcRx/f7rdyG1gY5MotRQmhf1ZhozVoAezB41FPqiYUhFjEI22WURt9aAH0ZMialkS2P4wmuQDW0NC3GKOCIqlbbZNLRCzqTm6rbQWNef69uBc0vFOwaOe/bzO3i843L/rd19c53P4cZ/7e37X71yXJEnShWJsRViSKWAtcAOwBFiVZMmM3W4B/lZVVwHfBr4+rjySJEkXknHOhF0D7KmqvVV1DNgArJixzwrgga69EVgeb/IlSZJeBcZZhM0D9g9tH+j6TrlPVR0HjgJXjDGTJEnSBaEXC/OTfCrJ9iTbp6enW8eRJEk6Z+Mswg4CC4a253d9p9wnySxgDvCXmQeqqnVVtayqls2dO3dMcSVJkl454yzCtgGLkyxKcjGwEtg0Y59NwCe79keArVXnelcrSZKkC9/YrphfVceT3AY8AkwB91XVrnRYIQAABRFJREFUziRfBbZX1SbgXuCHSfYAf2VQqEmSJE28sd62qKo2A5tn9H1lqP1v4KPjzCBJknQh6sXCfEmSpEljESZJktSARZgkSVIDFmGSJEkNWIRJkiQ1YBEmSZLUgEWYJElSA+nbBeqTTAN/HPPTXAkcHvNzaLwcw/5zDPvPMew3x+/8eHNVnfKei70rwl4JSbZX1bLWOXT2HMP+cwz7zzHsN8dv/DwdKUmS1IBFmCRJUgMWYae2rnUAnTPHsP8cw/5zDPvN8Rsz14RJkiQ14EyYJElSAxZhMyT5QJLfJ9mT5K7WeTSaJAuSPJ5kV5KdSVa3zqTRJZlK8lSSn7bOotEleV2SjUl+l2R3kve0zqTRJPlc9x76bJIfJXlt60yTyCJsSJIpYC1wA7AEWJVkSdtUGtFx4PNVtQS4FviMY9hLq4HdrUPorH0H2FJVbwWuxrHslSTzgDuAZVX1dmAKWNk21WSyCDvZNcCeqtpbVceADcCKxpk0gqo6VFVPdu1/MHjzn9c2lUaRZD5wI3BP6ywaXZI5wPuAewGq6lhVHWmbSmdhFjA7ySzgEuBPjfNMJIuwk80D9g9tH8B/4L2VZCGwFHiibRKN6G7gi8B/WgfRWVkETAP3d6eU70lyaetQOnNVdRD4JvAccAg4WlWPtk01mSzCNJGSXAb8GPhsVf29dR6dmSQfBP5cVb9pnUVnbRbwbuD7VbUU+Cfg+toeSXI5g7NAi4A3Apcm+XjbVJPJIuxkB4EFQ9vzuz71SJKLGBRg66vqodZ5NJLrgA8l2cdgOcD1SR5sG0kjOgAcqKoTM9AbGRRl6o/3A3+oqumqegl4CHhv40wTySLsZNuAxUkWJbmYwULETY0zaQRJwmAtyu6q+lbrPBpNVX2pquZX1UIGf39bq8pP4D1SVc8D+5O8petaDuxqGEmjew64Nskl3XvqcvxyxVjMah3gQlJVx5PcBjzC4Nsg91XVzsaxNJrrgE8AzyTZ0fV9uao2N8wkvdrcDqzvPszuBW5unEcjqKonkmwEnmTwjfOn8Or5Y+EV8yVJkhrwdKQkSVIDFmGSJEkNWIRJkiQ1YBEmSZLUgEWYJElSAxZhknovyctJdgw9ztsV2pMsTPLs+TqeJJ3gdcIkTYJ/VdW7WoeQpFE4EyZpYiXZl+QbSZ5J8uskV3X9C5NsTfJ0kseSvKnrf32SnyT5bfc4cauWqSQ/SLIzyaNJZnf735FkV3ecDY1epqSesgiTNAlmzzgdedPQ745W1TuA7wF3d33fBR6oqncC64E1Xf8a4OdVdTWD+x2euGPGYmBtVb0NOAJ8uOu/C1jaHefT43pxkiaTV8yX1HtJXqiqy07Rvw+4vqr2djd2f76qrkhyGHhDVb3U9R+qqiuTTAPzq+rFoWMsBH5WVYu77TuBi6rqa0m2AC8ADwMPV9ULY36pkiaIM2GSJl2dpj2KF4faL/O/9bQ3AmsZzJptS+I6W0lnzCJM0qS7aejnr7r2L4GVXftjwC+69mPArQBJppLMOd1Bk7wGWFBVjwN3AnOA/5uNk6TT8VObpEkwO8mOoe0tVXXiMhWXJ3mawWzWqq7vduD+JF8ApoGbu/7VwLoktzCY8boVOHSa55wCHuwKtQBrqurIeXtFkiaea8IkTaxuTdiyqjrcOoskzeTpSEmSpAacCZMkSWrAmTBJkqQGLMIkSZIasAiTJElqwCJMkiSpAYswSZKkBizCJEmSGvgvNH1KUmfpfQsAAAAASUVORK5CYII=\n","text/plain":["<Figure size 720x576 with 1 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"Ks_nfT9KX4bn"},"source":["### Regression problem"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"B9s447WRYg_K","executionInfo":{"status":"ok","timestamp":1630035821322,"user_tz":-420,"elapsed":25,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"3f152148-2591-4826-e5c3-44ec9b9e3fed"},"source":["import numpy as np\n","\n","X = tf.constant(np.arange(0, 100))\n","X"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<tf.Tensor: shape=(100,), dtype=int64, numpy=\n","array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16,\n","       17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33,\n","       34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50,\n","       51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67,\n","       68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84,\n","       85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99])>"]},"metadata":{},"execution_count":178}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"4xGj01LEYrdP","executionInfo":{"status":"ok","timestamp":1630035821323,"user_tz":-420,"elapsed":18,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"affda64f-4b89-40fd-9908-4b0908e642ed"},"source":["y = tf.constant(np.arange(0, 100)+10)\n","y"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<tf.Tensor: shape=(100,), dtype=int64, numpy=\n","array([ 10,  11,  12,  13,  14,  15,  16,  17,  18,  19,  20,  21,  22,\n","        23,  24,  25,  26,  27,  28,  29,  30,  31,  32,  33,  34,  35,\n","        36,  37,  38,  39,  40,  41,  42,  43,  44,  45,  46,  47,  48,\n","        49,  50,  51,  52,  53,  54,  55,  56,  57,  58,  59,  60,  61,\n","        62,  63,  64,  65,  66,  67,  68,  69,  70,  71,  72,  73,  74,\n","        75,  76,  77,  78,  79,  80,  81,  82,  83,  84,  85,  86,  87,\n","        88,  89,  90,  91,  92,  93,  94,  95,  96,  97,  98,  99, 100,\n","       101, 102, 103, 104, 105, 106, 107, 108, 109])>"]},"metadata":{},"execution_count":179}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"SlHZjM-WY9et","executionInfo":{"status":"ok","timestamp":1630035821323,"user_tz":-420,"elapsed":14,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"3a67e7d3-6ac5-4ef3-f53b-30805d3fcd56"},"source":["split_size = int(0.8 * len(X))\n","X_train = X[:split_size]\n","y_train = y[:split_size]\n","\n","X_test = X[split_size:]\n","y_test = y[split_size:]\n","\n","len(X_train), len(X_test), len(y_train), len(y_test)"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(80, 20, 80, 20)"]},"metadata":{},"execution_count":180}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"1SWHCcdNZPHF","executionInfo":{"status":"ok","timestamp":1630035821664,"user_tz":-420,"elapsed":350,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"2a689c9c-9619-42ec-8c17-fb90254e3c42"},"source":["from tensorflow.keras import layers\n","# Create a model\n","model_1 = tf.keras.Sequential([\n","  layers.Dense(100),\n","  layers.Dense(10),\n","  layers.Dense(1)\n","])\n","# Compile the model\n","model_1.compile(\n","    loss=\"mae\",\n","    optimizer=tf.keras.optimizers.Adam(),\n","    metrics=[\"mse\"]\n",")\n","# Fit the model\n","model_1_history = model_1.fit(X_train,\n","                              y_train,\n","                              validation_data=(X_test, y_test),\n","                              epochs=5)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Epoch 1/5\n","3/3 [==============================] - 0s 55ms/step - loss: 42.2303 - mse: 2137.3950 - val_loss: 73.5147 - val_mse: 5421.1855\n","Epoch 2/5\n","3/3 [==============================] - 0s 10ms/step - loss: 36.5655 - mse: 1588.9636 - val_loss: 59.8589 - val_mse: 3593.4492\n","Epoch 3/5\n","3/3 [==============================] - 0s 10ms/step - loss: 30.1282 - mse: 1042.0718 - val_loss: 46.2305 - val_mse: 2142.7500\n","Epoch 4/5\n","3/3 [==============================] - 0s 10ms/step - loss: 24.2274 - mse: 659.8851 - val_loss: 32.2626 - val_mse: 1042.9667\n","Epoch 5/5\n","3/3 [==============================] - 0s 13ms/step - loss: 17.8485 - mse: 341.8727 - val_loss: 18.0302 - val_mse: 325.3709\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"VRbBZlefa4Gs","executionInfo":{"status":"ok","timestamp":1630035821665,"user_tz":-420,"elapsed":9,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"542688e3-746b-4b29-af36-6d8202ee1719"},"source":["model_1.evaluate(X_test, y_test)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["1/1 [==============================] - 0s 12ms/step - loss: 18.0302 - mse: 325.3709\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["[18.0301570892334, 325.3708801269531]"]},"metadata":{},"execution_count":182}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"hZem-m18a8SS","executionInfo":{"status":"ok","timestamp":1630035822236,"user_tz":-420,"elapsed":574,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"7323e08c-ac17-45a8-ff28-d13af100ee51"},"source":["model_1_pred = model_1.predict(X_test)\n","y_pred = model_1_pred\n","y_pred"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([[72.848366],\n","       [73.755905],\n","       [74.66342 ],\n","       [75.570946],\n","       [76.47846 ],\n","       [77.38599 ],\n","       [78.29351 ],\n","       [79.20104 ],\n","       [80.10856 ],\n","       [81.01608 ],\n","       [81.92361 ],\n","       [82.83113 ],\n","       [83.738655],\n","       [84.64617 ],\n","       [85.5537  ],\n","       [86.46122 ],\n","       [87.368744],\n","       [88.27627 ],\n","       [89.183784],\n","       [90.09131 ]], dtype=float32)"]},"metadata":{},"execution_count":183}]},{"cell_type":"code","metadata":{"id":"q4AxOKTlaovW"},"source":["def plot_predictions(train_data=X_train,\n","                     train_label=y_train,\n","                     test_data=X_test,\n","                     test_label=y_test,\n","                     predictions=y_pred):\n","  \"\"\"\n","  Plot training data,test data and compares the predictions to ground truth \n","  \"\"\"\n","  plt.figure(figsize=(10,7))\n","  # plot data training in blue\n","  plt.scatter(train_data, train_label, c=\"b\", label=\"Training\")\n","  # plot data testing in green\n","  plt.scatter(test_data, test_label, c=\"g\", label=\"Testing\")\n","  # plt predictions in red\n","  plt.scatter(test_data, predictions, c=\"r\", label=\"Predictions\")\n","\n","  plt.legend()"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":428},"id":"U7QHc9SXa2Y0","executionInfo":{"status":"ok","timestamp":1630035822237,"user_tz":-420,"elapsed":25,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"f6f0e604-fdec-4701-a347-0fdd248f0303"},"source":["plot_predictions(predictions=model_1_pred)"],"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 720x504 with 1 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"yza-BawvbJyk"},"source":["### Classification Problems"]},{"cell_type":"markdown","metadata":{"id":"l4_5RnFFbtLY"},"source":["#### Binary Classification"]},{"cell_type":"code","metadata":{"id":"rv5PQlkabyKn"},"source":["from sklearn.datasets import make_circles\n","\n","# make 1000 samples\n","n_samples = 1000\n","# create circles\n","X, y = make_circles(n_samples=n_samples,\n","                    shuffle=True,\n","                    noise=0.03,\n","                    random_state=42)\n"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"yVwwrRWoc3JJ","executionInfo":{"status":"ok","timestamp":1630035822238,"user_tz":-420,"elapsed":24,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"25ef5e57-cbec-4d4a-e90d-6cfa3d49f8e8"},"source":["X.shape, y.shape"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["((1000, 2), (1000,))"]},"metadata":{},"execution_count":187}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":203},"id":"_FsQXlxic7_L","executionInfo":{"status":"ok","timestamp":1630035822239,"user_tz":-420,"elapsed":22,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"2f932a2b-8d70-40eb-9ec5-47bb39a1f147"},"source":["import pandas as pd\n","\n","df = pd.DataFrame({\"X1\":X[:,0],\n","                  \"X2\":X[:,1],\n","                  \"y\": y})\n","df.head()"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>X1</th>\n","      <th>X2</th>\n","      <th>y</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>0.754246</td>\n","      <td>0.231481</td>\n","      <td>1</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>-0.756159</td>\n","      <td>0.153259</td>\n","      <td>1</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>-0.815392</td>\n","      <td>0.173282</td>\n","      <td>1</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>-0.393731</td>\n","      <td>0.692883</td>\n","      <td>1</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>0.442208</td>\n","      <td>-0.896723</td>\n","      <td>0</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["         X1        X2  y\n","0  0.754246  0.231481  1\n","1 -0.756159  0.153259  1\n","2 -0.815392  0.173282  1\n","3 -0.393731  0.692883  1\n","4  0.442208 -0.896723  0"]},"metadata":{},"execution_count":188}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"R7p-6hiUdbhO","executionInfo":{"status":"ok","timestamp":1630035822239,"user_tz":-420,"elapsed":20,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"40bd1270-47cb-48e9-f5cb-2420eaabd701"},"source":["X[:,0].shape, X[:,1].shape"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["((1000,), (1000,))"]},"metadata":{},"execution_count":189}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":500},"id":"JMRmSqgHchBH","executionInfo":{"status":"ok","timestamp":1630035823616,"user_tz":-420,"elapsed":1393,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"19cb7da6-8dbc-48fa-fe0f-e22523fa049e"},"source":["import matplotlib.pyplot as plt\n","\n","plt.figure(figsize=(10,8))\n","plt.scatter(X[:,0], X[:,1], c=y, cmap=plt.cm.RdYlBu)"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.collections.PathCollection at 0x7f781c0f4b50>"]},"metadata":{},"execution_count":190},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 720x576 with 1 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"VywNAOYpdXnX"},"source":["##### Split data \n","\n","* using array slicing.\n","* using `sklearn.model_selection.train_test_splits()`"]},{"cell_type":"code","metadata":{"id":"IvsKlSavd7oW"},"source":["# (X_train,y_train), (X_test, y_test)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"hAB9OC8neCVg","executionInfo":{"status":"ok","timestamp":1630035823618,"user_tz":-420,"elapsed":21,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"9727e7a5-24db-4706-cdde-567a80afcf79"},"source":["split_size = int(0.8 * len(X))\n","X_train = X[:split_size]\n","y_train = y[:split_size]\n","\n","X_test = X[split_size:]\n","y_test = y[split_size:]\n","\n","len(X_train), len(X_test), len(y_train), len(y_test)"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(800, 200, 800, 200)"]},"metadata":{},"execution_count":192}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":282},"id":"CIjSUNV6eVU1","executionInfo":{"status":"ok","timestamp":1630035823618,"user_tz":-420,"elapsed":15,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"23988f7b-502c-4dee-fec9-ae6ef30f5c0b"},"source":["plt.subplot(1,2, 1)\n","plt.scatter(X_train[:,0], X_train[:, 1], c=y_train, cmap=plt.cm.RdYlBu)\n","\n","plt.subplot(1,2 ,2)\n","plt.scatter(X_test[:,0], X_test[:, 1], c=y_test, cmap=plt.cm.RdYlBu)"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.collections.PathCollection at 0x7f781bf29410>"]},"metadata":{},"execution_count":193},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 2 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"o3VZcdRaef9U","executionInfo":{"status":"ok","timestamp":1630035824145,"user_tz":-420,"elapsed":539,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"dcc8dca9-a5c6-4a44-ed41-e85951ad0ad5"},"source":["# create model\n","model_2 = tf.keras.Sequential([\n","  layers.Dense(100, activation=\"relu\"),\n","  layers.Dense(100, activation=\"relu\"),\n","  layers.Dense(1, activation=\"sigmoid\")\n","])\n","# compile the model\n","model_2.compile(\n","    loss=tf.keras.losses.BinaryCrossentropy(),\n","    optimizer=tf.keras.optimizers.Adam(),\n","    metrics=[\"accuracy\"]\n",")\n","# fit the model\n","model_2_history = model_2.fit(X_train,\n","                              y_train,\n","                              validation_data=(X_test, y_test),\n","                              epochs=10)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Epoch 1/10\n","25/25 [==============================] - 1s 8ms/step - loss: 0.6904 - accuracy: 0.4988 - val_loss: 0.6867 - val_accuracy: 0.5050\n","Epoch 2/10\n","25/25 [==============================] - 0s 3ms/step - loss: 0.6826 - accuracy: 0.4988 - val_loss: 0.6848 - val_accuracy: 0.5050\n","Epoch 3/10\n","25/25 [==============================] - 0s 3ms/step - loss: 0.6763 - accuracy: 0.5000 - val_loss: 0.6777 - val_accuracy: 0.5400\n","Epoch 4/10\n","25/25 [==============================] - 0s 3ms/step - loss: 0.6683 - accuracy: 0.5412 - val_loss: 0.6676 - val_accuracy: 0.5850\n","Epoch 5/10\n","25/25 [==============================] - 0s 3ms/step - loss: 0.6561 - accuracy: 0.6500 - val_loss: 0.6511 - val_accuracy: 0.5650\n","Epoch 6/10\n","25/25 [==============================] - 0s 3ms/step - loss: 0.6361 - accuracy: 0.5763 - val_loss: 0.6308 - val_accuracy: 0.7950\n","Epoch 7/10\n","25/25 [==============================] - 0s 3ms/step - loss: 0.6095 - accuracy: 0.8263 - val_loss: 0.5907 - val_accuracy: 0.8800\n","Epoch 8/10\n","25/25 [==============================] - 0s 4ms/step - loss: 0.5738 - accuracy: 0.8637 - val_loss: 0.5472 - val_accuracy: 0.9600\n","Epoch 9/10\n","25/25 [==============================] - 0s 3ms/step - loss: 0.5204 - accuracy: 0.9563 - val_loss: 0.4919 - val_accuracy: 0.9400\n","Epoch 10/10\n","25/25 [==============================] - 0s 3ms/step - loss: 0.4550 - accuracy: 0.9725 - val_loss: 0.4143 - val_accuracy: 0.9950\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"ISccq1_6fTwj"},"source":["def plot_decision_boundary(model, X, y):\n","  \"\"\"\n","  Plots the decision boundary created by a model predicting on X.\n","  This function has been adapted from two phenomenal resources:\n","   1. CS231n - https://cs231n.github.io/neural-networks-case-study/\n","   2. Made with ML basics - https://github.com/GokuMohandas/MadeWithML/blob/main/notebooks/08_Neural_Networks.ipynb\n","  \"\"\"\n","  # Define the axis boundaries of the plot and create a meshgrid\n","  x_min, x_max = X[:,0].min() - 0.1, X[:,0].max()+0.1\n","  y_min, y_max = X[:,1].min() - 0.1, X[:,1].max()+0.1\n","  xx, yy = np.meshgrid(np.linspace(x_min, x_max, 100),\n","                       np.linspace(y_min, y_max, 100))\n","  # Create X value (we're going to predictions on these):\n","  # ravel() is flatten an array\n","  # np.c_ is concat array along the second axis\n","  x_in = np.c_[xx.ravel(), yy.ravel()] # stack 2d array together\n","\n","  # make predictions\n","  y_pred = model.predict(x_in)\n","\n","  # check for multi-class:\n","  if len(y_pred[0]) > 1:\n","    print(\"doing multiclass classification\")\n","    # We have to reshape our predictions to get them ready for plotting\n","    y_pred = np.argmax(y_pred, axis=1).reshape(xx.shape)\n","  else:\n","    print(\"doing binary classificatioon\")\n","    y_pred = np.round(y_pred).reshape(xx.shape)\n","\n","  # plot the decision boundary\n","  plt.contourf(xx, yy, y_pred, cmap=plt.cm.RdYlBu, alpha=0.7)\n","  plt.scatter(X[:,0],X[:,1], c=y, s=40, cmap=plt.cm.RdYlBu)\n","  plt.xlim(xx.min(), xx.max())\n","  plt.ylim(yy.min(), yy.max())"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"vAoQRcJxfsZH","executionInfo":{"status":"ok","timestamp":1630035824146,"user_tz":-420,"elapsed":7,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"f1ddf245-094d-41e5-b736-cc4a953ff8da"},"source":["model_2.evaluate(X_test, y_test)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["7/7 [==============================] - 0s 2ms/step - loss: 0.4143 - accuracy: 0.9950\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["[0.4142822027206421, 0.9950000047683716]"]},"metadata":{},"execution_count":196}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":282},"id":"LXQ9Dtgefg5v","executionInfo":{"status":"ok","timestamp":1630035824731,"user_tz":-420,"elapsed":588,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"a90872f7-2d71-4e10-a066-8240f5a21125"},"source":["plot_decision_boundary(model_2, X_test, y_test)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["doing binary classificatioon\n"],"name":"stdout"},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"aKJtlCkpflRX"},"source":["#### Multiclass classification"]},{"cell_type":"code","metadata":{"id":"HgW_Ebaggncc"},"source":["# We done it in the previous section"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"jGGVeY_5guCg"},"source":["## Image classification\n","\n","For this problem, we'll using fashion-mnist dataset."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"3jf1sGd7hffG","executionInfo":{"status":"ok","timestamp":1630035825290,"user_tz":-420,"elapsed":566,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"08af2379-cc82-4301-e19c-0fb3ec316b4b"},"source":["from tensorflow.keras.datasets import fashion_mnist\n","\n","(X_train, y_train), (X_test, y_test) = fashion_mnist.load_data()\n","\n","X_train.shape, X_test.shape, y_train.shape, y_test.shape"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["((60000, 28, 28), (10000, 28, 28), (60000,), (10000,))"]},"metadata":{},"execution_count":199}]},{"cell_type":"code","metadata":{"id":"sMBTTgaKnfIP"},"source":["class_names = ['T-shirt/top', 'Trouser', 'Pullover', 'Dress', 'Coat',\n","               'Sandal', 'Shirt', 'Sneaker', 'Bag', 'Ankle boot']"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":281},"id":"lS9gOncSlAic","executionInfo":{"status":"ok","timestamp":1630035825838,"user_tz":-420,"elapsed":549,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"90a1b0e9-01dd-42c0-fd33-62f964f461a5"},"source":["import random \n","random_num = random.randint(0,len(X_train)-1)\n","img = X_train[random_num]\n","label = y_train[random_num]\n","plt.imshow(img)\n","plt.title(class_names[label])\n","plt.colorbar()\n","plt.show()"],"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 2 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"LpZ7KMrTnvNc"},"source":["#### Preprocess data"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"hLR7V0GYo_53","executionInfo":{"status":"ok","timestamp":1630035825839,"user_tz":-420,"elapsed":15,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"4f8b565e-d2d5-493b-babf-10cc97e3c01c"},"source":["len(class_names)"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["10"]},"metadata":{},"execution_count":202}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"yVf2UYtGpB-d","executionInfo":{"status":"ok","timestamp":1630035825839,"user_tz":-420,"elapsed":8,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"23d5d328-e48b-4787-b1fa-e61488695781"},"source":["y_train"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([9, 0, 0, ..., 3, 0, 5], dtype=uint8)"]},"metadata":{},"execution_count":203}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"SwJtTvLroMMU","executionInfo":{"status":"ok","timestamp":1630035826382,"user_tz":-420,"elapsed":547,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"9995893e-6aa9-47b4-f381-7b6ce94c9e36"},"source":["X_train = X_train/255.\n","y_train = tf.one_hot(y_train, depth=len(class_names))\n","\n","X_test = X_test/255.\n","y_test =tf.one_hot(y_test, depth=len(class_names))\n","\n","tf.reduce_max(X_train), tf.reduce_min(X_train), y_train[0]"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(<tf.Tensor: shape=(), dtype=float64, numpy=1.0>,\n"," <tf.Tensor: shape=(), dtype=float64, numpy=0.0>,\n"," <tf.Tensor: shape=(10,), dtype=float32, numpy=array([0., 0., 0., 0., 0., 0., 0., 0., 0., 1.], dtype=float32)>)"]},"metadata":{},"execution_count":204}]},{"cell_type":"markdown","metadata":{"id":"JfjdACHCo4bB"},"source":["#### Build and train Conv2D model"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"ANBJi15jpU0A","executionInfo":{"status":"ok","timestamp":1630035826383,"user_tz":-420,"elapsed":10,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"892d9424-476c-4f0a-e51b-44cf311a46e3"},"source":["X_train.shape"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(60000, 28, 28)"]},"metadata":{},"execution_count":205}]},{"cell_type":"code","metadata":{"id":"PRlp2BBEvMZ_"},"source":["# Create callback\n","# SAVE_DIR = \"save_dir\"\n","\n","def create_checkpoint_callback(model_name):\n","  return tf.keras.callbacks.ModelCheckpoint(filepath=f\"{model_name}/checkpoint.ckpt\",\n","                                                          #  monitor=\"val_accuracy\",\n","                                                           save_best_only=True,\n","                                                           save_weights_only=True,\n","                                                           save_freq=\"epoch\")"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"k6PTaNOOpOP0","executionInfo":{"status":"ok","timestamp":1630035908765,"user_tz":-420,"elapsed":82388,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"03f7c5b0-d81f-49d0-80e9-d61d02a39fbb"},"source":["from tensorflow.keras import layers\n","\n","# Build a model\n","inputs = layers.Input(shape=(28, 28,1), name=\"input_layer\")\n","\n","x = layers.Conv2D(filters=32,\n","                 kernel_size=3,\n","                 padding=\"same\",\n","                  activation=\"relu\")(inputs)\n","x = layers.Conv2D(filters=32,\n","                  kernel_size=3,\n","                  padding=\"same\",\n","                  activation=\"relu\")(x)\n","x = layers.MaxPooling2D()(x)\n","\n","x = layers.Flatten()(x)\n","\n","x = layers.Dense(100, activation='relu')(x)\n","x = layers.Dense(10, activation='relu')(x)\n","\n","outputs = layers.Dense(len(class_names), activation=\"softmax\")(x)\n","\n","model_3 = tf.keras.Model(inputs, outputs, name=\"model_3_conv2d\")\n","\n","# Compile the model\n","model_3.compile(\n","    loss=tf.keras.losses.CategoricalCrossentropy(),\n","    optimizer=tf.keras.optimizers.Adam(),\n","    metrics=[\"accuracy\"]\n",")\n","# Fit the model\n","model_3_history = model_3.fit(X_train,\n","                              y_train,\n","                              validation_data=(X_test, y_test),\n","                              epochs=10,\n","                              callbacks=[create_checkpoint_callback(model_3.name)])"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Epoch 1/10\n","1875/1875 [==============================] - 7s 3ms/step - loss: 0.4460 - accuracy: 0.8440 - val_loss: 0.3137 - val_accuracy: 0.8883\n","Epoch 2/10\n","1875/1875 [==============================] - 6s 3ms/step - loss: 0.2563 - accuracy: 0.9075 - val_loss: 0.3217 - val_accuracy: 0.8872\n","Epoch 3/10\n","1875/1875 [==============================] - 6s 3ms/step - loss: 0.2060 - accuracy: 0.9246 - val_loss: 0.2448 - val_accuracy: 0.9113\n","Epoch 4/10\n","1875/1875 [==============================] - 6s 3ms/step - loss: 0.1683 - accuracy: 0.9379 - val_loss: 0.2434 - val_accuracy: 0.9132\n","Epoch 5/10\n","1875/1875 [==============================] - 6s 3ms/step - loss: 0.1402 - accuracy: 0.9490 - val_loss: 0.2400 - val_accuracy: 0.9192\n","Epoch 6/10\n","1875/1875 [==============================] - 6s 3ms/step - loss: 0.1162 - accuracy: 0.9565 - val_loss: 0.2527 - val_accuracy: 0.9152\n","Epoch 7/10\n","1875/1875 [==============================] - 6s 3ms/step - loss: 0.0933 - accuracy: 0.9652 - val_loss: 0.2571 - val_accuracy: 0.9214\n","Epoch 8/10\n","1875/1875 [==============================] - 6s 3ms/step - loss: 0.0740 - accuracy: 0.9730 - val_loss: 0.2948 - val_accuracy: 0.9213\n","Epoch 9/10\n","1875/1875 [==============================] - 6s 3ms/step - loss: 0.0621 - accuracy: 0.9777 - val_loss: 0.3369 - val_accuracy: 0.9168\n","Epoch 10/10\n","1875/1875 [==============================] - 6s 3ms/step - loss: 0.0510 - accuracy: 0.9812 - val_loss: 0.3387 - val_accuracy: 0.9179\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"TUq4uYujgZII"},"source":["# model_3.load_weights(f\"{SAVE_DIR}/{model_3.name}\")"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"XA4lGBI2tt12","executionInfo":{"status":"ok","timestamp":1630035909923,"user_tz":-420,"elapsed":1170,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"270c1154-c6a7-48da-96d0-8c7b01693171"},"source":["model_3.evaluate(X_test, y_test)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["313/313 [==============================] - 1s 3ms/step - loss: 0.3387 - accuracy: 0.9179\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["[0.3386678993701935, 0.917900025844574]"]},"metadata":{},"execution_count":209}]},{"cell_type":"markdown","metadata":{"id":"MenMRVhWtFbl"},"source":["#### Use Another data set\n","\n","This time we'll using 10 classes from Food101 dataset"]},{"cell_type":"code","metadata":{"id":"pVCBEbMIgjRW","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1630035917167,"user_tz":-420,"elapsed":7246,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"11305b16-e60b-4338-da40-6a964ffb63d1"},"source":["import zipfile\n","\n","!wget https://storage.googleapis.com/ztm_tf_course/food_vision/10_food_classes_10_percent.zip\n","\n","zip_ref = zipfile.ZipFile(\"10_food_classes_10_percent.zip\")\n","zip_ref.extractall()\n","zip_ref.close()"],"execution_count":null,"outputs":[{"output_type":"stream","text":["--2021-08-27 03:45:09--  https://storage.googleapis.com/ztm_tf_course/food_vision/10_food_classes_10_percent.zip\n","Resolving storage.googleapis.com (storage.googleapis.com)... 