{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 基于 Keras Applications 的预训练模型在隐语联邦学习环境下的微调\n",
    "## 引言\n",
    "预训练模型加载和精调在机器学习中非常重要。一般来说，从头训练一个非常大的模型，不仅需要大量的算力资源，同时也需要耗费大量的时间。所以在传统的机器学习中，使用预训练模型，然后针对具体的任务做微调和迁移学习非常普遍。同样的，对于联邦学习来说，如果能够加载预训练模型进行微调和迁移学习，不仅能够节省参与方的算力资源，降低参与方的准入门槛，同时也能够加快模型的学习速度。\n",
    "\n",
    "得益于隐语联邦学习模块优异的兼容性，使得其可以直接加载TensorFlow.Keras的一系列[预训练模型](https://keras.io/api/applications/)；本教程将基于TensorFlow.Keras的[InceptionV3](https://arxiv.org/abs/1512.00567)的[微调教程](https://keras.io/api/applications/#finetune-inceptionv3-on-a-new-set-of-classes)展现如何基于TensorFlow.Keras的预训练模型在SecretFlow的框架下进行微调，充分展现SecretFlow的易用性。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 加载数据集\n",
    "### 数据集介绍\n",
    "Flower 数据集介绍：flower 数据集是一个包含了 5 种花卉（雏菊、蒲公英、玫瑰、向日葵、郁金香）共计 4323 张彩色图片的数据集。每种花卉都有多个角度和不同光照下的图片，每张图片的分辨率为 320x240。这个数据集常用于图像分类和机器学习算法的训练与测试。数据集中每个类别的数量分别是：daisy（633），dandelion（898），rose（641），sunflower（699），tulip（852）\n",
    "\n",
    "下载地址: http://download.tensorflow.org/example_images/flower_photos.tgz\n",
    "\n",
    "### 下载数据集并解压"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-10-11 07:11:54.892985: I tensorflow/core/util/port.cc:110] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.\n",
      "2023-10-11 07:11:55.019580: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.\n",
      "To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "2023-10-11 07:11:57.008960: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading data from https://secretflow-data.oss-accelerate.aliyuncs.com/datasets/tf_flowers/flower_photos.tgz\n",
      "67588319/67588319 [==============================] - 2s 0us/step\n"
     ]
    }
   ],
   "source": [
    "import tempfile\n",
    "import tensorflow as tf\n",
    "\n",
    "\n",
    "_temp_dir = tempfile.mkdtemp()\n",
    "path_to_flower_dataset = tf.keras.utils.get_file(\n",
    "    \"flower_photos\",\n",
    "    \"https://secretflow-data.oss-accelerate.aliyuncs.com/datasets/tf_flowers/flower_photos.tgz\",\n",
    "    untar=True,\n",
    "    cache_dir=_temp_dir,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 加载数据集"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 1201 files belonging to 5 classes.\n",
      "Using 961 files for training.\n",
      "Using 240 files for validation.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-10-11 07:12:05.321890: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1635] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 12653 MB memory:  -> device: 0, name: Tesla T4, pci bus id: 0000:3b:00.0, compute capability: 7.5\n",
      "2023-10-11 07:12:05.324020: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1635] Created device /job:localhost/replica:0/task:0/device:GPU:1 with 13775 MB memory:  -> device: 1, name: Tesla T4, pci bus id: 0000:3c:00.0, compute capability: 7.5\n"
     ]
    }
   ],
   "source": [
    "import math\n",
    "import tensorflow as tf\n",
    "\n",
    "img_height = 180\n",
    "img_width = 180\n",
    "batch_size = 32\n",
    "# In this example, we use the TensorFlow interface for development.\n",
    "data_set = tf.keras.utils.image_dataset_from_directory(\n",
    "    path_to_flower_dataset,\n",
    "    validation_split=0.2,\n",
    "    subset=\"both\",\n",
    "    seed=123,\n",
    "    image_size=(img_height, img_width),\n",
    "    batch_size=batch_size,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 划分数据集"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_set = data_set[0]\n",
    "test_set = data_set[1]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 查看数据集\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'tensorflow.python.data.ops.batch_op._BatchDataset'> <class 'tensorflow.python.data.ops.batch_op._BatchDataset'>\n"
     ]
    }
   ],
   "source": [
    "print(type(train_set), type(test_set))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-10-11 07:12:05.799561: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'Placeholder/_4' with dtype int32 and shape [961]\n",
      "\t [[{{node Placeholder/_4}}]]\n",
      "2023-10-11 07:12:05.800177: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'Placeholder/_0' with dtype string and shape [961]\n",
      "\t [[{{node Placeholder/_0}}]]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x.shape = (32, 180, 180, 3)\n",
      "y.shape = (32,)\n"
     ]
    }
   ],
   "source": [
    "x, y = next(iter(train_set))\n",
    "print(f\"x.shape = {x.shape}\")\n",
    "print(f\"y.shape = {y.shape}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 单机模式进行微调\n",
    "单机模式下进行预训练模型的微调，基本上参考TensorFlow.Keras的[官方教程](https://keras.io/api/applications/#finetune-inceptionv3-on-a-new-set-of-classes)，并根据数据集格式在编译模型的参数上作适当的修改，但影响不大；"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 微调顶部分类器"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from tensorflow.keras.applications.inception_v3 import InceptionV3\n",
    "from tensorflow.keras.preprocessing import image\n",
    "from tensorflow.keras.models import Model\n",
    "from tensorflow.keras.layers import Dense, GlobalAveragePooling2D\n",
    "\n",
    "\n",
    "# create the base pre-trained model\n",
    "base_model = InceptionV3(weights='imagenet', include_top=False)\n",
    "\n",
    "# add a global spatial average pooling layer\n",
    "x = base_model.output\n",
    "x = GlobalAveragePooling2D()(x)\n",
    "# let's add a fully-connected layer\n",
    "x = Dense(1024, activation='relu')(x)\n",
    "# and a logistic layer -- let's say we have 10 classes\n",
    "predictions = Dense(10, activation='softmax')(x)\n",
    "\n",
    "# this is the model we will train\n",
    "model = Model(inputs=base_model.input, outputs=predictions)\n",
    "\n",
    "# first: train only the top layers (which were randomly initialized)\n",
    "# i.e. freeze all convolutional InceptionV3 layers\n",
    "for layer in base_model.layers:\n",
    "    layer.trainable = False\n",
    "\n",
    "# compile the model (should be done *after* setting layers to non-trainable)\n",
    "model.compile(\n",
    "    optimizer='rmsprop',\n",
    "    loss='sparse_categorical_crossentropy',\n",
    "    metrics=[\"accuracy\"],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-10-11 07:12:13.444936: I tensorflow/compiler/xla/stream_executor/cuda/cuda_dnn.cc:424] Loaded cuDNN version 8600\n",
      "2023-10-11 07:12:14.278747: I tensorflow/tsl/platform/default/subprocess.cc:304] Start cannot spawn child process: No such file or directory\n",
      "2023-10-11 07:12:16.692342: I tensorflow/compiler/xla/service/service.cc:169] XLA service 0x561a8438d460 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:\n",
      "2023-10-11 07:12:16.692403: I tensorflow/compiler/xla/service/service.cc:177]   StreamExecutor device (0): Tesla T4, Compute Capability 7.5\n",
      "2023-10-11 07:12:16.692418: I tensorflow/compiler/xla/service/service.cc:177]   StreamExecutor device (1): Tesla T4, Compute Capability 7.5\n",
      "2023-10-11 07:12:17.022370: I tensorflow/compiler/mlir/tensorflow/utils/dump_mlir_util.cc:269] disabling MLIR crash reproducer, set env var `MLIR_CRASH_REPRODUCER_DIRECTORY` to enable.\n",
      "2023-10-11 07:12:18.304269: I tensorflow/tsl/platform/default/subprocess.cc:304] Start cannot spawn child process: No such file or directory\n",
      "2023-10-11 07:12:18.552866: I ./tensorflow/compiler/jit/device_compiler.h:180] Compiled cluster using XLA!  This line is logged at most once for the lifetime of the process.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "31/31 [==============================] - ETA: 0s - loss: 112.3790 - accuracy: 0.2206"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-10-11 07:12:21.562516: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'Placeholder/_4' with dtype int32 and shape [240]\n",
      "\t [[{{node Placeholder/_4}}]]\n",
      "2023-10-11 07:12:21.562776: I tensorflow/core/common_runtime/executor.cc:1197] [/device:CPU:0] (DEBUG INFO) Executor start aborting (this does not indicate an error and you can ignore this message): INVALID_ARGUMENT: You must feed a value for placeholder tensor 'Placeholder/_4' with dtype int32 and shape [240]\n",
      "\t [[{{node Placeholder/_4}}]]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "31/31 [==============================] - 15s 175ms/step - loss: 112.3790 - accuracy: 0.2206 - val_loss: 34.8170 - val_accuracy: 0.1500\n",
      "Epoch 2/50\n",
      "31/31 [==============================] - 2s 57ms/step - loss: 18.9387 - accuracy: 0.2737 - val_loss: 21.3622 - val_accuracy: 0.2625\n",
      "Epoch 3/50\n",
      "31/31 [==============================] - 2s 57ms/step - loss: 14.2041 - accuracy: 0.3091 - val_loss: 16.4007 - val_accuracy: 0.2625\n",
      "Epoch 4/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 11.6635 - accuracy: 0.2716 - val_loss: 14.6091 - val_accuracy: 0.2167\n",
      "Epoch 5/50\n",
      "31/31 [==============================] - 2s 57ms/step - loss: 8.6897 - accuracy: 0.3215 - val_loss: 7.3476 - val_accuracy: 0.4042\n",
      "Epoch 6/50\n",
      "31/31 [==============================] - 2s 57ms/step - loss: 8.0223 - accuracy: 0.2914 - val_loss: 3.7781 - val_accuracy: 0.3417\n",
      "Epoch 7/50\n",
      "31/31 [==============================] - 2s 57ms/step - loss: 6.0596 - accuracy: 0.3413 - val_loss: 6.9468 - val_accuracy: 0.2417\n",
      "Epoch 8/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 4.8735 - accuracy: 0.3486 - val_loss: 14.5767 - val_accuracy: 0.2583\n",
      "Epoch 9/50\n",
      "31/31 [==============================] - 2s 57ms/step - loss: 3.7563 - accuracy: 0.3954 - val_loss: 7.5974 - val_accuracy: 0.2625\n",
      "Epoch 10/50\n",
      "31/31 [==============================] - 2s 58ms/step - loss: 3.1183 - accuracy: 0.3871 - val_loss: 9.8358 - val_accuracy: 0.2583\n",
      "Epoch 11/50\n",
      "31/31 [==============================] - 2s 58ms/step - loss: 3.5958 - accuracy: 0.3725 - val_loss: 3.7865 - val_accuracy: 0.2792\n",
      "Epoch 12/50\n",
      "31/31 [==============================] - 2s 58ms/step - loss: 2.4401 - accuracy: 0.4287 - val_loss: 5.4358 - val_accuracy: 0.2833\n",
      "Epoch 13/50\n",
      "31/31 [==============================] - 2s 57ms/step - loss: 2.2799 - accuracy: 0.4194 - val_loss: 4.6291 - val_accuracy: 0.2208\n",
      "Epoch 14/50\n",
      "31/31 [==============================] - 2s 58ms/step - loss: 2.3558 - accuracy: 0.4204 - val_loss: 4.1438 - val_accuracy: 0.2125\n",
      "Epoch 15/50\n",
      "31/31 [==============================] - 2s 57ms/step - loss: 1.8603 - accuracy: 0.4828 - val_loss: 6.8592 - val_accuracy: 0.2917\n",
      "Epoch 16/50\n",
      "31/31 [==============================] - 2s 58ms/step - loss: 1.8960 - accuracy: 0.4880 - val_loss: 2.8224 - val_accuracy: 0.3042\n",
      "Epoch 17/50\n",
      "31/31 [==============================] - 2s 58ms/step - loss: 1.7029 - accuracy: 0.4984 - val_loss: 3.8373 - val_accuracy: 0.2250\n",
      "Epoch 18/50\n",
      "31/31 [==============================] - 2s 58ms/step - loss: 1.4418 - accuracy: 0.5120 - val_loss: 3.6055 - val_accuracy: 0.3292\n",
      "Epoch 19/50\n",
      "31/31 [==============================] - 2s 57ms/step - loss: 1.5190 - accuracy: 0.5245 - val_loss: 3.9276 - val_accuracy: 0.3458\n",
      "Epoch 20/50\n",
      "31/31 [==============================] - 2s 58ms/step - loss: 1.3073 - accuracy: 0.5619 - val_loss: 6.1296 - val_accuracy: 0.1875\n",
      "Epoch 21/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 1.3950 - accuracy: 0.5390 - val_loss: 3.7171 - val_accuracy: 0.2375\n",
      "Epoch 22/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 1.1851 - accuracy: 0.5640 - val_loss: 5.6681 - val_accuracy: 0.1708\n",
      "Epoch 23/50\n",
      "31/31 [==============================] - 2s 58ms/step - loss: 1.3379 - accuracy: 0.5838 - val_loss: 2.4407 - val_accuracy: 0.2625\n",
      "Epoch 24/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 1.1016 - accuracy: 0.6098 - val_loss: 3.4062 - val_accuracy: 0.2917\n",
      "Epoch 25/50\n",
      "31/31 [==============================] - 2s 58ms/step - loss: 1.1200 - accuracy: 0.5931 - val_loss: 2.9323 - val_accuracy: 0.3042\n",
      "Epoch 26/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 1.0031 - accuracy: 0.6389 - val_loss: 2.8517 - val_accuracy: 0.2708\n",
      "Epoch 27/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 1.0253 - accuracy: 0.6306 - val_loss: 4.8238 - val_accuracy: 0.3000\n",
      "Epoch 28/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 1.0335 - accuracy: 0.6576 - val_loss: 3.0405 - val_accuracy: 0.2958\n",
      "Epoch 29/50\n",
      "31/31 [==============================] - 2s 58ms/step - loss: 1.1657 - accuracy: 0.6181 - val_loss: 3.3026 - val_accuracy: 0.2375\n",
      "Epoch 30/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 0.9623 - accuracy: 0.6629 - val_loss: 3.2407 - val_accuracy: 0.2875\n",
      "Epoch 31/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 0.8993 - accuracy: 0.6920 - val_loss: 2.2036 - val_accuracy: 0.3917\n",
      "Epoch 32/50\n",
      "31/31 [==============================] - 2s 58ms/step - loss: 0.9263 - accuracy: 0.6816 - val_loss: 3.2231 - val_accuracy: 0.2917\n",