74.125.137.128, 142.250.101.128, 142.250.141.128, ...\n","Connecting to storage.googleapis.com (storage.googleapis.com)|74.125.137.128|:443... connected.\n","HTTP request sent, awaiting response... 200 OK\n","Length: 168546183 (161M) [application/zip]\n","Saving to: ‘10_food_classes_10_percent.zip.1’\n","\n","10_food_classes_10_ 100%[===================>] 160.74M   152MB/s    in 1.1s    \n","\n","2021-08-27 03:45:10 (152 MB/s) - ‘10_food_classes_10_percent.zip.1’ saved [168546183/168546183]\n","\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"X6uienXchIX6","executionInfo":{"status":"ok","timestamp":1630035917168,"user_tz":-420,"elapsed":26,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"5a2b8b0c-2b80-4d41-f417-c0bb143b12ad"},"source":["import os\n","\n","for dir_path, dir_names, file_names in os.walk(\"10_food_classes_10_percent\"):\n","  print(f\"There are {len(dir_names)} directories and {len(file_names)} in {dir_path}\")"],"execution_count":null,"outputs":[{"output_type":"stream","text":["There are 2 directories and 0 in 10_food_classes_10_percent\n","There are 10 directories and 0 in 10_food_classes_10_percent/train\n","There are 0 directories and 75 in 10_food_classes_10_percent/train/sushi\n","There are 0 directories and 75 in 10_food_classes_10_percent/train/fried_rice\n","There are 0 directories and 75 in 10_food_classes_10_percent/train/pizza\n","There are 0 directories and 75 in 10_food_classes_10_percent/train/ramen\n","There are 0 directories and 75 in 10_food_classes_10_percent/train/steak\n","There are 0 directories and 75 in 10_food_classes_10_percent/train/grilled_salmon\n","There are 0 directories and 75 in 10_food_classes_10_percent/train/chicken_curry\n","There are 0 directories and 75 in 10_food_classes_10_percent/train/chicken_wings\n","There are 0 directories and 75 in 10_food_classes_10_percent/train/ice_cream\n","There are 0 directories and 75 in 10_food_classes_10_percent/train/hamburger\n","There are 10 directories and 0 in 10_food_classes_10_percent/test\n","There are 0 directories and 250 in 10_food_classes_10_percent/test/sushi\n","There are 0 directories and 250 in 10_food_classes_10_percent/test/fried_rice\n","There are 0 directories and 250 in 10_food_classes_10_percent/test/pizza\n","There are 0 directories and 250 in 10_food_classes_10_percent/test/ramen\n","There are 0 directories and 250 in 10_food_classes_10_percent/test/steak\n","There are 0 directories and 250 in 10_food_classes_10_percent/test/grilled_salmon\n","There are 0 directories and 250 in 10_food_classes_10_percent/test/chicken_curry\n","There are 0 directories and 250 in 10_food_classes_10_percent/test/chicken_wings\n","There are 0 directories and 250 in 10_food_classes_10_percent/test/ice_cream\n","There are 0 directories and 250 in 10_food_classes_10_percent/test/hamburger\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"lvuQleIEhmUK"},"source":["# Set up train and test dir\n","train_dir = \"10_food_classes_10_percent/train\"\n","test_dir = \"10_food_classes_10_percent/test\""],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"Jzhk5HKth4az","executionInfo":{"status":"ok","timestamp":1630035917169,"user_tz":-420,"elapsed":23,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"9609abb2-a7f2-4bc1-ac93-ac70a62f7d17"},"source":["# Let's get the subdirectories\n","import pathlib\n","import numpy as np\n","\n","data_dir = pathlib.Path(train_dir)\n","class_names = sorted([item.name for item in data_dir.glob(\"*\")])\n","print(class_names)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["['chicken_curry', 'chicken_wings', 'fried_rice', 'grilled_salmon', 'hamburger', 'ice_cream', 'pizza', 'ramen', 'steak', 'sushi']\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"uJ9vxLOQiAH2"},"source":["# Let's visualize our images\n","import matplotlib.pyplot as plt\n","import matplotlib.image as mpimg\n","import random\n","import os\n","\n","def view_random_image(target_dir, target_class):\n","  # Set the target directory\n","  target_folder = os.path.join(target_dir, target_class)\n","\n","  # Get a random image path\n","  random_image = random.sample(os.listdir(target_folder),1)\n","  print(random_image)\n","  # Read in the image \n","  img = mpimg.imread(target_folder+\"/\"+random_image[0])\n","  plt.imshow(img)\n","  plt.title(target_class)\n","  plt.axis(\"off\");\n","\n","  print(f\"Image shape {img.shape}\")\n","  return img"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":298},"id":"F9GGSeeYXYaZ","executionInfo":{"status":"ok","timestamp":1630035917170,"user_tz":-420,"elapsed":19,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"24b28aed-dc7c-4351-9a0d-a9fb7a5e1452"},"source":["random_class = random.choice(class_names)\n","random_img = view_random_image(train_dir, random_class)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["['413789.jpg']\n","Image shape (512, 384, 3)\n"],"name":"stdout"},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"pRDw8HWxY-cu"},"source":["#### Use `image_dataset_from_directory`"]},{"cell_type":"code","metadata":{"id":"giKPyhqGZ_Ci"},"source":["IMG_SIZE = (224,224)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"rj7PefjibOst","executionInfo":{"status":"ok","timestamp":1630035917170,"user_tz":-420,"elapsed":14,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"9f8ee890-77b9-45b1-e22b-218ed981ff75"},"source":["IMG_SIZE + (3,)"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(224, 224, 3)"]},"metadata":{},"execution_count":217}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"ZZBOjA5-ZFko","executionInfo":{"status":"ok","timestamp":1630035917682,"user_tz":-420,"elapsed":523,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"61c89194-9ec7-4c6c-d909-7c6223dac47b"},"source":["import tensorflow as tf\n","\n","train_ds = tf.keras.preprocessing.image_dataset_from_directory(train_dir,\n","                                                               label_mode=\"categorical\",\n","                                                               image_size=IMG_SIZE)\n","test_ds = tf.keras.preprocessing.image_dataset_from_directory(test_dir,\n","                                                              label_mode=\"categorical\",\n","                                                              image_size=IMG_SIZE)\n","\n","train_ds, test_ds"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Found 750 files belonging to 10 classes.\n","Found 2500 files belonging to 10 classes.\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["(<BatchDataset shapes: ((None, 224, 224, 3), (None, 10)), types: (tf.float32, tf.float32)>,\n"," <BatchDataset shapes: ((None, 224, 224, 3), (None, 10)), types: (tf.float32, tf.float32)>)"]},"metadata":{},"execution_count":218}]},{"cell_type":"code","metadata":{"id":"WAcQ0Ty2dd_v"},"source":["# optimize performance of dataset\n","train_ds = train_ds.prefetch(tf.data.AUTOTUNE)\n","test_ds = test_ds.prefetch(tf.data.AUTOTUNE) "],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"DDHA4xN8Zwc6"},"source":["##### Create and fit Conv2D model"]},{"cell_type":"code","metadata":{"id":"I5IMURc1Z1wB","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1630035955332,"user_tz":-420,"elapsed":37656,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"c6796b9f-4945-4119-ca8c-6a2bc17fd07a"},"source":["from tensorflow.keras import layers\n","\n","# Create model\n","inputs = layers.Input(shape=(IMG_SIZE+(3,)), name=\"input_layer\")\n","\n","x = layers.Rescaling(1./255)(inputs)\n","\n","x = layers.Conv2D(filters=10,\n","                  kernel_size=3,\n","                  activation=\"relu\")(x)\n","x = layers.Conv2D(filters=10,\n","                  kernel_size=3,\n","                  activation=\"relu\")(x)\n","x = layers.MaxPool2D()(x)\n","\n","x = layers.Flatten()(x)\n","\n","outputs = layers.Dense(len(class_names), activation=\"softmax\")(x)\n","\n","model_4 = tf.keras.Model(inputs, outputs, name=\"model_4_Conv2D\")\n","\n","# Compile the model\n","model_4.compile(\n","    loss=tf.keras.losses.CategoricalCrossentropy(),\n","    optimizer=tf.keras.optimizers.Adam(),\n","    metrics=[\"accuracy\"]\n",")\n","\n","# Fit the model\n","model_4_history = model_4.fit(train_ds,\n","                              validation_data=test_ds,\n","                              epochs=5,\n","                              callbacks=[create_checkpoint_callback(model_4.name)])"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Epoch 1/5\n","24/24 [==============================] - 6s 240ms/step - loss: 7.6281 - accuracy: 0.0987 - val_loss: 2.3389 - val_accuracy: 0.1176\n","Epoch 2/5\n","24/24 [==============================] - 6s 235ms/step - loss: 2.2859 - accuracy: 0.1653 - val_loss: 2.2848 - val_accuracy: 0.1652\n","Epoch 3/5\n","24/24 [==============================] - 6s 239ms/step - loss: 2.1992 - accuracy: 0.2280 - val_loss: 2.2323 - val_accuracy: 0.1600\n","Epoch 4/5\n","24/24 [==============================] - 6s 233ms/step - loss: 1.9227 - accuracy: 0.3680 - val_loss: 2.3264 - val_accuracy: 0.1684\n","Epoch 5/5\n","24/24 [==============================] - 6s 233ms/step - loss: 1.5135 - accuracy: 0.5413 - val_loss: 2.4075 - val_accuracy: 0.1904\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"2z0hmEr0bTKI","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1630035959902,"user_tz":-420,"elapsed":4585,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"44c72e8a-a2fa-415d-b7b0-0ebdc7632d11"},"source":["# evaluate the model\n","model_4.evaluate(test_ds)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["79/79 [==============================] - 5s 53ms/step - loss: 2.4075 - accuracy: 0.1904\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["[2.40750789642334, 0.19040000438690186]"]},"metadata":{},"execution_count":221}]},{"cell_type":"code","metadata":{"id":"k5_mmKj5gZHv","colab":{"base_uri":"https://localhost:8080/","height":286},"executionInfo":{"status":"ok","timestamp":1630035960412,"user_tz":-420,"elapsed":526,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"7f12db1a-9aed-4ee9-ee9f-ba8925b06b70"},"source":["import pandas as pd\n","\n","pd.DataFrame(model_4_history.history).plot()"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.axes._subplots.AxesSubplot at 0x7f781ba52090>"]},"metadata":{},"execution_count":222},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"hsFVgCmjX60B"},"source":["##### Use `ImageDataGenerator`"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"yLMKFyWXY9vu","executionInfo":{"status":"ok","timestamp":1630035960413,"user_tz":-420,"elapsed":9,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"69c826b1-4935-40b7-f2cf-abc84ba48886"},"source":["import tensorflow as tf\n","from tensorflow.keras.preprocessing.image import ImageDataGenerator\n","\n","# set the seed \n","tf.random.set_seed(42)\n","\n","# Prepare dataset\n","train_datagen = ImageDataGenerator(rescale=1./255,\n","                                rotation_range=0.2,\n","                                shear_range=0.2,\n","                                zoom_range=0.2,\n","                                width_shift_range=0.2,\n","                                height_shift_range=0.2,\n","                                horizontal_flip=True)\n","test_datagen = ImageDataGenerator(rescale=1./255)\n","\n","train_ds = train_datagen.flow_from_directory(train_dir,\n","                                             target_size=IMG_SIZE,\n","                                             seed=42,\n","                                             shuffle=False)\n","\n","test_ds = test_datagen.flow_from_directory(test_dir,\n","                                             target_size=IMG_SIZE,\n","                                             seed=42,\n","                                             shuffle=False)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Found 750 images belonging to 10 classes.\n","Found 2500 images belonging to 10 classes.\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"7enMevfVY9GW"},"source":["# optimize performance dataset\n","# train_ds = tf.data.Dataset.from_generator(train_ds)\n","# test_ds = test_ds.batch(32).prefetch(tf.data.AUTOTUNE)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"UjlEbKCj5bsS","executionInfo":{"status":"ok","timestamp":1630035960414,"user_tz":-420,"elapsed":8,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"f8fd5913-92d6-4981-dd4c-e7bf3bea4961"},"source":["class_names"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["['chicken_curry',\n"," 'chicken_wings',\n"," 'fried_rice',\n"," 'grilled_salmon',\n"," 'hamburger',\n"," 'ice_cream',\n"," 'pizza',\n"," 'ramen',\n"," 'steak',\n"," 'sushi']"]},"metadata":{},"execution_count":225}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"CMbT9jJHimDs","executionInfo":{"status":"ok","timestamp":1630036061068,"user_tz":-420,"elapsed":100658,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"5fac8c7b-8158-4477-e385-c2c0ba2405f7"},"source":["from tensorflow.keras import layers\n","\n","# Create model\n","inputs = layers.Input(shape=(IMG_SIZE+(3,)), name=\"input_layer\")\n","\n","x = layers.Conv2D(filters=10,\n","                  kernel_size=3,\n","                  activation=\"relu\")(inputs)\n","x = layers.Conv2D(filters=10,\n","                  kernel_size=3,\n","                  activation=\"relu\")(x)\n","x = layers.MaxPool2D()(x)\n","\n","x = layers.Conv2D(filters=10,\n","                  kernel_size=3,\n","                  activation=\"relu\")(x)\n","x = layers.Conv2D(filters=10,\n","                  kernel_size=3,\n","                  activation=\"relu\")(x)\n","x = layers.MaxPool2D()(x)\n","\n","x = layers.Flatten()(x)\n","\n","outputs = layers.Dense(len(class_names), activation=\"softmax\")(x)\n","\n","model_5 = tf.keras.Model(inputs, outputs, name=\"model_5_Conv2D_data_aug\")\n","\n","# Compile the model\n","model_5.compile(\n","    loss=tf.keras.losses.CategoricalCrossentropy(),\n","    optimizer=tf.keras.optimizers.Adam(),\n","    metrics=[\"accuracy\"]\n",")\n","\n","# Fit the model\n","model_5_history = model_5.fit(train_ds,\n","                              validation_data=test_ds,\n","                              epochs=5,\n","                              callbacks=[create_checkpoint_callback(model_5.name)])"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Epoch 1/5\n","24/24 [==============================] - 21s 857ms/step - loss: 2.3800 - accuracy: 0.0320 - val_loss: 2.3027 - val_accuracy: 0.0996\n","Epoch 2/5\n","24/24 [==============================] - 20s 843ms/step - loss: 2.3029 - accuracy: 0.1000 - val_loss: 2.3026 - val_accuracy: 0.1000\n","Epoch 3/5\n","24/24 [==============================] - 20s 843ms/step - loss: 2.3033 - accuracy: 0.1000 - val_loss: 2.3026 - val_accuracy: 0.0996\n","Epoch 4/5\n","24/24 [==============================] - 20s 846ms/step - loss: 2.3034 - accuracy: 0.0960 - val_loss: 2.3026 - val_accuracy: 0.0992\n","Epoch 5/5\n","24/24 [==============================] - 20s 841ms/step - loss: 2.3034 - accuracy: 0.1000 - val_loss: 2.3026 - val_accuracy: 0.1000\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"3YTE2oFQYoCF","colab":{"base_uri":"https://localhost:8080/"},"outputId":"c9efcec7-caa8-4532-c7a7-a6d54cf7ae49"},"source":["# evaluate the model\n","# model_5.load_weights(f\"{SAVE_DIR}/{model_5.name}\")\n","model_5.evaluate(test_ds)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["79/79 [==============================] - 10s 132ms/step - loss: 2.3026 - accuracy: 0.1000\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"qNRJUqEA1yBD"},"source":["!wget https://raw.githubusercontent.com/mrdbourke/tensorflow-deep-learning/main/images/03-hamburger.jpeg\n","!wget https://raw.githubusercontent.com/mrdbourke/tensorflow-deep-learning/main/images/03-pizza-dad.jpeg\n","!wget https://raw.githubusercontent.com/mrdbourke/tensorflow-deep-learning/main/images/03-sushi.jpeg\n","!wget https://raw.githubusercontent.com/mrdbourke/tensorflow-deep-learning/main/images/03-steak.jpeg"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"q7gLVQjz12Fo"},"source":["files_test = [\"03-hamburger.jpeg\", \"03-pizza-dad.jpeg\", \"03-steak.jpeg\", \"03-sushi.jpeg\"]"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"YV_g_M4W13H3"},"source":["# Create a function to import and preprocess that image to be able to be used with our model\n","def load_and_prep_image(file_name, image_shape=224):\n","  \"\"\"\n","  Reads an image from file name, turns it into a tensor and reshapes it into (img_shape, image_shape, color_chanels)\n","  \"\"\"\n","  # Read an image \n","  img = tf.io.read_file(file_name)\n","  # Decode the read file into tensor\n","  img = tf.image.decode_image(img)\n","  # Resize the image\n","  img = tf.image.resize(img, size=(image_shape, image_shape))\n","  # Rescale the image (normalize)\n","  img = img/255\n","\n","  return img\n","def predict(model, class_names, file_name):\n","  img = load_and_prep_image(file_name)\n","  predict = model.predict(tf.expand_dims(img, axis=0))\n","  if len(predict[0])>1:\n","    predicted_class = class_names[tf.argmax(predict[0])]\n","  else:\n","    predicted_class = class_names[int(tf.round(predict[0]))]\n","  # print(predict, int(tf.round(predict)))\n","  plt.figure(figsize=(8,8))\n","  plt.imshow(img)\n","  plt.title(f\"Predicted: {predicted_class}\")\n","  plt.axis(False)\n","  plt.show()"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"xYsPv0k12HM_"},"source":["# predict(model_7, class_names, files_test[2])"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"oAGsVPqeZ0w9"},"source":["#### Transfer learning\n","\n","using EfficientNetB0-V2"]},{"cell_type":"code","metadata":{"id":"5W4z6YH-7GL2"},"source":["train_dir, test_dir"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"RmnttAJA2YvL"},"source":["from tensorflow.keras.preprocessing import image_dataset_from_directory\n","\n","print(train_dir, test_dir)\n","train_ds = image_dataset_from_directory(train_dir,\n","                                        label_mode=\"categorical\",\n","                                        image_size=IMG_SIZE)\n","test_ds = image_dataset_from_directory(test_dir,\n","                                       label_mode=\"categorical\",\n","                                       image_size=IMG_SIZE)\n","\n","train_ds = train_ds.prefetch(tf.data.AUTOTUNE)\n","test_ds = test_ds.prefetch(tf.data.AUTOTUNE)\n","\n","# train_ds, test_ds"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"vNYnnNSgJZBr"},"source":["# Set up data inputs\n","from tensorflow.keras.preprocessing.image import ImageDataGenerator\n","\n","IMG_SHAPE = (224, 224)\n","BATCH_SIZE = 32\n","EPOCHS = 5\n","\n","# Specify the train and the test directories\n","train_dir = \"10_food_classes_10_percent/train/\"\n","test_dir = \"10_food_classes_10_percent/test/\"\n","\n","train_datagen = ImageDataGenerator(rescale=1./255)\n","test_datagen = ImageDataGenerator(rescale=1./255)\n","\n","print(\"Training image:\")\n","train_data_10_percent = train_datagen.flow_from_directory(train_dir,\n","                                              target_size=IMG_SHAPE,\n","                                              batch_size=BATCH_SIZE,\n","                                              class_mode=\"categorical\")\n","print(\"Testing image:\")\n","test_data_10_percent = test_datagen.flow_from_directory(test_dir,\n","                                            target_size=IMG_SHAPE,\n","                                            batch_size=BATCH_SIZE,\n","                                            class_mode=\"categorical\")"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"zSgw_GkuvlxE"},"source":["import tensorflow as tf\n","import tensorflow_hub as hub\n","from tensorflow.keras import layers\n","\n","efficientnetb0_v2_url = \"https://tfhub.dev/google/imagenet/efficientnet_v2_imagenet1k_b0/feature_vector/2\"\n","# efficientnetb0_v1_url = \"https://tfhub.dev/tensorflow/efficientnet/b0/feature-vector/1\"\n","feature_extracter_layer = hub.KerasLayer(efficientnetb0_v2_url,\n","                                         trainable=False,\n","                                         \n","                                         name=\"feature_extractor_layer\")"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"72IIOCH6KZrX"},"source":["len(class_names)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"mXqMBX2mJ2Yc"},"source":["model_6_v2 = tf.keras.Sequential([\n","      feature_extracter_layer,\n","      layers.Dense(len(class_names), activation=\"softmax\", name=\"output_layer\")\n","])\n","\n","model_6_v2.compile(\n","    loss=\"categorical_crossentropy\",\n","    optimizer=tf.keras.optimizers.Adam(),\n","    metrics=[\"accuracy\"]\n",")\n","\n","model_6_v2_history = model_6_v2.fit(train_data_10_percent,\n","                              validation_data=test_data_10_percent,\n","                              epochs=5,\n","                              callbacks=[create_checkpoint_callback(model_6_v2.name)])"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"hSvDnveyOpI4"},"source":["model_6_v2.evaluate(test_data_10_percent)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"S2offuw_NDHY"},"source":["model_6.evaluate(test_data_10_percent)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"2nk7c6CH1MOy"},"source":["# Create model\n","inputs = layers.Input(shape=(IMG_SIZE+(3,)), name=\"input_layer\")\n","\n","x = layers.Rescaling(1./255)(inputs)\n","\n","x = feature_extracter_layer(x)\n","\n","outputs = layers.Dense(len(class_names), activation=\"softmax\")(x)\n","\n","model_7_V2 = tf.keras.Model(inputs, outputs, name=\"model_7_V2_transfer_learning_feature_extraction\")\n","\n","# Compile the model\n","model_7_V2.compile(\n","    loss=tf.keras.losses.CategoricalCrossentropy(),\n","    optimizer=tf.keras.optimizers.Adam(),\n","    metrics=[\"accuracy\"]\n",")\n","\n","# Fit the model\n","model_7_V2_history = model_7_V2.fit(train_ds,\n","                              validation_data=test_ds,\n","                              epochs=5,\n","                              callbacks=[create_checkpoint_callback(model_7_V2.name)])"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"3XT9cABNPijD"},"source":["model_7_V2.evaluate(test_ds)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"5rQkO0ryDX_e"},"source":["model_7.evaluate(test_ds)"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"ykFPqTv7NJEJ"},"source":["## Natural Languare Processing"]},{"cell_type":"markdown","metadata":{"id":"iIOtQjvQQGdF"},"source":["#### Binary Classification\n","\n","In this problem we'll use the `Twitter disaster datasets`"]},{"cell_type":"code","metadata":{"id":"_qUV1vimQOzM"},"source":["!wget https://github.com/BaoLocPham/Tensorflow_Deep_Learning/raw/main/Extras/nlp-getting-started.zip"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"yRFHHr-zRTSh"},"source":["import zipfile\n","\n","zip_ref = zipfile.ZipFile(\"nlp-getting-started.zip\")\n","zip_ref.extractall()\n","zip_ref.close()"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"rRk-casxRmQ2"},"source":["##### Explore the dataset"]},{"cell_type":"code","metadata":{"id":"QhDKNoHcR7wM"},"source":["import pandas as pd\n","\n","raw_df = pd.read_csv(\"train.csv\")\n","\n","raw_df.head()"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"tyzAq1iCTlKN"},"source":["raw_df[\"keyword\"].value_counts()"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"w7VcW3D9TqhA"},"source":["raw_df[\"location\"].value_counts()"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"g8NbCdiUSRCp"},"source":["data_df = raw_df.drop([\"id\",\"keyword\", \"location\"], axis=1)\n","data_df.head()"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"3hkAIZHtSp3n"},"source":["data_df[\"target\"].value_counts().plot(kind=\"barh\")"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"JviB0KVeSyHZ"},"source":["data_df_shuffled = data_df.sample(frac=1)\n","data_df_shuffled.head()"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"SWdz_POHULrU"},"source":["##### Split the dataset"]},{"cell_type":"code","metadata":{"id":"r3nJJgHzUaMS"},"source":["from sklearn.model_selection import train_test_split\n","\n","train_sentences, val_sentences, train_labels, val_labels = train_test_split(data_df_shuffled[\"text\"].to_numpy(),\n","                                                                            data_df_shuffled[\"target\"].to_numpy(),\n","                                                                            test_size=0.1,\n","                                                                            random_state=42)\n","\n","len(train_sentences), len(val_sentences), len(train_labels), len(val_labels)"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"8BG9EsscUdbe"},"source":["##### Build the text vectorizer and text embedding"]},{"cell_type":"code","metadata":{"id":"-4dz41tjVNpA"},"source":["train_sentences.shape"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"Ev1AozWqVvbC"},"source":["len_train_sentences = [len(sent.split(\" \")) for sent in train_sentences]\n","\n","len_df = pd.DataFrame({\"len\":len_train_sentences})\n","len_df.value_counts().plot(kind=\"hist\")"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"Br8ilOIRXIiU"},"source":["plt.hist(len_train_sentences, bins=20)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"wAxmvNmJVtNk"},"source":["# what is the len cover 95% of the training sentences\n","import numpy as np\n","\n","np.percentile(len_train_sentences, 95)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"ynDmsjdYX-JS"},"source":["# What is the number of words in the training sentences\n","\n","vocab = \" \".join(train_sentences)\n","vocab = set(vocab.split(\" \"))\n","len(vocab)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"ReTdBniKWgQM"},"source":["from tensorflow.keras.layers import TextVectorization\n","\n","MAX_LENGTH = 25\n","MAX_TOKEN = 10000\n","\n","text_vectorizer = TextVectorization(max_tokens=MAX_TOKEN,\n","                                         output_mode=\"int\",\n","                                         output_sequence_length=MAX_LENGTH)\n","\n","text_vectorizer.adapt(train_sentences)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"rQhMxVpZXgLS"},"source":["# Create a sample setences and tokenize it\n","sample_sentence = \"There's a flood in my street!