      "Epoch 33/50\n",
      "31/31 [==============================] - 2s 60ms/step - loss: 0.8958 - accuracy: 0.6930 - val_loss: 3.6673 - val_accuracy: 0.2583\n",
      "Epoch 34/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 0.8155 - accuracy: 0.7045 - val_loss: 3.7752 - val_accuracy: 0.2667\n",
      "Epoch 35/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 0.8687 - accuracy: 0.6982 - val_loss: 3.5233 - val_accuracy: 0.3708\n",
      "Epoch 36/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 0.9007 - accuracy: 0.7138 - val_loss: 2.5410 - val_accuracy: 0.3542\n",
      "Epoch 37/50\n",
      "31/31 [==============================] - 2s 60ms/step - loss: 0.7482 - accuracy: 0.7378 - val_loss: 3.7791 - val_accuracy: 0.3583\n",
      "Epoch 38/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 0.8524 - accuracy: 0.7534 - val_loss: 2.9299 - val_accuracy: 0.3375\n",
      "Epoch 39/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 0.6633 - accuracy: 0.7607 - val_loss: 3.8067 - val_accuracy: 0.3125\n",
      "Epoch 40/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 0.7970 - accuracy: 0.7336 - val_loss: 3.9137 - val_accuracy: 0.3333\n",
      "Epoch 41/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 0.7073 - accuracy: 0.7648 - val_loss: 2.9778 - val_accuracy: 0.3542\n",
      "Epoch 42/50\n",
      "31/31 [==============================] - 2s 60ms/step - loss: 0.6207 - accuracy: 0.7815 - val_loss: 3.2358 - val_accuracy: 0.3458\n",
      "Epoch 43/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 0.6773 - accuracy: 0.7700 - val_loss: 3.8493 - val_accuracy: 0.3292\n",
      "Epoch 44/50\n",
      "31/31 [==============================] - 2s 60ms/step - loss: 0.5552 - accuracy: 0.7950 - val_loss: 4.1360 - val_accuracy: 0.3292\n",
      "Epoch 45/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 0.5993 - accuracy: 0.8044 - val_loss: 3.3667 - val_accuracy: 0.3625\n",
      "Epoch 46/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 0.6991 - accuracy: 0.7877 - val_loss: 2.8805 - val_accuracy: 0.3625\n",
      "Epoch 47/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 0.6570 - accuracy: 0.7908 - val_loss: 3.2625 - val_accuracy: 0.3375\n",
      "Epoch 48/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 0.5095 - accuracy: 0.8293 - val_loss: 4.0133 - val_accuracy: 0.3750\n",
      "Epoch 49/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 0.5420 - accuracy: 0.8148 - val_loss: 3.1207 - val_accuracy: 0.3042\n",
      "Epoch 50/50\n",
      "31/31 [==============================] - 2s 59ms/step - loss: 0.5678 - accuracy: 0.8023 - val_loss: 3.8196 - val_accuracy: 0.3583\n"
     ]
    }
   ],
   "source": [
    "# train the model on the new data for a few epochs\n",
    "history = model.fit(train_set, validation_data=test_set, epochs=50)\n",
    "\n",
    "# at this point, the top layers are well trained and we can start fine-tuning\n",
    "# convolutional layers from inception V3. We will freeze the bottom N layers\n",
    "# and train the remaining top layers."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['loss', 'accuracy', 'val_loss', 'val_accuracy'])"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "history.history.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Draw accuracy values for training & validation\n",
    "plt.plot(history.history['accuracy'])\n",
    "plt.plot(history.history['val_accuracy'])\n",
    "plt.title('Model accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Test'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "# Draw loss for training & validation\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.title('Model loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Test'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 冻结底层网络层微调顶层网络层"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 input_1\n",
      "1 conv2d\n",
      "2 batch_normalization\n",
      "3 activation\n",
      "4 conv2d_1\n",
      "5 batch_normalization_1\n",
      "6 activation_1\n",
      "7 conv2d_2\n",
      "8 batch_normalization_2\n",
      "9 activation_2\n",
      "10 max_pooling2d\n",
      "11 conv2d_3\n",
      "12 batch_normalization_3\n",
      "13 activation_3\n",
      "14 conv2d_4\n",
      "15 batch_normalization_4\n",
      "16 activation_4\n",
      "17 max_pooling2d_1\n",
      "18 conv2d_8\n",
      "19 batch_normalization_8\n",
      "20 activation_8\n",
      "21 conv2d_6\n",
      "22 conv2d_9\n",
      "23 batch_normalization_6\n",
      "24 batch_normalization_9\n",
      "25 activation_6\n",
      "26 activation_9\n",
      "27 average_pooling2d\n",
      "28 conv2d_5\n",
      "29 conv2d_7\n",
      "30 conv2d_10\n",
      "31 conv2d_11\n",
      "32 batch_normalization_5\n",
      "33 batch_normalization_7\n",
      "34 batch_normalization_10\n",
      "35 batch_normalization_11\n",
      "36 activation_5\n",
      "37 activation_7\n",
      "38 activation_10\n",
      "39 activation_11\n",
      "40 mixed0\n",
      "41 conv2d_15\n",
      "42 batch_normalization_15\n",
      "43 activation_15\n",
      "44 conv2d_13\n",
      "45 conv2d_16\n",
      "46 batch_normalization_13\n",
      "47 batch_normalization_16\n",
      "48 activation_13\n",
      "49 activation_16\n",
      "50 average_pooling2d_1\n",
      "51 conv2d_12\n",
      "52 conv2d_14\n",
      "53 conv2d_17\n",
      "54 conv2d_18\n",
      "55 batch_normalization_12\n",
      "56 batch_normalization_14\n",
      "57 batch_normalization_17\n",
      "58 batch_normalization_18\n",
      "59 activation_12\n",
      "60 activation_14\n",
      "61 activation_17\n",
      "62 activation_18\n",
      "63 mixed1\n",
      "64 conv2d_22\n",
      "65 batch_normalization_22\n",
      "66 activation_22\n",
      "67 conv2d_20\n",
      "68 conv2d_23\n",
      "69 batch_normalization_20\n",
      "70 batch_normalization_23\n",
      "71 activation_20\n",
      "72 activation_23\n",
      "73 average_pooling2d_2\n",
      "74 conv2d_19\n",
      "75 conv2d_21\n",
      "76 conv2d_24\n",
      "77 conv2d_25\n",
      "78 batch_normalization_19\n",
      "79 batch_normalization_21\n",
      "80 batch_normalization_24\n",
      "81 batch_normalization_25\n",
      "82 activation_19\n",
      "83 activation_21\n",
      "84 activation_24\n",
      "85 activation_25\n",
      "86 mixed2\n",
      "87 conv2d_27\n",
      "88 batch_normalization_27\n",
      "89 activation_27\n",
      "90 conv2d_28\n",
      "91 batch_normalization_28\n",
      "92 activation_28\n",