\"\n","text_vectorizer([sample_sentence])"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"WmXabdbHYmpr"},"source":["# Choose a random sentence from the training data and tokenize it\n","random_sentence = random.choice(train_sentences)\n","print(f\"Origin text:\\n {random_sentence}\\\n","          \\n\\nVectorize version:\")\n","text_vectorizer([random_sentence])"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"VFLlgYzkZTSI"},"source":["# get the unique words in the vocabulary\n","words_in_vocab = text_vectorizer.get_vocabulary()\n","top_5_words = words_in_vocab[:5] # most common words\n","bottom_5_words = words_in_vocab[-5:] # least common words\n","print(f\"Number of words in vocab: {len(words_in_vocab)}\")\n","print(f\"5 most common words: {top_5_words}\")\n","print(f\"5 least common words: {bottom_5_words}\")"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"MqXomH1SYp1W"},"source":["from tensorflow.keras.layers import Embedding\n","\"\"\"\n","input_dim = the size of the vocabulary\n","output_dim = the size of the output embedding vector\n","input_length = length of the sentence being passed to the embedding layer\n","\"\"\"\n","text_embedding = Embedding(input_dim=MAX_TOKEN,\n","                           output_dim=128,\n","                           input_length=MAX_LENGTH,\n","                           embeddings_initializer=\"uniform\")"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"Aig_Qd8QZOTx"},"source":["# Get a random sentence in the training set \n","random_sentence = random.choice(train_sentences)\n","print(f\"Original text: \\n{random_sentence}\\\n","\\n\\nEmbedded version:\")\n","# Embed the ramdon sentence (turn it into dense vector of fixed size)\n","sample_vectorized = text_vectorizer([random_sentence])\n","sample_embedded = text_embedding(sample_vectorized)\n","sample_embedded"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"Agkipj5CZ4WZ"},"source":["##### Build and fit a model"]},{"cell_type":"markdown","metadata":{"id":"OMNQnxbDZ_Hf"},"source":["###### LSTM"]},{"cell_type":"code","metadata":{"id":"Od_twFdOaCSy"},"source":["from tensorflow.keras import layers\n","\n","# Create model\n","inputs = layers.Input(shape=(1,), dtype=tf.string, name=\"input_layer\")\n","\n","x = text_vectorizer(inputs)\n","x = text_embedding(x)\n","\n","x = layers.LSTM(64, return_sequences=True)(x)\n","x = layers.LSTM(64)(x)\n","\n","outputs = layers.Dense(1, activation=\"sigmoid\")(x)\n","\n","model_8 = tf.keras.Model(inputs, outputs, name=\"model_8_LSTM\")\n","\n","# Compile the model\n","model_8.compile(\n","    loss=tf.keras.losses.BinaryCrossentropy(),\n","    optimizer=tf.keras.optimizers.Adam(),\n","    metrics=[\"accuracy\"]\n",")\n","\n","# Fit the model\n","model_8_history = model_8.fit(train_sentences,\n","                              train_labels,\n","                              validation_data=(val_sentences, val_labels),\n","                              epochs=5,\n","                              callbacks=[create_checkpoint_callback(model_8.name)])"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"Sa355FZ8bdm1"},"source":["model_8.evaluate(val_sentences, val_labels)"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"4aZowjdXcEpT"},"source":["#### Multiclass Categorization\n","\n","In this problem, we'll use the data from [News category dataset](https://www.kaggle.com/rmisra/news-category-dataset)"]},{"cell_type":"code","metadata":{"id":"Q0d8K12ZfPS2","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1630036209572,"user_tz":-420,"elapsed":1254,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"6a6d131a-0121-492c-eca3-c8288b1dd382"},"source":["!wget https://github.com/BaoLocPham/Tensorflow_Deep_Learning/raw/main/Extras/News_Category_Dataset_v2.json.zip"],"execution_count":null,"outputs":[{"output_type":"stream","text":["--2021-08-27 03:50:07--  https://github.com/BaoLocPham/Tensorflow_Deep_Learning/raw/main/Extras/News_Category_Dataset_v2.json.zip\n","Resolving github.com (github.com)... 140.82.112.4\n","Connecting to github.com (github.com)|140.82.112.4|:443... connected.\n","HTTP request sent, awaiting response... 302 Found\n","Location: https://raw.githubusercontent.com/BaoLocPham/Tensorflow_Deep_Learning/main/Extras/News_Category_Dataset_v2.json.zip [following]\n","--2021-08-27 03:50:07--  https://raw.githubusercontent.com/BaoLocPham/Tensorflow_Deep_Learning/main/Extras/News_Category_Dataset_v2.json.zip\n","Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.111.133, 185.199.109.133, 185.199.110.133, ...\n","Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.111.133|:443... connected.\n","HTTP request sent, awaiting response... 200 OK\n","Length: 26677036 (25M) [application/zip]\n","Saving to: ‘News_Category_Dataset_v2.json.zip’\n","\n","News_Category_Datas 100%[===================>]  25.44M   137MB/s    in 0.2s    \n","\n","2021-08-27 03:50:08 (137 MB/s) - ‘News_Category_Dataset_v2.json.zip’ saved [26677036/26677036]\n","\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"fwHNfGoxhCcu"},"source":["import zipfile\n","\n","zip_ref = zipfile.ZipFile(\"News_Category_Dataset_v2.json.zip\")\n","zip_ref.extractall()\n","zip_ref.close()"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"ABIO9cOkmoXm","colab":{"base_uri":"https://localhost:8080/","height":287},"executionInfo":{"status":"ok","timestamp":1630038677064,"user_tz":-420,"elapsed":737,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"c08e9eee-873e-41f3-9438-8edd696a42f0"},"source":["import pandas as pd\n","\n","raw_df = pd.read_json(\"News_Category_Dataset_v2.json\", lines=True)\n","# raw_df.to_csv(\"News_Category_Dataset.csv\")\n","raw_df.head()"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>category</th>\n","      <th>headline</th>\n","      <th>authors</th>\n","      <th>link</th>\n","      <th>short_description</th>\n","      <th>date</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>CRIME</td>\n","      <td>There Were 2 Mass Shootings In Texas Last Week...</td>\n","      <td>Melissa Jeltsen</td>\n","      <td>https://www.huffingtonpost.com/entry/texas-ama...</td>\n","      <td>She left her husband. He killed their children...</td>\n","      <td>2018-05-26</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>ENTERTAINMENT</td>\n","      <td>Will Smith Joins Diplo And Nicky Jam For The 2...</td>\n","      <td>Andy McDonald</td>\n","      <td>https://www.huffingtonpost.com/entry/will-smit...</td>\n","      <td>Of course it has a song.</td>\n","      <td>2018-05-26</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>ENTERTAINMENT</td>\n","      <td>Hugh Grant Marries For The First Time At Age 57</td>\n","      <td>Ron Dicker</td>\n","      <td>https://www.huffingtonpost.com/entry/hugh-gran...</td>\n","      <td>The actor and his longtime girlfriend Anna Ebe...</td>\n","      <td>2018-05-26</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>ENTERTAINMENT</td>\n","      <td>Jim Carrey Blasts 'Castrato' Adam Schiff And D...</td>\n","      <td>Ron Dicker</td>\n","      <td>https://www.huffingtonpost.com/entry/jim-carre...</td>\n","      <td>The actor gives Dems an ass-kicking for not fi...</td>\n","      <td>2018-05-26</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>ENTERTAINMENT</td>\n","      <td>Julianna Margulies Uses Donald Trump Poop Bags...</td>\n","      <td>Ron Dicker</td>\n","      <td>https://www.huffingtonpost.com/entry/julianna-...</td>\n","      <td>The \"Dietland\" actress said using the bags is ...</td>\n","      <td>2018-05-26</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["        category  ...       date\n","0          CRIME  ... 2018-05-26\n","1  ENTERTAINMENT  ... 2018-05-26\n","2  ENTERTAINMENT  ... 2018-05-26\n","3  ENTERTAINMENT  ... 2018-05-26\n","4  ENTERTAINMENT  ... 2018-05-26\n","\n","[5 rows x 6 columns]"]},"metadata":{},"execution_count":45}]},{"cell_type":"code","metadata":{"id":"0rmjSFqK3vxx"},"source":[""],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"Nq2wT-vLyW2w"},"source":["raw_df.drop([\"authors\", \"link\", \"date\"], axis=1, inplace=True)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"GTxVcI6Dm7c3","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1630038677613,"user_tz":-420,"elapsed":18,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"b5499990-8c48-472d-a766-4f321870ad3d"},"source":["class_names = raw_df[\"category\"].unique()\n","class_names"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array(['CRIME', 'ENTERTAINMENT', 'WORLD NEWS', 'IMPACT', 'POLITICS',\n","       'WEIRD NEWS', 'BLACK VOICES', 'WOMEN', 'COMEDY', 'QUEER VOICES',\n","       'SPORTS', 'BUSINESS', 'TRAVEL', 'MEDIA', 'TECH', 'RELIGION',\n","       'SCIENCE', 'LATINO VOICES', 'EDUCATION', 'COLLEGE', 'PARENTS',\n","       'ARTS & CULTURE', 'STYLE', 'GREEN', 'TASTE', 'HEALTHY LIVING',\n","       'THE WORLDPOST', 'GOOD NEWS', 'WORLDPOST', 'FIFTY', 'ARTS',\n","       'WELLNESS', 'PARENTING', 'HOME & LIVING', 'STYLE & BEAUTY',\n","       'DIVORCE', 'WEDDINGS', 'FOOD & DRINK', 'MONEY', 'ENVIRONMENT',\n","       'CULTURE & ARTS'], dtype=object)"]},"metadata":{},"execution_count":47}]},{"cell_type":"code","metadata":{"id":"OmygdoKCu6-r"},"source":["raw_df[\"category\"] = raw_df[\"category\"].map(lambda x: \"WORLDPOST\" if x==\"THE WORLDPOST\" else x)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"2dKxm6zMuEOT"},"source":["raw_df[\"text\"] = raw_df[\"headline\"] + \" \" + raw_df[\"short_description\"]"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":52},"id":"M6VOnspTu3l6","executionInfo":{"status":"ok","timestamp":1630038677614,"user_tz":-420,"elapsed":15,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"e73eb879-d913-43bb-966f-f9f2f297f569"},"source":["raw_df[\"text\"][0]"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"application/vnd.google.colaboratory.intrinsic+json":{"type":"string"},"text/plain":["'There Were 2 Mass Shootings In Texas Last Week, But Only 1 On TV She left her husband. He killed their children. Just another day in America.'"]},"metadata":{},"execution_count":50}]},{"cell_type":"code","metadata":{"id":"N1NQpEXGv2Jz"},"source":["import regex as re\n","def clean_text(text):\n","  text = re.sub(r'[^\\w\\s]','',text)\n","  return text"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"nk4EdgPYxi02","executionInfo":{"status":"ok","timestamp":1630038677615,"user_tz":-420,"elapsed":14,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"842bed1a-81dc-4a56-9a04-b88b75c88b89"},"source":["import random \n","\n","random_int = random.randint(0, len(raw_df)-1)\n","random_text = raw_df[\"text\"][random_int]\n","print(f\"Text:{random_text}\\n Removed puntuation: {clean_text(random_text)}\")"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Text:Confessions of a Blue-Gened Blonde Unlike other people, I can't keep my depression in the closet. For one thing, I need the space for shoes. But more importantly, staying silent perpetuates the stigma that mental health isn't a topic for polite conversation. So might I propose we all stop being so polite?\n"," Removed puntuation: Confessions of a BlueGened Blonde Unlike other people I cant keep my depression in the closet For one thing I need the space for shoes But more importantly staying silent perpetuates the stigma that mental health isnt a topic for polite conversation So might I propose we all stop being so polite\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"6vFkG7T0x5n0","executionInfo":{"status":"ok","timestamp":1630039599274,"user_tz":-420,"elapsed":2195,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"cd8b4610-e2e4-4f79-cacc-fa312bdad470"},"source":["raw_df[\"text\"] = raw_df[\"text\"].map(clean_text)\n","raw_df[\"text\"] = raw_df[\"text\"].str.lower()\n","raw_df[\"text\"].head()"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["0    there were 2 mass shootings in texas last week...\n","1    will smith joins diplo and nicky jam for the 2...\n","2    hugh grant marries for the first time at age 5...\n","3    jim carrey blasts castrato adam schiff and dem...\n","4    julianna margulies uses donald trump poop bags...\n","Name: text, dtype: object"]},"metadata":{},"execution_count":86}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"TOml6LvoyJAc","executionInfo":{"status":"ok","timestamp":1630039603017,"user_tz":-420,"elapsed":310,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"3bc7831b-0426-4f4a-a4e6-9c6ec791b516"},"source":["# one hot encode labels\n","from sklearn.preprocessing import OneHotEncoder\n","\n","one_hot_encoder = OneHotEncoder(sparse=False)\n","label = one_hot_encoder.fit_transform(raw_df[\"category\"].to_numpy().reshape(-1,1))\n","\n","# raw_df[\"label\"] = label\n","class_names = np.array(one_hot_encoder.categories_, dtype=\"str\")[0]\n","class_names "],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array(['ARTS', 'ARTS & C', 'BLACK VO', 'BUSINESS', 'COLLEGE', 'COMEDY',\n","       'CRIME', 'CULTURE ', 'DIVORCE', 'EDUCATIO', 'ENTERTAI', 'ENVIRONM',\n","       'FIFTY', 'FOOD & D', 'GOOD NEW', 'GREEN', 'HEALTHY ', 'HOME & L',\n","       'IMPACT', 'LATINO V', 'MEDIA', 'MONEY', 'PARENTIN', 'PARENTS',\n","       'POLITICS', 'QUEER VO', 'RELIGION', 'SCIENCE', 'SPORTS', 'STYLE',\n","       'STYLE & ', 'TASTE', 'TECH', 'TRAVEL', 'WEDDINGS', 'WEIRD NE',\n","       'WELLNESS', 'WOMEN', 'WORLD NE', 'WORLDPOS'], dtype='<U8')"]},"metadata":{},"execution_count":87}]},{"cell_type":"markdown","metadata":{"id":"h87_cYEVyyo5"},"source":["##### Create data for training and validation"]},{"cell_type":"code","metadata":{"id":"yXXcKt_62aVU"},"source":["sentences = raw_df[\"text\"]\n","label = one_hot_encoder.fit_transform(raw_df[\"category\"].to_numpy().reshape(-1,1))"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"tQLSh0mf2x6w","executionInfo":{"status":"ok","timestamp":1630041019412,"user_tz":-420,"elapsed":404,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"9ec1c1d6-5258-4093-a9d1-9ada575fb291"},"source":["from sklearn.model_selection import train_test_split\n","\n","train_sentences, test_sentences, train_labels, test_labels = train_test_split(sentences.to_numpy(),\n","                                                                              label,\n","                                                                              test_size=0.2)\n","\n","len(train_sentences), len(test_sentences), len(train_labels), len(test_labels)"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(160682, 40171, 160682, 40171)"]},"metadata":{},"execution_count":118}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"gh4KZYXTAkla","executionInfo":{"status":"ok","timestamp":1630041020667,"user_tz":-420,"elapsed":2,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"ae69e655-16ea-4d02-8599-a29cb6579bda"},"source":["train_labels.shape"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(160682, 40)"]},"metadata":{},"execution_count":119}]},{"cell_type":"code","metadata":{"id":"hVyfKf6A_3pM"},"source":["train_features = tf.data.Dataset.from_tensor_slices(train_sentences)\n","train_labels = tf.data.Dataset.from_tensor_slices(train_labels)\n","\n","train_ds = tf.data.Dataset.zip((train_features, train_labels))\n","\n","test_features = tf.data.Dataset.from_tensor_slices(test_sentences)\n","test_labels = tf.data.Dataset.from_tensor_slices(test_labels)\n","\n","test_ds = tf.data.Dataset.zip((test_features, test_labels))"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"IEXXFkBbBChP"},"source":["train_ds = train_ds.batch(32).prefetch(tf.data.AUTOTUNE)\n","test_ds = test_ds.batch(32).prefetch(tf.data.AUTOTUNE)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"eyjJAKNm3TJX","executionInfo":{"status":"ok","timestamp":1630039606794,"user_tz":-420,"elapsed":314,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"14ab2f91-e0b2-48d2-fca0-ff725fad6af0"},"source":["import numpy as np\n","\n","train_sentences[0], class_names[np.argmax(train_labels[0])]"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["('katy perry drops new olympics single rise at the perfect moment this is a song thats been brewing inside me for years',\n"," 'ENTERTAI')"]},"metadata":{},"execution_count":90}]},{"cell_type":"markdown","metadata":{"id":"cAwdP-jP7U38"},"source":["##### Build the text vectorizer and text embedding"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":282},"id":"cDI4Aurq8NLL","executionInfo":{"status":"ok","timestamp":1630039763695,"user_tz":-420,"elapsed":992,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"913c42ab-6cd3-42c8-eb75-a056c5dc02d9"},"source":["len_train_sentences = [len(sent.split(\" \")) for sent in train_sentences]\n","\n","len_df = pd.DataFrame({\"len\":len_train_sentences})\n","len_df.value_counts().plot(kind=\"hist\")"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.axes._subplots.AxesSubplot at 0x7f0e818b4090>"]},"metadata":{},"execution_count":91},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"sV1xeNG99R57","executionInfo":{"status":"ok","timestamp":1630040064240,"user_tz":-420,"elapsed":314,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"bf4ad0c9-c4fc-4c4e-b500-d13ad8eb7ed9"},"source":["# what is the len cover 95% of the training sentences\n","import numpy as np\n","\n","np.percentile(len_train_sentences, 95)"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["56.0"]},"metadata":{},"execution_count":94}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"TPE3mj9U8PY8","executionInfo":{"status":"ok","timestamp":1630039919340,"user_tz":-420,"elapsed":1209,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"1d2fd60f-0bd5-4725-ef39-3607fddd1397"},"source":["vocab = \" \".join(train_sentences)\n","vocab = set(vocab.split(\" \"))\n","vocab_len = len(vocab)\n","vocab_len"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["102744"]},"metadata":{},"execution_count":93}]},{"cell_type":"code","metadata":{"id":"7ufB8qop82bZ"},"source":["from tensorflow.keras.layers import TextVectorization, Embedding\n","\n","MAX_TOKENS = 105000\n","MAX_LENGTH = 60\n","\n","text_vectorizer = TextVectorization(max_tokens=MAX_TOKENS,\n","                                    output_mode=\"int\",\n","                                    output_sequence_length=MAX_LENGTH)\n","text_vectorizer.adapt(train_sentences)\n","\n","text_embedding = Embedding(input_dim=MAX_TOKENS,\n","                           output_dim=256,\n","                           input_length=MAX_LENGTH,\n","                           embeddings_initializer=\"uniform\")"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"UniIDqKz-YK_","executionInfo":{"status":"ok","timestamp":1630042631880,"user_tz":-420,"elapsed":524,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"6305d7f8-548e-461a-8de2-001cf3d1145d"},"source":["# get the unique words in the vocabulary\n","words_in_vocab = text_vectorizer.get_vocabulary()\n","top_5_words = words_in_vocab[:5] # most common words\n","bottom_5_words = words_in_vocab[-5:] # least common words\n","print(f\"Number of words in vocab: {len(words_in_vocab)}\")\n","print(f\"5 most common words: {top_5_words}\")\n","print(f\"5 least common words: {bottom_5_words}\")"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Number of words in vocab: 102742\n","5 most common words: ['', '[UNK]', 'the', 'to', 'a']\n","5 least common words: ['0010', '001', '0002', '000064', '00']\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"qdUSRW66-iiH","executionInfo":{"status":"ok","timestamp":1630042631881,"user_tz":-420,"elapsed":8,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"28dd8bf2-5a13-4a50-cc20-1b29d7817e61"},"source":["# Get a random sentence in the training set \n","random_sentence = random.choice(train_sentences)\n","print(f\"Original text: \\n{random_sentence}\\\n","\\n\\nEmbedded version:\")\n","# Embed the ramdon sentence (turn it into dense vector of fixed size)\n","sample_vectorized = text_vectorizer([random_sentence])\n","sample_embedded = text_embedding(sample_vectorized)\n","sample_embedded"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Original text: \n","prince harry offers a perfect take on vets experiencing ptsd now this is how you talk about mental health and the military\n","\n","Embedded version:\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["<tf.Tensor: shape=(1, 60, 256), dtype=float32, numpy=\n","array([[[-0.02988741, -0.01978387, -0.00210118, ...,  0.03688402,\n","          0.04099676,  0.04347726],\n","        [-0.02565848, -0.01342086, -0.03138348, ..., -0.02096807,\n","         -0.01091192, -0.00520958],\n","        [ 0.03781096,  0.00968594,  0.04533634, ..., -0.02199518,\n","         -0.03617049, -0.00086548],\n","        ...,\n","        [ 0.02594527,  0.02468164,  0.00922592, ...,  0.01617893,\n","          0.04119625,  0.01296576],\n","        [ 0.02594527,  0.02468164,  0.00922592, ...,  0.01617893,\n","          0.04119625,  0.01296576],\n","        [ 0.02594527,  0.02468164,  0.00922592, ...,  0.01617893,\n","          0.04119625,  0.01296576]]], dtype=float32)>"]},"metadata":{},"execution_count":126}]},{"cell_type":"markdown","metadata":{"id":"jVFE-Nou5_wD"},"source":["##### Build and Fit the model"]},{"cell_type":"markdown","metadata":{"id":"qCaBq7Ol-r83"},"source":["###### LSTM"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"l4jxYJxO7IXZ","executionInfo":{"status":"ok","timestamp":1630043857978,"user_tz":-420,"elapsed":1160526,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"045d3292-afc7-45dd-e921-1a3d6827ea35"},"source":["from tensorflow.keras import layers\n","# Create the model\n","inputs = layers.Input(shape=(1,), dtype=tf.string, name=\"input_layer\")\n","\n","x = text_vectorizer(inputs)\n","x = text_embedding(x)\n","\n","x = layers.LSTM(64, return_sequences=True)(x)\n","x = layers.LSTM(64)(x)\n","\n","outputs = layers.Dense(len(class_names), activation=\"softmax\")(x)\n","\n","model_9 = tf.keras.Model(inputs, outputs, name=\"model_8_LSTM_multiclass\")\n","\n","# Compile the model\n","model_9.compile(\n","    loss=tf.keras.losses.CategoricalCrossentropy(),\n","    optimizer=tf.keras.optimizers.Adam(),\n","    metrics=[\"accuracy\"]\n",")\n","\n","# Fit the model\n","model_9_history = model_9.fit(train_ds,\n","                              validation_data=test_ds,\n","                              validation_steps=int(0.2 * len(test_ds)),\n","                              epochs=5)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Epoch 1/5\n","5022/5022 [==============================] - 228s 45ms/step - loss: 2.5274 - accuracy: 0.3407 - val_loss: 1.8650 - val_accuracy: 0.5134\n","Epoch 2/5\n","5022/5022 [==============================] - 224s 45ms/step - loss: 1.4981 - accuracy: 0.5962 - val_loss: 1.4178 - val_accuracy: 0.6130\n","Epoch 3/5\n","5022/5022 [==============================] - 223s 44ms/step - loss: 1.0140 - accuracy: 0.7152 - val_loss: 1.4825 - val_accuracy: 0.6172\n","Epoch 4/5\n","5022/5022 [==============================] - 223s 44ms/step - loss: 0.6855 - accuracy: 0.8052 - val_loss: 1.6933 - val_accuracy: 0.6011\n","Epoch 5/5\n","5022/5022 [==============================] - 223s 44ms/step - loss: 0.4590 - accuracy: 0.8694 - val_loss: 1.9582 - val_accuracy: 0.5690\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"0IWrUk7J8Glw","executionInfo":{"status":"ok","timestamp":1630042099081,"user_tz":-420,"elapsed":16589,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"f9cdea60-0616-4aea-975b-1edbd33f5bf6"},"source":["model_9.evaluate(test_ds)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["1256/1256 [==============================] - 17s 13ms/step - loss: 1.8221 - accuracy: 0.5874\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["[1.8220643997192383, 0.5873640179634094]"]},"metadata":{},"execution_count":123}]},{"cell_type":"markdown","metadata":{"id":"pg1CQBNB9RNJ"},"source":["#### Text Generation\n","\n","In this problem, we'll the Shakepeare dataset 🙂\n","\n","This part we reference to Tensorflow tutorial on generate text."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"idFoLY_9Ite7","executionInfo":{"status":"ok","timestamp":1630158673450,"user_tz":-420,"elapsed":4284,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"0a93ad5e-f537-43ab-dc4f-9f17b1c369c7"},"source":["import tensorflow as tf\n","path_to_file = tf.keras.utils.get_file('shakespeare.txt', 'https://storage.googleapis.com/download.tensorflow.org/data/shakespeare.txt')"],"execution_count":3,"outputs":[{"output_type":"stream","text":["Downloading data from https://storage.googleapis.com/download.tensorflow.org/data/shakespeare.txt\n","1122304/1115394 [==============================] - 0s 0us/step\n","1130496/1115394 [==============================] - 0s 0us/step\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"Fnmt_ccWJtO5"},"source":["##### Read the data"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"Pr_k48koKISl","executionInfo":{"status":"ok","timestamp":1630158673451,"user_tz":-420,"elapsed":32,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"01d7bddc-a647-4df4-aa43-82b5a787c737"},"source":["text = open(path_to_file, \"rb\").read().decode(encoding=\"utf8\")\n","# length of text is the number of characters in it\n","print(f'Length of text: {len(text)} characters')"],"execution_count":4,"outputs":[{"output_type":"stream","text":["Length of text: 1115394 characters\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"-xUb7btuKdns","executionInfo":{"status":"ok","timestamp":1630158673451,"user_tz":-420,"elapsed":25,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"80544db0-bd15-4a62-e2b6-34bfdaadbf42"},"source":["# Take a look at the first 250 characters in text\n","print(text[:250])"],"execution_count":5,"outputs":[{"output_type":"stream","text":["First Citizen:\n","Before we proceed any further, hear me speak.\n","\n","All:\n","Speak, speak.\n","\n","First Citizen:\n","You are all resolved rather to die than to famish?\n","\n","All:\n","Resolved. resolved.\n","\n","First Citizen:\n","First, you know Caius Marcius is chief enemy to the people.\n","\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"RAHXSu2uKflR","executionInfo":{"status":"ok","timestamp":1630158673451,"user_tz":-420,"elapsed":18,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"8c8f6ef4-4a2d-45f5-a0b8-a26a5ed983ff"},"source":["# The unique characters in the file\n","vocab = sorted(set(text))\n","print(f'{len(vocab)} unique characters')"],"execution_count":6,"outputs":[{"output_type":"stream","text":["65 unique characters\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"h9tGaEEwLlJ6","executionInfo":{"status":"ok","timestamp":1630158673880,"user_tz":-420,"elapsed":441,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"be2a9b97-6b31-4ebd-ae12-52116582efa6"},"source":["vocab"],"execution_count":7,"outputs":[{"output_type":"execute_result","data":{"text/plain":["['\\n',\n"," ' ',\n"," '!',\n"," '$',\n"," '&',\n"," \"'\",\n"," ',',\n"," '-',\n"," '.',\n"," '3',\n"," ':',\n"," ';',\n"," '?',\n"," 'A',\n"," 'B',\n"," 'C',\n"," 'D',\n"," 'E',\n"," 'F',\n"," 'G',\n"," 'H',\n"," 'I',\n"," 'J',\n"," 'K',\n"," 'L',\n"," 'M',\n"," 'N',\n"," 'O',\n"," 'P',\n"," 'Q',\n"," 'R',\n"," 'S',\n"," 'T',\n"," 'U',\n"," 'V',\n"," 'W',\n"," 'X',\n"," 'Y',\n"," 'Z',\n"," 'a',\n"," 'b',\n"," 'c',\n"," 'd',\n"," 'e',\n"," 'f',\n"," 'g',\n"," 'h',\n"," 'i',\n"," 'j',\n"," 'k',\n"," 'l',\n"," 'm',\n"," 'n',\n"," 'o',\n"," 'p',\n"," 'q',\n"," 'r',\n"," 's',\n"," 't',\n"," 'u',\n"," 'v',\n"," 'w',\n"," 'x',\n"," 'y',\n"," 'z']"]},"metadata":{},"execution_count":7}]},{"cell_type":"markdown","metadata":{"id":"R30B_cPQK_mo"},"source":["##### Preprocess the text\n","\n","Because the DL/ML model cannot understand the text, so we need to convert it in to numbers.\n","\n","The `preprocessing.StringLookup` layer can convert each character into a numeric ID. It just needs the text to be split into tokens first."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"-LT_4bTTLXNv","executionInfo":{"status":"ok","timestamp":1630158673881,"user_tz":-420,"elapsed":13,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"26a63c96-368d-4723-9aa4-00457ed90835"},"source":["example_texts = ['abcdefg', 'xyz']\n","\n","chars = tf.strings.unicode_split(example_texts, input_encoding='UTF-8')\n","chars"],"execution_count":8,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<tf.RaggedTensor [[b'a', b'b', b'c', b'd', b'e', b'f', b'g'], [b'x', b'y', b'z']]>"]},"metadata":{},"execution_count":8}]},{"cell_type":"code","metadata":{"id":"EndjMsJrMC-H","executionInfo":{"status":"ok","timestamp":1630158673881,"user_tz":-420,"elapsed":4,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}}},"source":["from tensorflow.keras import layers\n","ids_from_chars = layers.StringLookup(vocabulary=vocab,\n","                                     mask_token=None)"],"execution_count":9,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"66VrNi0jMhxM","executionInfo":{"status":"ok","timestamp":1630158674792,"user_tz":-420,"elapsed":5,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"a8c4c83b-bf7c-4ba2-957e-c4e086ec628f"},"source":["ids = ids_from_chars(chars)\n","ids"],"execution_count":10,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<tf.RaggedTensor [[40, 41, 42, 43, 44, 45, 46], [63, 64, 65]]>"]},"metadata":{},"execution_count":10}]},{"cell_type":"code","metadata":{"id":"3Q3fDnJcMtXx","executionInfo":{"status":"ok","timestamp":1630158675191,"user_tz":-420,"elapsed":4,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}}},"source":["## Revert from number to char\n","chars_from_ids = layers.StringLookup(vocabulary=ids_from_chars.get_vocabulary(),\n","                                     invert=True,\n","                                     mask_token=None)"],"execution_count":11,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"p5g8lpdhNHHi","executionInfo":{"status":"ok","timestamp":1630158675683,"user_tz":-420,"elapsed":15,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"c43f628f-2995-4c72-b0a3-833488a4874c"},"source":["chars = chars_from_ids(ids)\n","chars"],"execution_count":12,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<tf.RaggedTensor [[b'a', b'b', b'c', b'd', b'e', b'f', b'g'], [b'x', b'y', b'z']]>"]},"metadata":{},"execution_count":12}]},{"cell_type":"markdown","metadata":{"id":"JveRKrtyNKyC"},"source":["You can `tf.strings.reduce_join` to join the characters back into strings."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"xbnvt_fJNex_","executionInfo":{"status":"ok","timestamp":1630158676505,"user_tz":-420,"elapsed":4,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"ce8bdf85-412e-4678-8f96-705b5eaa305e"},"source":["tf.strings.reduce_join(chars, axis=-1)"],"execution_count":13,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<tf.Tensor: shape=(2,), dtype=string, numpy=array([b'abcdefg', b'xyz'], dtype=object)>"]},"metadata":{},"execution_count":13}]},{"cell_type":"code","metadata":{"id":"LLNbtpcENxdt","executionInfo":{"status":"ok","timestamp":1630158676906,"user_tz":-420,"elapsed":2,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}}},"source":["def text_from_ids(ids):\n","  return tf.strings.reduce_join(chars_from_ids(ids), axis=-1)"],"execution_count":14,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"shvBIaJeN8BK"},"source":["##### Create training example and targets"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"2x2A81cZOhBR","executionInfo":{"status":"ok","timestamp":1630158679140,"user_tz":-420,"elapsed":1460,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"e55f0b98-5538-4179-b8f4-dd7395828b80"},"source":["all_ids = ids_from_chars(tf.strings.unicode_split(text,\"UTF-8\"))\n","all_ids"],"execution_count":15,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<tf.Tensor: shape=(1115394,), dtype=int64, numpy=array([19, 48, 57, ..., 46,  9,  1])>"]},"metadata":{},"execution_count":15}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"8FXnYsdWOzyP","executionInfo":{"status":"ok","timestamp":1630158679141,"user_tz":-420,"elapsed":9,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"7eabeb7e-be9a-4764-fe99-5f2818fcce61"},"source":["ids_dataset = tf.data.Dataset.from_tensor_slices(all_ids)\n","ids_dataset"],"execution_count":16,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<TensorSliceDataset shapes: (), types: tf.int64>"]},"metadata":{},"execution_count":16}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"-l8rOgf5PMt7","executionInfo":{"status":"ok","timestamp":1630158679685,"user_tz":-420,"elapsed":547,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"86d0968e-d7fc-49a6-97ad-7f26ee5c88ed"},"source":["for id in ids_dataset.take(10):\n","  print(chars_from_ids(id).numpy().decode(\"utf-8\"))"],"execution_count":17,"outputs":[{"output_type":"stream","text":["F\n","i\n","r\n","s\n","t\n"," \n","C\n","i\n","t\n","i\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"Y0OnkAgvPemW","executionInfo":{"status":"ok","timestamp":1630158680017,"user_tz":-420,"elapsed":2,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}}},"source":["seq_length = 100\n","# examples_per_epoch = len(text)//seq_length\n","examples_per_epoch = 100"],"execution_count":18,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"wGkM2NItQc4K","executionInfo":{"status":"ok","timestamp":1630158680444,"user_tz":-420,"elapsed":16,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"4418dfb6-214d-4cc9-ed11-79eb25078f87"},"source":["len(text)"],"execution_count":19,"outputs":[{"output_type":"execute_result","data":{"text/plain":["1115394"]},"metadata":{},"execution_count":19}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"BlAVvFGQP_lC","executionInfo":{"status":"ok","timestamp":1630158680945,"user_tz":-420,"elapsed":2,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"6d24c358-d65a-4efb-dc92-dd8782d5fed3"},"source":["sequences = ids_dataset.batch(examples_per_epoch+1, drop_remainder=True)\n","\n","for seq in sequences.take(1):\n","  print(chars_from_ids(seq))"],"execution_count":20,"outputs":[{"output_type":"stream","text":["tf.Tensor(\n","[b'F' b'i' b'r' b's' b't' b' ' b'C' b'i' b't' b'i' b'z' b'e' b'n' b':'\n"," b'\\n' b'B' b'e' b'f' b'o' b'r' b'e' b' ' b'w' b'e' b' ' b'p' b'r' b'o'\n"," b'c' b'e' b'e' b'd' b' ' b'a' b'n' b'y' b' ' b'f' b'u' b'r' b't' b'h'\n"," b'e' b'r' b',' b' ' b'h' b'e' b'a' b'r' b' ' b'm' b'e' b' ' b's' b'p'\n"," b'e' b'a' b'k' b'.' b'\\n' b'\\n' b'A' b'l' b'l' b':' b'\\n' b'S' b'p' b'e'\n"," b'a' b'k' b',' b' ' b's' b'p' b'e' b'a' b'k' b'.' b'\\n' b'\\n' b'F' b'i'\n"," b'r' b's' b't' b' ' b'C' b'i' b't' b'i' b'z' b'e' b'n' b':' b'\\n' b'Y'\n"," b'o' b'u' b' '], shape=(101,), dtype=string)\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"CkXTsH7rQMUc","executionInfo":{"status":"ok","timestamp":1630158682039,"user_tz":-420,"elapsed":4,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"e52ef7de-2586-4363-d99b-b3d0dd9f43b8"},"source":["for seq in sequences.take(5):\n","  print(text_from_ids(seq).numpy())"],"execution_count":21,"outputs":[{"output_type":"stream","text":["b'First Citizen:\\nBefore we proceed any further, hear me speak.\\n\\nAll:\\nSpeak, speak.\\n\\nFirst Citizen:\\nYou '\n","b'are all resolved rather to die than to famish?\\n\\nAll:\\nResolved. resolved.\\n\\nFirst Citizen:\\nFirst, you k'\n","b\"now Caius Marcius is chief enemy to the people.\\n\\nAll:\\nWe know't, we know't.\\n\\nFirst Citizen:\\nLet us ki\"\n","b\"ll him, and we'll have corn at our own price.\\nIs't a verdict?\\n\\nAll:\\nNo more talking on't; let it be d\"\n","b'one: away, away!\\n\\nSecond Citizen:\\nOne word, good citizens.\\n\\nFirst Citizen:\\nWe are accounted poor citi'\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"yFseOwIdQ0CL"},"source":["For training you'll need a dataset of `(input, label)` pairs. Where `input` and label are sequences. At each time step the input is the current character and the label is the next character.\n","\n","Here's a function that takes a sequence as input, duplicates, and shifts it to align the input and label for each timestep:"]},{"cell_type":"code","metadata":{"id":"yf9fNPb8RNfd","executionInfo":{"status":"ok","timestamp":1630158683565,"user_tz":-420,"elapsed":2,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}}},"source":["def split_input_target(sequence):\n","  input_text = sequence[:-1]\n","  target_text = sequence[1:]\n","  return input_text, target_text"],"execution_count":22,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"ica9g2mqRZuG","executionInfo":{"status":"ok","timestamp":1630158684969,"user_tz":-420,"elapsed":12,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"b04d0501-5596-4493-c48d-aa7dc7a9b835"},"source":["sample_sequence = \"TensorFlow\"\n","\n","split_input_target(list(sample_sequence))"],"execution_count":23,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(['T', 'e', 'n', 's', 'o', 'r', 'F', 'l', 'o'],\n"," ['e', 'n', 's', 'o', 'r', 'F', 'l', 'o', 'w'])"]},"metadata":{},"execution_count":23}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"27xbhNKbRgYQ","executionInfo":{"status":"ok","timestamp":1630158686669,"user_tz":-420,"elapsed":503,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"e00d2c29-061a-4e6b-b1a1-dd5197a0e235"},"source":["dataset = sequences.map(split_input_target)\n","\n","for input_sample, target_sample in dataset.take(1):\n","  print(f\"Input: {text_from_ids(input_sample).numpy()}\")\n","  print(f\"Target: {text_from_ids(target_sample).numpy()}\")"],"execution_count":24,"outputs":[{"output_type":"stream","text":["Input: b'First Citizen:\\nBefore we proceed any further, hear me speak.\\n\\nAll:\\nSpeak, speak.\\n\\nFirst Citizen:\\nYou'\n","Target: b'irst Citizen:\\nBefore we proceed any further, hear me speak.\\n\\nAll:\\nSpeak, speak.\\n\\nFirst Citizen:\\nYou '\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"33kBgFoCR-p7"},"source":["###### Create training Batches"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"DNAvFhCNSVQH","executionInfo":{"status":"ok","timestamp":1630158688584,"user_tz":-420,"elapsed":3,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}},"outputId":"9652e596-c802-4f66-e0b0-33e7fdf22081"},"source":["# Batch size\n","BATCH_SIZE = 64\n","\n","# Buffer size to shuffle the dataset\n","# (TF data is designed to work with possibly infinite sequences,\n","# so it doesn't attempt to shuffle the entire sequence in memory. Instead,\n","# it maintains a buffer in which it shuffles elements).\n","BUFFER_SIZE = 10000\n","\n","dataset = (dataset.shuffle(BUFFER_SIZE)\n","                  .batch(BATCH_SIZE, drop_remainder=True)\n","                  .prefetch(tf.data.AUTOTUNE))\n","dataset"],"execution_count":25,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<PrefetchDataset shapes: ((64, 100), (64, 100)), types: (tf.int64, tf.int64)>"]},"metadata":{},"execution_count":25}]},{"cell_type":"markdown","metadata":{"id":"CQF7jHkRSyYK"},"source":["##### Build the model"]},{"cell_type":"code","metadata":{"id":"mGMtN6i2TQ0w","executionInfo":{"status":"ok","timestamp":1630158690755,"user_tz":-420,"elapsed":2,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}}},"source":["# Length of the vocabulary in chars\n","vocab_size = len(vocab)\n","\n","# Length of the sequence\n","seq_length = 100\n","\n","# The embedding dimension\n","embedding_dim = 256\n","\n","# Number of RNN units\n","rnn_units = 1024"],"execution_count":26,"outputs":[]},{"cell_type":"code","metadata":{"id":"P-tenLhvTtel","executionInfo":{"status":"ok","timestamp":1630158692481,"user_tz":-420,"elapsed":3,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}}},"source":["from tensorflow.keras import layers\n","# Create embedding layer\n","embedding_layer = layers.Embedding(input_dim=vocab_size,\n","                                   output_dim=embedding_dim,\n","                                   input_length=seq_length,\n","                                   embeddings_initializer=\"uniform\")"],"execution_count":27,"outputs":[]},{"cell_type":"code","metadata":{"id":"4N6dRAQqTR9i","executionInfo":{"status":"ok","timestamp":1630158853585,"user_tz":-420,"elapsed":1967,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}}},"source":["from tensorflow.keras import layers\n","inputs = layers.Input(batch_input_shape=[BATCH_SIZE, None], name=\"input_layer\")\n","\n","x = embedding_layer(inputs)\n","\n","x = layers.GRU(units=rnn_units,\n","               return_sequences=True,\n","               stateful=True)(x)\n","\n","outputs = layers.Dense(vocab_size)(x)\n","\n","model_10 = tf.keras.Model(inputs, outputs, name=\"model_10_text_generate\")\n"],"execution_count":38,"outputs":[]},{"cell_type":"code","metadata":{"id":"5Boxv2OEXLSE","executionInfo":{"status":"ok","timestamp":1630158766300,"user_tz":-420,"elapsed":1602,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}}},"source":["model_10 = tf.keras.Sequential([\n","     tf.keras.layers.Embedding(vocab_size, embedding_dim,\n","     batch_input_shape=[BATCH_SIZE, None]),\n","     tf.keras.layers.Dropout(0.2),\n","     tf.keras.layers.LSTM(rnn_units,\n","     return_sequences=True,\n","     stateful=True,\n","     recurrent_initializer=\"glorot_uniform\"),\n","     tf.keras.layers.Dropout(0.2), \n","     tf.keras.layers.LSTM(rnn_units,\n","     return_sequences=True,\n","     stateful=True,\n","     recurrent_initializer=\"glorot_uniform\"),\n","     tf.keras.layers.Dropout(0.2),\n","     tf.keras.layers.Dense(vocab_size)\n"," ])"],"execution_count":34,"outputs":[]},{"cell_type":"code","metadata":{"id":"C6DclrchU5h5","executionInfo":{"status":"ok","timestamp":1630158863044,"user_tz":-420,"elapsed":316,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}}},"source":["# Compile the model\n","model_10.compile(\n","    loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n","    optimizer=tf.keras.optimizers.Adam(),\n","    metrics=[\"accuracy\"]\n",")"],"execution_count":39,"outputs":[]},{"cell_type":"code","metadata":{"id":"l1FO862CWIPF","executionInfo":{"status":"ok","timestamp":1630158864255,"user_tz":-420,"elapsed":3,"user":{"displayName":"Lộc Bảo","photoUrl":"","userId":"01406170824357259025"}}},"source":["# # Directory where the checkpoints will be saved\n","# checkpoint_dir = './training_checkpoints'\n","# # Name of the checkpoint files\n","# checkpoint_prefix = os.path.join(checkpoint_dir, \"ckpt_{epoch}\")\n","checkpoint_prefix = model_10.name+\"/checkpoint.ckpt\"\n","checkpoint_callback = tf.keras.callbacks.ModelCheckpoint(\n","    filepath=checkpoint_prefix,\n","    monitor=\"accuracy\",\n","    save_best_only=True,\n","    save_weights_only=True)"],"execution_count":40,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"DkzNiuHiWBxr","outputId":"1dda6bf9-2c41-4356-8453-ec21cc411ca0"},"source":["EPOCHS=20\n","# Train the model\n","model_10_history = model_10.fit(dataset, epochs=EPOCHS, callbacks=[checkpoint_callback])"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Epoch 1/20\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"vMDFluf0WaaZ"},"source":["class MyModel(tf.keras.Model):\n","  def __init__(self, vocab_size, embedding_dim, rnn_units):\n","    super().__init__(self)\n","    self.embedding = tf.keras.layers.Embedding(vocab_size, embedding_dim)\n","    self.gru = tf.keras.layers.GRU(rnn_units,\n","                                   return_sequences=True,\n","                                   return_state=True)\n","    self.dense = tf.keras.layers.Dense(vocab_size)\n","\n","  def call(self, inputs, states=None, return_state=False, training=False):\n","    x = inputs\n","    x = self.embedding(x, training=training)\n","    if states is None:\n","      states = self.gru.get_initial_state(x)\n","    x, states = self.gru(x, initial_state=states, training=training)\n","    x = self.dense(x, training=training)\n","\n","    if return_state:\n","      return x, states\n","    else:\n","      return x"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"gLlkXNPgZlyU"},"source":["model = MyModel(\n","    # Be sure the vocabulary size matches the `StringLookup` layers.\n","    vocab_size=len(ids_from_chars.get_vocabulary()),\n","    embedding_dim=embedding_dim,\n","    rnn_units=rnn_units)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"o4lY8n0KZn9C","executionInfo":{"status":"ok","timestamp":1630047486231,"user_tz":-420,"elapsed":2266,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"4b8743de-9eb4-47fd-d918-4550776fc26d"},"source":["for input_example_batch, target_example_batch in dataset.take(1):\n","    example_batch_predictions = model(input_example_batch)\n","    print(example_batch_predictions.shape, \"# (batch_size, sequence_length, vocab_size)\")"],"execution_count":null,"outputs":[{"output_type":"stream","text":["(64, 100, 66) # (batch_size, sequence_length, vocab_size)\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"5zIIFtmrZrUa","executionInfo":{"status":"ok","timestamp":1630047512157,"user_tz":-420,"elapsed":349,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"1d353c4a-fdfe-4f58-fe38-57b7039a76ea"},"source":["sampled_indices = tf.random.categorical(example_batch_predictions[0], num_samples=1)\n","sampled_indices = tf.squeeze(sampled_indices, axis=-1).numpy()\n","sampled_indices"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([ 5,  7, 56, 61, 38, 65, 40, 31, 51, 10, 49, 46, 23,  2, 14, 45, 12,\n","       21, 35, 10, 20,  0, 64, 32, 22, 38, 38, 35, 31, 30, 38, 27, 63, 36,\n","       27, 58,  6, 64, 65,  1, 51,  4,  1, 21, 22, 12,  3, 59, 43,  8,  9,\n","        3, 51, 36, 56,  8, 51, 59, 30, 19, 51, 17, 55,  4, 52, 45,  4, 44,\n","        5, 62,  7, 50, 47, 55, 51, 44, 15, 24, 15,  8, 22, 54, 22, 11, 11,\n","       37, 64, 48, 21, 47, 21, 51,  4,  1, 17, 16,  9, 43, 63, 55])"]},"metadata":{},"execution_count":222}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"RqcjRljHZyLq","executionInfo":{"status":"ok","timestamp":1630047533376,"user_tz":-420,"elapsed":478,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"10d0acb6-7752-495a-9dee-165975067760"},"source":["print(\"Input:\\n\", text_from_ids(input_example_batch[0]).numpy())\n","print()\n","print(\"Next Char Predictions:\\n\", text_from_ids(sampled_indices).numpy())"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Input:\n"," b'g Adam Cupid, he that shot so trim,\\nWhen King Cophetua loved the beggar-maid!\\nHe heareth not, he sti'\n","\n","Next Char Predictions:\n"," b\"&,qvYzaRl3jgJ Af;HV3G[UNK]ySIYYVRQYNxWNs'yz\\nl$\\nHI;!td-.!lWq-ltQFlDp$mf$e&w,khpleBKB-IoI::XyiHhHl$\\nDC.dxp\"\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"WvgwWR6JZ8bu"},"source":["Train the model"]},{"cell_type":"code","metadata":{"id":"uBTewYyZZ3UA"},"source":["loss = tf.losses.SparseCategoricalCrossentropy(from_logits=True)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"1xw7zMGcaCRp","executionInfo":{"status":"ok","timestamp":1630047579588,"user_tz":-420,"elapsed":3,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"02a7509e-17ff-4a6a-fbb1-56ff1749a28e"},"source":["example_batch_loss = loss(target_example_batch, example_batch_predictions)\n","mean_loss = example_batch_loss.numpy().mean()\n","print(\"Prediction shape: \", example_batch_predictions.shape, \" # (batch_size, sequence_length, vocab_size)\")\n","print(\"Mean loss:        \", mean_loss)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Prediction shape:  (64, 100, 66)  # (batch_size, sequence_length, vocab_size)\n","Mean loss:         4.189515\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"bUBKObUOaClj"},"source":["model.compile(optimizer='adam', loss=loss)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"MAC7QaXcaFUC"},"source":["import os\n","# Directory where the checkpoints will be saved\n","checkpoint_dir = './training_checkpoints'\n","# Name of the checkpoint files\n","checkpoint_prefix = os.path.join(checkpoint_dir, \"ckpt_{epoch}\")\n","\n","checkpoint_callback = tf.keras.callbacks.ModelCheckpoint(\n","    filepath=checkpoint_prefix,\n","    save_weights_only=True)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"HMP6KEoeaIpQ","executionInfo":{"status":"ok","timestamp":1630048257274,"user_tz":-420,"elapsed":610982,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"9f5295f9-c7d1-4583-9843-22fecd2fcbb2"},"source":["model_10_history = model.fit(dataset, epochs=EPOCHS, callbacks=[checkpoint_callback])"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Epoch 1/20\n","172/172 [==============================] - 26s 134ms/step - loss: 2.7276\n","Epoch 2/20\n","172/172 [==============================] - 24s 134ms/step - loss: 1.9962\n","Epoch 3/20\n","172/172 [==============================] - 25s 133ms/step - loss: 1.7163\n","Epoch 4/20\n","172/172 [==============================] - 25s 134ms/step - loss: 1.5542\n","Epoch 5/20\n","172/172 [==============================] - 25s 134ms/step - loss: 1.4551\n","Epoch 6/20\n","172/172 [==============================] - 25s 134ms/step - loss: 1.3867\n","Epoch 7/20\n","172/172 [==============================] - 25s 134ms/step - loss: 1.3330\n","Epoch 8/20\n","172/172 [==============================] - 25s 135ms/step - loss: 1.2883\n","Epoch 9/20\n","172/172 [==============================] - 25s 136ms/step - loss: 1.2474\n","Epoch 10/20\n","172/172 [==============================] - 25s 135ms/step - loss: 1.2074\n","Epoch 11/20\n","172/172 [==============================] - 25s 135ms/step - loss: 1.1677\n","Epoch 12/20\n","172/172 [==============================] - 25s 135ms/step - loss: 1.1259\n","Epoch 13/20\n","172/172 [==============================] - 25s 135ms/step - loss: 1.0834\n","Epoch 14/20\n","172/172 [==============================] - 25s 134ms/step - loss: 1.0372\n","Epoch 15/20\n","172/172 [==============================] - 25s 134ms/step - loss: 0.9884\n","Epoch 16/20\n","172/172 [==============================] - 25s 135ms/step - loss: 0.9368\n","Epoch 17/20\n","172/172 [==============================] - 25s 135ms/step - loss: 0.8846\n","Epoch 18/20\n","172/172 [==============================] - 25s 134ms/step - loss: 0.8308\n","Epoch 19/20\n","172/172 [==============================] - 25s 135ms/step - loss: 0.7782\n","Epoch 20/20\n","172/172 [==============================] - 25s 135ms/step - loss: 0.7296\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"53ATZxY9aTAX"},"source":["## Time series\n","\n","Data sauce: https://www.coindesk.com/price/bitcoin\n","\n","The latest data this notebook use is 21/8/2021."]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"hNHgxz9xcja_","executionInfo":{"status":"ok","timestamp":1630050638311,"user_tz":-420,"elapsed":1327,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"98e7697a-c39e-45db-cb30-275c817e24c2"},"source":["!wget https://raw.githubusercontent.com/BaoLocPham/Tensorflow_Deep_Learning/main/Extras/BTC_USD_2013-10-01_2021-08-21-CoinDesk.csv"],"execution_count":null,"outputs":[{"output_type":"stream","text":["--2021-08-27 07:50:36--  https://raw.githubusercontent.com/BaoLocPham/Tensorflow_Deep_Learning/main/Extras/BTC_USD_2013-10-01_2021-08-21-CoinDesk.csv\n","Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.108.133, 185.199.109.133, 185.199.110.133, ...\n","Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.108.133|:443... connected.\n","HTTP request sent, awaiting response... 200 OK\n","Length: 183567 (179K) [text/plain]\n","Saving to: ‘BTC_USD_2013-10-01_2021-08-21-CoinDesk.csv’\n","\n","\r          BTC_USD_2   0%[                    ]       0  --.-KB/s               \rBTC_USD_2013-10-01_ 100%[===================>] 179.26K  --.-KB/s    in 0.02s   \n","\n","2021-08-27 07:50:36 (7.24 MB/s) - ‘BTC_USD_2013-10-01_2021-08-21-CoinDesk.csv’ saved [183567/183567]\n","\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":234},"id":"cNRR5MeElWia","executionInfo":{"status":"ok","timestamp":1630051348524,"user_tz":-420,"elapsed":337,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"d0764cdf-4c87-4c0e-9f9c-112c84f373c7"},"source":["import pandas as pd\n","\n","data_df = pd.read_csv(\"BTC_USD_2013-10-01_2021-08-21-CoinDesk.csv\",\n","                       parse_dates=[\"Date\"],\n","                       index_col=[\"Date\"])\n","\n","data_df.head()"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>Currency</th>\n","      <th>Closing Price (USD)</th>\n","      <th>24h Open (USD)</th>\n","      <th>24h High (USD)</th>\n","      <th>24h Low (USD)</th>\n","    </tr>\n","    <tr>\n","      <th>Date</th>\n","      <th></th>\n","      <th></th>\n","      <th></th>\n","      <th></th>\n","      <th></th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>2013-10-01</th>\n","      <td>BTC</td>\n","      <td>123.65499</td>\n","      <td>124.30466</td>\n","      <td>124.75166</td>\n","      <td>122.56349</td>\n","    </tr>\n","    <tr>\n","      <th>2013-10-02</th>\n","      <td>BTC</td>\n","      <td>125.45500</td>\n","      <td>123.65499</td>\n","      <td>125.75850</td>\n","      <td>123.63383</td>\n","    </tr>\n","    <tr>\n","      <th>2013-10-03</th>\n","      <td>BTC</td>\n","      <td>108.58483</td>\n","      <td>125.45500</td>\n","      <td>125.66566</td>\n","      <td>83.32833</td>\n","    </tr>\n","    <tr>\n","      <th>2013-10-04</th>\n","      <td>BTC</td>\n","      <td>118.67466</td>\n","      <td>108.58483</td>\n","      <td>118.67500</td>\n","      <td>107.05816</td>\n","    </tr>\n","    <tr>\n","      <th>2013-10-05</th>\n","      <td>BTC</td>\n","      <td>121.33866</td>\n","      <td>118.67466</td>\n","      <td>121.93633</td>\n","      <td>118.00566</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["           Currency  Closing Price (USD)  ...  