      "93 conv2d_26\n",
      "94 conv2d_29\n",
      "95 batch_normalization_26\n",
      "96 batch_normalization_29\n",
      "97 activation_26\n",
      "98 activation_29\n",
      "99 max_pooling2d_2\n",
      "100 mixed3\n",
      "101 conv2d_34\n",
      "102 batch_normalization_34\n",
      "103 activation_34\n",
      "104 conv2d_35\n",
      "105 batch_normalization_35\n",
      "106 activation_35\n",
      "107 conv2d_31\n",
      "108 conv2d_36\n",
      "109 batch_normalization_31\n",
      "110 batch_normalization_36\n",
      "111 activation_31\n",
      "112 activation_36\n",
      "113 conv2d_32\n",
      "114 conv2d_37\n",
      "115 batch_normalization_32\n",
      "116 batch_normalization_37\n",
      "117 activation_32\n",
      "118 activation_37\n",
      "119 average_pooling2d_3\n",
      "120 conv2d_30\n",
      "121 conv2d_33\n",
      "122 conv2d_38\n",
      "123 conv2d_39\n",
      "124 batch_normalization_30\n",
      "125 batch_normalization_33\n",
      "126 batch_normalization_38\n",
      "127 batch_normalization_39\n",
      "128 activation_30\n",
      "129 activation_33\n",
      "130 activation_38\n",
      "131 activation_39\n",
      "132 mixed4\n",
      "133 conv2d_44\n",
      "134 batch_normalization_44\n",
      "135 activation_44\n",
      "136 conv2d_45\n",
      "137 batch_normalization_45\n",
      "138 activation_45\n",
      "139 conv2d_41\n",
      "140 conv2d_46\n",
      "141 batch_normalization_41\n",
      "142 batch_normalization_46\n",
      "143 activation_41\n",
      "144 activation_46\n",
      "145 conv2d_42\n",
      "146 conv2d_47\n",
      "147 batch_normalization_42\n",
      "148 batch_normalization_47\n",
      "149 activation_42\n",
      "150 activation_47\n",
      "151 average_pooling2d_4\n",
      "152 conv2d_40\n",
      "153 conv2d_43\n",
      "154 conv2d_48\n",
      "155 conv2d_49\n",
      "156 batch_normalization_40\n",
      "157 batch_normalization_43\n",
      "158 batch_normalization_48\n",
      "159 batch_normalization_49\n",
      "160 activation_40\n",
      "161 activation_43\n",
      "162 activation_48\n",
      "163 activation_49\n",
      "164 mixed5\n",
      "165 conv2d_54\n",
      "166 batch_normalization_54\n",
      "167 activation_54\n",
      "168 conv2d_55\n",
      "169 batch_normalization_55\n",
      "170 activation_55\n",
      "171 conv2d_51\n",
      "172 conv2d_56\n",
      "173 batch_normalization_51\n",
      "174 batch_normalization_56\n",
      "175 activation_51\n",
      "176 activation_56\n",
      "177 conv2d_52\n",
      "178 conv2d_57\n",
      "179 batch_normalization_52\n",
      "180 batch_normalization_57\n",
      "181 activation_52\n",
      "182 activation_57\n",
      "183 average_pooling2d_5\n",
      "184 conv2d_50\n",
      "185 conv2d_53\n",
      "186 conv2d_58\n",
      "187 conv2d_59\n",
      "188 batch_normalization_50\n",
      "189 batch_normalization_53\n",
      "190 batch_normalization_58\n",
      "191 batch_normalization_59\n",
      "192 activation_50\n",
      "193 activation_53\n",
      "194 activation_58\n",
      "195 activation_59\n",
      "196 mixed6\n",
      "197 conv2d_64\n",
      "198 batch_normalization_64\n",
      "199 activation_64\n",
      "200 conv2d_65\n",
      "201 batch_normalization_65\n",
      "202 activation_65\n",
      "203 conv2d_61\n",
      "204 conv2d_66\n",
      "205 batch_normalization_61\n",
      "206 batch_normalization_66\n",
      "207 activation_61\n",
      "208 activation_66\n",
      "209 conv2d_62\n",
      "210 conv2d_67\n",
      "211 batch_normalization_62\n",
      "212 batch_normalization_67\n",
      "213 activation_62\n",
      "214 activation_67\n",
      "215 average_pooling2d_6\n",
      "216 conv2d_60\n",
      "217 conv2d_63\n",
      "218 conv2d_68\n",
      "219 conv2d_69\n",
      "220 batch_normalization_60\n",
      "221 batch_normalization_63\n",
      "222 batch_normalization_68\n",
      "223 batch_normalization_69\n",
      "224 activation_60\n",
      "225 activation_63\n",
      "226 activation_68\n",
      "227 activation_69\n",
      "228 mixed7\n",
      "229 conv2d_72\n",
      "230 batch_normalization_72\n",
      "231 activation_72\n",
      "232 conv2d_73\n",
      "233 batch_normalization_73\n",
      "234 activation_73\n",
      "235 conv2d_70\n",
      "236 conv2d_74\n",
      "237 batch_normalization_70\n",
      "238 batch_normalization_74\n",
      "239 activation_70\n",
      "240 activation_74\n",
      "241 conv2d_71\n",
      "242 conv2d_75\n",
      "243 batch_normalization_71\n",
      "244 batch_normalization_75\n",
      "245 activation_71\n",
      "246 activation_75\n",
      "247 max_pooling2d_3\n",
      "248 mixed8\n",
      "249 conv2d_80\n",
      "250 batch_normalization_80\n",
      "251 activation_80\n",
      "252 conv2d_77\n",
      "253 conv2d_81\n",
      "254 batch_normalization_77\n",
      "255 batch_normalization_81\n",
      "256 activation_77\n",
      "257 activation_81\n",
      "258 conv2d_78\n",
      "259 conv2d_79\n",
      "260 conv2d_82\n",
      "261 conv2d_83\n",
      "262 average_pooling2d_7\n",
      "263 conv2d_76\n",
      "264 batch_normalization_78\n",
      "265 batch_normalization_79\n",
      "266 batch_normalization_82\n",
      "267 batch_normalization_83\n",
      "268 conv2d_84\n",
      "269 batch_normalization_76\n",
      "270 activation_78\n",
      "271 activation_79\n",
      "272 activation_82\n",
      "273 activation_83\n",
      "274 batch_normalization_84\n",
      "275 activation_76\n",
      "276 mixed9_0\n",
      "277 concatenate\n",
      "278 activation_84\n",
      "279 mixed9\n",
      "280 conv2d_89\n",
      "281 batch_normalization_89\n",
      "282 activation_89\n",
      "283 conv2d_86\n",
      "284 conv2d_90\n",
      "285 batch_normalization_86\n",
      "286 batch_normalization_90\n",
      "287 activation_86\n",
      "288 activation_90\n",
      "289 conv2d_87\n",
      "290 conv2d_88\n",
      "291 conv2d_91\n",
      "292 conv2d_92\n",
      "293 average_pooling2d_8\n",
      "294 conv2d_85\n",
      "295 batch_normalization_87\n",
      "296 batch_normalization_88\n",
      "297 batch_normalization_91\n",
      "298 batch_normalization_92\n",
      "299 conv2d_93\n",
      "300 batch_normalization_85\n",
      "301 activation_87\n",
      "302 activation_88\n",
      "303 activation_91\n",
      "304 activation_92\n",
      "305 batch_normalization_93\n",
      "306 activation_85\n",
      "307 mixed9_1\n",
      "308 concatenate_1\n",
      "309 activation_93\n",