24h High (USD)  24h Low (USD)\n","Date                                      ...                               \n","2013-10-01      BTC            123.65499  ...       124.75166      122.56349\n","2013-10-02      BTC            125.45500  ...       125.75850      123.63383\n","2013-10-03      BTC            108.58483  ...       125.66566       83.32833\n","2013-10-04      BTC            118.67466  ...       118.67500      107.05816\n","2013-10-05      BTC            121.33866  ...       121.93633      118.00566\n","\n","[5 rows x 5 columns]"]},"metadata":{},"execution_count":245}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"ZEGTN4IEl5aV","executionInfo":{"status":"ok","timestamp":1630051349784,"user_tz":-420,"elapsed":3,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"b7998c09-46e1-407c-905d-b0c46e86f813"},"source":["data_df.info()"],"execution_count":null,"outputs":[{"output_type":"stream","text":["<class 'pandas.core.frame.DataFrame'>\n","DatetimeIndex: 2882 entries, 2013-10-01 to 2021-08-21\n","Data columns (total 5 columns):\n"," #   Column               Non-Null Count  Dtype  \n","---  ------               --------------  -----  \n"," 0   Currency             2882 non-null   object \n"," 1   Closing Price (USD)  2882 non-null   float64\n"," 2   24h Open (USD)       2882 non-null   float64\n"," 3   24h High (USD)       2882 non-null   float64\n"," 4   24h Low (USD)        2882 non-null   float64\n","dtypes: float64(4), object(1)\n","memory usage: 135.1+ KB\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":234},"id":"wmilfxeSmP8l","executionInfo":{"status":"ok","timestamp":1630051351667,"user_tz":-420,"elapsed":338,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"620c2259-c6b9-4b54-b06e-e66173e37439"},"source":["bitcoin_prices = pd.DataFrame(data_df[\"Closing Price (USD)\"]).rename(columns={\"Closing Price (USD)\":\"Price\"})\n","bitcoin_prices.head()"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>Price</th>\n","    </tr>\n","    <tr>\n","      <th>Date</th>\n","      <th></th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>2013-10-01</th>\n","      <td>123.65499</td>\n","    </tr>\n","    <tr>\n","      <th>2013-10-02</th>\n","      <td>125.45500</td>\n","    </tr>\n","    <tr>\n","      <th>2013-10-03</th>\n","      <td>108.58483</td>\n","    </tr>\n","    <tr>\n","      <th>2013-10-04</th>\n","      <td>118.67466</td>\n","    </tr>\n","    <tr>\n","      <th>2013-10-05</th>\n","      <td>121.33866</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                Price\n","Date                 \n","2013-10-01  123.65499\n","2013-10-02  125.45500\n","2013-10-03  108.58483\n","2013-10-04  118.67466\n","2013-10-05  121.33866"]},"metadata":{},"execution_count":247}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":515},"id":"dbT8vrqYndkv","executionInfo":{"status":"ok","timestamp":1630051359859,"user_tz":-420,"elapsed":507,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"8e62c198-3dc8-49c1-93e7-477b7d927ede"},"source":["# Plot our data\n","import matplotlib.pyplot as plt\n","\n","bitcoin_prices.plot(figsize=(10,8))\n","plt.ylabel(\"Bitcoins price\")\n","plt.title(\"Price of Bitcoin from 1 Otc 2013 to 21 Aug 2021\", fontsize=16)\n","plt.show()"],"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 720x576 with 1 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"zV9SZO1jodgK"},"source":["##### Preprocess data"]},{"cell_type":"code","metadata":{"id":"B6XnAxD3r8IV"},"source":["# get bitcoins dates and prices array\n","dates = bitcoin_prices.index.to_numpy()\n","prices = bitcoin_prices[\"Price\"].to_numpy()"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"6J1GQoIDsreR"},"source":["WINDOW_SIZE = 7\n","HORIZON = 1"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":234},"id":"EtunXuyzsMTY","executionInfo":{"status":"ok","timestamp":1630052532700,"user_tz":-420,"elapsed":328,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"7965446a-ee5a-480b-c669-e27140c7ba87"},"source":["bitcoin_prices_ds = bitcoin_prices.copy()\n","\n","for i in range(WINDOW_SIZE):\n","  bitcoin_prices_ds[f\"Price+{i+1}\"] = bitcoin_prices[\"Price\"].shift(i+1)\n","bitcoin_prices_ds.head()"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>Price</th>\n","      <th>Price+1</th>\n","      <th>Price+2</th>\n","      <th>Price+3</th>\n","      <th>Price+4</th>\n","      <th>Price+5</th>\n","      <th>Price+6</th>\n","      <th>Price+7</th>\n","    </tr>\n","    <tr>\n","      <th>Date</th>\n","      <th></th>\n","      <th></th>\n","      <th></th>\n","      <th></th>\n","      <th></th>\n","      <th></th>\n","      <th></th>\n","      <th></th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>2013-10-01</th>\n","      <td>123.65499</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","    </tr>\n","    <tr>\n","      <th>2013-10-02</th>\n","      <td>125.45500</td>\n","      <td>123.65499</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","    </tr>\n","    <tr>\n","      <th>2013-10-03</th>\n","      <td>108.58483</td>\n","      <td>125.45500</td>\n","      <td>123.65499</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","    </tr>\n","    <tr>\n","      <th>2013-10-04</th>\n","      <td>118.67466</td>\n","      <td>108.58483</td>\n","      <td>125.45500</td>\n","      <td>123.65499</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","    </tr>\n","    <tr>\n","      <th>2013-10-05</th>\n","      <td>121.33866</td>\n","      <td>118.67466</td>\n","      <td>108.58483</td>\n","      <td>125.45500</td>\n","      <td>123.65499</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","      <td>NaN</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                Price    Price+1    Price+2  ...  Price+5  Price+6  Price+7\n","Date                                         ...                           \n","2013-10-01  123.65499        NaN        NaN  ...      NaN      NaN      NaN\n","2013-10-02  125.45500  123.65499        NaN  ...      NaN      NaN      NaN\n","2013-10-03  108.58483  125.45500  123.65499  ...      NaN      NaN      NaN\n","2013-10-04  118.67466  108.58483  125.45500  ...      NaN      NaN      NaN\n","2013-10-05  121.33866  118.67466  108.58483  ...      NaN      NaN      NaN\n","\n","[5 rows x 8 columns]"]},"metadata":{},"execution_count":253}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"Wx92pP5espKv","executionInfo":{"status":"ok","timestamp":1630052746110,"user_tz":-420,"elapsed":532,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"a383850a-dd4f-4086-9f7e-1054e0e1c4a3"},"source":["# Make X (features) and y (labels)\n","X = bitcoin_prices_ds.dropna().drop([\"Price\"], axis=1).astype(np.float32)\n","y = bitcoin_prices_ds.dropna()[\"Price\"].astype(np.float32)\n","\n","# Split data in train and test set\n","split_size = int(0.8 * len(X))\n","\n","X_train, y_train = X[:split_size], y[:split_size]\n","X_test, y_test = X[split_size:], y[split_size:]\n","\n","X_train.shape, X_test.shape, y_train.shape, y_test.shape"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["((2300, 7), (575, 7), (2300,), (575,))"]},"metadata":{},"execution_count":254}]},{"cell_type":"markdown","metadata":{"id":"6BTFGtTrtv2F"},"source":["##### Build and Fit model"]},{"cell_type":"markdown","metadata":{"id":"M1M_FsMCt1oP"},"source":["###### Conv1D"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"vFBi5QRUt7_x","executionInfo":{"status":"ok","timestamp":1630053401165,"user_tz":-420,"elapsed":41627,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"e20923c6-9798-4dc3-9503-baf13532924e"},"source":["import tensorflow as tf\n","from tensorflow.keras import layers\n","\n","model_11 = tf.keras.Sequential([\n","  layers.Lambda(lambda x: tf.expand_dims(x, axis=1)),\n","  layers.Conv1D(64,\n","                3,\n","                padding=\"causal\",\n","                activation=\"relu\"),\n","  layers.Dense(HORIZON)\n","], name=\"model_11_timeseries_conv1d\")\n","\n","# Compile model\n","model_11.compile(\n","    loss=\"mae\",\n","    optimizer=tf.keras.optimizers.Adam(),\n","    metrics=[\"mae\",\"mse\"]\n",")\n","\n","# Fit the model\n","model_11_history = model_11.fit(X_train,\n","                                y_train,\n","                                validation_data=(X_test, y_test),\n","                                epochs=100,\n","                                callbacks=[create_checkpoint_callback(model_11.name)])"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Epoch 1/100\n","72/72 [==============================] - 1s 8ms/step - loss: 614.6157 - mae: 614.6157 - mse: 2161472.7500 - val_loss: 1242.7042 - val_mae: 1242.7042 - val_mse: 4008170.7500\n","Epoch 2/100\n","72/72 [==============================] - 0s 5ms/step - loss: 175.5749 - mae: 175.5749 - mse: 157533.1406 - val_loss: 1178.5653 - val_mae: 1178.5653 - val_mse: 3722637.2500\n","Epoch 3/100\n","72/72 [==============================] - 0s 6ms/step - loss: 168.5308 - mae: 168.5308 - mse: 147045.2656 - val_loss: 1132.5740 - val_mae: 1132.5740 - val_mse: 3436972.7500\n","Epoch 4/100\n","72/72 [==============================] - 0s 6ms/step - loss: 160.2467 - mae: 160.2467 - mse: 132323.6562 - val_loss: 1084.9133 - val_mae: 1084.9133 - val_mse: 3149296.7500\n","Epoch 5/100\n","72/72 [==============================] - 0s 5ms/step - loss: 154.1229 - mae: 154.1229 - mse: 123689.0312 - val_loss: 1053.8153 - val_mae: 1053.8153 - val_mse: 2842061.7500\n","Epoch 6/100\n","72/72 [==============================] - 0s 6ms/step - loss: 149.9594 - mae: 149.9594 - mse: 118283.0859 - val_loss: 1111.0377 - val_mae: 1111.0377 - val_mse: 3280233.5000\n","Epoch 7/100\n","72/72 [==============================] - 0s 5ms/step - loss: 148.0216 - mae: 148.0216 - mse: 111634.1328 - val_loss: 976.0928 - val_mae: 976.0928 - val_mse: 2492185.2500\n","Epoch 8/100\n","72/72 [==============================] - 0s 5ms/step - loss: 137.5817 - mae: 137.5817 - mse: 103001.3281 - val_loss: 986.2731 - val_mae: 986.2731 - val_mse: 2421195.0000\n","Epoch 9/100\n","72/72 [==============================] - 0s 6ms/step - loss: 140.9859 - mae: 140.9859 - mse: 102137.7891 - val_loss: 944.4949 - val_mae: 944.4949 - val_mse: 2253428.0000\n","Epoch 10/100\n","72/72 [==============================] - 0s 5ms/step - loss: 135.2110 - mae: 135.2110 - mse: 95595.7500 - val_loss: 904.6028 - val_mae: 904.6028 - val_mse: 2123699.7500\n","Epoch 11/100\n","72/72 [==============================] - 0s 5ms/step - loss: 134.1078 - mae: 134.1078 - mse: 94529.0000 - val_loss: 892.1868 - val_mae: 892.1868 - val_mse: 2052869.7500\n","Epoch 12/100\n","72/72 [==============================] - 0s 6ms/step - loss: 131.9740 - mae: 131.9740 - mse: 92058.7891 - val_loss: 878.5240 - val_mae: 878.5240 - val_mse: 2077303.5000\n","Epoch 13/100\n","72/72 [==============================] - 0s 5ms/step - loss: 133.0147 - mae: 133.0147 - mse: 92997.3828 - val_loss: 867.2388 - val_mae: 867.2388 - val_mse: 1953527.5000\n","Epoch 14/100\n","72/72 [==============================] - 0s 5ms/step - loss: 126.4689 - mae: 126.4689 - mse: 88352.7891 - val_loss: 920.2789 - val_mae: 920.2789 - val_mse: 2095238.8750\n","Epoch 15/100\n","72/72 [==============================] - 0s 6ms/step - loss: 123.0613 - mae: 123.0613 - mse: 83630.0391 - val_loss: 829.5771 - val_mae: 829.5771 - val_mse: 1850643.7500\n","Epoch 16/100\n","72/72 [==============================] - 0s 5ms/step - loss: 129.3268 - mae: 129.3268 - mse: 88094.6719 - val_loss: 853.1009 - val_mae: 853.1009 - val_mse: 1969992.7500\n","Epoch 17/100\n","72/72 [==============================] - 0s 6ms/step - loss: 121.8377 - mae: 121.8377 - mse: 81135.4688 - val_loss: 1012.3956 - val_mae: 1012.3956 - val_mse: 2386700.7500\n","Epoch 18/100\n","72/72 [==============================] - 0s 5ms/step - loss: 120.6378 - mae: 120.6378 - mse: 80222.7812 - val_loss: 829.7302 - val_mae: 829.7302 - val_mse: 1877411.8750\n","Epoch 19/100\n","72/72 [==============================] - 0s 5ms/step - loss: 118.7516 - mae: 118.7516 - mse: 78247.2969 - val_loss: 816.9943 - val_mae: 816.9943 - val_mse: 1777415.2500\n","Epoch 20/100\n","72/72 [==============================] - 0s 5ms/step - loss: 122.1463 - mae: 122.1463 - mse: 81814.5625 - val_loss: 798.6764 - val_mae: 798.6764 - val_mse: 1745173.8750\n","Epoch 21/100\n","72/72 [==============================] - 0s 5ms/step - loss: 127.1342 - mae: 127.1342 - mse: 84279.7500 - val_loss: 838.1213 - val_mae: 838.1213 - val_mse: 1901565.3750\n","Epoch 22/100\n","72/72 [==============================] - 0s 6ms/step - loss: 118.6854 - mae: 118.6854 - mse: 79104.6797 - val_loss: 795.2349 - val_mae: 795.2349 - val_mse: 1743587.7500\n","Epoch 23/100\n","72/72 [==============================] - 0s 5ms/step - loss: 117.8973 - mae: 117.8973 - mse: 77749.3672 - val_loss: 815.4985 - val_mae: 815.4985 - val_mse: 1755202.1250\n","Epoch 24/100\n","72/72 [==============================] - 0s 5ms/step - loss: 120.4848 - mae: 120.4848 - mse: 80611.0938 - val_loss: 1026.9548 - val_mae: 1026.9548 - val_mse: 2408503.2500\n","Epoch 25/100\n","72/72 [==============================] - 0s 5ms/step - loss: 118.7152 - mae: 118.7152 - mse: 77051.1406 - val_loss: 904.9764 - val_mae: 904.9764 - val_mse: 2115205.0000\n","Epoch 26/100\n","72/72 [==============================] - 0s 5ms/step - loss: 127.5716 - mae: 127.5716 - mse: 80757.7422 - val_loss: 791.5551 - val_mae: 791.5551 - val_mse: 1724984.3750\n","Epoch 27/100\n","72/72 [==============================] - 0s 6ms/step - loss: 121.5461 - mae: 121.5461 - mse: 79912.6484 - val_loss: 957.8674 - val_mae: 957.8674 - val_mse: 2309966.2500\n","Epoch 28/100\n","72/72 [==============================] - 0s 5ms/step - loss: 121.1825 - mae: 121.1825 - mse: 81258.0000 - val_loss: 819.7735 - val_mae: 819.7735 - val_mse: 1756093.8750\n","Epoch 29/100\n","72/72 [==============================] - 0s 6ms/step - loss: 117.6826 - mae: 117.6826 - mse: 75872.4609 - val_loss: 774.4570 - val_mae: 774.4570 - val_mse: 1646321.5000\n","Epoch 30/100\n","72/72 [==============================] - 0s 5ms/step - loss: 117.6340 - mae: 117.6340 - mse: 77335.7578 - val_loss: 816.1813 - val_mae: 816.1813 - val_mse: 1741859.6250\n","Epoch 31/100\n","72/72 [==============================] - 0s 5ms/step - loss: 113.0255 - mae: 113.0255 - mse: 74252.2500 - val_loss: 879.5677 - val_mae: 879.5677 - val_mse: 1998102.8750\n","Epoch 32/100\n","72/72 [==============================] - 0s 5ms/step - loss: 113.9206 - mae: 113.9206 - mse: 71991.7031 - val_loss: 848.5192 - val_mae: 848.5192 - val_mse: 1903237.1250\n","Epoch 33/100\n","72/72 [==============================] - 0s 5ms/step - loss: 116.7674 - mae: 116.7674 - mse: 75280.2500 - val_loss: 790.7396 - val_mae: 790.7396 - val_mse: 1674656.8750\n","Epoch 34/100\n","72/72 [==============================] - 0s 5ms/step - loss: 122.6911 - mae: 122.6911 - mse: 74974.7812 - val_loss: 775.3099 - val_mae: 775.3099 - val_mse: 1663314.2500\n","Epoch 35/100\n","72/72 [==============================] - 0s 5ms/step - loss: 111.9028 - mae: 111.9028 - mse: 70346.7422 - val_loss: 764.5463 - val_mae: 764.5463 - val_mse: 1618440.5000\n","Epoch 36/100\n","72/72 [==============================] - 0s 5ms/step - loss: 122.9731 - mae: 122.9731 - mse: 80130.0078 - val_loss: 755.6406 - val_mae: 755.6406 - val_mse: 1608783.5000\n","Epoch 37/100\n","72/72 [==============================] - 0s 5ms/step - loss: 116.0143 - mae: 116.0143 - mse: 73730.4922 - val_loss: 807.4422 - val_mae: 807.4422 - val_mse: 1715085.7500\n","Epoch 38/100\n","72/72 [==============================] - 0s 5ms/step - loss: 114.0964 - mae: 114.0964 - mse: 71650.9297 - val_loss: 769.2487 - val_mae: 769.2487 - val_mse: 1623107.8750\n","Epoch 39/100\n","72/72 [==============================] - 0s 6ms/step - loss: 113.7422 - mae: 113.7422 - mse: 73969.9453 - val_loss: 749.5640 - val_mae: 749.5640 - val_mse: 1590180.1250\n","Epoch 40/100\n","72/72 [==============================] - 0s 5ms/step - loss: 112.5666 - mae: 112.5666 - mse: 72108.2344 - val_loss: 1085.5486 - val_mae: 1085.5486 - val_mse: 2779712.0000\n","Epoch 41/100\n","72/72 [==============================] - 0s 5ms/step - loss: 120.8737 - mae: 120.8737 - mse: 75782.5469 - val_loss: 793.5712 - val_mae: 793.5712 - val_mse: 1676225.5000\n","Epoch 42/100\n","72/72 [==============================] - 0s 6ms/step - loss: 122.6445 - mae: 122.6445 - mse: 77773.6875 - val_loss: 751.8911 - val_mae: 751.8911 - val_mse: 1595545.6250\n","Epoch 43/100\n","72/72 [==============================] - 0s 5ms/step - loss: 110.9496 - mae: 110.9496 - mse: 71686.4297 - val_loss: 772.9129 - val_mae: 772.9129 - val_mse: 1627905.7500\n","Epoch 44/100\n","72/72 [==============================] - 0s 5ms/step - loss: 115.1343 - mae: 115.1343 - mse: 72201.6328 - val_loss: 805.1144 - val_mae: 805.1144 - val_mse: 1698990.5000\n","Epoch 45/100\n","72/72 [==============================] - 0s 5ms/step - loss: 110.6958 - mae: 110.6958 - mse: 71129.5312 - val_loss: 743.9867 - val_mae: 743.9867 - val_mse: 1571284.0000\n","Epoch 46/100\n","72/72 [==============================] - 0s 5ms/step - loss: 112.5385 - mae: 112.5385 - mse: 71682.9062 - val_loss: 747.3276 - val_mae: 747.3276 - val_mse: 1580716.0000\n","Epoch 47/100\n","72/72 [==============================] - 0s 6ms/step - loss: 121.1405 - mae: 121.1405 - mse: 80991.8984 - val_loss: 751.0188 - val_mae: 751.0188 - val_mse: 1590292.6250\n","Epoch 48/100\n","72/72 [==============================] - 0s 5ms/step - loss: 115.3148 - mae: 115.3148 - mse: 73910.4062 - val_loss: 738.6165 - val_mae: 738.6165 - val_mse: 1553284.3750\n","Epoch 49/100\n","72/72 [==============================] - 0s 5ms/step - loss: 111.3701 - mae: 111.3701 - mse: 71430.7422 - val_loss: 814.2628 - val_mae: 814.2628 - val_mse: 1720237.3750\n","Epoch 50/100\n","72/72 [==============================] - 0s 6ms/step - loss: 117.0571 - mae: 117.0571 - mse: 76310.3203 - val_loss: 755.8821 - val_mae: 755.8821 - val_mse: 1590923.1250\n","Epoch 51/100\n","72/72 [==============================] - 0s 5ms/step - loss: 112.7996 - mae: 112.7996 - mse: 72642.3203 - val_loss: 741.3145 - val_mae: 741.3145 - val_mse: 1559005.0000\n","Epoch 52/100\n","72/72 [==============================] - 0s 5ms/step - loss: 110.0012 - mae: 110.0012 - mse: 70867.5859 - val_loss: 739.4976 - val_mae: 739.4976 - val_mse: 1557574.2500\n","Epoch 53/100\n","72/72 [==============================] - 0s 5ms/step - loss: 117.4945 - mae: 117.4945 - mse: 74311.4766 - val_loss: 854.5791 - val_mae: 854.5791 - val_mse: 1904808.1250\n","Epoch 54/100\n","72/72 [==============================] - 0s 5ms/step - loss: 114.5156 - mae: 114.5156 - mse: 69617.7266 - val_loss: 931.2373 - val_mae: 931.2373 - val_mse: 2052093.3750\n","Epoch 55/100\n","72/72 [==============================] - 0s 6ms/step - loss: 111.9847 - mae: 111.9847 - mse: 70956.6484 - val_loss: 749.2933 - val_mae: 749.2933 - val_mse: 1574013.3750\n","Epoch 56/100\n","72/72 [==============================] - 0s 6ms/step - loss: 112.3521 - mae: 112.3521 - mse: 71564.3125 - val_loss: 741.2443 - val_mae: 741.2443 - val_mse: 1562458.1250\n","Epoch 57/100\n","72/72 [==============================] - 0s 5ms/step - loss: 109.3991 - mae: 109.3991 - mse: 68753.2266 - val_loss: 752.8886 - val_mae: 752.8886 - val_mse: 1593718.8750\n","Epoch 58/100\n","72/72 [==============================] - 0s 6ms/step - loss: 114.9481 - mae: 114.9481 - mse: 73698.3672 - val_loss: 774.4450 - val_mae: 774.4450 - val_mse: 1625428.2500\n","Epoch 59/100\n","72/72 [==============================] - 0s 6ms/step - loss: 112.9934 - mae: 112.9934 - mse: 71924.4844 - val_loss: 827.8801 - val_mae: 827.8801 - val_mse: 1799912.3750\n","Epoch 60/100\n","72/72 [==============================] - 0s 6ms/step - loss: 113.2683 - mae: 113.2683 - mse: 72310.8594 - val_loss: 800.7916 - val_mae: 800.7916 - val_mse: 1685989.1250\n","Epoch 61/100\n","72/72 [==============================] - 0s 5ms/step - loss: 114.3941 - mae: 114.3941 - mse: 74116.9219 - val_loss: 736.0428 - val_mae: 736.0428 - val_mse: 1547057.2500\n","Epoch 62/100\n","72/72 [==============================] - 0s 5ms/step - loss: 110.1019 - mae: 110.1019 - mse: 68784.1250 - val_loss: 843.4028 - val_mae: 843.4028 - val_mse: 1790287.7500\n","Epoch 63/100\n","72/72 [==============================] - 0s 5ms/step - loss: 109.7553 - mae: 109.7553 - mse: 70183.0938 - val_loss: 759.3615 - val_mae: 759.3615 - val_mse: 1611635.7500\n","Epoch 64/100\n","72/72 [==============================] - 0s 5ms/step - loss: 112.1018 - mae: 112.1018 - mse: 70659.3516 - val_loss: 783.7443 - val_mae: 783.7443 - val_mse: 1688880.2500\n","Epoch 65/100\n","72/72 [==============================] - 0s 6ms/step - loss: 108.4254 - mae: 108.4254 - mse: 68499.9688 - val_loss: 739.7853 - val_mae: 739.7853 - val_mse: 1557794.6250\n","Epoch 66/100\n","72/72 [==============================] - 0s 5ms/step - loss: 108.7251 - mae: 108.7251 - mse: 67546.5547 - val_loss: 754.0783 - val_mae: 754.0783 - val_mse: 1599178.1250\n","Epoch 67/100\n","72/72 [==============================] - 0s 5ms/step - loss: 114.5058 - mae: 114.5058 - mse: 70602.2422 - val_loss: 733.4193 - val_mae: 733.4193 - val_mse: 1536945.7500\n","Epoch 68/100\n","72/72 [==============================] - 0s 6ms/step - loss: 115.7112 - mae: 115.7112 - mse: 75303.3438 - val_loss: 736.5015 - val_mae: 736.5015 - val_mse: 1549591.7500\n","Epoch 69/100\n","72/72 [==============================] - 0s 5ms/step - loss: 115.6527 - mae: 115.6527 - mse: 74236.4141 - val_loss: 740.6400 - val_mae: 740.6400 - val_mse: 1562211.7500\n","Epoch 70/100\n","72/72 [==============================] - 0s 6ms/step - loss: 111.5195 - mae: 111.5195 - mse: 70818.0625 - val_loss: 748.8092 - val_mae: 748.8092 - val_mse: 1583725.8750\n","Epoch 71/100\n","72/72 [==============================] - 0s 6ms/step - loss: 110.2919 - mae: 110.2919 - mse: 70020.4922 - val_loss: 782.3090 - val_mae: 782.3090 - val_mse: 1642768.8750\n","Epoch 72/100\n","72/72 [==============================] - 0s 5ms/step - loss: 111.3615 - mae: 111.3615 - mse: 69935.5625 - val_loss: 736.8748 - val_mae: 736.8748 - val_mse: 1542114.6250\n","Epoch 73/100\n","72/72 [==============================] - 0s 6ms/step - loss: 123.0455 - mae: 123.0455 - mse: 79125.9766 - val_loss: 735.3514 - val_mae: 735.3514 - val_mse: 1546376.0000\n","Epoch 74/100\n","72/72 [==============================] - 0s 6ms/step - loss: 108.2294 - mae: 108.2294 - mse: 68449.3984 - val_loss: 739.4819 - val_mae: 739.4819 - val_mse: 1558257.3750\n","Epoch 75/100\n","72/72 [==============================] - 0s 6ms/step - loss: 116.4074 - mae: 116.4074 - mse: 74349.6250 - val_loss: 820.3811 - val_mae: 820.3811 - val_mse: 1782254.8750\n","Epoch 76/100\n","72/72 [==============================] - 0s 6ms/step - loss: 111.8116 - mae: 111.8116 - mse: 71993.1250 - val_loss: 743.2713 - val_mae: 743.2713 - val_mse: 1565871.0000\n","Epoch 77/100\n","72/72 [==============================] - 0s 6ms/step - loss: 110.8903 - mae: 110.8903 - mse: 71101.8203 - val_loss: 797.1416 - val_mae: 797.1416 - val_mse: 1675480.1250\n","Epoch 78/100\n","72/72 [==============================] - 0s 5ms/step - loss: 114.4614 - mae: 114.4614 - mse: 72635.7500 - val_loss: 744.0034 - val_mae: 744.0034 - val_mse: 1571904.5000\n","Epoch 79/100\n","72/72 [==============================] - 0s 5ms/step - loss: 114.6727 - mae: 114.6727 - mse: 72899.6406 - val_loss: 807.4010 - val_mae: 807.4010 - val_mse: 1736566.8750\n","Epoch 80/100\n","72/72 [==============================] - 0s 5ms/step - loss: 115.9497 - mae: 115.9497 - mse: 72990.1719 - val_loss: 750.4838 - val_mae: 750.4838 - val_mse: 1576199.2500\n","Epoch 81/100\n","72/72 [==============================] - 0s 6ms/step - loss: 110.0240 - mae: 110.0240 - mse: 70506.0234 - val_loss: 781.6802 - val_mae: 781.6802 - val_mse: 1638547.3750\n","Epoch 82/100\n","72/72 [==============================] - 0s 5ms/step - loss: 115.6292 - mae: 115.6292 - mse: 75849.5078 - val_loss: 775.2158 - val_mae: 775.2158 - val_mse: 1660926.5000\n","Epoch 83/100\n","72/72 [==============================] - 0s 5ms/step - loss: 111.9470 - mae: 111.9470 - mse: 69547.1172 - val_loss: 734.4750 - val_mae: 734.4750 - val_mse: 1542035.6250\n","Epoch 84/100\n","72/72 [==============================] - 0s 5ms/step - loss: 111.0202 - mae: 111.0202 - mse: 70569.6406 - val_loss: 772.3024 - val_mae: 772.3024 - val_mse: 1645596.5000\n","Epoch 85/100\n","72/72 [==============================] - 0s 5ms/step - loss: 115.3405 - mae: 115.3405 - mse: 72116.5859 - val_loss: 733.1739 - val_mae: 733.1739 - val_mse: 1534513.1250\n","Epoch 86/100\n","72/72 [==============================] - 0s 5ms/step - loss: 116.3801 - mae: 116.3801 - mse: 73810.5703 - val_loss: 751.3033 - val_mae: 751.3033 - val_mse: 1591644.0000\n","Epoch 87/100\n","72/72 [==============================] - 0s 6ms/step - loss: 108.2442 - mae: 108.2442 - mse: 68011.0312 - val_loss: 742.3723 - val_mae: 742.3723 - val_mse: 1568656.6250\n","Epoch 88/100\n","72/72 [==============================] - 0s 5ms/step - loss: 112.4901 - mae: 112.4901 - mse: 70228.8516 - val_loss: 755.1449 - val_mae: 755.1449 - val_mse: 1583440.8750\n","Epoch 89/100\n","72/72 [==============================] - 0s 6ms/step - loss: 111.2611 - mae: 111.2611 - mse: 71487.8125 - val_loss: 841.6678 - val_mae: 841.6678 - val_mse: 1787221.8750\n","Epoch 90/100\n","72/72 [==============================] - 0s 6ms/step - loss: 108.5631 - mae: 108.5631 - mse: 68547.1172 - val_loss: 733.2198 - val_mae: 733.2198 - val_mse: 1538952.6250\n","Epoch 91/100\n","72/72 [==============================] - 0s 6ms/step - loss: 107.7878 - mae: 107.7878 - mse: 68608.3203 - val_loss: 742.8192 - val_mae: 742.8192 - val_mse: 1570357.3750\n","Epoch 92/100\n","72/72 [==============================] - 0s 6ms/step - loss: 108.3571 - mae: 108.3571 - mse: 68104.9844 - val_loss: 731.7509 - val_mae: 731.7509 - val_mse: 1537135.2500\n","Epoch 93/100\n","72/72 [==============================] - 0s 6ms/step - loss: 110.2940 - mae: 110.2940 - mse: 67336.8203 - val_loss: 733.7688 - val_mae: 733.7688 - val_mse: 1543114.6250\n","Epoch 94/100\n","72/72 [==============================] - 0s 6ms/step - loss: 110.3550 - mae: 110.3550 - mse: 69303.6406 - val_loss: 745.7553 - val_mae: 745.7553 - val_mse: 1562961.7500\n","Epoch 95/100\n","72/72 [==============================] - 0s 6ms/step - loss: 114.9851 - mae: 114.9851 - mse: 73586.6328 - val_loss: 737.9631 - val_mae: 737.9631 - val_mse: 1551233.5000\n","Epoch 96/100\n","72/72 [==============================] - 0s 6ms/step - loss: 112.9743 - mae: 112.9743 - mse: 70900.0781 - val_loss: 789.7216 - val_mae: 789.7216 - val_mse: 1700705.8750\n","Epoch 97/100\n","72/72 [==============================] - 0s 5ms/step - loss: 114.2373 - mae: 114.2373 - mse: 72022.5859 - val_loss: 891.1962 - val_mae: 891.1962 - val_mse: 1931597.8750\n","Epoch 98/100\n","72/72 [==============================] - 0s 5ms/step - loss: 114.2034 - mae: 114.2034 - mse: 70945.9609 - val_loss: 815.3397 - val_mae: 815.3397 - val_mse: 1778063.6250\n","Epoch 99/100\n","72/72 [==============================] - 0s 5ms/step - loss: 113.3866 - mae: 113.3866 - mse: 71213.4531 - val_loss: 783.9146 - val_mae: 783.9146 - val_mse: 1643760.2500\n","Epoch 100/100\n","72/72 [==============================] - 0s 5ms/step - loss: 111.7523 - mae: 111.7523 - mse: 70866.0859 - val_loss: 769.8317 - val_mae: 769.8317 - val_mse: 1646260.5000\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"oOcUDI9PvEIC","executionInfo":{"status":"ok","timestamp":1630053402803,"user_tz":-420,"elapsed":616,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"a167b400-66eb-4b37-d871-ac0a8be5b483"},"source":["model_11.evaluate(X_test, y_test)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["18/18 [==============================] - 0s 4ms/step - loss: 769.8317 - mae: 769.8317 - mse: 1646260.5000\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["[769.8317260742188, 769.8317260742188, 1646260.5]"]},"metadata":{},"execution_count":263}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"tWGEvWgwwM6N","executionInfo":{"status":"ok","timestamp":1630053501698,"user_tz":-420,"elapsed":648,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"212bade5-16e9-4c50-e678-da138a5e5b5d"},"source":["model_11_pred = model_11.predict(X_test)\n","model_11_pred.shape"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(575, 1, 1)"]},"metadata":{},"execution_count":266}]},{"cell_type":"markdown","metadata":{"id":"YVcBRmtGzCr-"},"source":["###### RNN"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"3uSxi-Jiz8W6","executionInfo":{"status":"ok","timestamp":1630054716744,"user_tz":-420,"elapsed":83786,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"dec3afce-397b-4d8e-bfc6-213dab5a197a"},"source":["import tensorflow as tf\n","from tensorflow.keras import layers\n","\n","model_12 = tf.keras.Sequential([\n","  layers.Lambda(lambda x: tf.expand_dims(x, axis=1)),\n","  layers.LSTM(128, activation=\"relu\"),\n","  layers.Dense(HORIZON)\n","], name=\"model_12_timeseries_conv1d\")\n","\n","# Compile model\n","model_12.compile(\n","    loss=\"mae\",\n","    optimizer=tf.keras.optimizers.Adam(),\n","    metrics=[\"mae\",\"mse\"]\n",")\n","\n","# Fit the model\n","model_12_history = model_12.fit(X_train,\n","                                y_train,\n","                                validation_data=(X_test, y_test),\n","                                epochs=100,\n","                                callbacks=[create_checkpoint_callback(model_12.name)])"],"execution_count":null,"outputs":[{"output_type":"stream","text":["WARNING:tensorflow:Layer lstm_17 will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU.\n","Epoch 1/100\n","72/72 [==============================] - 2s 13ms/step - loss: 1026.2882 - mae: 1026.2882 - mse: 4610462.5000 - val_loss: 1442.9316 - val_mae: 1442.9316 - val_mse: 5740814.0000\n","Epoch 2/100\n","72/72 [==============================] - 1s 9ms/step - loss: 203.9400 - mae: 203.9400 - mse: 209989.7188 - val_loss: 1408.9698 - val_mae: 1408.9698 - val_mse: 5072966.5000\n","Epoch 3/100\n","72/72 [==============================] - 1s 9ms/step - loss: 198.2693 - mae: 198.2693 - mse: 193740.7969 - val_loss: 1450.5195 - val_mae: 1450.5195 - val_mse: 5067226.0000\n","Epoch 4/100\n","72/72 [==============================] - 1s 9ms/step - loss: 186.2885 - mae: 186.2885 - mse: 179248.4531 - val_loss: 1305.2961 - val_mae: 1305.2961 - val_mse: 4420998.5000\n","Epoch 5/100\n","72/72 [==============================] - 1s 9ms/step - loss: 186.1210 - mae: 186.1210 - mse: 175591.8750 - val_loss: 1288.1019 - val_mae: 1288.1019 - val_mse: 4222062.5000\n","Epoch 6/100\n","72/72 [==============================] - 1s 9ms/step - loss: 185.6789 - mae: 185.6789 - mse: 168282.8594 - val_loss: 1221.6522 - val_mae: 1221.6522 - val_mse: 3922099.0000\n","Epoch 7/100\n","72/72 [==============================] - 1s 9ms/step - loss: 173.0833 - mae: 173.0833 - mse: 154303.8281 - val_loss: 1178.2135 - val_mae: 1178.2135 - val_mse: 3733199.2500\n","Epoch 8/100\n","72/72 [==============================] - 1s 9ms/step - loss: 167.1607 - mae: 167.1607 - mse: 143210.9219 - val_loss: 1325.8922 - val_mae: 1325.8922 - val_mse: 4028437.2500\n","Epoch 9/100\n","72/72 [==============================] - 1s 9ms/step - loss: 163.4415 - mae: 163.4415 - mse: 139321.1875 - val_loss: 1107.9547 - val_mae: 1107.9547 - val_mse: 3295406.5000\n","Epoch 10/100\n","72/72 [==============================] - 1s 9ms/step - loss: 159.0353 - mae: 159.0353 - mse: 129658.7656 - val_loss: 1276.1880 - val_mae: 1276.1880 - val_mse: 3686374.0000\n","Epoch 11/100\n","72/72 [==============================] - 1s 9ms/step - loss: 154.0426 - mae: 154.0426 - mse: 123559.0547 - val_loss: 1066.4669 - val_mae: 1066.4669 - val_mse: 2905584.7500\n","Epoch 12/100\n","72/72 [==============================] - 1s 9ms/step - loss: 148.8539 - mae: 148.8539 - mse: 116747.7578 - val_loss: 1021.6285 - val_mae: 1021.6285 - val_mse: 2714909.5000\n","Epoch 13/100\n","72/72 [==============================] - 1s 9ms/step - loss: 152.6696 - mae: 152.6696 - mse: 117550.0547 - val_loss: 1052.3127 - val_mae: 1052.3127 - val_mse: 2717483.5000\n","Epoch 14/100\n","72/72 [==============================] - 1s 9ms/step - loss: 146.2918 - mae: 146.2918 - mse: 110175.7812 - val_loss: 1051.0876 - val_mae: 1051.0876 - val_mse: 2675293.0000\n","Epoch 15/100\n","72/72 [==============================] - 1s 9ms/step - loss: 141.3336 - mae: 141.3336 - mse: 103498.0391 - val_loss: 975.5906 - val_mae: 975.5906 - val_mse: 2523081.0000\n","Epoch 16/100\n","72/72 [==============================] - 1s 9ms/step - loss: 140.7292 - mae: 140.7292 - mse: 101705.6719 - val_loss: 944.9571 - val_mae: 944.9571 - val_mse: 2315818.0000\n","Epoch 17/100\n","72/72 [==============================] - 1s 9ms/step - loss: 139.0196 - mae: 139.0196 - mse: 100177.4922 - val_loss: 931.3334 - val_mae: 931.3334 - val_mse: 2233606.0000\n","Epoch 18/100\n","72/72 [==============================] - 1s 9ms/step - loss: 133.8853 - mae: 133.8853 - mse: 95833.6953 - val_loss: 916.6119 - val_mae: 916.6119 - val_mse: 2165392.7500\n","Epoch 19/100\n","72/72 [==============================] - 1s 9ms/step - loss: 132.4507 - mae: 132.4507 - mse: 95041.6641 - val_loss: 1043.2096 - val_mae: 1043.2096 - val_mse: 2835044.5000\n","Epoch 20/100\n","72/72 [==============================] - 1s 9ms/step - loss: 139.1185 - mae: 139.1185 - mse: 100393.9531 - val_loss: 976.1929 - val_mae: 976.1929 - val_mse: 2323734.0000\n","Epoch 21/100\n","72/72 [==============================] - 1s 10ms/step - loss: 132.4190 - mae: 132.4190 - mse: 92936.5078 - val_loss: 894.4122 - val_mae: 894.4122 - val_mse: 2139379.5000\n","Epoch 22/100\n","72/72 [==============================] - 1s 9ms/step - loss: 126.4129 - mae: 126.4129 - mse: 88568.4453 - val_loss: 891.9963 - val_mae: 891.9963 - val_mse: 2032442.3750\n","Epoch 23/100\n","72/72 [==============================] - 1s 9ms/step - loss: 128.8799 - mae: 128.8799 - mse: 88091.3438 - val_loss: 911.3646 - val_mae: 911.3646 - val_mse: 2082301.3750\n","Epoch 24/100\n","72/72 [==============================] - 1s 9ms/step - loss: 127.0017 - mae: 127.0017 - mse: 87041.0469 - val_loss: 955.7239 - val_mae: 955.7239 - val_mse: 2222838.7500\n","Epoch 25/100\n","72/72 [==============================] - 1s 9ms/step - loss: 128.8149 - mae: 128.8149 - mse: 85559.1797 - val_loss: 845.9072 - val_mae: 845.9072 - val_mse: 1893119.3750\n","Epoch 26/100\n","72/72 [==============================] - 1s 9ms/step - loss: 126.8640 - mae: 126.8640 - mse: 86135.3203 - val_loss: 843.7654 - val_mae: 843.7654 - val_mse: 1879010.1250\n","Epoch 27/100\n","72/72 [==============================] - 1s 9ms/step - loss: 127.3452 - mae: 127.3452 - mse: 83642.4453 - val_loss: 838.7847 - val_mae: 838.7847 - val_mse: 1861041.7500\n","Epoch 28/100\n","72/72 [==============================] - 1s 9ms/step - loss: 124.6696 - mae: 124.6696 - mse: 85366.9844 - val_loss: 842.9875 - val_mae: 842.9875 - val_mse: 1864201.5000\n","Epoch 29/100\n","72/72 [==============================] - 1s 9ms/step - loss: 121.9210 - mae: 121.9210 - mse: 80229.2344 - val_loss: 949.1964 - val_mae: 949.1964 - val_mse: 2184931.0000\n","Epoch 30/100\n","72/72 [==============================] - 1s 9ms/step - loss: 120.1651 - mae: 120.1651 - mse: 80505.6016 - val_loss: 821.2162 - val_mae: 821.2162 - val_mse: 1802204.5000\n","Epoch 31/100\n","72/72 [==============================] - 1s 9ms/step - loss: 137.2688 - mae: 137.2688 - mse: 90556.9141 - val_loss: 1295.9017 - val_mae: 1295.9017 - val_mse: 3874739.7500\n","Epoch 32/100\n","72/72 [==============================] - 1s 9ms/step - loss: 123.0780 - mae: 123.0780 - mse: 84375.7812 - val_loss: 830.9355 - val_mae: 830.9355 - val_mse: 1872380.8750\n","Epoch 33/100\n","72/72 [==============================] - 1s 9ms/step - loss: 120.6266 - mae: 120.6266 - mse: 80229.1172 - val_loss: 805.9028 - val_mae: 805.9028 - val_mse: 1768335.0000\n","Epoch 34/100\n","72/72 [==============================] - 1s 9ms/step - loss: 121.1560 - mae: 121.1560 - mse: 80020.4922 - val_loss: 837.6220 - val_mae: 837.6220 - val_mse: 1831078.0000\n","Epoch 35/100\n","72/72 [==============================] - 1s 9ms/step - loss: 119.2379 - mae: 119.2379 - mse: 79432.2109 - val_loss: 821.3906 - val_mae: 821.3906 - val_mse: 1782506.6250\n","Epoch 36/100\n","72/72 [==============================] - 1s 9ms/step - loss: 120.7990 - mae: 120.7990 - mse: 78293.5234 - val_loss: 828.7652 - val_mae: 828.7652 - val_mse: 1854590.5000\n","Epoch 37/100\n","72/72 [==============================] - 1s 9ms/step - loss: 119.3988 - mae: 119.3988 - mse: 78545.9219 - val_loss: 797.7250 - val_mae: 797.7250 - val_mse: 1745116.0000\n","Epoch 38/100\n","72/72 [==============================] - 1s 9ms/step - loss: 119.2137 - mae: 119.2137 - mse: 78010.1719 - val_loss: 850.3937 - val_mae: 850.3937 - val_mse: 1855774.5000\n","Epoch 39/100\n","72/72 [==============================] - 1s 9ms/step - loss: 121.2411 - mae: 121.2411 - mse: 78577.2812 - val_loss: 788.1565 - val_mae: 788.1565 - val_mse: 1709919.7500\n","Epoch 40/100\n","72/72 [==============================] - 1s 9ms/step - loss: 121.3022 - mae: 121.3022 - mse: 79176.6875 - val_loss: 879.2565 - val_mae: 879.2565 - val_mse: 2013734.7500\n","Epoch 41/100\n","72/72 [==============================] - 1s 10ms/step - loss: 126.3830 - mae: 126.3830 - mse: 84399.5625 - val_loss: 799.7679 - val_mae: 799.7679 - val_mse: 1711514.7500\n","Epoch 42/100\n","72/72 [==============================] - 1s 9ms/step - loss: 119.5768 - mae: 119.5768 - mse: 78113.2422 - val_loss: 789.8646 - val_mae: 789.8646 - val_mse: 1715563.1250\n","Epoch 43/100\n","72/72 [==============================] - 1s 9ms/step - loss: 119.0731 - mae: 119.0731 - mse: 76218.7344 - val_loss: 786.1385 - val_mae: 786.1385 - val_mse: 1684216.7500\n","Epoch 44/100\n","72/72 [==============================] - 1s 9ms/step - loss: 118.4691 - mae: 118.4691 - mse: 77465.2656 - val_loss: 792.8167 - val_mae: 792.8167 - val_mse: 1690651.8750\n","Epoch 45/100\n","72/72 [==============================] - 1s 9ms/step - loss: 119.2971 - mae: 119.2971 - mse: 77699.0312 - val_loss: 850.0552 - val_mae: 850.0552 - val_mse: 1845050.3750\n","Epoch 46/100\n","72/72 [==============================] - 1s 10ms/step - loss: 124.9411 - mae: 124.9411 - mse: 81461.9609 - val_loss: 780.5667 - val_mae: 780.5667 - val_mse: 1684371.5000\n","Epoch 47/100\n","72/72 [==============================] - 1s 9ms/step - loss: 116.6305 - mae: 116.6305 - mse: 75804.9609 - val_loss: 882.8779 - val_mae: 882.8779 - val_mse: 1936964.1250\n","Epoch 48/100\n","72/72 [==============================] - 1s 10ms/step - loss: 117.6456 - mae: 117.6456 - mse: 76583.8516 - val_loss: 812.7902 - val_mae: 812.7902 - val_mse: 1775466.6250\n","Epoch 49/100\n","72/72 [==============================] - 1s 9ms/step - loss: 114.8765 - mae: 114.8765 - mse: 73800.6953 - val_loss: 871.2551 - val_mae: 871.2551 - val_mse: 1898514.5000\n","Epoch 50/100\n","72/72 [==============================] - 1s 9ms/step - loss: 119.8777 - mae: 119.8777 - mse: 75219.2891 - val_loss: 886.7770 - val_mae: 886.7770 - val_mse: 1944125.2500\n","Epoch 51/100\n","72/72 [==============================] - 1s 9ms/step - loss: 116.6218 - mae: 116.6218 - mse: 75218.4531 - val_loss: 764.5287 - val_mae: 764.5287 - val_mse: 1629573.3750\n","Epoch 52/100\n","72/72 [==============================] - 1s 10ms/step - loss: 113.4520 - mae: 113.4520 - mse: 74331.8438 - val_loss: 763.9804 - val_mae: 763.9804 - val_mse: 1629496.3750\n","Epoch 53/100\n","72/72 [==============================] - 1s 9ms/step - loss: 115.4158 - mae: 115.4158 - mse: 75608.6562 - val_loss: 783.4832 - val_mae: 783.4832 - val_mse: 1681163.0000\n","Epoch 54/100\n","72/72 [==============================] - 1s 9ms/step - loss: 115.7159 - mae: 115.7159 - mse: 74921.2109 - val_loss: 767.5028 - val_mae: 767.5028 - val_mse: 1622101.8750\n","Epoch 55/100\n","72/72 [==============================] - 1s 12ms/step - loss: 112.3216 - mae: 112.3216 - mse: 73961.7422 - val_loss: 758.4400 - val_mae: 758.4400 - val_mse: 1607597.5000\n","Epoch 56/100\n","72/72 [==============================] - 1s 10ms/step - loss: 112.9053 - mae: 112.9053 - mse: 72482.1719 - val_loss: 792.2749 - val_mae: 792.2749 - val_mse: 1699887.1250\n","Epoch 57/100\n","72/72 [==============================] - 1s 9ms/step - loss: 112.5270 - mae: 112.5270 - mse: 73800.6250 - val_loss: 757.6525 - val_mae: 757.6525 - val_mse: 1601010.8750\n","Epoch 58/100\n","72/72 [==============================] - 1s 10ms/step - loss: 111.9071 - mae: 111.9071 - mse: 71857.8750 - val_loss: 827.7628 - val_mae: 827.7628 - val_mse: 1764711.2500\n","Epoch 59/100\n","72/72 [==============================] - 1s 9ms/step - loss: 119.0281 - mae: 119.0281 - mse: 77151.7266 - val_loss: 940.4958 - val_mae: 940.4958 - val_mse: 2107188.2500\n","Epoch 60/100\n","72/72 [==============================] - 1s 9ms/step - loss: 114.6931 - mae: 114.6931 - mse: 70874.4375 - val_loss: 989.9706 - val_mae: 989.9706 - val_mse: 2283315.7500\n","Epoch 61/100\n","72/72 [==============================] - 1s 9ms/step - loss: 119.1633 - mae: 119.1633 - mse: 78431.7109 - val_loss: 869.7864 - val_mae: 869.7864 - val_mse: 1880294.8750\n","Epoch 62/100\n","72/72 [==============================] - 1s 9ms/step - loss: 115.7275 - mae: 115.7275 - mse: 74803.5938 - val_loss: 798.4624 - val_mae: 798.4624 - val_mse: 1709747.3750\n","Epoch 63/100\n","72/72 [==============================] - 1s 10ms/step - loss: 115.1491 - mae: 115.1491 - mse: 75451.2812 - val_loss: 759.9907 - val_mae: 759.9907 - val_mse: 1610704.3750\n","Epoch 64/100\n","72/72 [==============================] - 1s 9ms/step - loss: 111.3562 - mae: 111.3562 - mse: 72298.1641 - val_loss: 817.1317 - val_mae: 817.1317 - val_mse: 1734083.8750\n","Epoch 65/100\n","72/72 [==============================] - 1s 9ms/step - loss: 112.7971 - mae: 112.7971 - mse: 70709.9219 - val_loss: 747.9771 - val_mae: 747.9771 - val_mse: 1581699.5000\n","Epoch 66/100\n","72/72 [==============================] - 1s 9ms/step - loss: 110.9498 - mae: 110.9498 - mse: 71512.7266 - val_loss: 749.2474 - val_mae: 749.2474 - val_mse: 1577378.1250\n","Epoch 67/100\n","72/72 [==============================] - 1s 10ms/step - loss: 112.8183 - mae: 112.8183 - mse: 73098.1328 - val_loss: 763.2220 - val_mae: 763.2220 - val_mse: 1605120.2500\n","Epoch 68/100\n","72/72 [==============================] - 1s 10ms/step - loss: 111.3743 - mae: 111.3743 - mse: 71613.8906 - val_loss: 745.1490 - val_mae: 745.1490 - val_mse: 1572711.8750\n","Epoch 69/100\n","72/72 [==============================] - 1s 9ms/step - loss: 112.6847 - mae: 112.6847 - mse: 72056.0781 - val_loss: 798.0152 - val_mae: 798.0152 - val_mse: 1705375.2500\n","Epoch 70/100\n","72/72 [==============================] - 1s 9ms/step - loss: 114.5391 - mae: 114.5391 - mse: 74581.8359 - val_loss: 823.0783 - val_mae: 823.0783 - val_mse: 1746398.6250\n","Epoch 71/100\n","72/72 [==============================] - 1s 10ms/step - loss: 111.5236 - mae: 111.5236 - mse: 70531.2969 - val_loss: 744.7511 - val_mae: 744.7511 - val_mse: 1567117.3750\n","Epoch 72/100\n","72/72 [==============================] - 1s 9ms/step - loss: 112.4445 - mae: 112.4445 - mse: 72028.2500 - val_loss: 824.8604 - val_mae: 824.8604 - val_mse: 1750456.5000\n","Epoch 73/100\n","72/72 [==============================] - 1s 9ms/step - loss: 116.2553 - mae: 116.2553 - mse: 75998.2578 - val_loss: 743.9653 - val_mae: 743.9653 - val_mse: 1566012.6250\n","Epoch 74/100\n","72/72 [==============================] - 1s 10ms/step - loss: 111.3740 - mae: 111.3740 - mse: 70932.7969 - val_loss: 743.9230 - val_mae: 743.9230 - val_mse: 1564646.0000\n","Epoch 75/100\n","72/72 [==============================] - 1s 9ms/step - loss: 111.0935 - mae: 111.0935 - mse: 72642.6328 - val_loss: 742.1802 - val_mae: 742.1802 - val_mse: 1563207.6250\n","Epoch 76/100\n","72/72 [==============================] - 1s 10ms/step - loss: 111.5162 - mae: 111.5162 - mse: 69680.0234 - val_loss: 818.8529 - val_mae: 818.8529 - val_mse: 1726811.8750\n","Epoch 77/100\n","72/72 [==============================] - 1s 10ms/step - loss: 112.6039 - mae: 112.6039 - mse: 72636.5469 - val_loss: 763.5221 - val_mae: 763.5221 - val_mse: 1601767.0000\n","Epoch 78/100\n","72/72 [==============================] - 1s 10ms/step - loss: 112.1333 - mae: 112.1333 - mse: 71577.6875 - val_loss: 852.1160 - val_mae: 852.1160 - val_mse: 1870757.3750\n","Epoch 79/100\n","72/72 [==============================] - 1s 9ms/step - loss: 110.3991 - mae: 110.3991 - mse: 69943.6484 - val_loss: 739.9709 - val_mae: 739.9709 - val_mse: 1556680.5000\n","Epoch 80/100\n","72/72 [==============================] - 1s 10ms/step - loss: 109.6685 - mae: 109.6685 - mse: 69219.7188 - val_loss: 810.8018 - val_mae: 810.8018 - val_mse: 1710784.7500\n","Epoch 81/100\n","72/72 [==============================] - 1s 10ms/step - loss: 112.1841 - mae: 112.1841 - mse: 72500.7891 - val_loss: 800.1325 - val_mae: 800.1325 - val_mse: 1708698.1250\n","Epoch 82/100\n","72/72 [==============================] - 1s 9ms/step - loss: 112.6623 - mae: 112.6623 - mse: 71422.1016 - val_loss: 739.1899 - val_mae: 739.1899 - val_mse: 1551940.6250\n","Epoch 83/100\n","72/72 [==============================] - 1s 10ms/step - loss: 108.9917 - mae: 108.9917 - mse: 69235.8828 - val_loss: 738.1684 - val_mae: 738.1684 - val_mse: 1550255.2500\n","Epoch 84/100\n","72/72 [==============================] - 1s 10ms/step - loss: 112.0200 - mae: 112.0200 - mse: 71680.6328 - val_loss: 748.4071 - val_mae: 748.4071 - val_mse: 1579421.7500\n","Epoch 85/100\n","72/72 [==============================] - 1s 10ms/step - loss: 114.6637 - mae: 114.6637 - mse: 75452.4141 - val_loss: 741.7859 - val_mae: 741.7859 - val_mse: 1554957.7500\n","Epoch 86/100\n","72/72 [==============================] - 1s 10ms/step - loss: 110.7016 - mae: 110.7016 - mse: 71804.0781 - val_loss: 739.3774 - val_mae: 739.3774 - val_mse: 1551757.0000\n","Epoch 87/100\n","72/72 [==============================] - 1s 10ms/step - loss: 114.6655 - mae: 114.6655 - mse: 72002.8281 - val_loss: 930.5051 - val_mae: 930.5051 - val_mse: 2141598.2500\n","Epoch 88/100\n","72/72 [==============================] - 1s 10ms/step - loss: 114.5451 - mae: 114.5451 - mse: 72607.8672 - val_loss: 796.0022 - val_mae: 796.0022 - val_mse: 1702027.2500\n","Epoch 89/100\n","72/72 [==============================] - 1s 10ms/step - loss: 109.2005 - mae: 109.2005 - mse: 69102.8594 - val_loss: 734.2189 - val_mae: 734.2189 - val_mse: 1538250.8750\n","Epoch 90/100\n","72/72 [==============================] - 1s 10ms/step - loss: 111.7326 - mae: 111.7326 - mse: 69473.9844 - val_loss: 802.4585 - val_mae: 802.4585 - val_mse: 1685441.5000\n","Epoch 91/100\n","72/72 [==============================] - 1s 10ms/step - loss: 113.2562 - mae: 113.2562 - mse: 70083.6562 - val_loss: 786.1472 - val_mae: 786.1472 - val_mse: 1645693.1250\n","Epoch 92/100\n","72/72 [==============================] - 1s 10ms/step - loss: 109.3141 - mae: 109.3141 - mse: 69362.9609 - val_loss: 817.9207 - val_mae: 817.9207 - val_mse: 1714555.3750\n","Epoch 93/100\n","72/72 [==============================] - 1s 10ms/step - loss: 112.1319 - mae: 112.1319 - mse: 71989.5547 - val_loss: 740.1094 - val_mae: 740.1094 - val_mse: 1552140.8750\n","Epoch 94/100\n","72/72 [==============================] - 1s 9ms/step - loss: 112.4329 - mae: 112.4329 - mse: 73161.6094 - val_loss: 747.5308 - val_mae: 747.5308 - val_mse: 1574021.1250\n","Epoch 95/100\n","72/72 [==============================] - 1s 10ms/step - loss: 110.7324 - mae: 110.7324 - mse: 70775.0156 - val_loss: 751.9614 - val_mae: 751.9614 - val_mse: 1570574.5000\n","Epoch 96/100\n","72/72 [==============================] - 1s 9ms/step - loss: 108.9629 - mae: 108.9629 - mse: 69826.9531 - val_loss: 757.7120 - val_mae: 757.7120 - val_mse: 1579657.1250\n","Epoch 97/100\n","72/72 [==============================] - 1s 10ms/step - loss: 112.4630 - mae: 112.4630 - mse: 68430.2578 - val_loss: 772.7535 - val_mae: 772.7535 - val_mse: 1612696.8750\n","Epoch 98/100\n","72/72 [==============================] - 1s 10ms/step - loss: 113.9022 - mae: 113.9022 - mse: 72309.2344 - val_loss: 820.2136 - val_mae: 820.2136 - val_mse: 1783134.0000\n","Epoch 99/100\n","72/72 [==============================] - 1s 10ms/step - loss: 115.1185 - mae: 115.1185 - mse: 71977.6484 - val_loss: 809.0808 - val_mae: 809.0808 - val_mse: 1745587.6250\n","Epoch 100/100\n","72/72 [==============================] - 1s 10ms/step - loss: 109.7332 - mae: 109.7332 - mse: 69693.9375 - val_loss: 736.1164 - val_mae: 736.1164 - val_mse: 1541232.0000\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"IzIaDYev0NK3"},"source":["# Plot prediction"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"vV3xOSRD1bHh"},"source":["model_12_pred = model_11.predict(X_test)\n","model_12_pred = tf.squeeze(model_12_pred)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":486},"id":"kkoE8oi21jeX","executionInfo":{"status":"ok","timestamp":1630055325078,"user_tz":-420,"elapsed":618,"user":{"displayName":"Bao Loc Pham","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GiWbogwajI_dxHURuHYrlX7OCoYRQ4QSfMg2o2L2w=s64","userId":"14790014142134469910"}},"outputId":"395ca51f-6908-4090-ca81-c3cfdc979425"},"source":["plt.figure(figsize=(10,8))\n","\n","offset = 500\n","\n","plt.plot(dates[-len(X_test)+offset:], prices[-len(X_test)+offset:], \"-\", label=\"Test data\")\n","plt.plot(dates[-len(X_test)+offset:], model_12_pred[offset:], \".