      "310 mixed10\n"
     ]
    }
   ],
   "source": [
    "# let's visualize layer names and layer indices to see how many layers\n",
    "# we should freeze:\n",
    "for i, layer in enumerate(base_model.layers):\n",
    "    print(i, layer.name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# we chose to train the top 2 inception blocks, i.e. we will freeze\n",
    "# the first 249 layers and unfreeze the rest:\n",
    "for layer in model.layers[:249]:\n",
    "    layer.trainable = False\n",
    "for layer in model.layers[249:]:\n",
    "    layer.trainable = True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# we need to recompile the model for these modifications to take effect\n",
    "# we use SGD with a low learning rate\n",
    "from tensorflow.keras.optimizers import SGD\n",
    "\n",
    "model.compile(\n",
    "    optimizer=SGD(learning_rate=0.0001, momentum=0.9),\n",
    "    loss='sparse_categorical_crossentropy',\n",
    "    metrics=[\"accuracy\"],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "31/31 [==============================] - 13s 118ms/step - loss: 1.7043 - accuracy: 0.2882 - val_loss: 1.6911 - val_accuracy: 0.3458\n",
      "Epoch 2/50\n",
      "31/31 [==============================] - 2s 71ms/step - loss: 1.6528 - accuracy: 0.3330 - val_loss: 1.6635 - val_accuracy: 0.3375\n",
      "Epoch 3/50\n",
      "31/31 [==============================] - 2s 70ms/step - loss: 1.6001 - accuracy: 0.3673 - val_loss: 1.6358 - val_accuracy: 0.3250\n",
      "Epoch 4/50\n",
      "31/31 [==============================] - 2s 71ms/step - loss: 1.5633 - accuracy: 0.3975 - val_loss: 1.6324 - val_accuracy: 0.3208\n",
      "Epoch 5/50\n",
      "31/31 [==============================] - 2s 71ms/step - loss: 1.5255 - accuracy: 0.3975 - val_loss: 1.6274 - val_accuracy: 0.3250\n",
      "Epoch 6/50\n",
      "31/31 [==============================] - 2s 71ms/step - loss: 1.4992 - accuracy: 0.4048 - val_loss: 1.6040 - val_accuracy: 0.3167\n",
      "Epoch 7/50\n",
      "31/31 [==============================] - 2s 71ms/step - loss: 1.4658 - accuracy: 0.4287 - val_loss: 1.5823 - val_accuracy: 0.3292\n",
      "Epoch 8/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 1.4350 - accuracy: 0.4454 - val_loss: 1.5678 - val_accuracy: 0.3208\n",
      "Epoch 9/50\n",
      "31/31 [==============================] - 2s 71ms/step - loss: 1.4175 - accuracy: 0.4443 - val_loss: 1.5482 - val_accuracy: 0.3250\n",
      "Epoch 10/50\n",
      "31/31 [==============================] - 2s 71ms/step - loss: 1.3867 - accuracy: 0.4589 - val_loss: 1.5394 - val_accuracy: 0.3292\n",
      "Epoch 11/50\n",
      "31/31 [==============================] - 2s 71ms/step - loss: 1.3550 - accuracy: 0.4880 - val_loss: 1.5272 - val_accuracy: 0.3167\n",
      "Epoch 12/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 1.3254 - accuracy: 0.4932 - val_loss: 1.5242 - val_accuracy: 0.3250\n",
      "Epoch 13/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 1.3148 - accuracy: 0.5120 - val_loss: 1.5063 - val_accuracy: 0.3458\n",
      "Epoch 14/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 1.2890 - accuracy: 0.5297 - val_loss: 1.4933 - val_accuracy: 0.3500\n",
      "Epoch 15/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 1.2470 - accuracy: 0.5609 - val_loss: 1.4905 - val_accuracy: 0.3333\n",
      "Epoch 16/50\n",
      "31/31 [==============================] - 2s 71ms/step - loss: 1.2174 - accuracy: 0.5525 - val_loss: 1.4791 - val_accuracy: 0.3458\n",
      "Epoch 17/50\n",
      "31/31 [==============================] - 2s 71ms/step - loss: 1.1906 - accuracy: 0.5723 - val_loss: 1.4681 - val_accuracy: 0.3500\n",
      "Epoch 18/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 1.1686 - accuracy: 0.5817 - val_loss: 1.4669 - val_accuracy: 0.3500\n",
      "Epoch 19/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 1.1449 - accuracy: 0.6046 - val_loss: 1.4621 - val_accuracy: 0.3625\n",
      "Epoch 20/50\n",
      "31/31 [==============================] - 2s 70ms/step - loss: 1.1219 - accuracy: 0.6181 - val_loss: 1.4552 - val_accuracy: 0.3667\n",
      "Epoch 21/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 1.0932 - accuracy: 0.6327 - val_loss: 1.4428 - val_accuracy: 0.3708\n",
      "Epoch 22/50\n",
      "31/31 [==============================] - 2s 71ms/step - loss: 1.0565 - accuracy: 0.6462 - val_loss: 1.4397 - val_accuracy: 0.3625\n",
      "Epoch 23/50\n",
      "31/31 [==============================] - 2s 71ms/step - loss: 1.0192 - accuracy: 0.6722 - val_loss: 1.4278 - val_accuracy: 0.3792\n",
      "Epoch 24/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 0.9906 - accuracy: 0.6785 - val_loss: 1.4345 - val_accuracy: 0.3833\n",
      "Epoch 25/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 0.9632 - accuracy: 0.6930 - val_loss: 1.4148 - val_accuracy: 0.3875\n",
      "Epoch 26/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 0.9382 - accuracy: 0.7242 - val_loss: 1.4085 - val_accuracy: 0.4000\n",
      "Epoch 27/50\n",
      "31/31 [==============================] - 2s 73ms/step - loss: 0.9137 - accuracy: 0.7471 - val_loss: 1.4146 - val_accuracy: 0.3958\n",
      "Epoch 28/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 0.8796 - accuracy: 0.7409 - val_loss: 1.4115 - val_accuracy: 0.4042\n",
      "Epoch 29/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 0.8471 - accuracy: 0.7419 - val_loss: 1.4027 - val_accuracy: 0.4167\n",
      "Epoch 30/50\n",
      "31/31 [==============================] - 2s 73ms/step - loss: 0.8274 - accuracy: 0.7607 - val_loss: 1.3981 - val_accuracy: 0.3958\n",
      "Epoch 31/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 0.8057 - accuracy: 0.7638 - val_loss: 1.3978 - val_accuracy: 0.4125\n",
      "Epoch 32/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 0.7804 - accuracy: 0.7950 - val_loss: 1.4013 - val_accuracy: 0.4083\n",
      "Epoch 33/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 0.7360 - accuracy: 0.8158 - val_loss: 1.3971 - val_accuracy: 0.4125\n",
      "Epoch 34/50\n",
      "31/31 [==============================] - 2s 73ms/step - loss: 0.7070 - accuracy: 0.8158 - val_loss: 1.3992 - val_accuracy: 0.3875\n",
      "Epoch 35/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 0.6836 - accuracy: 0.8252 - val_loss: 1.3861 - val_accuracy: 0.3958\n",
      "Epoch 36/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 0.6650 - accuracy: 0.8408 - val_loss: 1.3846 - val_accuracy: 0.4042\n",
      "Epoch 37/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 0.6598 - accuracy: 0.8252 - val_loss: 1.4019 - val_accuracy: 0.4250\n",