-\", label=\"Model predict\")\n","plt.legend()\n","plt.show();\n","# plt.plot(dates[-offset])\n","# plt.plot(dates[:])"],"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAmIAAAHVCAYAAABScZe2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeXyVZ53//9d1su8JSYBsEAq0hQAJSyld6CpQuzLqjK2tdtRa/VVHR622Hf22LqOOD8ep1rpMx62OdnFqa+lCW7pQoBslEPYCYc0GZD1JyH7O9fvjvpOc7AvZ834+HjzOOdd93fe5Yn2QD9f1uT6XsdYiIiIiIiPPM9oDEBEREZmsFIiJiIiIjBIFYiIiIiKjRIGYiIiIyChRICYiIiIyShSIiYiIiIySfgVixphjxpjdxpg8Y8w2t22KMWaDMeaQ+5rgthtjzEPGmHxjzC5jzJKA59zu9j9kjLk9oH2p+/x8914z1D+oiIiIyFgzkBmxK621OdbaZe7ne4HXrLVzgdfczwAfBua6f+4Efg1O4AY8AFwILAceaA3e3D6fC7jvmkH/RCIiIiLjxNksTd4EPOq+fxRYG9D+J+t4F4g3xqQAa4AN1toKa20lsAG4xr0Wa6191zrVZf8U8CwRERGRCSu4n/0s8IoxxgL/ba19BJhmrS1xr58Eprnv04CCgHsL3bbe2gu7ae9VUlKSzczM7OfwRUREREZPbm5umbU2uXN7fwOxS621RcaYqcAGY8wHgRettdYN0oaVMeZOnOVOZsyYwbZt24b7K0VERETOmjHmeHft/VqatNYWua+ngWdwcrxOucuKuK+n3e5FQEbA7eluW2/t6d20dzeOR6y1y6y1y5KTuwSVIiIiIuNKn4GYMSbKGBPT+h5YDewB1gGtOx9vB551368DPuXunlwBeN0lzJeB1caYBDdJfzXwsnut2hizwt0t+amAZ4mIiIhMWP1ZmpwGPONWlAgGHrPWvmSMeR/4qzHms8Bx4J/c/i8C1wL5QB3waQBrbYUx5vvA+26/71lrK9z3dwF/BCKA9e4fERERkQnNOBsVx59ly5bZzjlizc3NFBYW0tDQMEqjkp6Eh4eTnp5OSEjIaA9FRERkxBljcgNKgLXpb7L+uFBYWEhMTAyZmZmoJuzYYa2lvLycwsJCZs2aNdrDERERGTMm1BFHDQ0NJCYmKggbY4wxJCYmaqZSRESkkwkViAEKwsYo/XcRERHpasIFYqOpvLycnJwccnJymD59OmlpaW2fm5qa+rx/48aNvP322/36rszMTMrKynrt88Mf/rBfzxIREZHRoUBsCCUmJpKXl0deXh5f+MIX+OpXv9r2OTQ0tM/7BxKI9YcCMRERkbFNgdgwy83N5fLLL2fp0qWsWbOGkhLnVKiHHnqI+fPns2jRIm6++WaOHTvGb37zGx588EFycnLYvHlzh+eUl5ezevVqsrKyuOOOOwjc7bp27VqWLl1KVlYWjzzyCAD33nsv9fX15OTkcOutt/bYT0REREbPhCpfsX//fubNmwfAd5/by77i6iH9zvmpsTxwQ1a/+n7nO98hKiqKZ555hmeffZbk5GSefPJJXn75ZX7/+9+TmprK0aNHCQsLo6qqivj4eL7zne8QHR3N3Xff3eV5X/7yl0lKSuL+++/nhRde4Prrr6e0tJSkpCQqKiqYMmUK9fX1XHDBBbz55pskJiYSHR1NbW1t2zN66jdSAv/7iIiITCaTonzFWNPY2MiePXtYtWoVAD6fj5SUFAAWLVrErbfeytq1a1m7dm2fz9q0aRNPP/00ANdddx0JCQlt1x566CGeeeYZAAoKCjh06FC3AVZ/+4mIiMjImLCBWH9nroaTtZasrCzeeeedLtdeeOEFNm3axHPPPccPfvADdu/ePajv2LhxI6+++irvvPMOkZGRXHHFFd2WiehvPxERERk5yhEbRmFhYZSWlrYFYs3Nzezduxe/309BQQFXXnklP/7xj/F6vdTW1hITE0NNTU23z7rssst47LHHAFi/fj2VlZUAeL1eEhISiIyM5IMPPuDdd99tuyckJITm5uY++4mIiMjoUCA2jDweD0899RT33HMP2dnZ5OTk8Pbbb+Pz+bjttttYuHAhixcv5stf/jLx8fHccMMNPPPMM90m6z/wwANs2rSJrKwsnn76aWbMmAHANddcQ0tLC/PmzePee+9lxYoVbffceeedbUugvfUTERGR0TFhk/Vl7NF/HxERmax6StbXjJiIiIhMSmcaW6htbBnVMSgQExERkUnp2bxiFjzwMsVV9aM2BgViIiIiMikVVtYR7DFMiw0ftTEoEBMREZFJqaCyntT4CII8ZtTGoEBMREREJqXCyjrSEyJGdQwKxERERGRSKqysJyMhclTHoEBsiBljuO2229o+t7S0kJyczPXXXz+g52RmZlJWVnbWfc7WsWPHWLBgAQDbtm3jy1/+cq/9f/jDHw7reERERIZCQ7OP0ppGzYhNNFFRUezZs4f6emcHxoYNG0hLSxvlUXXV0jLw7brLli3joYce6rWPAjERERkPCiud39MZUzQjNroKtsLmnzqvQ+Taa6/lhRdeAODxxx/nlltuabtWUVHB2rVrWbRoEStWrGDXrl0AlJeXs3r1arKysrjjjjsILLT75z//meXLl5OTk8PnP/95fD5fr98fHR3NV7/6VbKysrj66qspLS0F4IorruBf//VfWbZsGT//+c/Jzc3l8ssvZ+nSpaxZs4aSkhIAcnNzyc7OJjs7m1/+8pdtz924cWPbzF5tbS2f/vSnWbhwIYsWLeJvf/sb9957L/X19eTk5HDrrbcOwf+SIiIiw6Ogsg5g1GfEJuyh36y/F072cZB2YzWc2gPWD8YD0xZAWGzP/acvhA//R59fffPNN/O9732P66+/nl27dvGZz3ym7ciiBx54gMWLF/P3v/+d119/nU996lPk5eXx3e9+l0svvZT777+fF154gd/97neAU43+ySef5K233iIkJIS77rqLv/zlL3zqU5/q8fvPnDnDsmXLePDBB/ne977Hd7/7XR5++GEAmpqa2LZtG83NzVx++eU8++yzJCcn8+STT/Ktb32L3//+93z605/m4Ycf5rLLLuMb3/hGt9/x/e9/n7i4uLbDyisrK/noRz/Kww8/TF5eXp//G4mIiIym1hmx9FHOEZu4gVh/NHidIAyc1wZv74FYPy1atIhjx47x+OOPc+2113a4tmXLFv72t78BcNVVV1FeXk51dTWbNm3i6aefBuC6664jISEBgNdee43c3FwuuOACAOrr65k6dWqv3+/xePj4xz8OwG233cZHPvKRtmut7QcOHGDPnj2sWrUKAJ/PR0pKClVVVVRVVXHZZZcB8MlPfpL169d3+Y5XX32VJ554ou1z63hFRETGg8LKOkKDPEyNCRvVcUzcQKwfM1cUbIVHbwRfEwSFwkd/CxnLh+Trb7zxRu6++242btxIeXn5oJ9jreX222/nRz/60aCfYUx7fZSoqKi252ZlZfHOO+906FtVVTXo7xERERkvCivqSUuIwDOKNcRgsueIZSyH29fBVd9yXocoCAP4zGc+wwMPPMDChQs7tK9cuZK//OUvgJNzlZSURGxsLJdddhmPPfYYAOvXr6eyshKAq6++mqeeeorTp08DTo7Z8ePHe/1uv9/PU089BcBjjz3GpZde2qXPeeedR2lpaVsg1tzczN69e4mPjyc+Pp4tW7YAtI21s1WrVnXIH2sdb0hICM3Nzb2OT0REZLSNhRpiMNkDMXCCr5VfH9IgDCA9Pb3bUg/f+c53yM3NZdGiRdx77708+uijgJM7tmnTJrKysnj66aeZMWMGAPPnz+ff//3fWb16NYsWLWLVqlVtSfU9iYqKYuvWrSxYsIDXX3+d+++/v0uf0NBQnnrqKe655x6ys7PJycnh7bffBuAPf/gDX/ziF8nJyemwaSDQt7/9bSorK1mwYAHZ2dm88cYbANx5550sWrRIyfoiIjKmFVbWj3p+GIDp6RftWLds2TK7bdu2Dm379+9n3rx5ozSisSM6Opra2trRHkYX+u8jIiJjwZnGFrIeeJlvrDmPL145Z0S+0xiTa61d1rldM2IiIiIyqRRVjY0aYqBAbEIai7NhIiIiY0VBxdioIQYKxERERGSSaa8hpkBsyI3XnLeJTv9dRERkrCisrCMs2ENy9OjWEIMJFoiFh4dTXl6uX/pjjLWW8vJywsPDR3soIiIiFFTUk54Q0aHO5miZUAVd09PTKSwsbDtbUcaO8PBw0tPTR3sYIiIiFFbVjYlEfZhggVhISAizZs0a7WGIiIjIGFZQUU9ORvxoDwOYYEuTIiIiIr2pbmjGW988Joq5ggIxERERmUSK3B2TGQrEREREREbWWKohBgrEREREZBJprSE2VpL1FYiJiIjIpFFQWUdkaBAJkSGjPRRggu2aFBERkTGiYCsc2wyZKyFjeXt7SxMceQOK82D2lR2vjYDCyrFTQwwUiImIiMhQK9gKf7wefI1gPBCXAb5maKiC5rr2fpt/Are/ADMuHLGhFVbWj5lEfdDSpIiIiAySt76Z53cVdz3R5thmJwgDsH4IiYA5V8Gyz8A5VwLubJSvGZ75PJTlj8h4rbUUVtSNmUR9GEAgZowJMsbsMMY8737ebIzJc/8UG2P+7rZfYYzxBly7P+AZ1xhjDhhj8o0x9wa0zzLGvOe2P2mMCR3KH1JERESG3rq8Ir702A427DvV8ULmSmcmDCA4Am78Bdz0S1jzA7jy3yA4HEwQeEKg9hT8+mLY+GNoaRzW8VbXt1DT2DJmEvVhYEuTXwH2A7EA1tqVrReMMX8Dng3ou9lae33gzcaYIOCXwCqgEHjfGLPOWrsP+DHwoLX2CWPMb4DPAr8exM8jIiIiI+RkdQMA/7XhIB+aNw2Px53pyljuLEeGuEFYYB5YxnK4fV17/lj8DHjpPtj4Q9j9f7D8Tmiq6ZpbNgQKKsdW6Qro54yYMSYduA74bTfXYoGrgL/38ZjlQL619oi1tgl4ArjJONlyVwFPuf0eBdb2b/giIiIyWkprnBmsD07W8Pzuko4X6yrgnCu6D6YylsPKrzuvMdPhH/8At/3NCcDWfwNe+z48eqOTazaECtsCsbEzI9bfpcmfAd8E/N1cWwu8Zq2tDmi7yBiz0xiz3hiT5balAQUBfQrdtkSgylrb0qm9C2PMncaYbcaYbTrYW0REZHSV1jQyPyWW86fH8LMNB2nxuWFCg9cJqmK7/XXevTkfgqWfcT9Y8DU5s2ZDqHCMVdWHfgRixpjrgdPW2tweutwCPB7weTsw01qbDfyCvmfK+s1a+4i1dpm1dllycvJQPVZEREQGobS2kWmxYXxt1bkcKTvD09uLnAveQuc1Ln1gD5x9JQSFOe+Nx1meHEIFFXXEhAUTGzF2ikb0Z0bsEuBGY8wxnOXEq4wxfwYwxiThLDm+0NrZWlttra11378IhLj9ioCMgOemu23lQLwxJrhTu4iIiIxhpTWNJMeEsWr+NLLT4/j5a4dobPGB1/01PtBALGM5/PPzEDcDIhMhbemQjrewsp60MVRDDPoRiFlr77PWpltrM4Gbgdettbe5lz8GPG+tbWjtb4yZ7uZ9YYxZ7n5HOfA+MNfdIRnqPmuddfa8vuE+C+B2Oib+i4iIyBjj91vKapuYGhOOMYavrz6Poqp6nny/ALxuJtJAAzFwgrE1P4Dak/DBC333H4DCyvoxtWMSzr6O2M10XJYEJ6DaY4zZCTwE3GwdLcCXgJdxdl/+1Vq7173nHuBrxph8nJyx353luERERGQYVdQ14fNbkmOcpcSVc5NYPmsKv3g9n+bKQvAEQ/S0wT38/Ouc3ZTvDl0BBWstBZVjq4YYDLCyvrV2I7Ax4PMV3fR5GHi4h/tfBF7spv0IzhKniIiIjAOtOyZbAzFjDHevPo9/+u93OHr4AOfGpIAnaHAP9wTBhV+Al/8NindA6uKzHm9lXTN1Tb4xlagPqqwvIiIig9A5EANYPmsKl52bTPWpo7TE9LxjssXnp9nXXSGGAIs/CaExQzYrVlAx9mqIgQIxERERGYS2QCw6rEP73avPJdlfRn5jXJd7iqrq+cnLH7DiR6/xif95t/cvCI+FxbfBnr9BdUnvffuhtXTFWKohBgrEREREZBBKa7vOiAEsSo0lzVPBW6fDqTzThN9v2XyolM/9aRsrf/w6v954mLiIEN4/Vkn+6drev+TCz4PfB+93qSc/YG3FXKeMrRmxsVNIQ0RERMaN0ppGokKDiArrFEqcKSWYFo63JPDlJ3ZQWFnP0bIzTIkK5fOXz+bWC2cQGuThwh+9xrqdxXxt1bk9f8mUWU7i/rbfw2V3O0cmDVJBZR1xESHEhocM+hnDQTNiIiIiMmCtNcS6cIu5zph1LpsPlZEQGcKDH8/mnfuu4p5rzic9IZKpseFcPDuRdXlFOFWserHiLqivgJ1PnNV4ndIVY2s2DBSIiYiIyCCcrmnoPhCrdgKxT665hDfuvoKn77qEf1icTlhwxx2UN2ancqy8jt1F3t6/aObFMH2Rk7TfV9DWi4KKOtLjx1Z+GCgQExERkUHoeUbMqaofljiDWUlRPd5/TVYKoUEens0r7v2LjIGLvghlB+Dwa4Maq7WWwsr6MbdjEhSIiYiIyCCU1jR22TEJOEuTIZEQkdDr/XGRIVxxXjLP7SzG5+9jpivrI05x2Dd+BJt/CgVbBzTWstomGlv8Y66qPigQExERkQFqaPZR3dDS89JkbJozk9WHG3NSOV3TyHtHy3vvGBwK534YirbB6/8Oj944oGCsoHJs1hADBWIiIiIyQGU9lK4AnBmxuJ6LuQa6+vxpRIUGsa6v5UmAqCTn1frB1wTHNvd3uG01xDQjJiIiIuNeazHXqTHhXS96i/p92HdEaBBrsqbz4u4SGlt8vXc+dw0YN2wJCoXMlf0eb2tV/bR4zYiJiIjIOHe6m+ONAGhpgtpTENu/QAzghpxUqhta2HSwrPeOGcsh+xbAwC2POZ/7qbCynilRoV1rnhVsHVTO2VBSICYiIjJZDFHg0d05kwDUFAO23zNiAJfOSWJKVCjP5hX13XnRPznP97X0f7A4VfUzOueHFWyFR2+A1/7deR2lYEyBmIiIyGRw4j34/Rp47Xvw6PVnFXiU1jRiDEyJCu14wS1d0d8cMYCQIA/XLUzh1f2nONPYR4CVvhw8IXB8y4DG65Su6JQfdmwztDQCfvA1DyjnbCgpEBMREZkMPnjBSXQHJwB54etQfnhQjyqtbWRKZCghQZ3CCLeqPnEZA3rejTmpNDT72bDvVO8dQyMhbSkc638g5vdbirqrIZa5sn1n5wBzzoaSAjEREZHJIDzGeTUe8ARD6Qfw8AXw9y9CxdGely27ae+xmKtbVZ/Y/s+IASydkUBafET/liczL4HiPGis6dezS2sbafL5Se+8YzJtKQSHQ+oSuH3dgHLOhpIO/RYREZkMGrxOAHb5PXDOFRA/E976Gbz/O9j5uNPH+sET5CTFx0yHyuOw92nw+yA4DG5/DjKW937OZESCM3M1AB6P4YbsVH67+QgVZ5q6LnkGyrzUDQzfgzkf6vPZJyp6qCF2chc01zlV+0cpCAPNiImIiEwORTsgJRsu/6YTeMRMg2t+BF/Z6cwOWR9gwd8CO/7XCXb2/M35jHV2RLp5VL0ebzSARP1AN2an0uK3vLi7pPeOGRc6AWU/lyf3FVcDcN60mI4XWu+feclAhzqkFIiJiIhMdH4fFO9wAq7OYlNgzQ+cZToT5Lx++mV4oBI+85LzGcDjgcyVWGt7WZosGlDpikDzUmKYOzW67+KuoVGQuhiOvdWv5+YVVJEcE0ZKXKeaZ8e2wJTZzs8/ihSIiYiITHSlB6D5TPeBGDgzZLc/B1d9y3mduaJje2wqTM2CjOVU17fQ5PP3cM5kwaBnxIwx3JSTytZjFRRX1ffeOfNSKN4OTWf6fO7OgipyMuIxgUcu+X1w/B3nOaNMgZiIiMhEV7zdeU1d0nOfjOWw8utd86UylsP510PFYfD7KK1tALqpIdZY4+ShDaB0RWc3ZKcC8NzOPmbFZl7qLJn2UYLDW9fMkbIz5GTEd7xwcjc0ehWIiYiIyAgoyoWwWEicM7j705ZCUy2UHey5qn5rDbFBLk0CzEyMYmFaHK/u76OMxYwLnWXUPvLEdhZWAXQNxMZIfhgoEBMREZn4inKdvCrPIH/tty5pFuUGnDPZKRBrLV0xyKXJVpfOTWLHiarei7uGxTgbD473nieWV1CFMbAwPa7jheNvwZRzzmr2bqgoEBMREZnImhvg1N6e88P6Y8psCIvrEIgldz7wexBV9btzyewkWvyWrccqeu+YeakTYDbV9dhlZ0EVs5OjiQ0PaW/0+5xAbAzMhoECMRERkQnl/7YVsD6wBMTJ3U4+VVov+WF98XggbXFbIBYa7CE2vFMpUm+hUyw25ux2IS6dmUBokId3Dpf33jHzUvA1QeH73V621pJXUEV2eqdlyVN7nFy2Uaqk35kCMRERkQnCWsuP1n/Avzy+g3ePuIFMUa7zejYzYq33n9pLlbea5OiwjrsQwSldET0dgkK6v7+fIkKDWDIznrfyy3rvOGOFE/j1sDxZWFlP+ZkmcmZ0zg9z+2dqRkxERESGULG3gYozTVjgrr9sp6CizgnEYlKcEhRnI3UJ+FuIrNzXQzHXgiHLubpkdhL7SqqpPNPUc6fwOJi+qMd6YnkFbqJ+5xmxY1sgIfOsc9mGigIxERGRCWJ3oReAn3xsEc0+P5/70zb8RblnPxsGbc+YWrN3yKvqd3bxnCSshXeO9GN5svB9Jw+uk50FVYQFezg/JaCivt/vzKCNgbIVrRSIiYiITBB7i70EeQzXLkzh4U8s4eSpEjwVh7Gpi3u9r8Rbz49f+oBT1V0DmjaxKRCTSmbD/q6BmLVuVf2hmRFblB5HVGgQbx/uY3ly5iXga4SibV0u5RVUsSAtjpCggFDn9F5oqBoz+WGgQExERGTC2F3kZe7UaMJDgrj83GR+cKFTAuLZ09O77e/3W/733eOs+q9N/HrjYZ7KLez1+f7UJZzvO9S1qn5dObQ0QFzGkPwcIUEeLjwnkbfz+5gRm3kRYLosTzb7/Owp9nZN1B9D9cNaKRATERGZAKy17CnysiCtvWbWtVOc3ZP3bwvllb0nO/TPP13Lxx95h//39z1kZ8SREhfetrTZkzPJ2czynCI9vNPMmbe1htjQ1eW6eHYiR8rOUOLt5bijiASYvgCOdyzseuBkDQ3N/m4S9bdA/EyIH5qAcSgoEBMREZkATlY3UFbbxMKAQMwUbcefOJfM9FS++mQeB0/V0NTi5xevHeLan2/m4KlafvKxRfz5sxeydGYCe4p7D8RKYxcAcE7TwY4XWgOxIVqaBLh4dhIAb/U5K3apc9RRS2NbU1tF/cAZsbb8sLGzLAkKxERERCaE1tmsthkxa6FoG560JTzyyWVEhgVzx6PbuPHhLfx0w0FWZU1jw9cu4x+XZWCMYUFaHIWV9b3uVCwIPxe/NaSc2dfxQnVrMdehm2k6f3oMU6JC+84Ty7zEWRYt2t7WlHeiiilRoWRMiWjvd3of1FeOmbIVrRSIiYiITAB7irx4DMxPiXUaqouh9hSkLWV6XDi/uW0pJ70NVNU18z+fWsYvP7GEqQHV8Vtn0vYWV/f4HScbQzlsU0mo3N3xgrcAgsIgKmnIfh6Px3DRbCdPzFrbc8fWfK+A5cmdhVVkp8d1rHU2BvPDQIGYiIjIhOAk6scQERrkNBS7M0Ru2YmlMxN47euX8+rXL2fV/Gld7s9KdQK43pYnS2sa2WlnE346z5lxa+UtcvLDOhd5PUuXzE7iZHUDR8rO9NwpcgpMzWpL2K9paObQ6VpyMhI69ju+BeJnQMLMIR3j2VIgJiIiMs5Za9ldVN0hUZ+iXPAEw7QFbU0ZUyKJDgvu5gkQHxlKekIEu4t6D8QOBM3FnDndnhcGQ1q6ItDFsxMBeLvP444ugeNvw5s/4WjeG1gL2RkB/1v4/U6gNnPs1A9rpUBMRERknDtV3UhZbSML0mLbG4tynSAsJLznGztZmBbH3t4CsdpGiqLmtz+/lbdwWCrVz0yMJC0+grf7Ou4oeppTT+yNHzLvldtYYg6SkxGQqF+6H+orxlQh11YKxERERMa51lmsth2Tfj8U5w24ov6CtDiOlddR3dDc7fXT1Y1Ux54HQaHtgZivBWpKhiUQM8Zw8exE3jlSjt/fS55YS2s5DT9B/kY+ErWb+MjQ9utt50sqEBMREZEh1pao7+Z5UZ4PjdUDDsRa88T2FnWfsF9a28iU2GiYvrB9l2JNCVj/sCxNAlw8J5Gqumb2lfS8iYC5qyE4HDAY4J9anoM3fgSNtc71Y5udHZ1jLD8MFIiJiIiMe3uKvMxOjiYy1M3/ap2tSlsyoOcsaNs52f3yZGlNo3O8UdpSKN4Bft+wlK4I1FpPrNcyFhnL4fbnqLnkPu5q+gpF066AN/8DfrEENtwP+a9C8vnDMr6zpUBMRERknNtd5O1QyJXi7RAaDUnnDug5SdFhToX9bvLEzjS2UNfkaw/Ems9A6YFhqaofaFpsOHOmRvdd2DVjOW+lfIr1/gupuu4R+OyrEJkEb/0cmuvgyEan8OsYo0BMRERkHDtd3cDpmsauOyZTF4MnaMDPW5AWx55uArHSGqdyfXJ0WPuSZ/H2Yamq39klsxPZerSCphZ/r/3yCryEBBnmpcRCxgWw4KO0hTrW7yxRjjH9DsSMMUHGmB3GmOfdz380xhw1xuS5f3LcdmOMecgYk2+M2WWMWRLwjNuNMYfcP7cHtC81xux273nImCEuRCIiIjJBtSXqp7uBWEsjnNw94GXJVgtS4zhSdoYzjS0d2ktr3UAsJgymzIawOCfgqy5y3ofHdve4IXHR7CTqm33kFVT12i+voJL5KbGEh7gB6KyVEBwGJsjZYDDGjjeCgc2IfQXY36ntG9baHPdPntv2YWCu++dO4NcAxpgpwAPAhcBy4AFjTGu1tV8Dnwu475pB/CwiIiKTzu4iLyawov6pPeBrghIZdycAACAASURBVNRBBmJpsVhLl+T41hmxqbFh4PFA2mInEPMWDtuyZKuLzknEY3rPE/P5LbsLvWQHlq3IWA63r4OrvuW8Ziwf1nEORr8CMWNMOnAd8Nt+dL8J+JN1vAvEG2NSgDXABmtthbW2EtgAXONei7XWvmudMwz+BKwdzA8jIiIy2ewp8nJOUhRRrYVaizpW1B+o1lyzzsuTp6udEhHJ0WHtzz+119mhOQylKwLFRYawIC2Ot3vJE8s/XcuZJl/H+mHgBF8rvz4mgzDo/4zYz4BvAp0XZ3/gLj8+aIxx/8uQBhQE9Cl023prL+ymvQtjzJ3GmG3GmG2lpaX9HLqIiMjE1SVR/+DLEBLlnDU5CFNjw0mOCeuSsF9a20iQx5DQWp8rbSn4W6Ds4LDmh7W6aHYiOwoqqWtq6fb6TnfZMrtzIDbG9RmIGWOuB05ba3M7XboPOB+4AJgC3DP0w+vIWvuItXaZtXZZcnLycH+diIjImHa6poFT1QGJ+vtfgPwNzo7GP9006F2CC1Jju9QSK61pJCk6FI/HTeMOnHEb5qVJcM6dbPbZHndP7iioIjY8mFmJUcM+lqHU/YFTHV0C3GiMuRYIB2KNMX+21t7mXm80xvwBuNv9XAQEFhNJd9uKgCs6tW9029O76S8iIiK9aA2WFqbFOUn667/RftHX5OwSHMSS3MK0ON48WEp9k6/tEPG2GmKtYqY7M2HVRcNWQyzQBZlTCA/x8Lk/bSMzMZIlMxJYMjOBJTMSOG96DDsLqsjOiG8PFMeJPgMxa+19OLNfGGOuAO621t5mjEmx1pa4OxzXAnvcW9YBXzLGPIGTmO91+70M/DAgQX81cJ+1tsIYU22MWQG8B3wK+MUQ/owiIiITUmuiflZaHLz0TScoCgp1Cq2exS7BrLQ4/Bb2n6xmyQzn13ZpbWN7flirhEznOxt6qXo/RCJCg3jmrkvYdLCU3OOVbDpUxtM7nHmbqNAg6pp9fOnKOcM+jqHWnxmxnvzFGJMMGCAP+ILb/iJwLZAP1AGfBnADru8D77v9vmetrXDf3wX8EYgA1rt/REREpBe7i7zMSooiev9fYdvv4OIvw7wbnJmwzJWDTlBvzTnbW+RtD8RqGslKCchFK9gKBe857zd8G1Jzhj0hfl5KrFMjDLDWUlBRz/YTleQer+TgqRquXZgyrN8/HAYUiFlrN+IsJ2KtvaqHPhb4Yg/Xfg/8vpv2bcCCgYxFRERksttT5OUfUsrh+a86gdfVD0BQ8FkHRClx4UyJCmWPu/Tp81vKaps6Lk0e2wzWPYjb1zLoZdDBMsYwIzGSGYmRrF08/Dlqw+VsZsRERERklJTVNlLnLeP/M9+ByET42B+cIGwIGGPISo1t2zlZWdeEz287BmKZK53lT1/TmC2WOh4oEBMRERmHdhdU8POQXxLVVAa3rYfooa0msDAtjkc2HaGxxdd+vFFgINZaLPUsl0EnOwViIiIi403BVua8dB8ZQTupX/WfRKQvG/KvWJAWR4vfcuBkDVV1zUCnQAyc4EsB2FlRICYiIjKeFGyFP15Hhq8JHx4i0hYNy9e0V9ivJizYKTvaZdeknLWBnDUpIiIio+3YZicvq9XxLcPyNekJEcSGB7On2NvxwG8ZUpoRExERGU9mXorF2bBog0KGLUneGMOCtDj2FHmJCAkiKjSo/TxLGTKaERMRERlPoqdigFf8y9i3+s/DmqO1MC2OD0pqKK6q12zYMFEgJiIiMp6U7ATgly1rycy5cli/Kistjiafn3ePlCsQGyYKxERERMYJv9/ywY4ttBBES9L5xIaHDOv3LUh1qthX1jUrEBsmCsRERETGgV2FVXzsN29z6sB7nAieyU9vuXDYvzMzMYpoNy9MOyaHh7LuRERExrDy2kZ+8vIBntxWQGJkKBdGFBI2/1qMO1s1nDwew/zUWLYerdCM2DBRICYiIjIGtfj8/O+7x/mvDQepb/Jxx6Wz+PIFkYT/qsI5YHuELEyLY+vRCqbGhI/Yd04mCsRERETGoP/edISfvHyAlXOTeOCG+cyZGgMfvOBcTMkesXEsSHNm3jQjNjwUiImIiIxBbx4sZVF6HH/6zHKMMU5jyU4wHpiWNWLjuOr8adx64QyWZSaM2HdOJkrWFxERGWOafX52FlSxbOaU9iAMnEAs6VwIjRqxscRFhPCDf1hIzDDv0JysFIiJiIiMMfuKq2ls8bN0ZqdZqJKdkDJy+WEy/BSIiYiIjDG5xysBWDIzvr2x5hTUlIxofpgMPwViIiIiY0zuiUrS4iNIiYtob3Qr6isQm1gUiImIiIwx249XsqS7ZUmA6QtHfkAybBSIiYiIjCHFVfWUeBtYOiO+44WSPJgyG8KHv5CrjBwFYiIiImNIa37Y0plTOl4o2aVlyQlIgZiIiMgYknu8koiQIOalxLQ31lWA98SIVtSXkaFATEREZAzZfqKSnIx4goMCfkWX5DmvmhGbcBSIiYiIjBF1TS3sLa7uvn4YwPRFIz8oGVYKxERERMaIXYVefH7bfSAWPwMip3R/o4xbCsRERETGiNZE/cVddkzu1LLkBKVATEREZIzYfrySOVOjiY8MbW9s8ELFEQViE5QCMRERkTHAWkvuiUqWzui0LHlyt/OasnjkByXDToGYiIjIGHCk7AxVdc1d88OKW3dMKlF/IlIgJiIiMga0H/TdTaJ+TCpETx2FUclwUyAmIiIyBuQeqyQ+MoRzkqI6XlCi/oSmQExERGQMyD1RyZIZCXg8pr2x6QyUHVQgNoEpEBMRERllVXVN5J+u7ZofdnIPYBWITWAKxEREREbZjhNVACzpvGOy9WgjnTE5YSkQExERGWW5xysJ8hiyM+I6XijZCVHJEJMyOgOTYadATEREZJTlHq9kfkoskaHBHS+0Juob0/2NMu4pEBMRERlFLT4/eQVVXfPDmhvg9H7lh01wCsRERERG0Qcna6hv9nUNxHb/FawPQiJHZ2AyIhSIiYiIjKLWQq4dArGCrfD815z3m37ifJYJSYGYiIjIKMo9XklKXDip8RHtjcc2g7/Zee9rdj7LhKRATEREZBTlHq/seqxR5krATdAPCnU/y0SkQExERGSUnPQ2UFRVz9LO9cOmLwLjgZmXwu3rIGP56AxQhl2/AzFjTJAxZocx5nn381+MMQeMMXuMMb83xoS47VcYY7zGmDz3z/0Bz7jGvSffGHNvQPssY8x7bvuTxpjQofwhRUREOijYCpt/Ouq5V9tP9HDQ96m9TqL+hZ9XEDbBDWRG7CvA/oDPfwHOBxYCEcAdAdc2W2tz3D/fAyeQA34JfBiYD9xijJnv9v8x8KC1dg5QCXx2MD+MiIhInwq2wh+vg9e+D4/eOKrB2K5CLyFBhvkpsR0vFG93XlMXj/ygZET1KxAzxqQD1wG/bW2z1r5oXcBWIL2PxywH8q21R6y1TcATwE3GGANcBTzl9nsUWDuwH0NERKSfjm4CXxNgnddRTITfU+TlvOkxhAZ3+nVcnAeRSRDX169WGe/6OyP2M+CbgL/zBXdJ8pPASwHNFxljdhpj1htjsty2NKAgoE+h25YIVFlrWzq1d2GMudMYs80Ys620tLSfQxcREQkQHLA7MSh41BLhrbXsLvKyMC2u68XiHc5smCrqT3h9BmLGmOuB09ba3B66/ArYZK1t/SfFdmCmtTYb+AXw9yEZKWCtfcRau8xauyw5OXmoHisiIpPJibednYgAF31p1HKwCivr8dY3s6BzINZUB6X7ddD3JNGfGbFLgBuNMcdwlhOvMsb8GcAY8wCQDHyttbO1ttpaW+u+fxEIMcYkAUVARsBz0922ciDeGBPcqV1ERGRoVZfAgfVOEnxYHNSVj9pQ9hR5AViQ2ikQO7kbrF/5YZNEn4GYtfY+a226tTYTuBl43Vp7mzHmDmANcIu1tm3J0hgz3c37whiz3P2OcuB9YK67QzLUfdY6N8fsDeBj7iNuB54dsp9QRESkVd6fnd2ISz8N6UuhsKfFnuG3p9hLsMdw3vSYjheKdzivCsQmhbOpI/YbYBrwTqcyFR8D9hhjdgIPATe7Of0twJeAl3F2X/7VWrvXvece4GvGmHycnLHfncW4REREuvL7IfdPMOsySJwNacvg9F5oOjMqw9ldVM3caTGEhwR1vFC8A6KnQUzKqIxLRlZw313aWWs3Ahvd993ea619GHi4h2svAi92034EZ1eliIjI8Dj8OnhPwKrvOp/TL3CWAIt3QOalIzoUay17irx8aN7UrhdL8pSoP4mosr6IiEwOuX9wSkKcf73zOW2p81q4bcSHUuJtoOJMU9cdk421UHpAy5KTiAIxERGZ+GpOOkn6OZ+AYHfHZFQiTDkHCt8f8eHsdhP1szoHYid3AVaB2CSiQExERCa+Hf/rJun/c8f2tGXOjJi1IzqcPUVegjzdVdR3E/VTVLpislAgJiIiE1vnJP1A6RdA7UmoHtmqSXuKvMxJju4+UT82DWKmjeh4ZPQoEBMRkYntiJuk33k2DJwSFjCieWJORf3qroVcob2ivkwaCsRERGRi29aapH9D12vTFkJQ2IjmiZ2qbqSstpGFaZ2WJRu8UJ6vivqTjAIxERGZkNbtLOYr/7Me2zlJP1BwKKRkQ9HIFXZtq6jfeUasZKfzqhmxSUWBmIiITDhNLX7+48X9pB17GtNdkn6g9GVQnAe+5hEZ2+4iL8bA/NSeEvUViE0mCsRERGTCeTaviOnVO7kz6AUq47O6JukHSl8GLfVwam/PfYbQ3mIvs5OjiQztVBe9eAfEz3DKasikoUBMREQmFJ/f8ubrL/BE2A+I95whxnsACrb2fEPaMue1aGQS9ncXebsWcgUnEFPZiklHgZiIiEwor+w9yQzvdkJoAcBYPxzb3PMN8TMgauqI7Jw8XdPAqerGrvlhdRVQeUz5YZOQAjEREZkwrLX8auNhjkc7M0sWaLLBtMy4pOebjHGWJ0cgENtbVA3Ags75YSV5zqsCsUlHgZiIiEwYW/LL2F3k5YZl52CA4tQ13Nr0bxwImdf7jenLoPwQ1FcO6/h6PNqoNVFfpSsmHQViIiIyYfzqjcNMiw3jQ8FOKQj/h3/Cdnsuuwq9vd/Ylic2vGUsdhd5OScpiuiwzon6eZAwCyIShvX7ZexRICYiIhPC9hOVvHOknM+tPIfgI69B6hLS02cQHxnCzoKq3m9OWwKYYV+e3Fvk7aGifp6WJScpBWIiIjIh/OqNw8RHhnDLgminUv7c1RhjWJQeT15fgVhYDEydN6yBWHltI8Xehq47Js+UOUcwKRCblBSIiYjIuHfgZA2v7j/F7RdlElW4Cawf5q4CICc9joOnaqhraun9IWlLnRIW1g7LGPcUO4n6WZ2PNipWov5kpkBMRETGvd+8eZjI0CD++eJMOPQKRCa2BTbZGfH4Lex1A6EepV/gJOtXHBmWMbYebZSV2kOifkr2sHyvjG0KxEREZFwrqKhj3c5iPrF8BgkRwZD/Ksz5EHiCAFiUHg/Qd55YupuwP0zLk7sLvcxMjCQuIqTjheIdkDgXwmO7v1EmNAViIiIyrv33psN4DNyx8hwnqKkrhzmr2q4nx4SRFh/Rd55Y8vkQ6uaXDYM9xT0l6u/QsuQkpkBMRETGrdM1Dfx1WyEfXZLO9LhwZ1kSA3Ou7tAvOyOOnYV9BGKeICcgGoajjirPNFFYWd81Uf/AeqgphsgpQ/6dMj4oEBMRkXHruZ0lNLX4+dxl5zgN+RucXK9OgU12ejwFFfVUnGnq/YHpF8DJ3dBcP6Tj3FPs5IctCMwPK9gKf/2U837bH3o/D1MmLAViIiIybr17pJyZiZHMTo6G2lIo2t62WzJQdoabJ9bXrFj6MvC3QMmuIR3nntajjQJ3TB7bDL5m572/pffzMGXCUiAmIiLjkt9v2Xq0ghWzEp2Gw68BtttAbEFaHMb0I2G/tcL+pv8c0hmqPUVe0hMiiI8MbW+ceWn7+6BQyFw5ZN8n44cCMRERGZf2n6zGW9/MitnuMuShVyBqKkzvWgYiOiyYuVOj+w7Eqo4DBvJfgUdvHLJgbE+xt2t+WFg0YOH86+H2dZCxfEi+S8YXBWIiIjIuvXukAoALZyWC3wf5r7llK7r/1ZadHs/OQi+2t4KtxzYD7nVf05AsF3rrmzleXtd1x+SBF53X6/5LQdgkpkBMRETGpdb8sNT4CKf2V0NVt8uSrbIz4qlwdy/2KHOls0wIzi7KIVgu3HGiEqD7HZNpyyBm2ll/h4xfCsQmo4KtsPmn2qEjIuOWz29570g5F53j5ocdegVMEMy+ssd7stP7kbCfsRw+8X9gPDD/piGZqXpl3ykiQ4NYPitgJ2fNSSjKhfM+fNbPl/EteLQHICOsYCv88Xpnp05wKNz+nKbERWTc2V9STXVDCytaA7H8DZBxIUQk9HjPedNjCA32sLOgiusXpfb88NlXwKzLnDIWZ8nvt2zYd4orzksmPCSo/cLBl9xBXXvW3yHjm2bEJptjm8HXCPihpQGevQu2/g9UHHWua7ZMRMaBd4+UA3DhOVOc2aWSnTD3Q73eExrsISs1lp0F3r6/YO4aKP0AKo+f1Th3FFRRWtPI6vnTO144sB7iZ8LUeWf1fBn/NCM22Uxb4L4xTkJrYy28eLfTFJsGtafA+iEoTLt4RGTMevdIBZmJkaTERcCOvzmNc1f3eV92ejxPvl9Ai89PcFAvcxHnroGX73OWPJd/btDjfGXfSYI9hivPn9re2HQGjmyEpZ8GYwb9bJkYNCM22biH4LLkk/Dpl+DrH8C/bIdrfgzB4U5RQesfst1CIiJDzee3bD1a3r4seWgDxKQE/EOzZ9kZcdQ3+8gvre29Y+JsmDK7fQlxEKy1vLL3FBfNTux40PeRjc6KhPLDBAVik09rtehV32+f7UqcDSu+AP/wGydBFVRcUETGloC0iQ75Ycffdpb5UrL7NbvUlrDfVz0xcGbFjm52ZrAG4dDpWo6WnWF1VudlyRchLA5mXjyo58rEoqXJyebkLicvISK+67WM5ZB9C+Q9Brc8oWVJERkbjm2BP93k1ArzeIiespL/DPHz4R1AwRZnFv/w606w1sffW5mJUcSGB5NX4OXjF/TxveeugXd/BUc3DWr26pW9JwFYPT+gPIXfBwdecspsBIX0cKdMJpoRm2xKdkHKop6vL/gIYEFpCyIyVrz/OydtAgt+H9PL32Vl8D7CSnc7QRg4AU4/0ik8HsOi9Pj+zYjNuBhCY+Dgy4Ma9st7T5GTEc+02PD2xqJcqCvTsqS0USA2mTTWQMXhbo//aJN+AWDgxHsjNiwRkV6dKXVeTRA2OII77Ld5cMEz8Im/QnCEUz9sAOkU2RlxHDhVQ0Ozr/eOwaFOKYtDr0Bv1fi7UVxVz+4iL2u6W5b0BDsnAIigpcnJ5eQe57W3GbHwOJiWBQXvjsyYRER609wARdvhvOsgfSlHohaz5a/1/Hx2ImSkObu7j212grB+plNkp8fj81v2FntZOnNK753nroH9z8GpPTB9Yb+H3bYsmdWpav6B9TDzku7TQ2RS0ozYZHLSTdSf3ksgBs5fZoXbnKl+EZHRdGQjNJ+BCz4DK7/OG2cyAfd8SXD+vlr59QHltOZkOEFQXr/qibklMQa4PPny3lPMmRrN7OTo9sbyw05tMhVxlQAKxCaTkl0QlQwx03vvl7ECGqvh9P6RGZeISE8+eA7CYiHzMsAp5DorKYrpceF93NizqbHhTI8N71+eWMw0SF3sLE/2U+WZJrYeq2BN59mwtmr61wxgtDLRKRCbTE7udGbD+triPeNC51XLkyIymnwtzlLe3NUQHOqcL3m0ghXn9LGc2A+LZ8Tz3tFyfP5+5H7NXQOF70NdRb+e/doHp/H5bffV9KdmQULmwAcsE5YCsQmqtrGFHScqeedwORsPnOaVXSfwn9rPfjL587vHOVbWS12c+JkQPV0J+yIyugreg7pymHc94JwvWRN4vuRZuCknlVPVjbx58HTfnc9d7ezOzH+1X89+Ze9JpseGsyg9rr2xrsKpeabdktJJvwMxY0yQMWaHMeZ59/MsY8x7xph8Y8yTxphQtz3M/ZzvXs8MeMZ9bvsBY8yagPZr3LZ8Y8y9Q/fjTV7/+kQe//Crt7nlf97ln//wPj9/fB0e28LD+6P49t/38P3n9/V8szFOvkWBAjERGUUfPO8ct+buMGw7X3LW2QdiV8+bRlJ0GI+9V9B355TFEDW1X1X265t8bDpUyuqsaZjA1Yf8V8H6lB8mXQxkRuwrQGDS0I+BB621c4BK4LNu+2eBSrf9Qbcfxpj5wM1AFnAN8Cs3uAsCfgl8GJgP3OL2lUFqbPHxVn4Z1y1M4fHPreDpuy7mkVVO4cBv3/FxbshOJa+gCtvbduwZK6DquHOYrojISLMW9j8P51wBYTHA0OSHtQoJ8vCPy9J548BpTnobeu/s8TgFWPNfdZZLe/HmwVIamv3dl62IdvPNRAL0KxAzxqQD1wG/dT8b4CrgKbfLo8Ba9/1N7mfc61e7/W8CnrDWNlprjwL5wHL3T7619oi1tgl4wu0rg7TjRBX1zT7WLk7jotmJLJmRQFr9IQiNISVzPsszEyg/00RRVX3PD8lY4byeUJ6YiIyCk7vAe6JtWXIo88Na3XxBBj6/5f+29WNW7Nw10OCFwq29dntl30niIkJYPitgnC1NcOBliE2Dom1nOWqZaPo7I/Yz4JuAW8KYRKDKWtv6T4NCIM19nwYUALjXvW7/tvZO9/TULoO05VAZQR7T8S+sk7ucGjgeD4vcs9Z2FfaydXv6QucQcC1Pisho+OAF5+xbdylvKPPDWs1MjOKSOYk88X4B/r6S9s+5EjwhvZaxaPb5eW3/aa4+fyohQQG/Xrf9DlrqoDgPHr3ROYpJxNVnIGaMuR44ba3NHYHx9DWWO40x24wx20pLS0d7OGPWlvwycjLiiQl3zzHz+5xirm4h1/NTYggN8vS+dTs4FNKWKhATkRFTWFlHfZNbv3D/8zDjIohKAuCdw05+2FAGYgA3XzCDoqp6NueX9d4xPBZmXtRrGYv3j1bgrW/uesj3vr+7b/zga+rXUUwyefRnRuwS4EZjzDGcZcOrgJ8D8caY1sr86UCR+74IyABwr8cB5YHtne7pqb0La+0j1tpl1tplycnJ/Rj65OOtb2ZXYRWXzElqb6w44hREdAu5hgUHMS8lhp2FfdTQybgQSnZCU90wjlhExAnCrvzPjSz+/it863fr4PRe6me37zB890g55yRFdTy3cQiszprGlKhQHn/vRN+d566B0/vglfu7ndV6ee9JwkM8XH5up99P1SXO7N4Aj2KSyaHPQMxae5+1Nt1am4mTbP+6tfZW4A3gY26324Fn3ffr3M+411+3Tlb4OuBmd1flLGAusBV4H5jr7sIMdb9j3ZD8dJPQu0fK8Vu4NDAQK9npvAYcbZSdEc/uQm/vNXRmrHAO2i3ePkyjFRFxPJtXTLPPsjYnjWnFGwC45qVYbvvte/zvO8fYerSCC4d4Ngycf5h+dEkar+4/xemaPpL2Y1Kc17cf6rLEaK3llX2nWDk3mYjQoPZ7vIXOxqcL7oCrvuUcyTSAUwBk4jubOmL3AF8zxuTj5ID9zm3/HZDotn8NuBfAWrsX+CuwD3gJ+KK11ufmkX0JeBlnV+Zf3b4yCG/llxEZGtR2hAfg5IcFhULy+W1Ni9LjOdPk40hpbc8PS7/AedXypIgMs3V5xSydmcB/fHQR/5J6kLop87n2sgsp9tbz/57dS01jy5Am6ge6efkMWvyWp3ILe+9Yecx9Y7ssMR46XUuJt4EPzZva8Z7815zXZZ8Z8FFMMjkM6NBva+1GYKP7/gjOjsfOfRqAf+zh/h8AP+im/UXgxYGMRbq3Jb+MC2dNITQ4IMYu2QVT50FQSFtTToZTaDCvoIq502K6f1jkFEg6T4VdRWRY7S+p5sCpGr5/UxbUnsYUvEfkFfdxzxXnc88155N/uoa8Ai/XLkwZlu+fnRzN8llTePL9Ar5w2Ww8nh5OH5m1EjzBzkpBUEiHJcZtxyoBWN65xln+q85uyYB/CIsEUmX9CaS4qp4jpWc65odZ6+6Y7HjQ9zlJ0USHBfe+cxKc444K3gO/v/d+IiKD9GxeMUEe4wRaH7wA2LayFQBzpsbwsaXpHXciDrFPLJ/B8fI63nGLxnYrYznc8JDzfsVdHWa3th2vIDEqlMzEyPb+vmbn0PI5V/d9tJxMWgrEJpC33F0/l84NCMSqi50jQlKyO/T1eAwL0+L6kbC/AhqqoOzgUA9XRAS/37Iur4jL5iaRGB3mVNNPmAVTR7au9zULphMXEcLjW/tI2s/5hHNW5MldHZpzj1eydGZCx2r6hdugsbrtZACR7igQm0C25JeRFB3KeYFLja1/WXSaEQNYlBHH/pJqGlt8PT80o/UAcC1PisjQ23a8kmJvA2sXp0FDNRx5E86/bsRnkMJDgvjIkjRe3nuS8trGnjsaA/PXOjNd9c5yZGlNI8fL61iWmdCxb/6rzk7JWZcP38Bl3FMgNkFYa3krv4xL5iR1/BdZyU7AwLSsLvdkp8fT7LPsL6np+cGJsyEySYGYiAyLv+cVERESxIfmTXNqdPmbYd4NozKWW5bPoNlneXp7txWU2s2/yckT+8BJbc49XgHA0pmdNhPkv+osX0bEd36CSBsFYhPEgVM1lNU2dcwPAydRP3EOhEV3uSc7o7XCfi/Lk8Y4s2I66khEhlhTi58Xd5ewOmsaUWHBsP1RCImCPorcD5dzp8WwdGYCj79/ovezeFMXQ9wM2OdUbdp2rJLQYA8L0mLb+9SWQkmekx8m0gsFYhPElkNufljnQOzkrg71wwKlxoWTFB1KXm8V9sH5F13FYecvFhGRIbLpYClVdc3clJMKx7bA0U1O8en/XTtqxwDdfEEGRM/BrQAAIABJREFUR0rPsPVoRc+djIH5N8Lh16HBy/vHK8lOjyMsOKB+2OHXndfZCsSkdwrEJoi38ss4JzmK1PiI9sa6CvAWdEnUb2WMITs9vh87J90DwPs47FZEZCD+nldEQmQIK+cmw76AOt6jeAzQ9YtSiQkP5sn3+zgIfP5a8DfTtO9F9hZ5u1+WjEyElJzhG6xMCArEJoCmFj/vHa3ofjYMuk3Ub7UoPZ7DpbXUNDT3/AUpOU5BWC1PisgQqW1s4dX9p7huUYpTliLSDWSMZ1SPAYoIDeKarOls2Heq941MaUshNo3a7U/R4rcsmxmQqO/3w+HXnNkwj37NSu/0/5AJIK+giromX/f5YdDjjBhAdkYc1sLuol5mxULCnZwIJeyLyBB5Ze9JGpr9rM1Jcxr8LYCBK+4b9WOAPrxwOjWNLbyd30tNMY8H5t9EbNEmoqljaWAgVpLnlA1S2QrpBwViE8CW/DI8BlZ0Poft5C6ITW//l2Y3FqW3Juz3sTyZsRyKcmHjj0ctd0NEJo5n84pJT4hoD2BKP3B2aV/+zVE/BuiSOUlEhwXz0p6TvXecfxPBtomb4/eTEBXa3t56rNHsq4ZvkDJhKBCbALYcKmVRejxxESEdL5T0nKjfakpUKDOmRLKzr4T9iETnX6xv/keXw25FRAairLaRLfll3Jid2l5up/TAmDkGKCw4iKvOn8or+07S4uv5VBF/2gWcJoGbQrd1vJD/qpPSEZ08zCOViUCB2DhX3dDMzkJv1/yw/5+9945u67rSt58DgGDvXeykrC5RVrMlW5Ys27EdtxQ7sZ3itMkvySSZSU8+p2c8kzKZTPqk20kcl9iJe5ctS5Ysq3eKKuwdJFjABqLc748DECCBC4AkKJLiedbiAnXvuRcHJEVs7v3ud48MQtfZkPowL6sKU8NnxBwD8lFzz6iQVqFQzH2ePdaKy61JE1cA5wh0nYfsxTO7MT9uXJFH96AjZPfkuc5BnnOuZ2n/XrD3y4ND3bKxSZUlFRGiArE5zls1VlxuLVAf1n5SBk1hMmIgjV2be4aw2EK4SV/yNimihRkV0ioUirnPE0eaWZKXzCLvFBDredBcsyYjBrBlcTZxMQZeOKlfnjxQ183zrsswue3SjBbkZADNrQIxRcSoQGyOs/tcJ3ExBtaUjHNurpJGg5EM647I2LVogxxyC3Drz2dcw6FQKOYm9V0DHG7o8WXDQOrDYFYFYglmE1sWZfPCiTbc7uDmrgfqrdTEr0BLzBk1d+XcKxCbCoXrL+BuFXMZFYjNcd4418mGssyxRoKN++DNX8nP//EvYfVcKwpSMAjC68Q2fQYQYK2Z2qYVCsW85akjLQDcWrnAd9BSDQjIumRmNqXDjSvy6bDZOazzu/FgfTerS7MQS2+RGbGRQSnUL98CRtMF3q1irqICsTlMW+8w5zr6uXLhuG7Jul0yzQ8R6bkSzCYW5SZzNJxOLDkPijdC1VOh1ykUCoUOTx1tYUNZxljzactpSC+FmHjd62aCbUtziDEKXjjRGnBudNB3SbqcPekYhD0/A1uLKksqJoQKxOYwu895xxqN68wp3QwI+RGhnmtVYSpHm3pCz1cDOdaj/QR0npvcphUKxbyld8jB2Y5+ti4e9ztrFnVM+pMSF8MVC7N4/kRbwO9G76DvdaXpUHKFdNHf9T/ypJovqZgAKhCbwxxt6iE51sSSvOSxJxZcKh9LN0dsjFhZlEbPoING61DohUtvkY9eDZpCoVBEyOnWPgCW5vsNx3Y5ofPsrOqY9OfGFXk0dQ9xsqVvzHHfoO9UWYZccjO47JCYDX0tM7RbxVxEBWJzmLbeYRakxWMwiLEn+loADVbdEbGovtJj7HoklGAfILVQilBPqUBMoVBMjNNtNgCW5vkFYt214HbMyowYwHXL8jAIAsxdD4wf9O3d/0Cn8lpUTAgViM1h2vuGyU2NCzzR0yAf04ojvtfivGTMJgPHwgn2QeohWo+CtTbi+ysUCkVVax/pCTHkpsT6Do52TM7OjFhGopnLyjJ53k8nNuxwcbJl3KDvkQGkJERTXouKCaECsTlMW98wucmxgSd6G+VjalHE94oxGli+IIWj4TJi4FeefDri+ysUCkVVax9L81N8bvrgC8SyFs3MpiLgxpV5nLcMcK5DZvSONvbgcI0b9F2+BUxxIIzKa1ExIVQgNkdxuTUsNjt5QTNi3kCscEL3rCxM40RzX8iRHoDsbspfrcqTCoUiYlxujep2G0v8y5IghfqpxRCbNDMbi4Drl+cB8PxxWZ48UN8NMHbQd9EGqcnddu+MDy1XzC1UIDZH6ey349YgN0WnNJmcD6Yg2bIQVBalMuRwcc7SH37xstug+YAv6FMoFIoQ1HUNMOxwszR/XHOR5fSsLUt6yU2JY01xGs97dGIH67upyE4cO+gbZPC1+QsqCFNMCBWIzVHaeocBnUCst2FCZUkvq7yC/YYIdWKgypMKhSIiqoJ1TLpds7pj0p8bV+RzqrWP+q4BDtZ3s85fH6ZQTAEViM1R2vtkIJYXNCPWOCGhvpeyzERykmN56VR7+MWZFZC7QpUnFQpFRJxutWE0CBbm+JUge+rBOTxrOyb9uWGFLE/+8rVz9A45WFuaHuYKhSIyVCA2R/EGYrmp48qPbjf0NkHaxDNiBoPgPeuKeK26g6buwfAXLLsNGvdCX6DrtEKhUPhT1dpHRXYicTF+49g6Zt+MST2KMhJYviCFvx9sAhgr1FcopoAKxOYobX3DGA2CzMRxgVh/m/TkmURpEuDODfK6R/ZHoP3ylidPPzOp51IoFPMHb8fkGEatK2Zvx6Q/N67IQ9MgM9FMWVbiTG9HcZGgArE5SnufnZzkWIzjzVxHPcRKJnXfwvQEti7K5pH9jTjCdU9mL5Z/yarypEKhCEHP4AgtvcPBOyaTF0Bc6sxsbILcsCIfgDUl6WMtOBSKKaACsTlKe9+wTsekJ5M1idKkl7svK6HDZmd7VUf4xUtvhfrd0B/BWoVCMS8ZddSfgx2T/izMSeJjV5bxwY2T+0NXoQiGCsTmKG29w2Pdqb30ejJiE/QQ8+fqxdnkp8bx4Fv14Rcvuw00typPKhQKXbwdk8vGdEy6ofPMnNCH+fP1m5ex+ZLs8AsVighRgdgcpb1vWKdjsgESssA8ef2CyWjgveuL2HW2k4auMKL93OWQXAC7f6ZmqykUiqBUtfaRkWgm238SSG8jOAbnVEZMoZgOVCA2BxkacdE37NSZM9k4pbKklzvXF2M0CB7a3xB6YdN+GGiXg3sfuEUFYwqFIoDTbTaW5iePG21ULR/nWEZMoYg2KhCbg7R5rSuSdTJik+yY9CcvNY5tS3L4+4FGRpwhRPt1u2RpEtSgW4VCEYDT5aa6zcbSAKH+7B72rVBcKFQgNgcZNXMdnxHTNI+H2MTNXINx92XFdPaP8NKpNv1FpZvlgFsAYVCDbhUKxRjqugawO90sCbCuqIbEHEhQDvWK+Y0KxOYgo2au4zViA53gHIpaIHbVJdkUpsfzt7dClCeLNsA9T4M5Gcq2qhlrCoViDFWtF0fHpEIxXahAbA7imzM5rmty1EMsOoGY0SC4a0Mxe853URNqEHjRBvkxoCwsFArFWKpa+zCNH22kaTIjpvRhCoUKxOYi7X12Es1GkuNixp4Yta6YukbMyx3rCjEZBA/tCyPaz1kqW9Hdrqg9t0KhmPvI0UZJxJr8Rhv1tcCITWXEFApUIDYnmU4z1/HkJMfxtuW5PHawiWFHiCArZ6kc3ttdF7XnVigUcx9vx+QYLHNnxqRCMd2oQGwO0qYbiDXIUSFRHhdy94YSugcdvHAihGg/Z6l87KiK6nMrFIq5S8/gCK29w0FmTCrrCoXCiwrE5iDtfcOBHZMgDRJTo6MP82dTRSalmQmhRftZnhKDCsQUCoWHUx5H/cCOydOQkAlJyqFeoVCB2BxD0zQ6+uzkBBtv1NMQ1bKkF4NHtL+vzsrZdlvwRbFJctB4x6moP79CoZib6HdMKqG+QuFFBWKzncZ9sOvHo4711oERRlzuwPFGmuZx1Y9+RgzgttUFAOw+16m/KGeZT/uhUCjmPadb+8hKMpPjbz6tacq6QqHwI2wgJoSIE0LsE0IcFUKcFEJ8x3N8lxDiiOejRQjxhOf4ViFEr9+5b/rd6wYhRLUQ4pwQ4qt+x8uEEG95jj8ihDBPx4udczTug/tvgu3fHR0f1N5nBwgMxIZ7ZBdSFDsm/clNiSU51kRN54D+opwl0HkWXI5p2YNCoZhbVLX1sWS8o35/h/x9pTJiCgUQWUbMDmzTNK0SWA3cIIS4XNO0zZqmrdY0bTXwJvAPv2t2ec9pmvZdACGEEfglcCOwDLhLCLHMs/4HwE80TVsIdAMfjcqrm+ucflaODYLR8UFeM9ec8YHYqIfY9ARiQgjKc5I4H8pPLGcZuB3QdX5a9qBQKOYOTpebM+39ITomVUZMoYAIAjFN4n33jfF8aN7zQogUYBvwRJhbbQDOaZpWo2naCPAwcJuQU2C3AY951j0AvGNCr+JixO2G86/5/u0ZH9SmN95o1LpiekqTABVZidRYQmTEvH/hKp2YQjHvqe0cYMTpVh2TCkUYItKICSGMQogjQAfwsqZpb/mdfgewXdO0Pr9jGz2lzOeFEMs9xwqARr81TZ5jmUCPpmnOccfnN2/+HNqOwpWfh+R8mW0q2uDLiCWPE+v3er6009A16aU8O5HW3mEGR5zBF2QtkgGj6pxUKOY9ox2TwYZ9x6VCUu4M7EqhmH1EFIhpmubylCALgQ1CiBV+p+8CHvL79yGgxFPK/DnhM2URI4T4uBDigBDigMViidZtZx8tR2D792DpLXDNN2HFu+VfkU477X3DZCWZiTGO+9b1NEBM4rQO0C3PliNKdLNiMXGQUQEWFYgpFPOdqlYbMcZxo40AmvZDbLJ8VCgUE+ua1DStB3gNuAFACJGFLDk+67emz1vK1DTtOSDGs64Z8BcwFXqOdQFpQgjTuOPBnv+3mqat0zRtXXb2Reo/MzIIj38MErPhlp+BEFCyCVx2aD5EW28IM9e0Irl+mijPTgQIoxNbojJiCoWC021ytJHZ5Pc207gP2o5BbxM8cOtoN7hCMZ+JpGsyWwiR5vk8HrgO8HoU3A48o2nasN/6PI/uCyHEBs9zdAH7gUs8HZJm4E7gKU3TNGRwd7vnFvcAT0bjxc0VznXYePev98iy40v3Qtc5eOf/+bJbxRvlY/0btPfZAzsmQQZi09Qx6aU0MxEhQmTEQJZQrTXgGNZfo1AoLnqqWvtYNl4fdvZl3+eeBiSFYr4TSUYsH3hNCHEMGUy9rGnaM55zdzK2LAkyoDohhDgK/Ay40yP4dwKfBl4EqoBHNU076bnmK8DnhRDnkJqxP0zlRc017nu2ioP13Zzb9Sgc+CNs+gyUb/EtSMiAnOVQL4O1gI5JkBqxaRTqA8TFGClMjw9tYZG9BDS3HACuUCjmJdaBEdr77CwZ3zGZXiIfhQGMZijdfOE3p1DMMkzhFmiadgy4VOfc1iDHfgH8Qmf9c8BzQY7XIEuc8469NV28Vm0hm25WH/4G5K2CbV8PXFiyCe3I3+gZ+HBgRsxug6HuabOu8Kc8K4macBYWIMuT+aumfT8KhWL2cdoj1A/omDR5fndt+LjUvhbNy1/7CsUYlLP+DKJpGt9//jTXJdbyRMJ9xDgH4N1/AFOQ8UUlmxCOAVaIWvJSx533WldMc2kSpE6sxjKA260FX5BZAYYYJdhXKOYxp/QCMWutfLzmWyoIUyg8qEAsyjx9tIXaUKU7P1482YapaS+/cX+TAncLQnPjHLQGX1xyBQAbDKcDS5Ne64q0ksluO2IqspMYcrhG/cwCMMZA1iVKsK9QzGOqWm1kJcWSlTTuj0ZrDSQvAHPCzGxMoZiFqEAsijhdbv7t4cO8//dv0dlvD7v2J88f50fx92PQXAAINLpPvhb8guRc+pNK2WA4HVianGZXfX+8nZOhBftLVSCmUMxjzrTbAh31QQZiGeUXfkMKxSxGBWJRxDowgluD5p4hPvnXg4w43bpr/7G3mm/2fZsydwMYYtCEEQcmjses1L2mOeVSGYgljxvF2dMgha+JOdF6KbpUeL3EOkPpxJZCTz3YQ6xRKBQXLfVdA5RmJgaesNZARtmF35BCMYtRgVgU6bDJLNjNq/LZX9fNt546gXTnGMtgTzsrXnkflxur0N75G/jwc2hX38tHtG/w+mCp7v2r41aRKgZJs43rSOxtlPoww/R/O3OSY0k0G8OMOloqHzurp30/CoVidtE76KBv2Elxxrjy43AfDHSojJhCMQ4ViEURi6cc+eEryvjk1goe2tfIX/bWj13U28Twb99GubuB89f8FlF5JxRtwHDVF3AuWMfx5l7d+x/yzEgXDW+OPeE1c70ACCEozw43/NsTiKnypEIx72iwDgJQND4Q6/YI9VUgplCMQQViUcTiyYjlJMfyxbct5polOXzn6VPsOdcpFxx/HPcvNxI70MJP83/Ios13jLl+RUEqp1r7cLqClzRPD6XSYciB+t1jT/RMv4eYPxXZYYZ/p5fKNnUViCkU8w5vIBaQEbPWyMfMigu8I4VidqMCsSjiDcSykmIxGgT/e+dqyrIS+dTfDtF++EV4/COIkT6MaNy1MfCvwlWFqQw73JzXCXLa++zUJlZC/R7wljwdQzLdP43DvsdTnp1Ec88QQyOu4AsMRsherAIxhWIe4suIxY894Q3E0pVGTKHwRwViUcRis5McayLebAQgOS6G339wHZoG+567HwABxAgXxX0HA65fWZAKELQ8qWka7X3DdGSsgwELdJ6VJ3qb5OMFKk2Cr3MypE1HzjIViCkU85AG6yAZiWaS42LGnrDWQFIuxCYFv1ChmKeoQCyKdPbbyU4e65tTmpXIL+6+lOohGbw4NYEwBR/tUZaVRILZyPGmnoBzNruTwREXA/mXyQPe8uSodcUFzIhlRdA5mb0EbC0wFPhaFArFxUujdTBQHwbQpawrFIpgqEAsilhs9kADQ2DzJdlctbQQgDcLPozhnqeDukobDYLlC1KCZsTae6WBanzuJfKvyvo98kTvhXPV91KWJYPK8x1hMmIAltP6axQKxUVHg3UwUB8GykNModBBBWJRxBIkI+ZlfVo/rphkLv/oj0OO9tAT7Lf3Sf1ZXmo8lGySGTFNkxkxYYTk/Oi9kDDEm40UpMWH8RJbIh87Tl2YTSkUihnH6XLT3DNE8Xh92MgA9LepQEyhCIIKxKKIxaYfiIneJozpxcQYQ3/J9QT73pFCuSlxctxRX7M0Te1phNQCMIad3x5VysN1TqYWgTkJOlRGTKGYL7T2DuNya0E6JpV1hUKhhwrEosSww4Vt2KkbiNHbGJGg3ivYPzZOJ9buCcTyUuNG505Sv0dmxC5gx6SXiuwkaiz9QQ1rARDCM+pIZcQUivmCroeYt2NSBWIKRQAqEIsSXuuK7CAaMcCTuQofiHkF+yfG6cTa+4ZJjY8hLsYohfDx6bI82XthPcS8lGcnMjDiGi2ZBiV7idKIKRTzCH0PsfPyUY03UigCUIFYlPAO+Q6aERvuBXtvRBkxPcF+W+8wuSmeexsMMitWsxNsrRfUusLLaOdkSIf9ZdJqo99ygXalUChmkgbrICaDID81iIdYQhbEpc7MxhSKWYwKxKKEv5lrAD0T62wMJthv7xuW+jAvJZugtwE09wXtmPRSkePpnAzpJeYR7FuUn5hCMR9osA5SmB6P0SDGnrDWKkd9hUIHFYhFCUuojJjXYiLCEmIwwX57n5288YGYlxkoTealxJFgNobPiIES7CsU8wRdDzFlXaFQ6KICsSjhzYhlJpkDT04wIzZesO9ya1j67WMzYrkrweT5hTdkndymp4AQgrKsMJ2TSblSy6YE+wrFvCCoh5hjSHZ5q0BMoQiKCsSihMVmJyPRHNyeorcBjLGQmB3RvcYL9jv77bjcGrmpfoFYyyFwyU5K/vkJaNw31ZcwYcqzkzgfKiMmBKQUwrntM7I/hUJx4egdctAz6FDWFQrFBFGBWJSw2OxhOiYLpcg+AsYL9ketK/wzYnW7wOsc4XLIf19gKrITae4ZYtihM/y7cZ/MhvU2wAO3TCwYa9wHu36sAjiFYo7QqNsx6bWuUB2TCkUwLqwL6EVMsDmTo/Q2TbizcUVBKg/ta8DpctPW6zVz9bt/6WYwxYJrBIzBZ1dON+XZSWga1HUNsCQvJXBB3S7ZTADgHIaH7oIV74KyLVB6JXSekWtKN/umDbgccOYleOzD4HbITOI9T4WcRnBR0bgv8GuiUMwBGpWHmEIxKVQgFiUs/XbWlSQGP9nbCJe8bUL3W1WYyp92S8F+0IxY0QYZoMzgm3a5Z+ZkjUUnECvdDKY4cNlBGCC9FA7/Ffb9FvB2VWnyXFop2PtgsAtfqg8ZaNbtmh9BSd0bcP/N8nNT3PwKQBVznlEPscwggVh8htSLKhSKAFQgFgU0TfMM/A4i1HcMQ3/7hC0m/AX77X12jAZB5vjSZ9GGGX2jLs/2Dv/W0YkFCxadI9B8AHZ8H2pfl+s0txzRtPRmSMoD5xDs+bnn+Mxk+2aEY48yGoTOpwBUcVHQYB0kLSGGlLiYsSdUx6RCERIViEWBfruTYYc7eGmyr1k+TrA06S/YHxhxkZ0UG+jNM8MkmE3kp8ZRE8pLbHywaDJL641tX4cH9vlKq7f9cuy6/g44+jB84In5E4xofoPeDYb5E4AqLgqCdkyCDMSKN174DSkUcwQViEWB0fFGwQKxngb5OMGMmL9gPzHWNLZjchbhnTk5YcKVVgvWwtGHIPPi+0t65xkL+alxXJKbPPaEtRYyKmSJ1pwEhetnZoMKxSRotA6yvGCcc75jWGpkVUZModBFdU1GAd+cySDB0qiZ68Td770O+809Q+Sl6DQCzDDl2dJLTHf4dyiKNsDmLwTPeCXlykdb69Q2OMtwuzX+9cFDfOupk2NPOIagaR8svhGu/Q5018L5V2dmkwrFBHG5NZq6hwIzYj31gKZc9RWKEKhALAp09o8AehmxRilGTymY8H29Dvs1loGxZq6ziPKsRGx25+hkgaiRnC8fbe3Rve8Mc97Sj83u5K1aK72DDt+Jpv2yTFu6GVbeLrVye34+cxtV9iGKCdDaO4TTrYWwrlAZMYVCDxWIRQGLTXY16o43Ss4HY0zguTCs9Evzz9pALFsO/z7fEUInNhmS8+TjRZYRO9zom5bwWnWH70TtLhmwl2yUtiSX/T+oeQ1aj134TTbug/vfDtu/Bw/cqoIxRVgawnqIqUBModBDBWJRwNIvuxrT4oMEWz2Nkx7K7RXswzjrilmEt3OypnMSOrFQjJYm26J73xnmaGMPybEmcpJjeemU32ur2wX5qyHOE3yv+4jUib35iwu/ydpd0s8Nzde9qVCEQNfMteu8/JlW1hUKhS4qEIsCXusKQ7Cuxt6GSenDwCfYh9mbEVuQGk9cjCH0zMkgOF1uuvrtDI3ouPKbzJCQedFlxI409lBZlMY1S3N5vdqC3emCkUFoOgBlfl2S8Wmw5oNw4nEpdr6QpCzwfW40qe5NRVgarIMYDYL88U1F1hrZgCJmV8e3QjGbUF2TUcBi03HVd7ugr2XSGTGQgv39dd3kpc5Osb7BICjLCt45Oexw8cKJNnaesdA1MELP4Ajdgw66B0ewDTsBKMtK5LUvbg1+8+R86cF2kTA04uJ0m41PbClnXUkGD+1rYM/5Lq42nZRTBMYHPJd/Et76Dez9NVx/34XbaL9fpm7th+ePfYhi0jRYhyhIi8c0ftautUZ1/yoUYVCBWBSw9OvMmbS1gts56YwYwE0r8znV0kdhehB/nllCeXbi6IBykIL0h95q4PFDTXQPOshOjiU/NY60BDOlWYmkJ5hJjY/hSGMPO89acLm14B5pyXkXVUbsZEsvLrfG6qJ0NlZkkmg28vKpdq5O2QXCCMWXj70grRiWvxMOPgBbvuwrW0aJcx02GqyDbFuSO+7EdshdIcdSeTU+CkUIgnqIOUekRnbVe2dmUwrFHEEFYnpMYOZfp22EZflBRvz0eKwrUosnvY11pRk88v9mtxliRVYizx9v5YnDzTy8v4G9NVZMBsHbludy94YSNlVkBi3bPrCnjtfPWOgeHCErWCCbnAftJwOPz1GOeIT6lUWpxMUY2bI4m5dPtXNf7i5EwRqITQ68aNNn4MRjMhi74rNR3c99z1bxZk0Xx799PTHeTIbdBg17YeOnYGQAjjwk31BNQaZGKBQemqyDvG153tiDPQ3SpFgJ9RWKkCiNWDAa98GfIusac7s1/YHfU/AQm0tU5CTh1uDfHzlCU/cQX7p+MXu+to1fvW8tV16SFVw7B2R6RkJ1eew/AkjKk6VJt46ObI5xuLGHgrR4cpKljua6ZbkM2Hqh+ZC+DmvBaii7SpYnnTpfp0kw7HDxZk0Xww43Va19vhO1u2SZdOG1UH41OAakv5lCoUO/3UnXwEiQjsnz8lEFYgpFSFQgFow6z5tRBF1jPUMOnG4teEZn1FW/cHr2OUvYtiSHf9lcxv0fXs/OL13Nv169cDTYCEVGoicQG9DxIEvOk39RD3RGc7szxtHGHlYXpY3+++rFOWwwnkFoTii9Uv/CTZ8FWwuc/EfU9rK/zsqwQ45UOlTf7TtxfjvEJELR5bJ5QBigZkfUnldx8aHbMektayszV4UiJCoQC0bpZjB6AisReuZfyPFGvY2y88+cOB27nDUkx8Vw703L2Lo4Rzf7FQxv8KqbERs1dZ37OrHOfjtN3UNUFvl0XmkJZt6ZXoMDU6A+zJ+F10L2UtjxA9gZHZPV16stmE0GspJiOdjQ4ztx7hWZgTOZpSatYC2cf23Kz6e4eAnpIRabIn8HKhQKXVQgFoyiDXDP09LLKrUoZNctZOt7AAAgAElEQVSPb7yRjqv+FDomL3a8GTHrQLhAbO57iR3xBDuri8b6KW0ynOKIu5zavmBXeRACltwE3TXw6n9ExWT19TMWLivLYENZui8j1nUeuutg4TW+heVXQ8shGOoOeh+FImRGLKNMWVcoFGFQgZgexZfBlq/IN7/mg7rLLP2hXPWbLnp92FRITzAjBHTpjUdKvnjmTR5t6sFoEKwo8GvqGO4jy1bFXvcyXj4VJtgcnczgnrLJanPPEGc7+tmyKJs1xek09wzR3jcsuyVhbCBWcbUsD9e9MennU1zcNFgHSYkzkZowztDaWqP0YQpFBKhALBQr75B6mQN/0l3SadOZM6lpsjQ5hY7Jix2jQZCeYKZLLyPmdde/CLzEjjT2sCg3mQSzX6Nyw16E5qIlbR0vnwrzGiu2yTI5gNE8JZPV16stAGxdnM2aEpmhO1TfLfVhGeVj3zwL1sn/A6o8qdChwTpIcea4bJjLAd31KhBTKCJABWKhiEuBVXdId3Od0oyl305cjIGk2HFOIINWcAxe9EL9qZKZaNbXiBljIDF7zmfE3G4tQKgPQN1OMJrJX7GFg/Xd+plBkOXy1e+Tn9/51ymZrL5+poOCtHgqspNYviAFs8nA0bp2qN0JFdeMXWwyy0aCGhWIKYIT1EOspwE0l3TVVygUIVGBWDjWfQScQ3D0kaCnva76YrwOotfTMalKkyHJSDTra8TAY+o6tzVitV0D9A07WV00zpC1dhcUrmfbyhLcGmw/3RH8Bl5W3y0fHcOT3ovD5Wb3uS6uWpSNEIJYk5GVBakMnt8t/3BYeG3gRRVXyzJTd/2kn1dxceJ2azRZhygK0IfVykeVEVMowhI2EBNCxAkh9gkhjgohTgohvuM5fr8QolYIccTzsdpzXAghfiaEOCeEOCaEWON3r3uEEGc9H/f4HV8rhDjuueZnIiCqmUHyK2HBGjj4J1luHIecM6kj1Acl1g9DVlIsnXr2FSAF+3M8IxZUqD/UA23HoPRKli9IoSAtnpdOhilPFqwDU7zMXE2SQ/Xd9NudbFmUPXpsTXEaRV1vohligttolF8tH5WNhWIc7bZhRlxufesKFYgpFGGJJCNmB7ZpmlYJrAZuEEJ4e+2/pGnaas/HEc+xG4FLPB8fB34NIITIAL4FXAZsAL4lhPC+M/0a+Be/626Y8iuLJus+ApbT0PBmwCmLTWe80aiZq9KIhSJsRiwpF2xzWyN2tKmHRLORhTlJvoMNb0oRfOlmhBBcuzSHN85Z9IeggywTlmycUiD2+hkLJoPgioU+S4G1JelcKY5gy10PsUmBF2UvlgGxKk8qxtHQpdMxWb8HDDGyC1ehUIQkbCCmSbwTnWM8H4GpIR+3AX/2XLcXSBNC5APXAy9rmmbVNK0beBkZ1OUDKZqm7dU0TQP+DLxjCq8p+qx4F8SmBhXtW/Rc9Xsapcg5Pj3wnGKUjEQzPYMOHC538AXJ+TDQAS7n9G2icR+88p2o+HMF40hjDysLU8fO06zdJb3qPNYo1y3LY9jhZtdZS+iblV0FliroD1PG1OH1MxbWlqSTHOfrcFuXYWepoZGqRB3dmRBQvhVqXge3zvdJMS8J6iHWuA+qnpSm2H++bdr+XykUFwsRacSEEEYhxBGgAxlMveU5dZ+n/PgTIYQ3GikAGv0ub/IcC3W8KcjxYPv4uBDigBDigMUS5g0rmpgTofK9cOoJGOgaPexwuekeHNE3c00rUh46YcjyjDnqHtTzEvO660/T97txH9x/E7zxP/DAzVF/0xh2uKhq7QvwD6NupxTcx8gJBJeVZ5AcZwrfPVl2lXycRFaswzbMyZY+tizOHnM8q11aU7wyslz/4vKrYcgqy6kKhYdG6yAGAQvS4n0Hz7wg/8/ClK1WFIr5QESBmKZpLk3TVgOFwAYhxArga8ASYD2QAXxl2nbp28dvNU1bp2nauuzs7PAXRJO1H5a/VI48OHrIOjCCpul4iPU0KH1YBGQkzrC7ft0u2WoP4HRE/U3jVGsfDpc2Vqg/aIW2E2MsKGKMBrYtyWH76Q5c7hAJ57xKmZ2dRCC284wcFeWvDwPg3HZ6jZk81ZqOFkQHCUD5FvmoypMKPxqsgyxIi/cNjQdfOVIYp2y1olDMBybUNalpWg/wGnCDpmmtnvKjHfgTUvcF0Az4RyCFnmOhjhcGOT67yF0m5+8dvH+0PBPSVd+bEVOExDv4W99df5q9xPzfJAyhx1lNhqBC/fo9gCZnOfpx3bJcrAMjHKwP4WJvNEHpFZMKxF4/YyE7OZZl+X6msm4XnH+VjtwraLeN0NKr05GZnAc5y5RgXzGGAOsKWzucfg4WXgfb7oV7npqS1YpCMR+IpGsyWwiR5vk8HrgOOO3RduHpcHwHcMJzyVPABz3dk5cDvZqmtQIvAm8TQqR7RPpvA170nOsTQlzuudcHgSej+zKjxLqPgPW8LCvhC8SyxmfE7P3Sd0xlxMKSOTr4e4YyYikFjEoeF90Q9TeNI4095KXEkZfqNwT9+N/BYPKVbzxsWZRNjFHwajgbi7KroLvW15kbAS63xq6zFq66JHus1UrLYRjuIXbx24BxA8DHU3411L8JjqGIn1dxcdNgHRobiO3+qawc3PgD2PwFFYQpFBEQSUYsH3hNCHEM2I/UiD0DPCiEOA4cB7KA//Csfw6oAc4BvwM+BaBpmhX4nuce+4Hveo7hWfN7zzXngeen/tKmgWW3SfH9gT8CITJiqmMyYjJHB3/rWFgk5gBi+rzEvBme2BRwhrDRmCRHm3rGDPqm7g049SS4nfDX28do0pLjYqgsTOPNmq4gd/LDqxObQBn1WFMPPYOOAH0Y514BBPlrbiQuxhA6G1dxNbjsQbuHFfOPwREnnf12n4eYrR0O/AFWvRcylZGrQhEppnALNE07Blwa5Pg2nfUa8K865/4I/DHI8QPAinB7mXFi4qS7+d5fwyvfwdC/EogN1IgpD7GISYuPwSBClCaNJkjKmXJGzOXWuH9PHbevLSQ13m8mXs0O6d5fdBl0npnSc4zHOjBCfdcgd673C8hf/jajGTivkNkva7CxIpNfvnaOvmEHKXHjZvd5yV4KCZmyPOk1eQ3D62csGARsXpg19sS5V6BgLTHJWVQWpnG4IUQgVrJJWhLU7JAjlxTzmkarzIyOZsR2/1TqLa/64gzuSqGYeyhn/YmyYI0c3fHGT7jt6Ce4Mq6GuBjj2DXKVT9iDAZBRqKZTj2xPkTFS2z3uU6+98wpXjzhl1nTNBlUlG+F9FLPWJZQziwT42ijVx/mGW10/DFo3i/LkjpC5o3lmbg12F9rHX87H14tW+3OiPe7o9pCZVEa6Z5SMCCbBpoPjrrprylJ52RLH8MOHS8zc6IMWNXcSQXjrCtsbTIbVnmnyoYpFBNEBWITpafO84mGQXOwxVwdZE2jzBwk5V3Inc1ZpKnr9Lrr7/AMum71F6N3VEmPsvKtsozsHJ60P1cwjjT2IASsLEwFyxl46rMykLnnaV0h85qSdMwmA2+ej6A82dfsczAPQffACEebegK7Jff/XurUPDq8NcXpON0ax5p69W9WsVVaWAx0hn1excXNmEDMmw3b/IUZ3pVCMfdQgdhEKd0sjTgBDUF9ckDVVmrEUgtk5kIRlszEWH37CojKvMnXz8gAq63PT2ju1YeVbYG0Evl5T8OUnsefI409LMpJJknY4dEPytL27X+SJT4dIXNcjJE1xZHoxDx2ErWvh93HrnOdaNo424rGfbDj+/LzF74KjftYUywzd4dClSe9446e//LEPdca98GuH18Yg88L+VzzlPquAZJjTaS5uqRuVmXDFIpJoSKFiVK0AT70DOStQgNESn7gmt4mpQ+bABlJ4QZ/50tD10m66zdaBzlvGQDGZcRqdkDmQllC9jZW9ERnsLWmaVKoX5gCz35Bjsh69+9lgB6GjeVZnGrto0fP5BbkG17ygohsLF6vtpCWEMOqwjTfwaqnZYkdRrVqmUmxlGYmhBbsez3XTjwOD9wSeaCz/w/wx+vh1f+AB26d3gCpcR/88QbY/r3pf655Sle/nSePtLC2NB2htGEKxZRQgdhkKNoAdz2ESzNwR0/g2CN6GlXH5ATISjTTqdc1CR4vMU2WESfBjjOyLLkwJ4k2byDmHJEdjN4MT5QDsfquQXoGHbxbvAZHH4ItX4lY4L5pYSaaBntrQujEhJDlydpdIccOud0ar5+xsPmS7LEjlro9r3OcVm1NcTqHG7r1jV0b9jD6a8M5LHVv4TjwJ3jui7IMqrmn32396MOeIFNTzu7TxA9fqGbA7uSbWzLg4J+g8i414FuhmCQqEJskwwn5/MF1I6u6X5JeTF6cI1LPpDJiEZORGEvfsJMRZ4h5kzBpndjr1R0UZcSzsTzTlxFrPgCOAakPAznsOiHTF6BMkSONPSwTdayv+i/5HFu+HPG1lYVpxMcY2RuJjcVgp5w9qUNVWx+d/faxZcmBLjj7ElxyfYBWbU1JOp39I6MdcQGUbgZTLAjPr45Df4YzLwVf6xyBZz4Pz/y7bHLBEwhOt9t661Hf5wajcnaPMocaunnkQCMfvbKM8urfebJhShumUEwWFYhNEovNzv85b2XYnA4vfcPXvdbXDGiQWhjyeoWPzEjmTcKkdGJ2p4s957vYuiiH/LQ4eoccDI44ZVlSGKD0St/itJKoaMQ0TaPz2As8aP5PRGwSvOv3MiCIELPJwLrSdPacDyOI9zrzhyhPemdXXrXIz7Zi/+/BOQTXfTdAq7amWE4AONigk40r2iADt21fhzv/BlmXwEPvhX2/G7tuoBP+8g7ZSXfFv8FHX4Jl75Ado+9/PCpGny+ebONn28+OPdiwVwbZaz8MpjippVOmolHD5db4xhMnyEuJ43NF5+T3/ZJrVTZMoZgCKhCbJJZ+OzYSaFj5GVn6OPOiPDFq5qoyYpEy6q4/DfMmD9R1MzjiYuvibPI97vZtvcMyEFuwBuL9dFNpxVMqTbrcGs8db+VrP/sDH6r5ImmiH2Hvly74E2RjRSZn2vtDl2zTiiG9TDcQc7s1HjvYxKaKTHKSPc7+I4Ow7zew6EbIWRJwzeK8ZBLNRg7V9+g/b9EGGcAtuQk+/LzMrD33RXj+q9J5/7kvwa82SmuMd/1eBnwGI6x8tzSynUBQqofbrfEfz57ip9vPMmD30w6+dp/0hbv+Phn4Ne336doUU+bBt+o52dLHjzfaifvnh2QJuGaH0uEpFFNABWKTxOuqP7L6HsiogJe/KcXkysx1woy66+tZWCRmy+zVJLzEdlR3YDYa2FiRSV5KPAAWiwWaDvjKkl7SS+T3L4TmKhgDdif3767l6v/ewacePERZ/2GMwi0LcW7npDRKG8szASIrT9a9EbSRYW9NF03dQ7x3vd/P4pEHYbBLZqmCYDQIVhenhe6c9Cc2Ce58EC7/FLz1a7j/7bDvt7K54qYfw6o7fGuLN8nHKGi2dp3rpNE6hMutccDbXFC7SwalV35Oep4tvRmGe6B+95SfTwGd/XZ+9GI1Vy7MYpPxlPzZBvmzp3R4CsWkUYHYJPEGYjmpSXDdd6CzGg7/2ZcRU6XJiMlIDDP422CUo44mkRHbUW1hQ1kGCWbT6LxHV+0b8i/58q1jF6cVg9sR8fP0DTv40Yun2fT9V/n206fITDLz6/et4WPv/4BHDSUmrYdaWZBKUqyJPZH4idn7oO1owKlHDzSSHGfi+uWe0q7LCW/+Ago3QPHlurdcU5xOVWvf2ExTKAxGuOG/YNGNaN75mcIQOKg9MRNylkPd1AOjv71VT3pCDCaDkMGqpsGO/5Lefes+IhdVXAOmeDj97JSfTwHff/40ww4X3751OSLeO8TeMP2aP4XiIkcFYpPEYrMjhCeIWHIzFF0Or/2XNAlNypOCZkVEZHk0YiHd9SfhJdbcM8TZjn62euYr5qXIQCypZZd8gx6vHUorlY8R6sR+8vIZfrXjPJeXZ/D4Jzfyz09dwY0r8zGmezzJFt0Q1LQ1EkxGAxvKMtgbSSAGAeXJ3iEHz59o47bVC3yTH6qegu46mQ3zH/w9jjUl6bg1OSczUo439fLDgbdj12JwE3xqAAClV8gy1hTKhe19w7xS1cEd64qoLEqTgVjt6zLztfkLECMzn5gTYOE1MhCL4sSE+ciBOiuPHWziY5vLWZiTJMeBGWJkE8okf8YVCoVEBWKTxNJvJyPBjMlokG9q198n7RVOPaH0YRMkJS4Go0FE4K4/sUBsR7W0u/AGYvFmI2kJMeR1viVNVccHyxO0sKhus1FZmMZvPrCOtSUZvhPddfJxw8em9Aa1sTyTms4Bn+VGMJJyZJPBoT+P0ek8dbQFu9PNe9d5XpOmSffzzIWw+O0hn3dNkcx2HArlJ4ZsSthzrpP3//4tbvnFG/ylOZeP8S2ezvyw/ptzyRWyW7XlSMh7h+LR/Y243Bp3bSjm8vIMjjX14Np+H6QUwJoPjl285CbZQNNyaNLPN99xutx848mTLEiN4zPbFsrS/aknYdH1cPXXVBCmUEwRFYhNkk6bfeyw78J1sPyd8vORQSVenQAGgyA9wRzeXb9/ooGYhYK0eCqyk0aPLU8aIMdeF1iWBF8AHWFGrL5rkJLMhMATXnF+etmE9juejRVSJ/ZmTYjuycZ90NckRx35Gaz+/UAjS/KSWVGQItfV7oTWI7Dps2EnPqQmxLAwJ4nf7qzhA394i28+eYI/vlHLa9Ud1HUOMOJ08+LJNt7xqz3c/fu3ON1m46s3LmHPV7eRuvgKfjhwk/6bc8kV8rH+jci+COMc8l1ujYf3N7KpIpOyrEQuL8/kCo5ibN7nyYbFjb1+0Q3SK63qmcieTxHAX/bWU9XaxzduXkaC2SQbIGytsOy2md6aQnFRYJrpDcxWGq2DNHYPsqkiK+h5S/+4QAxkl9bJf0LHKenorVL2EZOVZKYrpLt+nsdd3wHGmLD3G3G62XOuk9suLUD4leG2xpySn5RvDbwoJl4OGI8gIzbidNPaO0RJRhCnfGut1EhNsWFjaX4KqfExvHm+i3deqqM5rNsFbk/ZzTkMh/9ClWkJx5p6+dYty3yvffdP5Wtb9d6Invvem5byxOFmajsH+OehZmx+ejEhZIKtOCOB/3jHCm5fWzha/qwsTOXZY6109tvJSgpSnk/KhqzFUid25edCb6JxnxT/uxwe24t/sNOxlOaeIb72dtnxubY4jWTTY/Sa80i99AOB90jIkOXQ08/Ctd+K6LUrfHTYhvmfl85w1aJsbljh0RqeekKOeVt0w8xuTqG4SFCBmA5ffuwY1e02Xv7cVaNdff5YbHbKMhPHHrSel2/A/u7hKhCLCDn4O0wgBlIAHkEjxIF6KwMjLraOG3S91nUUKylk5K4IfmFaSUSmrk3dg7g1KB7/MwCyNJlaCCZz2PuEwmgQXFaWEXrupNdg1WWX0dGhP+OubaLE+C7esdoTJLYdh/Pb4ZpvBWaMdLh6cQ5XL84BZAmya2CEus4BajoHaOgaZFFeMm9fkSdL8354xygda+ph25JcnT1fAcf+LpsHjCbOttu494kT/PftlRT7Zxhrd/m0ZG4nPHQX7Sn3sCDxSt62TP48JNS/ymrDeX4Z+1n+Ve/rveRmORuz86z0PVNEzG9er2HY6eLb3qDeW5ZceA3Epcz09hSKiwJVmtTh27cuxzbs4NtPnwo4p2kalvGlSfANBBchxMqKoGQmxdIVcsyR10sssvLk69UWYoyCTQv9MpqaxqLBg7zhWo7drSPeTiuOqDRZbx0E0C9NTrEs6WVjRSaN1iEaPc8XgL/B6j1P47zqq1R07+Jl8xdIP/C/Mpj55ydlc4K3m3CCCCHISoplXWkG71lXxBevX8ytlQsCgjCAFQWpCAFHG3v1b1hyBYzYRjs9nzvexr5aK59+6BB2p8u3LtkbyAkwxjKSXsGdXb/iBcO/YT78J2nb8fRnscVk8bOu9fpdnktuko9VT0/i1c9fRpxu/nm4mWuX5lLuLe83H5SaO1WWVCiihgrEdFicl8xntl3C00dbePHk2Dd/m92J3ekOLL2Mvineq8qSEyQzMYLSJEQciO2otrC+NIOkWL+kr6WapJFO3nCvoKNPJ+hLL5FD28MMGPcGRiUZQQIxay2kl0a0z3B4S+Mhs2Jeg9WyzbyY9SGuGf5veoq2SXPTB26B9uPSlqPzTFT2FIqkWBMLs5M4Fqrj0jvNwGNjsb/OSkqciWNNvXz/+dO+dS1H5B80V30JPvQMv170R+4auRdzVik8+3m4/yawtZHo6mW5dt7nJzae1EJYcKmysZggr1V3YB0Y4fa1fhnoU0/I78niG2duYwrFRYYKxELwya0VLMtP4etPnKDHb/yO10MsICMGvjdFFYRNiMxEM7Zh59iMiD9J3kAsvMdXS88Q1e22sfMVAWpeA2C3a4Vv5uR40kqkx5itJeRz1HcNEh9jDPwZsNvk/MeM6GTEFuUmkZloDm9j4eHRA424U4vI/NBDng5CT+ZP0y6Y6eaqwjSONfXqDw5PzpPdm/W7cbjcHGro5p2XFvChTaX8aXcdL5xoA8cwHH9UZl623YurYD2P7G/AVLGFuI+/DJXvG72d0NxsMlaFNr9dcrMcfdQX+vuq8PHYwSaykmJ9/480TZYlK7ZBXOrMbk6huIhQgVgIYowGfnj7KroHRvjuM74SZWeoQEwxKTK88yYHdPylErNkyTeCjNjOMxYAtno0TqOc/CfO2DRyRTetvTpDrb0WFmF0YvVdgxRnJIxpBJDX1cnHKJUmhRBcXp7JmzVd+oGNh5aeIXaetXD72kKMBgGXfkCWJC9wqbyyKJWugRGae3S+xiDLk/VvcrJJjqBaX5bB196+hJUFqXz5saN0HvwHDPfCahlw7ajuoKV3mLs3FMtugXUfGn1twmimK/uy8IEYqKxYhHT223ntdAfvWlPgK0E3H5KG1aosqVBEFRWIhWFFQSqf3FrBPw4189pp6Utl6VeBWLTJTJRfS93Zigaj7PqLIBDbUW0hPzWORbk+2wrq90DjWxjtPTxo/k+0hreCX+w1Yw2jE2uwDowVlnuxeq0rSsPuM1I2VmTS2jtMXZeOTszD4web0DS4Y62nW3OGSuWVo4L9EDqx0ivB3kvNCfl92FCaQazJyC/vXoOmQeP236KlFMqh3cDf3mogOzmWa5d5dGPjXlvGkis41tSrrxPLXiyzcCoQi4gnj7TgdGvjypL/lCauYXzoFArFxFCBWAR8ettCFuUm8bV/HKdv2OErTQZrz1dMisykMGOOICIvMYfLze5znWxdnD02W3X0IQAEEIOT5DadQCylUK4KYWGhaRoN1sHg+jBvRixKpUnw8xMLUZ50uzX+frCJjeWZYwPEGSiVL8lPJsYoQjvze/zERmp2UZKZQI5n6kFxZgI/uzGTypEj7Ey8HgwGWnqGeK26g/esKyTGv0HA77VdXp45du7keISQWbG6XTAU4RzNeYqmafz9QCOVhaksyk32HvSUJa+G+LSZ3aBCcZGhArEIiDUZ+dHtlXTYhvnPZ6uw2OzEGAWp8eH9rBSRkemZN6k7+BsiGnN0sL4bm90ZqA8b8mRnhBGnMHFQLA9+A5NZOrSHyIh12OwMO9z6HZPx6VHV0JRnJZKTHBtSsP9WrZUG6yDvWT/zM05jTUaW5qdwLFTnZGoBWnopOV37WV+aMebU1cPbMQiNe+tW8dLJNh7e34gG3Lm+WPd2a0vSfXMn9Vhys7TBOPPSBF/R/OJkSx+n22xjs2Eth+X/CVWWVCiijvIRi5DKojT+5apyfvN6DRXZiWQmxmIw6M/rU0wMb2kyrLt+o04my8OOagsmg+CKcbYVNB+E4o1wyXX84HgGh0cq9G+SVhxSI1bvKREG9RCzRs+6wosQgk0VmbxxTurEAnRp+AZ837giP6rPPVlWFaby5OEW3G5N9/9JX+5lXGp9ls4SvwyL2w1HHsRduoU0WwVf/PtRYmOMXHVJNkXBMpAeEswm39xJPQrWyqaP009DZWTGtvORxw42YTYauKVyge/gqSelqa4qSyoUUUdlxCbA565dRHlWIuctA0ofFmVS4k2YDCKMhUU+DHaBU3/NjuoO1pakkxznl61sOSzHAF36Adj8Bfqz14Se35heEjIjVt81AEhn+QC666JalvSysSKTzn47P3qxmt/trOFPu2v5y956Ht7XwKMHGnnueCu3VvoN+J5hVhWmYbM7qekc0F1z0rySdNHPppQO38H6N6CnHsOaD4zqxSw2O3dfpp8N83J5eQbHQ+nEDAbpKXZuOzhCNBLMY0acbp480sx1y3NJS/AY5GqatK0o3yonFSgUiqiiArEJEBdj5Ie3r0IIJdSPNkII6a4fLiMG0l0/CK29Q5xuswV2S55+RnYOeryP8lPj6LAN43S5gz9PWrE0rdQJ+BqsgxgEFKTFjz3hcsqusigK9b1ctSibpFgTv9pxnvueq+I7T5/iG0+c4Kv/OM6XHzvGiMvNXRvCBysXiko/h309Xh5YCEBBz0HfwcN/hdhUWHozJZmJ/PzuS3n3mkKuWZKjcxcfl5dn4nRrHAw1rHzpzeAYhKf/Tc2DDcKrp9vpHnSMLUu2HpV/YKiypEIxLajS5ARZV5rBj++oZMH4N2HFlMlMig2tEUvyM3VNC5zj+Kqnq/WapePetKuegZJNo3/N56bG4dZk92t+apDvY1oJoMksWkZ5wOn6rkEWpMVjNo37O6a3UWqQolyaBMhPjefot96Gw+XG4XLjdGk43PLR6dKIjTGQmxLZ+KILwcKcJBLMRo419fKuNcF1ay81x/JpUy6Z9bvh8k9Iu4pTT0rLihj5fdm6OCcwsNbBXyd21XiNoBej5w+oY4/AqaeU8fI4HjvYRE5yLJv9S/unnpR/yHgtQBQKRVRRgdgk0HtjUUyNyN31g5u6bq/qoHIn3ewAACAASURBVCgjnkty/GwrOs9CZzWs/+joofxUGbC09g7rBGJ+XmLBAjHroI5Qv04+TkNpEuTsSaPBOGvKj6EwGgQrFqRypDF4Rqy5Z4jmniG6y9aTWb9Hlr9OPC4Hl1/6vqDXhCPBbGJVYWponVjjXmTvrCbnc6p5sKNYbHZeq7bwL5vLfd5h3rJk2VWqLKlQTBOqNKmYNWQmmcOI9fXnTQ6NuNh9rpNrluSOFbN75wt65w0CeSky+NLViYXxEmu0DlKcEWzYd/Q9xOYyqwpTOdXax4gzsAS8v9YKQNwlW+QkAku1LEvmLIMFayb9nJeXZ4b2E/MOSfdScuWkn+ti44nDzbjcGrevLfAdPPowWGsgv3LmNqZQXOSoQEwxa8hINIf2EUvIlJ1bQbzEdp/rxO50B5YlTz8j39hTfVlM/4xYUJIXyFJMEC8x27AD68BI8IyYtVaWvpIXBJ6bh6wqSmPE6eZMuy3g3L46K8mxJvJXXSsP7P+d7Gy99P3S82uShNWJFW2Ae56GRTeC5pbl0Ehp3Ae7fnxRass0TeOxg02sLkpjYY7HO6xxHzz1afn5W/93Ub5uhWI2oAIxxawhM9FMv93JsENn3qTBIHViQTJi20+3k2g2cllZpu9gb7N8c186VtuSlhBDrMlAe59OIGY0ycAtSEbMa12ha+aaXiL3qaCyUHqpBTN23V9rZU1JOsbMMhm47v+DDLJXTc1WIiI/saIN8N6/SC3fK98Gt87Pmz81O+FPb4dX74MHbr3ogpITzX1Ut9u4Y52f7OLcdql5BHA5LtisUoVivqHeMRSzhkzPpILQ7vq5ARoxTdPYXtXBVYuyxwroveNsltwyZr0QgvzUOP2MGOh6iTVYvR5iOmauqiw5SnFGAmkJMQHGrt0DI5zt6GdDWYbMfmUvATQoukzOFJ0CibER6MQAjDFwzTeg4yQc/3votW43PP8lcDvkQHjXyEUXlPz9YCNmk4GbV/llc0f65aMwXNBZpQrFfEMFYopZQ0ZiJGOO8gMyYiea++iw2blmae7YtaefhqxFkL0o4DZ5qXG06Q3+Bl0vsdGM2HgzV00Da920dEzOVYQQrCpMC8iI7a+T+rD1pRkys+QNapr2RyXTFFYn5mXZO6X26dX7wBmiW/fV74LltO/fF1lQYne6ePJIC9cvz/NNC3G7ofp5yFoC276uuksVimlEBWKKWUOWZ96k7uBvCDrmaPvpdoSArYv9LAsGrVC3W7flPj81PkxGrERq0cYZfzZYB8hMNJMUO67heNAKI7Zp65icq1QWpnK2o5+hEV/5b3+dFbPRwKrCVBmEaR4xv9sVlUxTRH5iIEvI134HehtkaTQYh/8Kb/wE1n4YNnr0Ujf/70UVlGyv6qB3aJx32NmXwHoetnzpgs8qVSjmGyoQU8waMhIjKU3mwZB1TAZje1UHlxalkeU/hP3MC7KMtDR4IJaXGkd73zButxb8edI8nZO9TWMO13cN6pclQZUmx7GqMA2XW+Nki688ua+um8qiVGnDUbpZZpiEMWqZpoh0Yl4qrpaO8Tt/FCjcr3sDnv53KL8a3v4juPLzUsfWcXLKe5xNPH6widyUWK709w7b+0s5c1WZuCoU044KxBSzhkxPRiykhYW/qSvQ3jfM8ebewLJk1TPyjUTHCiEvJQ6HS9P3LfP3EvOjvmswuFDf6g3EVEbMH59gXwY5gyNOTjb3+gZ9F22QZa9t90at/JUYa2JNcTpPHG6mb9gR/oJrvw1DVpy7fsrXnzjOieZe6DoPj7xf+sjdcb/UlCVmwsLrpKYsEoH/HKCz386OMxbecWkBRu9M0LbjULsTNnxcvm6FQjGtqEBMMWtIjjURY4xg3iSMBmJB3fRHBuD8dukdpmOFkOexsAjvJeYLxEacblp7h/RnTPpfpwAgJyWOvJS40VFHhxt6cLo11pf5mYMWbYh6+etrb19Cu83Od546FX7xgkth+btwv/lLXtx7lMffOA4P3gEIuPsRiPcbSr7qPbJZ5CIR6z91pAWXW+Pd/ibVe/8PYhJg7T0ztzGFYh6hAjHFrEEIQWZiLF3hNGIw6iW2vaqDgrR4Fucm+9ac2y4d2pfeEuQGEp+XmI5gPylPlsr8ArGm7kHcGhSPF+qDLE0m54+O5lH4WFWYyjFPRmxfrRUhZPlwOrm0OJ1/3VrB44eaeOFE8EkM/uws+gTC5eD75j/wsaqPoPXUw51/C9T8Lb4RYlPg6CPTtPMLy+OHmlhZkMoi7/+f/g44/iisvhvip/d7pFAoJCoQU8wqwpq6+mXEhh0u3jhn4ZqlOWPd9E8/A/EZULxJ9zajGTE9LzGDAVKLxnRO1lu9HZM6pUlVlgxKZVEatZ0D9A462F9nZWleCilx01/y+sw1l7CyIJWv/eM4HTb9xoym7kE+/UIP+8yXcY3hEAW0AwIMQUZJxcTDsluh6ikYGZy+zV8ATrf1cbKlj3et8XPS3/8Hac9x2SdnbmMKxTxDBWKKWUVmkpnOkO76GVLYfepJTrz1MsMO91h9mHNECvUX3yiNWXXISozFZBD6pUkI8BJrtIYzcy3Vv9c8ZpVHJ3aooZvDDT3SP+wCEGM08JP3VjI44uKrjx9H0wIbM5wuN//28BHcGixftQ7vCs3t1i8/rrpTemxVPzd9m78A/PNQMyaD4NZKj3eYYxj2/x4uuR6yFs7s5hSKeYQKxBSzisxEM9aBEKXJpv3S7qB+N2teuYtPmp/nsmK/smTdLtn9pmNb4cVgEOSmxIUOxMZ5idV3DRIfYyQ7OXbsOscQ2FqUdYUOqwqkxupv+xoYcrh8Qv0LwMKcZL5ywxJePd3Bw/sbA87/5JUzHKzv5r53riBt9S0IUzxODDiESb+Ds+QK2QhybO6WJ50uN/883MzWxTmjRsqceEzO/dz4qZndnEIxz1CBmGJWkZkUG7prsm7XqABf4OYrhr8Q97+L4clPS23Yvt+BIQZik/Xv4SEid/3BTrBLh/H6rkGKMxLGlkHBlzVTpcmgpCbEUJaVyCtV7QCsL7uw2qMPbSrlioWZfO+ZU9R3DYwef+NsJ7/acZ73rCvkttUFox2cOxZ8nHuc9zKctzb4DQ0GWHmH/Hnrt1ygVxFddp/vosNm9w341jR481eQsxzKtszs5hSKeYYKxBSzioxEM4MjrjEGoGMo3QzGWDRhZFgzc2TJ52HR9XDyCfjru+DM83IUzYN3hHVpz0uN09eIgZ+XmMykNFgHdDzE6uSjKk3qsqowFU2D0swEcpLjLuhzGwyCH91eidEg+PyjR3G5NSw2O5979AgV2Ul8+9blvsVFGzBc9Xn2OhZyoC6EIWzlndKn7sTj0/8CpoHHDzaRGh/D1Us83ca1r0t/tMs/OaWh6wqFYuKEDcSEEHFCiH1CiKNCiJNCiO94jj8ohKgWQpwQQvxRCBHjOb5VCNErhDji+fim371u8FxzTgjxVb/jZUKItzzHHxFCmKfjxSpmP153/S698qQna/FmySd438j/R8FNX4V3/Ra+dG7swOgI5gHmpcTR2jsUVDsE+AKx7no0TaPBquMh5jVzVaVJXVYVyvLkhSxL+rMgLZ7v3racg/Xd/N/r5/nC34/SO+TgF3dfSoJ5rJbw8vJMYoyCnWdDZLtylkLeyjlZnrQNO3jxZBu3Vi4g1mSUf7A890WIS5OZPoVCcUGJJCNmB7ZpmlYJrAZuEEJcDjwILAFWAvHAx/yu2aVp2mrPx3cBhBBG4JfAjcAy4C4hxDLP+h8AP9E0bSHQDXx06i9NMReJyF2/aAM/GLgJV+EGn14rJg7WfwxM8RG7tOelxjHscNM7pGP6Oeol1kCHzc6ww63fMWlOhoTMcC9v3nJpsQzELiufua/RO1YX8PaVefzoxWp2nrHwzZuXsSQvJWBdgtnEupIMdp4JU3ZcdSe0HILOs9O04+nh+eNt2J1u2S3ZuA8euFm+hpF+aDs209tTKOYdYQMxTdLv+WeM50PTNO05zzkN2AcU6t5EsgE4p2lajaZpI8DDwG1CCm62AY951j3A/9/encdXVd77Hv/8Mg9AIAMJkJAQBgVUFBFBBQVxrHW4pVdtj0Or9Z62Hm/HV+1pz/F28NrT03t6b2t7Wqu2eNqXQ9U6FaSocLCVQQ4KEiYZkhCGJCSMCUlI9nP/WGvDJslOAiSsvcP3/XrllZ1nPWutZ62Vnfz2M8Ktp3At0g/0ZHb9mkNNrNmxnznnDj1xw0nO0j4sy5vzK2o/scw8L7DbX3Fsse/O5xAr95ol1aQT1UVFg/nt5y7h1guHB1YGM+PRW8+ncEg6N08azmcvHRk178xxeWzcc4iarpquz/sUWELc1Yq9uLqK0rxMLiwa7NUah5cLc67fTFQrEk961EfMzBLN7EOgBljknFsRsS0ZuAt4M2KX6X5T5gIzC3fAGAFEDluq8tNygP3OudZ26Z2V4wEzW2Vmq2pr47OTrHQtJzPcNBk9EFt8bDb9/I4bT2KW9m5n1zfzOuzvrzjWyTtq02R2SbfnO5uZGbPOGUpSYrDdUodkpvDO16/i/91xYcdBFxFmjvPWXVz68d7oBxs0zOvYvvZ5L4iJAzvqG1m5vZ5PTS70rj8jz99ivbbWp4icnB79VXTOtTnnLsSr9ZpqZudFbP4lsNQ5F/4otRoo9psyfw680luFdc494Zyb4pybkpeX1/0OEnfCQ+m7ml3/7Q01DM9KY/yw7kdGduX47PrdzyVWWd9IYoIxYki7mfNDIW/UpEZMxo2UpIQugzCA8QWDyB2Q2n3z5KQ7vClOKpf3Ygn7zsurd2IGt140wgse1zzr9Q278lu9ttaniJyck/p46pzbDywGrgcws0eAPOBrEXkOhpsynXPzgWQzywV2AkURhyv00+qAwWaW1C5dzkKZKYmkJCVE7SNWc7CJJZtquXZiQbf/TLuTNzCVBOtidn3w+p7t3YxVvc/wwWkkt6/RObQb2po1YrKfSUgwZozN5a9b9hIKdVHbde5NkJgKC7/T7SjdoDnnePmDKqaX5jBicDpsWgCV78HV/wSzvq0gTCQgPRk1mWdmg/3X6cA1wEYzux+4DrjTOReKyF/g9/vCzKb656gD3gfG+iMkU4A7gNf8PmaLgbn+Ie4BXu2tC5T44q03mRK1aXLesnKOhkLce1nJaZ8rOTGBvIGp7Im23uSOlbDpTWht4sHKrzA7s7xjHo2Y7LdmjsulvqGFsl0Ho2eqWQ+hVtj1XzDvkzEdjK2u3EdFXSP/bXIhtLXCW49AzliYrMW9RYLUkxqxYcBiM1uLF0wtcs69AfwKyAeWtZumYi6wzszWAD8D7vD79LcCDwILgQ3AC865Mn+fbwFfM7MteH3Gnuql65M4lDMgpdOmyYbmVn6/vJLrJhRQkttJp/lTUJCVHr1psvxdb64oIJmjXJGwrmOeej8QU9NkvzNjrNf9octpLMrfhfDCSK1NsH1p3xfsFL34XztJT07khvMKYPU82LsZrvkeJPb9up8iEl30xfh8zrm1wEWdpHe6r3PuceDxKNvmAx0WaHPObcMbVSlCdmZqp02Tf1y1gwNHjvKFmaW9dq6CQalsq23ofGN48ti2ZsyFGBWq7JhnX7k3XUZWd4OGJd7kDkhl4vBB/OfmWr48K8rai/7vCK1NgIPqss7zBexA41HeWLuLG84rIJMjsOQxGDkdzrkx6KKJnPU0s77EnNzMFPa2m76itS3EU3/bzsXFQ7i4uPeWyBmWlR591KQ/HUb1lG8wv/USxtQugm1LTsyzbzsMLlKtQj81Y2weqyv2cbi5tfMM4SlTrv4nmHAblL0My35xZgvZhaNtIea9V85VP1lMQ3Mrn502Ev72M2iohWt/qClXRGKAAjGJOdmZKR1qxBaWVbOj/ghfmNG7TYAFWWkcam7lUFOUSV2LprJ65Of5WuuXaM4qhVe+BEf2H99ev13Nkv3YzHG5tIYcy7bWRc8UnjJl7lMw/mZY+I9Q9qczV8hOOOd4a3011/3fpTzyWhnjhw3i9X+4gouHNMOyx2HibVA4JdAyiohHgZjEnJwBqRw52kZji1cL4ZzjiaVbKc7J4JoJBb16rvAUFtVdjJysqGukiVTabv0VHNoDC751fGN4Mlfply4uHkJGSmL301gAJCR6y20VTYOX/wdUvNf3BezEup0H+MxvVnD/M6sAePLuKfzh/kuZODwLlvxvaDsKV/9zN0cRkTNFgZjEnGOTuvrNk++X72NN1QHuv2IUiQm925RSMKj7ucQq6xvIyUwhY9SlMPMbsPY5WP8qNB2AI/UaMdmPpSYlMq00p+sO+5GS0+HOZ7355569E2o39W0BIzjneOTVdXzy8b+ycc9Bvn/LRBZ+ZSZzJuR7U73UbIAPfu8tBZbde/0sReT0KBCTmHNsmSO/efKJpdsYkpHM3IuLutrtlHS7zBFejdjI8BqTM78Jwy+C178Clf4CE2qa7Ndmjs2loq7x2OoK3crIhr970es3+LtPwlvfOyPTWlTtO8K8ZRXcMmk4S745i7unl5w4791rD0FCMoyZ0+dlEZGeUyAmMSfbrxGrb2hma+1h3tpQzV3TiklPSez1cw0d5M3kH7XDPl4gdmxpo8RkuO0JONoIL/nr3DcfjrqvxL+Z48LTWHSx3FF7Q0rg6kegoRr++m/w9PXw0hdg5W9g43zYvRY+fguW/p/Og7QdK+HdKNuiCK+HevslI8lKbzd45MPnoGoltLXA838X0/OdiZxtup2+QuRMy/WXOdp7uIVF62tISUrgruklfXKutOREcjJTos6u39IaYveBI4zMiZieIm8cXPx5WPFL7+c/fxVyx2hm8n5qVG4mhUPSWbq5lrumFfd8x4YavM+6IW8+unUvwkcvdJ53SCkMLoSMHG+C2E0LwIW8qTF6uPRQuV9jV5LbyXqom97wXzgvGCt/V7+vIjFCNWISc8I1Yh9XH+Kl1VV8avII8gam9tn5CrLSotaIVe1rJOQ6Wew7M+f467aj/sSe0h+ZGTPG5rFsax1H20Ld7xBWMgOSUr155pLS4d4F8PXNcP87cP6ngXB/R4PEJGhthj0feTVloVYvEAsHTT1QUddAalIC+QPTOrkI/zO3JWpxb5EYoxoxiTkZKYmkJSfw++WVtLSGuO+Kvu1YXDAojV1RArGKeq+5pzinXSA2aiYkpXlBmP6x9XtXjsvl2ZWV/OQvm3ho9lgyU3vwpzM8x1j5u97vR7gGamA+uAdgwxteoJWYArf84vj2HSvhdzd5a5hCj3+3yusaKc7JIKGzAS2NeyH3HJh0+4llEZHAKRCTmOOtN5nKzv1HmDN+KGOGDujT8xVkpbG6ct8JadUHm3h9zS7+uKoK4Hhn/bCiqXDP6x3/yUq/NOvcodx0wTB+/Z/beHn1Tr5+zTg+PaWo+1G8RVM7/92IFqSFt937hreQeNX7XqDWAxV1DRTndLL0l/Nn/B9/kzffmYjEFAViEpNyBqSwc/8RvjCj74fZD8tKY1/jUfYcaOKdjTW8tmYnK7bX4xxMHD6IH9wykaGdNfdE+ycr/U5qUiKPf2Yyn7u8nkf/vIGHX/6I3/6tnG/feC5XjsvzpoeIwjnX+faufn+KpsJn/wiPT4H534DP/wUSovckCYUcFXWNXOkPLDjBoT3eNCv553V3mSISAAViEpPG5Q8kPTmRqaOy+/xcBf4UFpf96G1CDkpzM3lo9lhuvnA4o/P6tjZO4svFxdm89MXLWLBuDz9asJF7f/s+V4zJ5R9mj+Fom6OyvpHK+kZ21DeyY5/3uqU1xIyxucwZn8/V4/OP9YHsVvpguOYH8Mrfw4e/h8l3R81ac6iZ5tZQ5zVi4fUv8yeewhWLSF9TICYx6V/nXkBbKEpNQi+7dFQ2F40czCUl2dw8aTgThw86I+eV+GRm3Hj+MOaMz+c/llfw83c+5vYnlh/bnpxoFA7JoCg7gwsKs2gLweKNNSwsqybBvNn654zP55oJ+ZR2F+hPugNWz4NFj8C5N3lzlHUiPGKyQ19GgOp13vehE07pekWkbykQk5hkZiQlnplgqCg7gz996fIzci7pP1KSErjvilHMnVzI0o9ryRuYysjsDPIHpXXoO+acY93OgyzaUM1b66t5bMFGHluwkavOyePpey7pvIM9eIty3/gT+PVMeOcHcNNPO80Wnmy2JFqN2KARUYM4EQmWAjERkdOQlZHMJycN7zKPmXF+YRbnF2bxtWvGUbWvkT+sqOTfl2zl1TU7ue2iwug7F5wHUx+AFb+Ci+6CEZM7ZCmvayQ50Y6tnXqC6jI1S4rEMM0jJiJyhhUOyeCb157DBYVZ/PjNTRxpaet6h1nfhsw8+PPXIdRxLrOKugaKhmSQlNjuT3prC+zdpEBMJIYpEBMRCUBCgvHdT0xg94EmfvPutq4zp2XBtT+EXavhg2c6bC7f29h5/7C9m73JYTViUiRmKRATEQnI1FHZ3HBeAf++ZCvVUZbZOuaC/w7Fl8PC78I7Pzy2XqRzLvocYhoxKRLzFIiJiATo4RvOpS3k+NeFm7rOaAaT74GWQ7D0JzDvZtixkr2HW2hoaaMk2ojJxBTIGdM3hReR06ZATEQkQMU5mdx7eQkvra5i3c4DXWc+WOW/OL54d3jEZHFulBqxvHMgMbl3Cy0ivUaBmIhIwL48awxDMlL4wRvrcc5Fz1gyAxL8we6JyVAyg/I6fz3U9gvTgz9i8vw+KLGI9BYFYiIiActKT+arc8ayYns9f1lfHT1j0VS4+Rfe6+lfhqKpVNY1kGDeSMwTNOyFw3vUP0wkxikQExGJAXdOHcnYoQN4bP4GWlo7TlFxzKTbIXMo7CsHvDnERgxJJyWp3Z9zddQXiQsKxEREYkBSYgLf+cR4yusaeWZZefSMZjB6NmxbAqEQFXUN0WfUB01dIRLjFIiJiMSIq84Zysxxefzs7Y/Z19ASPePo2dBYB3vWUl4XZQ6x6jKv5mxAXt8VWEROmwIxEZEY8t1PjOdwcys/f2dL9EylVwFwZOMiDhw5GqVGbJ2aJUXigAIxEZEYMi5/INdOKODNdbujj6AcmA/559P68dsAHSdzbWuF2o0KxETigAIxEZEYc/mYHHYdaKKyvjF6ptGzyNzzPuk0dZzMtX4btDapf5hIHFAgJiISY6aPzgFg2da66JlGzybBtXJpwgaK2s8hVqMRkyLxQoGYiEiMGZ03gLyBqSzb1kUgNnI6Ry2FG9LXk5aceOK26jKwRG9WfRGJaQrERERijJkxrTSHZVvrovcTS06jLPk8LrePOm6rLoPccZCU2rcFFZHTpkBMRCQGTSvNpuZQM9v3NkTNs7j1PApbK+FA1YkbNGJSJG4oEBMRiUHTS/1+YlGaJw81HeXNIxO8H7YuPr6h6QDsr1QgJhInFIiJiMSgUbmZ5A9Kjdphv6KukU2uiKa0PNj6zvENNRu87xoxKRIXFIiJiMQgM2N6aQ7Lt9V32k+soq4RMJqKrjy23BHgNUuCasRE4oQCMRGRGDV9dA57DzezpeZwh23ldV7fsbRz58CRetizxttQXQZpg2HQ8DNZVBE5RQrERERi1DS/n9jyTvqJVdQ1kDcwlbRzrvYSws2T1WVes6TZmSqmiJwGBWIiIjFqZHYGw7PSOu2wX17XSHF2BgwYCgXnex32QyGoXq9mSZE4okBMRCRGmRnTRnv9xEKhE/uJVdQ1HF9jcvRsqFwOtRug5ZACMZE4okBMRCSGTS/Nob6hhc01h46lHWlpo/pg8/E1JkfPhtBRWPFr72eNmBSJGwrERERiWGfrToYXAy/O9WvEiqZBUhqseQ4wGHrumS6miJwiBWIiIjGscEgGRdnpJwRi4RGTx2rEktOg+HJoa4bsUkjJDKKoInIKug3EzCzNzFaa2RozKzOz7/npo8xshZltMbPnzSzFT0/1f97iby+JONa3/fRNZnZdRPr1ftoWM3u49y9TRCR+TRuVw4rtx/uJVfiBWHF2RMA1erb3PTEFdqw800UUkVPUkxqxZmC2c24ScCFwvZlNA/4F+KlzbgywD7jPz38fsM9P/6mfDzObANwBTASuB35pZolmlgj8ArgBmADc6ecVERG85skDR46yYc9BwBsxOSQjmayM5OOZMvO877UbYd7NCsZE4kS3gZjzhGcTTPa/HDAbeNFPnwfc6r++xf8Zf/vVZmZ++nPOuWbn3HZgCzDV/9rinNvmnGsBnvPziogIHfuJnTBiMuxAFWCAg7YWKH/3zBZSRE5Jj/qI+TVXHwI1wCJgK7DfOdfqZ6kCRvivRwA7APztB4CcyPR2+0RL76wcD5jZKjNbVVtb25Oii4jEvWFZ6ZTkZByb2LV8byPF4f5hYaNmeB32LdFrniyZEUBJReRkJfUkk3OuDbjQzAYDfwICGZLjnHsCeAJgypQpHRdfExHpp6aPzuGNtbs50tLGrgNHKM4pPDFD0VS45zWvJqxkhveziMS8HgViYc65/Wa2GJgODDazJL/WqxDY6WfbCRQBVWaWBGQBdRHpYZH7REsXERG85Y6eXbmDN8t241zEiMlIRVMVgInEmZ6Mmszza8Iws3TgGmADsBiY62e7B3jVf/2a/zP+9necc85Pv8MfVTkKGAusBN4HxvqjMFPwOvS/1hsXJyLSX0z31518bqXXk6NDHzERiUs9qREbBszzRzcmAC84594ws/XAc2b2Q+AD4Ck//1PAf5jZFqAeL7DCOVdmZi8A64FW4Mt+kydm9iCwEEgEnnbOlfXaFYqI9ANDB6VRmpfJiu31QJQaMRGJO90GYs65tcBFnaRvwxvx2D69Cfh0lGM9CjzaSfp8YH4PyisictaaXprDttoGBqYmkZ2ZEnRxRKQXaGZ9EZE4EZ7Gojg3A29WIBGJdwrERETixDS/n5j6h4n0Hyc1alJERIKTOyCVB2aWcumo7KCLIiK9RIGYiEgc+ccbxwddBBHpRWqa/iNLewAABoJJREFUFBEREQmIAjERERGRgCgQExEREQmIAjERERGRgCgQExEREQmIAjERERGRgCgQExEREQmIAjERERGRgCgQExEREQmIAjERERGRgCgQExEREQmIAjERERGRgCgQExEREQmIAjERERGRgCgQExEREQmIAjERERGRgCgQExEREQmIAjERERGRgJhzLugynBIzqwUqgi5HjMsF9gZdCDlpem7B0v3vX/Q8g6X7f1yxcy6vfWLcBmLSPTNb5ZybEnQ55OTouQVL979/0fMMlu5/99Q0KSIiIhIQBWIiIiIiAVEg1r89EXQB5JTouQVL979/0fMMlu5/N9RHTERERCQgqhETERERCYgCsTPIzIrMbLGZrTezMjP7n356tpktMrOP/e9D/PTPmtlaM/vIzN4zs0kRx3razGrMbF0357zezDaZ2RYzezgi3czsUTPbbGYbzOyhKPs/6O/rzCw3Iv0qMztgZh/6X/98uvcnVsXYc3s34p7vMrNXouw/ysxW+Ps/b2YpfvpMM1ttZq1mNrc37k9fitN7f9a/Z7oSY8/0av/98KGZ/dXMxkTZ/2L//FvM7GdmZn76p/1rCJlZXIwMjNP7/6iZ7TCzw+3S7zWz2oj31P2nc28C45zT1xn6AoYBk/3XA4HNwATgx8DDfvrDwL/4ry8DhvivbwBWRBxrJjAZWNfF+RKBrUApkAKsASb42z4HPAMk+D8PjXKMi4ASoBzIjUi/Cngj6Ht6tj23dvleAu6OcowXgDv8178Cvui/LgEu8J/93KDvbT+992f9eyZenql/7vH+6y8Bv4tyjJXANMCABcANfvp44BxgCTAl6Hvbj+//NL/ch9ul3ws8HvQ9Pd0v1YidQc653c651f7rQ8AGYARwCzDPzzYPuNXP855zbp+fvhwojDjWUqC+m1NOBbY457Y551qA5/xzAXwR+L5zLuQfryZKmT9wzpWfzHX2NzH23AAws0HAbKBDrYz/aX028GInZSt3zq0FQt1fefDi7d775znr3zNdibFn6oBB/ussYFf7nc1sGDDIObfcef/9n4ko2wbn3KaeXnssiLf7759nuXNud48vMs4oEAuImZXgfXJeAeRH/JLtAfI72eU+vE9iJ2MEsCPi5yo/DWA0cLuZrTKzBWY29iSPDTDdzNb4+088hf3jTgw8t7Bbgbedcwc72T8H2O+ca+1i/7gTJ/e+O2fde6YrMfBM7wfmm1kVcBfwoyj7V0XZP67Fyf3vzqf8ptMXzazoFPYPnAKxAJjZALymja+0/2Puf+Jy7fLPwnsDfKsXi5EKNDlvxuPfAE+f5P6r8ZZrmAT8nCi1A/1JjDy3sDuBZ/vguDGpn9z7s+4905UYeaZfBW50zhUCvwX+rRePHdP6yf1/HShxzl0ALOJ4jV5cUSB2hplZMt4v/x+ccy/7ydV+9Xe4GrwmIv8FwJPALc65um6OXRTRafHvgZ1A5CeEQj8NvE8l4fP/Ca/fEGa20N//ya7O5Zw76Jw77L+eDyRHdkzub2LoueHf56nAnyPSIp9bHTDYzJI62z/exNm9j+pse890JRaeqZnlAZOccyv89OeBy8wsMWL/7/v7F7bf/xQvPSbE2f2PyjlX55xr9n98Eri4B5cfc5K6zyK9xe+78xSwwTkXGfm/BtyDVy17D/Cqn38kXrB0l3Nuc3fHd87tAC6MOF8SMNbMRuG9Ge4APuNvfgWYBWwHrsTrNIlz7roeXksBUO2cc2Y2FS+o7/INGq9i7LkBzMXr9N0UcYwTnpuZLfbzPRdZtngTj/e+i2s5a94zXYmhZ7oPyDKzcf5xr/HL1Ba5v3+Mg2Y2Da8J7268Gs24FI/3v4trGRbRnHozXn+3+ONiYMTA2fIFXIFX3bsW+ND/uhGvT8/bwMfAW0C2n/9JvF/WcN5VEcd6FtgNHMWr3bovyjlvxAuytgLfiUgfjPep/iNgGd4nk872f8g/fiteR8on/fQHgTK8ETDLgcuCvr9nw3Pzty0Bru+mzKV4I722AH8EUv30S/zzNuAFAWVB399+eO/P+vdMvDxT4Da8v4Fr/GdbGmX/KcA6f//HOT4Z+m3+eZuBamBh0Pe3n97/H/vHD/nf/5ef/ljEe2oxcG7Q9/dUvjSzvoiIiEhA1EdMREREJCAKxEREREQCokBMREREJCAKxEREREQCokBMREREJCAKxEREREQCokBMREREJCAKxEREREQC8v8Bg++i/7+4SCIAAAAASUVORK5CYII=\n","text/plain":["<Figure size 720x576 with 1 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","metadata":{"id":"i_vzV5uI16rA"},"source":[""],"execution_count":null,"outputs":[]}]}