      "Epoch 38/50\n",
      "31/31 [==============================] - 2s 73ms/step - loss: 0.6042 - accuracy: 0.8522 - val_loss: 1.3979 - val_accuracy: 0.4208\n",
      "Epoch 39/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 0.5765 - accuracy: 0.8793 - val_loss: 1.3919 - val_accuracy: 0.4042\n",
      "Epoch 40/50\n",
      "31/31 [==============================] - 2s 73ms/step - loss: 0.5533 - accuracy: 0.8907 - val_loss: 1.3877 - val_accuracy: 0.4042\n",
      "Epoch 41/50\n",
      "31/31 [==============================] - 2s 71ms/step - loss: 0.5341 - accuracy: 0.8772 - val_loss: 1.4072 - val_accuracy: 0.4083\n",
      "Epoch 42/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 0.5130 - accuracy: 0.9043 - val_loss: 1.4061 - val_accuracy: 0.3833\n",
      "Epoch 43/50\n",
      "31/31 [==============================] - 2s 73ms/step - loss: 0.4917 - accuracy: 0.9084 - val_loss: 1.4121 - val_accuracy: 0.3750\n",
      "Epoch 44/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 0.4768 - accuracy: 0.9136 - val_loss: 1.4326 - val_accuracy: 0.4042\n",
      "Epoch 45/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 0.4641 - accuracy: 0.9084 - val_loss: 1.4192 - val_accuracy: 0.4083\n",
      "Epoch 46/50\n",
      "31/31 [==============================] - 2s 71ms/step - loss: 0.4416 - accuracy: 0.9282 - val_loss: 1.4249 - val_accuracy: 0.4042\n",
      "Epoch 47/50\n",
      "31/31 [==============================] - 2s 71ms/step - loss: 0.4154 - accuracy: 0.9251 - val_loss: 1.4281 - val_accuracy: 0.3958\n",
      "Epoch 48/50\n",
      "31/31 [==============================] - 2s 72ms/step - loss: 0.3910 - accuracy: 0.9396 - val_loss: 1.4334 - val_accuracy: 0.4000\n",
      "Epoch 49/50\n",
      "31/31 [==============================] - 2s 71ms/step - loss: 0.3940 - accuracy: 0.9396 - val_loss: 1.4281 - val_accuracy: 0.4000\n",
      "Epoch 50/50\n",
      "31/31 [==============================] - 2s 71ms/step - loss: 0.3821 - accuracy: 0.9490 - val_loss: 1.4365 - val_accuracy: 0.4250\n"
     ]
    }
   ],
   "source": [
    "# we train our model again (this time fine-tuning the top 2 inception blocks\n",
    "# alongside the top Dense layers\n",
    "history = model.fit(train_set, validation_data=test_set, epochs=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjcAAAHHCAYAAABDUnkqAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/OQEPoAAAACXBIWXMAAA9hAAAPYQGoP6dpAABhvUlEQVR4nO3dd3gU5d7G8e/spncgFQi9gwSkFwuKUhTBhopKsxwUbOg5B18VxKNir2BDBQsKiIJYAAGlKlID0jtEIAkB0knbnfePhWAEA4Ekk93cn+uaK9nZmd3fjuDePPMUwzRNExEREREPYbO6ABEREZHSpHAjIiIiHkXhRkRERDyKwo2IiIh4FIUbERER8SgKNyIiIuJRFG5ERETEoyjciIiIiEdRuBERERGPonAjIhWeYRg8/fTTJT5v7969GIbB5MmTiz1u0aJFGIbBokWLzqs+EalYFG5E5JxMnjwZwzAwDINly5ad9rxpmsTGxmIYBtdee60FFYqIuCjciEiJ+Pn58cUXX5y2f/Hixfz555/4+vpaUJWIyCkKNyJSIr179+arr76ioKCgyP4vvviCNm3aEB0dbVFlIiIuCjciUiK33XYbR44cYf78+YX78vLymDFjBgMGDDjjOVlZWTz66KPExsbi6+tL48aNeeWVVzBNs8hxubm5PPLII0RERBAcHMx1113Hn3/+ecbXPHDgAEOHDiUqKgpfX1+aN2/Oxx9/XHofFPjqq69o06YN/v7+hIeHc8cdd3DgwIEixyQmJjJkyBBq1qyJr68vMTEx9O3bl7179xYes3r1anr06EF4eDj+/v7UrVuXoUOHlmqtInKKl9UFiIh7qVOnDp06deLLL7+kV69eAMyZM4e0tDRuvfVW3nrrrSLHm6bJddddxy+//MJdd91Fq1atmDdvHv/+9785cOAAr7/+euGxd999N59//jkDBgygc+fO/Pzzz1xzzTWn1ZCUlETHjh0xDIMRI0YQERHBnDlzuOuuu0hPT+fhhx++4M85efJkhgwZQrt27Rg3bhxJSUm8+eabLF++nHXr1hEWFgbAjTfeyKZNm3jggQeoU6cOycnJzJ8/n/379xc+vvrqq4mIiGDUqFGEhYWxd+9evvnmmwuuUUT+gSkicg4mTZpkAuaqVavM8ePHm8HBwWZ2drZpmqZ58803m926dTNN0zRr165tXnPNNYXnzZo1ywTMZ599tsjr3XTTTaZhGObOnTtN0zTN+Ph4EzDvv//+IscNGDDABMwxY8YU7rvrrrvMmJgYMyUlpcixt956qxkaGlpY1549e0zAnDRpUrGf7ZdffjEB85dffjFN0zTz8vLMyMhIs0WLFubx48cLj/v+++9NwBw9erRpmqZ57NgxEzBffvnlf3ztmTNnFl43ESkfui0lIiXWv39/jh8/zvfff09GRgbff//9P96S+vHHH7Hb7Tz44INF9j/66KOYpsmcOXMKjwNOO+7vrTCmafL111/Tp08fTNMkJSWlcOvRowdpaWmsXbv2gj7f6tWrSU5O5v7778fPz69w/zXXXEOTJk344YcfAPD398fHx4dFixZx7NixM77WyRae77//nvz8/AuqS0TOjcKNiJRYREQE3bt354svvuCbb77B4XBw0003nfHYffv2Ub16dYKDg4vsb9q0aeHzJ3/abDbq169f5LjGjRsXeXz48GFSU1P54IMPiIiIKLINGTIEgOTk5Av6fCdr+vt7AzRp0qTweV9fX1588UXmzJlDVFQUl156KS+99BKJiYmFx1922WXceOONjB07lvDwcPr27cukSZPIzc29oBpF5J+pz42InJcBAwZwzz33kJiYSK9evQpbKMqa0+kE4I477mDQoEFnPKZly5blUgu4Wpb69OnDrFmzmDdvHk899RTjxo3j559/pnXr1hiGwYwZM1ixYgXfffcd8+bNY+jQobz66qusWLGCoKCgcqtVpLJQy42InJfrr78em83GihUr/vGWFEDt2rU5ePAgGRkZRfZv3bq18PmTP51OJ7t27Spy3LZt24o8PjmSyuFw0L179zNukZGRF/TZTtb09/c+ue/k8yfVr1+fRx99lJ9++omNGzeSl5fHq6++WuSYjh078txzz7F69WqmTJnCpk2bmDp16gXVKSJnpnAjIuclKCiId999l6effpo+ffr843G9e/fG4XAwfvz4Ivtff/11DMMoHHF18uffR1u98cYbRR7b7XZuvPFGvv76azZu3Hja+x0+fPh8Pk4Rbdu2JTIykvfee6/I7aM5c+awZcuWwhFc2dnZ5OTkFDm3fv36BAcHF5537Nix04a8t2rVCkC3pkTKiG5Lich5+6fbQn/Vp08funXrxhNPPMHevXuJi4vjp59+4ttvv+Xhhx8u7GPTqlUrbrvtNt555x3S0tLo3LkzCxcuZOfOnae95gsvvMAvv/xChw4duOeee2jWrBlHjx5l7dq1LFiwgKNHj17Q5/L29ubFF19kyJAhXHbZZdx2222FQ8Hr1KnDI488AsD27du58sor6d+/P82aNcPLy4uZM2eSlJTErbfeCsAnn3zCO++8w/XXX0/9+vXJyMhg4sSJhISE0Lt37wuqU0TOTOFGRMqUzWZj9uzZjB49mmnTpjFp0iTq1KnDyy+/zKOPPlrk2I8//piIiAimTJnCrFmzuOKKK/jhhx+IjY0tclxUVBQrV67kmWee4ZtvvuGdd96hWrVqNG/enBdffLFU6h48eDABAQG88MIL/Pe//yUwMJDrr7+eF198sbB/UWxsLLfddhsLFy7ks88+w8vLiyZNmjB9+nRuvPFGwNWheOXKlUydOpWkpCRCQ0Np3749U6ZMoW7duqVSq4gUZZh/by8VERERcWPqcyMiIiIeReFGREREPIrCjYiIiHgUhRsRERHxKAo3IiIi4lEUbkRERMSjVLp5bpxOJwcPHiQ4OBjDMKwuR0RERM6BaZpkZGRQvXp1bLbi22YqXbg5ePDgaROCiYiIiHtISEigZs2axR5T6cJNcHAw4Lo4ISEhFlcjIiIi5yI9PZ3Y2NjC7/HiVLpwc/JWVEhIiMKNiIiImzmXLiXqUCwiIiIeReFGREREPIrCjYiIiHiUStfn5lw5HA7y8/OtLsNteXt7Y7fbrS5DREQqIYWbvzFNk8TERFJTU60uxe2FhYURHR2t+YRERKRcKdz8zclgExkZSUBAgL6Yz4NpmmRnZ5OcnAxATEyMxRWJiEhlonDzFw6HozDYVKtWzepy3Jq/vz8AycnJREZG6haViIiUG3Uo/ouTfWwCAgIsrsQznLyO6rskIiLlSeHmDHQrqnToOoqIiBUUbkRERMSjKNzIP6pTpw5vvPGG1WWIiIiUiMKNBzAMo9jt6aefPq/XXbVqFffee2/pFisiIlLGNFqqFOUVOHE4nfj7lO9lPXToUOHv06ZNY/To0Wzbtq1wX1BQUOHvpmnicDjw8jp7jREREaVbqIiISDlQy00pycotYEdyBvuOZONwOsv1vaOjowu30NBQDMMofLx161aCg4OZM2cObdq0wdfXl2XLlrFr1y769u1LVFQUQUFBtGvXjgULFhR53b/fljIMgw8//JDrr7+egIAAGjZsyOzZs8v1s4qIiJyNws1ZmKZJdl7BWTen6SS/wEl6Tj67krPO6ZyzbaZpltrnGDVqFC+88AJbtmyhZcuWZGZm0rt3bxYuXMi6devo2bMnffr0Yf/+/cW+ztixY+nfvz8bNmygd+/e3H777Rw9erTU6hQREblQui11FsfzHTQbPc+S9978TA8CSukW1zPPPMNVV11V+Lhq1arExcUVPv7f//7HzJkzmT17NiNGjPjH1xk8eDC33XYbAM8//zxvvfUWK1eupGfPnqVSp4iIyIVSy00l0bZt2yKPMzMzeeyxx2jatClhYWEEBQWxZcuWs7bctGzZsvD3wMBAQkJCCpdZEBERqQjUcnMW/t52Nj/T4+wH5mbC0V2uX/FhvzMCL19/6lQ7//Wp/L1Lb8mCwMDAIo8fe+wx5s+fzyuvvEKDBg3w9/fnpptuIi8vr9jX8fb2LvLYMAyc5dzHSEREpDgKN2dhGMa53RryCQOfRnBsLwHOAkLMJP50hpOd50dEsG+Z11lSy5cvZ/DgwVx//fWAqyVn79691hYlIiJSCnRbqjT5BkNEY/AJxG44qW1Lxkj/k+N5FW9tpYYNG/LNN98QHx/P+vXrGTBggFpgRETEIyjclDa7D1RriBkUCUC4kQ4pO3Hm51pcWFGvvfYaVapUoXPnzvTp04cePXpw8cUXW12WiIjIBTPM0hxv7AbS09MJDQ0lLS2NkJCQIs/l5OSwZ88e6tati5+f3wW/lyM7FTN1H144cWDHXrU2+IVe8Ou6i9K+niIiUnkV9/39d2q5KUP2gDByQhuQbfpixwFHd0P6QahceVJERKRcKdyUsaDAQI4F1CbFPJEyM5PgyE5wVLx+OCIiIp5A4aYcxIQGcMQeyX5nJE5skJcJh7e6ho+LiIhIqVK4KQc2m0FsVX/SCGKHszoFNl9wFsCRHZCRpNtUIiIipUjhppwE+HgRFeJLLt5sd1TH4VfF9UTGQTi6xxV2RERE5IIp3JSjiGBfAn28KDBhb0E4ZmgsYEBuGhzeBnnZVpcoIiLi9hRuypFhGNSs6o/NMMjKKyDFGQzhjVxz4zjyIGU7ZKXoNpWIiMgFULgpZ75edmLCXHO+JKbnkGP4umY19g0FTEhLgNT94HRYW6iIiIib0tpSFqga4EP68QIycvJJOJpN/cggbFXrQmayqw/O8aOQlwW+QWDzAps32P/201Z6i2qKiIh4EoUbCxiGQc0q/mxPKuB4voPk9FyiQ/0gOAp8AuDYXnDkQnYxSzYYthNhxwe8fMDLF+y+rt/tvgo/IiJSaem2lEW87TZqhPkDcDgjl+zcE6OlfIMhoimExkJwDASEu5Zs8A50BRnjxH8y0+kKQHkZGFVqYwRHYwRUwfAJxLB7YRhG4fb0k6POu07DMJg1a9YFfloREZHyo5YbC4UF+JCeU0Bqdh4Jx47TMDIIm81w3XoKDP/nE50OcOaDowAcuRzavg4K8sGRy7SZ3zH65XfYtuSbwsODAgNcw81Da4Lduxw+mYiIiHXUcmOx6qF+eNtt5BY4SEzPObeTbHbw8nP1yQmoRnTDVkQ3bUd0i66ExjbFsNmJbtaF6MZtia7blKnfzqNphyvwCwyiSeOGvDNhQuFL5eXlMWLECGJiYvDz86N27dqMGzcOgDp16gBw/fXXYxhG4WMREZGKTC03Z2OakF928894ATUDnew9ks2RY9kE23IJ9jvRuuIdAIZxfi/sEwAEMOWbHxn92kTGP/c4rZvWY93GrdzznycJ9Pdl0NC7eeutt5g9ezbTp0+nVq1aJCQkkJCQAMCqVauIjIxk0qRJ9OzZE7td/XhERKTiszTcLFmyhJdffpk1a9Zw6NAhZs6cSb9+/Yo9Jzc3l2eeeYbPP/+cxMREYmJiGD16NEOHDi2bIvOz4fnqZfPaJwQDF53pif87CD6BF/TaY8aM4dVXX+OGW2+FzGTq1qrJ5u17eP+dtxnU/zr279tHw4YN6dq1K4ZhULt27cJzIyIiAAgLCyM6OvqC6hARESkvloabrKws4uLiGDp0KDfccMM5ndO/f3+SkpL46KOPaNCgAYcOHcLpdJZxpe4pKyuLXbt2cdddd3HPPfcU7i8oyCc0OAjSDzC4Xzeu+uILGjduTM+ePbn22mu5+uqrLaxaRETkwlgabnr16kWvXr3O+fi5c+eyePFidu/eTdWqVQHKvh+Id4CrBaUcZOUWsCclCxOoVcWfUO+AC3q9zEzXquMTJ06kQ4cOp54wTex56WDYubhpXfb8+i1zft3IguWr6d+/P927d2fGjBkX9N4iIiJWcas+N7Nnz6Zt27a89NJLfPbZZwQGBnLdddfxv//9D39//zOek5ubS27uqfli0tPTS/amhnHBt4bOVaAPhDt9SM7I5UC2QUCQibfXefa5AaKioqhevTq7d+/m9ttvP/2AgjxISyAEuKVHB27pfQk39e1Nz+tu5OjRo1StWhVvb28cDs2WLCIi7sOtws3u3btZtmwZfn5+zJw5k5SUFO6//36OHDnCpEmTznjOuHHjGDt2bDlXev4iQ/zIyHFN7rc7JYt6EYF4289/UNvYsWN58MEHCQ0NpWfPnuTm5rJ69WqOHTvGyJEjeW3yLGLCQ2ldLxIbTr768jOiIyMIC3KFxTp16rBw4UK6dOmCr68vVapUKa2PKiIiUibcaii40+nEMAymTJlC+/bt6d27N6+99hqffPIJx48fP+M5jz/+OGlpaYXbyZFAFZXNMKhdLaBwePiew1kUOM6/T9Hdd9/Nhx9+yKRJk7jooou47LLLmDx5MnXr1gUgOCSEl96YQNteA2h37UD2Jhzkx8/exJayDTIO8erLLzN//nxiY2Np3bp1aX1MERGRMmOYZsVYgtowjLOOlho0aBDLly9n586dhfu2bNlCs2bN2L59Ow0bNjzr+6SnpxMaGkpaWhohISFFnsvJyWHPnj3UrVsXPz+/8/4spSH3RMtNvsOJn7edeuGBeF1AC845y89xLd6Z5+qvg93HNfmfX2iJX6oiXU8REXFvxX1//51btdx06dKFgwcPFnaUBdi+fTs2m42aNWtaWFnp8/1LoMnJd7An5cJacM6Ztx9UawBV6rjWrnLkwdHdcGQXFBSz1pWIiEgFYWm4yczMJD4+nvj4eAD27NlDfHw8+/fvB1y3lAYOHFh4/IABA6hWrRpDhgxh8+bNLFmyhH//+98MHTr0HzsUu7PCgGOzcfxkwCmPYe+GAf5VILIpBEUCBuSmQ/IWV9DJSXdNbigiIlIBWRpuVq9eTevWrQv7cowcOZLWrVszevRoAA4dOlQYdACCgoKYP38+qamptG3blttvv50+ffrw1ltvWVJ/efDztlMv4lTA2ZuSjaO85vWx2SGkBkQ0AZ9gwIScNDi6C5I3Q0YSOPLLpxYREZFzVGH63JQXd+lz83fH8x3sOZxJgdMkwMeLuuGB2G3nP0z8vOQfh6wUOH4MzJPDww1Xf5zAcPAJKrJcREW+niIi4l48ts9NeamIec/f214YaLLzCtibkoXDWc51evtDWCxENYfQWq4JDjEhJxWO7HTdtspMcq1aTsW8jiIi4vkUbv7C29u1YGV2dtktlHkh/H28qHci4GTlFbD3iAUBB1y3qwKrQURjCG8MAeFg2MCRC+kHXbessg6TneXq+H3yuoqIiJQHt5rEr6zZ7XbCwsJITk4GICAgAON8V+UuIwZQPcjOn8eOk5mVx+78XGqEBWAr71tUhWzgFwE+VV0djbOPYObnkZ2ZQPKxDMK887EbjS2qTUREKiOFm785ufr1yYBTUTkLnKRk5uI0IcnbRtVAn4oRxEwb5BVA9hHC9nxH9I4v4Pc4uGos1Lvc6upERKQSULj5G8MwiImJITIykvz8ij0S6HjCMR7/5g/yCpxc2iiCJ3o3LZ+J/s6BtzMPe/4fsC8QDsXDp32h/hXQ/WmIibO6PBER8WAaLeXmFm8/zD2frCbP4eSG1jV45eY4C29RnUHmYVj6Cqz6CJwnwmKLm6DVAIjtAL5B1tYnIiJuoSTf3wo3HuCnTYncN2UtDqfJbe1r8fz1LSrGLaq/OroHfnkO/vjq1D7DDtVbQe3OULsL1OromjxQRETkbxRuiuGJ4QZg9vqDPDR1HaYJQ7vU5alrm1a8gANwaD2seA/2LYPU/X970oCoFq6wU6cL1L1UYUdERACFm2J5argBmL46gf/M2ADAiG4NeKxHBR+llJoA+36Ffctd25GdRZ+3ebv66TS/Hpr0Pq/FO0VExDMo3BTDk8MNwGe/7eWpbzcB8NjVjRhxxdlXSq8wMpJg/6+wdznsWQIp2049Z/eB+le6gk7jXuDnef/tRETknyncFMPTww3A+4t3MW7OVgCeurYZd3Wta3FF5yl5K2yeBRu/+VvQ8YWGV7mCTp2urkkE7Rr4JyLiyRRuilEZwg3AGwu288aCHQA8eU1T7upat2L2wTlXyVtg00xX0Dmy4/Tn/au61rcKjHD9DPjL7+ENoe5lRda9EhER96JwU4zKEm5M0+SFOVt5f8luAHo0j+LFG1sSFuBjcWUXyDQhaZMr6Gz+9kQ/nXP4Ixx1EVz2b2jSB2wVYy4gERE5dwo3xags4QZcAeejZXt4ce5W8h0mMaF+vHFLKzrUq2Z1aaXH6YDso5B1GLJTXD+z/vZz92LIy3AdH9HUFXKa9XOtkSUiIm5B4aYYlSncnLTxQBoPfLmOPSlZ2Ax44IqGPHBFgwozm3GZyz4Kv7/nGoKem+baF94ILnkMWtyo/joiIm5A4aYYlTHcAGTlFjD62018vfZPANrXqcrrt7aiRpi/xZWVo+OpsPID+G0C5KS69lWt5wo5LfuDXauXi4hUVAo3xais4eakWesO8OSsjWTmFhDq782LN7akZ4toq8sqXznpsGoi/Doejh917QurDa3vhJY3Q5U6lpYnIiKnU7gpRmUPNwD7jmTx4JfrWP+n6xbNHR1r8eQ1zfDzrmR9UHIzYfVHsPwtV3+dk2p1crXkNL9eMySLiFQQCjfFULhxyStw8tr87by3eBcATWNC+GRIOyJD/CyuzAJ52a7RV39Md3U+Pjn6yu4DDa+GlrdAox7g5WtpmSIilZnCTTEUbopauuMwj0xbT0pmLrFV/fn8rg7UrhZodVnWST8If8yADdMgaeOp/X6h0Kyva7RVQFUIqOaaWyfgxOYbonl0RETKkMJNMRRuTrf/SDZ3fvw7+45kEx7ky6dD29Osuq4NiRtdrTkbvoKMg8Ufa/M6EXaquVY6b3Kta10sn4ByKVVExNMp3BRD4ebMkjNyGPTxKrYcSifYz4uPB7ejXZ2qVpdVMTgdsHcZbJ8LmUmQfcQ1vDz7qKtDcn72mc/zDoAGV7omDmzUA/zDyrVsERFPonBTDIWbf5Z2PJ+7P1nFqr3H8PWy8e4dF3NFkyiry6r48o+fCDtHXOFn50LY+j2kJZw6xuYFdS6Bpte6WnWCK9kINRGRC6RwUwyFm+Idz3Mw/Iu1/Lw1GbvN4JWbW3J965pWl+V+TBMOrXeFnC3fw+EtRZ+v2c61unmjXhDZVP11RETOQuGmGAo3Z5fvcPKfGRuYue4AAGP6NGNIFzddWbyiOLILtnznCjt/rir6XFhtaNwbGveE2l00maCIyBko3BRD4ebcOJ0mz3y/mcm/7gXgwSsa8MhVjdx7ZfGKIv2Qq//OtjmwexE4ck895xvq6qfTuDfU6QI+QeDl5wo8uvYiUokp3BRD4ebcmabJ2z/v5LX52wG4s2Ntnr6uOXabvmRLTV6WK+Bs+xG2z3Mt9Hkmhs0Vck5u3id++gRBWKxrVuW/biE1tDCoiHgUhZtiKNyU3Ge/7WX07E2YJlzZJJK3bmtNoK8Wmyx1TiccWOMKOtvmnN5PpyRsXhBW61TYiWwG0S0hqjn4BpVWxSIi5UbhphgKN+fnhw2HeGR6PHkFTprFhPDR4LbEhFaiRTetYJpQkAsFOae2/Jyij3PSIXU/HNsDx/a6ttT94Mj7hxc1XIuFRl/k2mLiXD+DonTbS0QqNIWbYijcnL+1+49xzyerOZKVR2SwLx8NasdFNUOtLkv+zumAjEOnws6RXa7ZlhP/cO0/k8AI11D1i26CBt211ISIVDgKN8VQuLkwCUezGTp5FTuSM/H3tvPGra3o0VxztriNzMOQ9Icr6JzcUraD6Tx1jG+oaz6eFjdC3cvArluQImI9hZtiKNxcuPScfIZPWcvSHSkYBvxfr6bcfUldjaRyV3nZkLjBNVR94zdFl5oICHetqdXiRtdq6TabdXWKSKWmcFMMhZvSUeBwMmb2Jqb8vh+A29rX4pm+zfG268vPrTmdsP832Pg1bJ7lmnX5pODqUPdSCIqEwHDXrazAiFO/B4S7RnGJiJQBhZtiKNyUHtM0+Xj5Xp79YTOmCV0bhDPh9osJ9dckdB7BUQB7FruCzpbvIDf97Of4BEOVE5MSNuvrGp2lFj2RyiU3EzISIbxBqb6swk0xFG5K3/zNSTw0dR3ZeQ4aRAbxTN/mdKpXTbepPEl+Duz6GVK2QVbKie3wiS0FslPOPEKran1XyGnW1zUy62x/JvJzXEPgE/9wjfpqeDXEti+bzyRipZNfvRXh/5P5Oa7BBxkHIaSma/oIL59zPz/9ECSsgP0ntsQ/IKIJ3P9rqZapcFMMhZuysfFAGnd/sprE9BwA2tSuwohuDbi8cYRCTmVgmq6WnczDcHAtbP4WdswvOvtyWG1odh006wc12riC0clOzSdHc6XsANNR9LWbXAtXjoGIRuX6kUTKRGYy/PwsxH/hChGtboOWt0JojbJ9X6fD9Q+GI7vgyE7XdvTE76kJwF+igGFzzZNVrYHrHyjVGkC1eq6fITVcgxD2r4CE3123sVP3n/5+YbVgxOpSHXmpcFMMhZuyk5yRw9sLdzJtdQJ5Ba7RN82rhzCiWwN6NI/GppmNK5fcDNjx06mgk5996jnvgKKP/8q/KkS3AL8w11pcphMMO1x8J1z+uFZUF/eUnwO/vwtLXoW8jL89aUC9y6HVAFeY9wk499fNSYO0P13/WMg82Zqa/LfHh123iZz5//w6viGuv1tpByA/q2SfzbBBVAuo1RFiO7h+hpb+gssKN8VQuCl7yek5TFy6mym/7yc7z/Wv8IaRQdzfrT59WlbHS52OK5+8LNi5wBV0ts+DvEzAgGr1Xf9TPDmpYPRFEBxzqqk+eSssHOuatRlcoajTCOjyIPgGW/ZxRM6Zabr+3M8fDan7XPuqt3a1Rqb9Ceu/hH3LTx3vEwzN+7mCTq1Orr8LpumaoyplOxze7ro9fPL3zMRzr8Xu6/o7V63+X1pkTmyB4X95r8QTrTonWnaO7j7xc4+rNdY7EGq2dYWYWh2hRlvwK/vvU4WbYijclJ9jWXlMWr6HSb/uJSOnAIBaVQO47/L63NSmpkZWVVb5Oa7/cVapAz6B53bOvl9dXw4nV1QPCIfL/gttBpesb4BIeTq4Dub+H+w/0fckOMYValreUnRahaN7YP1UV9A5GYDA9XckoJrrdm1xHfoDqkHgiVGMQZF/Gcl4YguKdM1CHlLjwqZzcDpct9UCIyyZ/0rhphgKN+UvPSefz37bx0fL9nA0y9XptH3dqrx/RxuqBOqLSc6RabpGbS0c6/pXJECVunDxQFfrT1Qz1/+8S9LH6/gxV+vQ4S2uYe8B4UWHtweGu5rr1W9MSiL9ECx8BtZ/4Xrs5Q9dHnK1OBYX6E9OxbD+C9g060QL5wmGHarWhfBGri2iMYQ3hvCG5dJqUhG4TbhZsmQJL7/8MmvWrOHQoUPMnDmTfv36ndO5y5cv57LLLqNFixbEx8ef83sq3FgnO6+AL1cm8Mb87WTkFlCnWgAfD25HvQgt5Cgl4MiHtZ/CohdcfQv+yi/UtUhoZDNX2Ils7vqJAYe3QfJmOLwVkre4fv7TchR/Zfc5FXSCoqFmO6jd2dUpWvP6yPFjrj9bh7eeCMpbXR1tT/Ypa3mLq7WmpB2GT97KNU1XkKlar9Ivi+I24WbOnDksX76cNm3acMMNN5xzuElNTaVNmzY0aNCApKQkhRs3sy0xg6GTV3Eg9ThhAd68d0cbOtarZnVZ4m5yM2HdZ65bVUmb4cgOcBaU/HVCakJkU1ez/fGjRYe4//Vfzn9n93H1Najd2bXFtlc/IHeVmwmrPnSFFC9f8PL7558Fx0+FmcPbIDPpzK8Z2wF6jIOabcr3s3gwtwk3f2UYxjmHm1tvvZWGDRtit9uZNWuWwo0bOpyRyz2friY+IRVvu8G4G1pyU5vS710vlUhBrqtvQtImSN7kCjzJmyH9gOv54BjX3BuRTU/8bOb6F3FxTfp52a45fE7O7XNsj+u2wb5fT/9SM+yuuXxqd3a9blht14SGITXLpn+CabpaDQpyXX0qbPbSf4+ykn3UFQyCIl23Eq1qAXPkw5rJsPhFV6A9XyE1Xf/NI5ue+Nkcalys25mlrCTf3263It6kSZPYvXs3n3/+Oc8+++xZj8/NzSU399RcG+np5zDLqpS5iGBfpt7bkUenr+eHPw7x2Ffr2ZuSxcirGmnIuJwfL1/XEPLoFkX3Hz/m+ulfpeSv6RMAPrVcc3ac1OFfrmBxdLdrlMu+X10/U/e75vg5uLboaxh21y2Jk2EnrI7r9fxCAMM1jNawub4IC38/seWdmOk1M+n0n5lJpyZONOwQUt01/Da0piswhNaE0FjXewdGum6T5Ga4OqbmZri2nLRTv+cfd7U+Ne4F3v4lv1bnIu0A/DbeFSj+OhVAQDVX/SE1T/ysfupzRDR23RYszaBgmrBpJvz8P9d/R3D132o1wPVcQY4rNBb5eeJ3m93VzyXiREiuRH1e3IlbhZsdO3YwatQoli5dipfXuZU+btw4xo4dW8aVyfnw87bz9m2tqRseyPhfdjL+l53sOZLFqzfH4eftRv8KlYrtfELN2RjGqSG1Fw907UtNcLXqJPzuGv2Sus8VeBx5rp+p+2Hv0jKoxeaa+DAtwbVdiN/fdXWgbnada2K52l1KZ7HUI7tg+RsQ/+WpuVaCol3hquC4qzN39hHXRI5n4l/lRJj4S+tIRFNXy09JQ8/uxbBgjGskE7iC02X/hYsHaeSdB3Gb21IOh4OOHTty1113MWzYMACefvrps96WOlPLTWxsrG5LVTAz1vzJ499sIN9h0io2jIkD2xIRXLk7z4kHcDpdLSyp++DYvqI/84+7Jigs3My/PXa6WlCCo139gYKiTv3+1312b9d7pP15aks/cOL3BFdrSXaKa44g3xBXvyDfYFdrg28w+Iae6CtkwtYfigakkJrQ8mZXp9jIpiX//EmbYOlrsOkb1+cBqN0VLhkJ9a9wPT5+DNIPntj+/MvvB1zX6theisye+1d+Ya7Wkyq1T7RW1XC1Vp383S/sVPg5tAEWPA27Froe+wRB5weh03Dw1aAGd+CRfW5SU1OpUqUKdvupf9E7nU5M08Rut/PTTz9xxRVXnPV91Oem4lqx+wj/+mwNacfzqRHmz6Qh7WgUpQ6aIhfM6Ty3FpiTQ5E3TIVN30Ju2qnnolu6Qk6NNq7bdd4ntpO/231OBYmEVbD0Vdg+59T5DXu4Qk2tjiWrPf+4qy/V4W2uIfuHt7lGux3bcyow/RPvQFfI8a/ialEDsHlD26Fw6b8hKKJktYilPDLcOJ1ONm/eXGTfO++8w88//8yMGTOoW7cugYFnnxBM4aZi2304k7s+Wc2elCyCfb14/842dG4QbnVZIpVPfg5snwsbprmW0TjbSDTDfiLw+P2lc67hmm2360iIaVn69R05EXr+3mKVfsB1m+vvWtwEVzzhGlYtbsdtOhRnZmayc+fOwsd79uwhPj6eqlWrUqtWLR5//HEOHDjAp59+is1mo0WLoh0FIyMj8fPzO22/uK96EUF8c19n/vXZGlbuPcqgSSt54YaW3KiRVCLly9vPFUya94OsI65bS5tmueYGys92jSTLzz7Vh8Z0uNZMyssAmxfE3QpdHoHwBmVX38klO84k/7jr9lban66aoy+CqOZlU4tUOJaGm9WrV9OtW7fCxyNHjgRg0KBBTJ48mUOHDrF//xlWGxWPViXQh0/vas+/Z2zgu/UHefSr9fx57DgPXtlAK4yLWCGwGrS/x7X9nSPfNeFc/vEToSfLNeze6ls+3v6nOn1LpVNhbkuVF92Wch9Op8lL87bx3uJdANzcpibP33CR1qQSEamESvL9rW8JqbBsNoNRvZrwbL8W2Az4as2fDJ28ioycfKtLExGRCkzhRiq8OzrWZuLAtvh721m6I4Wb3/uNQ2nHrS5LREQqKIUbcQtXNo1i2r86Eh7ky9bEDK6f8CubD2q2aREROZ3CjbiNljXDmHl/ZxpEBpGYnkP/939jyfYLWA9GREQ8ksKNuJXYqgF8PawzHepWJTO3gCGTVzFp+R4qWb94EREphsKNuJ3QAG8+vas9N7SugcNpMva7zYz6+g9yCxxWlyYiIhWAwo24JV8vO6/2j+P/ejfBZsC01QncPvF3Dmfknv1kERHxaAo34rYMw+DeS+vz0eB2BPt5sXrfMfqOX8bGA2lnP1lERDyWwo24vW6NI5k1vAv1wgM5mJbDTe/9ynfrD1pdloiIWEThRjxC/YggZg7vwqWNIsjJd/LAl+t4Zd42nE51NBYRqWwUbsRjhPp7M2lwO+65pC4A43/Zyb2frdGMxiIilYzCjXgUu83giWua8erNcfh42ViwJYlr317Ga/O3sz4hVS05IiKVgBbOFI+1bv8x/vXZGpL/MoIqPMiXyxtHcEWTSLo2DCfEz9vCCkVE5FyV5Ptb4UY8Wlp2PvM2JfLz1mSW7UwhM7eg8Dkvm0HbOlW4okkkVzaNon5EkIWViohIcRRuiqFwU3nlFThZvfcoP29N5udtyew+nFXk+ds71OLJa5rh72O3qEIREfknCjfFULiRk/YdyXIFna3JLN2RAkD9iEDevLU1LWqEWlydiIj8lcJNMRRu5EyW7jjMo9PXk5yRi7fd4N89GnN313rYbIbVpYmICCX7/tZoKRHgkoYRzH34Uq5uFkW+w+T5H7dy58e/k5iWY3VpIiJSQgo3IidUDfTh/TvbMO6Gi/D3trN85xF6vrmEuRsPWV2aiIiUgMKNyF8YhsFt7Wvx/YNduahGKKnZ+Qz7fC3/nbGBrL+MtBIRkYpL4UbkDOpHBPH1fZ257/L6GCdWHb/27WVsS8ywujQRETkLhRuRf+DjZeO/PZvwxd0diQn1Y09KFgM//p1DacetLk1ERIqhcCNyFp3qV2POQ5fQKCqIpPRchkxaVWQyQBERqVgUbkTOQViADx8Pbkd4kC9bEzMYPmUtBQ6n1WWJiMgZKNyInKOaVQL4aFBb/LxtLN5+mDGzN1HJpokSEXELCjciJRAXG8abt7bGMGDK7/v5cOkeq0sSEZG/UbgRKaEezaN5ondTAJ6fs0Xz4IiIVDAKNyLn4a6udRnYqTamCQ9Piyc+IdXqkkRE5ASFG5HzYBgGo69txhVNIsnJd3L3J6tIOJptdVkiIoLCjch587LbePu21jSLCSElM48hk1eRdjzf6rJERCo9hRuRCxDo68XHg9sRHeLHzuRM7vt8DXkFGiIuImIlhRuRCxQd6sfHg9sR6GPn111HeGR6PDn5DqvLEhGptBRuREpBs+ohjL/9YrxsBj9sOMQtH6wgKT3H6rJERColhRuRUtKtcSSfDm1PWIA36xNS6fP2MtbtP2Z1WSIilY7CjUgp6twgnNnDu9IoKojkjFxu+WAF36z90+qyREQqFYUbkVJWq1oA39zfhe5No8grcDJy+nqe/3ELDqeWahARKQ8KNyJlIMjXiw/ubMMDVzQA4IMluxmqoeIiIuVC4UakjNhsBo9e3Zi3b2tduNjm9ROWs+twptWliYh4NIUbkTLWJ646M4Z1pnqoH7tTsug3YTmLtiVbXZaIiMdSuBEpBy1qhPLtiK60rV2FjJwChkxexcvztlLg0IR/IiKlzdJws2TJEvr06UP16tUxDINZs2YVe/w333zDVVddRUREBCEhIXTq1Il58+aVT7EiFygi2Jcp93RgQIdamCZM+GUXt36wgoOpx60uTUTEo1gabrKysoiLi2PChAnndPySJUu46qqr+PHHH1mzZg3dunWjT58+rFu3rowrFSkdvl52nr/+IsYPaE2wrxer9x2j91tLmb85yerSREQ8hmGaZoUYn2oYBjNnzqRfv34lOq958+bccsstjB49+pyOT09PJzQ0lLS0NEJCQs6jUpHSsf9INiO+XMuGP9MAGNKlDqN6NcHXy25xZSIiFU9Jvr/dus+N0+kkIyODqlWr/uMxubm5pKenF9lEKoJa1QKYMawzd3WtC8Ck5Xu56d3f2Hcky+LKRETcm1uHm1deeYXMzEz69+//j8eMGzeO0NDQwi02NrYcKxQpno+XjaeubcaHA9sSFuDNHwfSuOatZcxef9Dq0kRE3JbbhpsvvviCsWPHMn36dCIjI//xuMcff5y0tLTCLSEhoRyrFDk33ZtF8eODl9C2dhUycwt48Mt1jPp6A+k5mvRPRKSk3DLcTJ06lbvvvpvp06fTvXv3Yo/19fUlJCSkyCZSEVUP82fqvR0Z0a0BhgFTVyVw5auL+Tb+ABWka5yIiFtwu3Dz5ZdfMmTIEL788kuuueYaq8sRKVVedhuP9WjMlLs7UDc8kMMZuTw0NZ4BE39nZ3KG1eWJiLgFS8NNZmYm8fHxxMfHA7Bnzx7i4+PZv38/4LqlNHDgwMLjv/jiCwYOHMirr75Khw4dSExMJDExkbS0NCvKFykzneuHM/fhS3j0qkb4etn4bfcRer25lBfnbiU7r8Dq8kREKjRLh4IvWrSIbt26nbZ/0KBBTJ48mcGDB7N3714WLVoEwOWXX87ixYv/8fhzoaHg4m4SjmYz9rtNLNjiWrKhRpg/o/s04+pmURiGYXF1IiLloyTf3xVmnpvyonAj7mr+5iSenr2JAydmNO7WOIKx17WgVrUAiysTESl7lWaeG5HK5KpmUSwYeRnDu9XH227wy7bDXPX6Yt5dtEtrVImI/IXCjYgb8fex8+8eTZj78KV0bRBOboGTF+du5fp3fmVroiaoFBEBhRsRt1Q/IojP7mrPKzfHEeLnxR8H0ujz9jLeWLCdvAK14ohI5aZwI+KmDMPgpjY1WTDyMq5uFkW+w+SNBTu4bvwyNvyZanV5IiKWUbgRcXORIX68f2cbxg9oTdVAH7YmZtBvwnJemLOVnHyH1eWJiJQ7hRsRD2AYBte2rM78Ry7lurjqOE14b/Euer+5lNV7j1pdnohIuVK4EfEg1YJ8eeu21kwc2JbIYF92p2Rx8/u/MebbjWRonSoRqSQUbkQ80FXNopg/8jL6t62JacInv+3jqteW8NOmRKtLExEpcwo3Ih4q1N+bl26KY8rdHahdLYDE9Bzu/WwN932+hqT0HKvLExEpMwo3Ih6uS4Nw5j18KfddXh+7zWDOxkS6v7qYz1fsw+msVBOUi0gloXAjUgn4edv5b88mfP9AV+Jiw8jILeDJWRvp//5v7EjSauMi4lkUbkQqkaYxIXxzX2fG9GlGgI+d1fuO0futpbw2fzu5BRo2LiKeQeFGpJKx2wyGdKnL/JGXcWWTSPIdJm8t3EHf8cvZfyTb6vJERC6Ywo1IJVUjzJ8PB7XlndsvptqJyf/6jF/Gom3JVpcmInJBFG5EKjHDMOh9UQzfP9iVVrFhpB3PZ8jkVYz/eYc6G4uI21K4ERFiQv2Z9q+O3Na+FqYJr/y0nWGfr9HEfyLilhRuRAQAXy874264iBduuAgfu42fNifRd8JydiZrNJWIuBeFGxEp4tb2tZg+rBMxoX7sPpxF3/HLmbtRMxuLiPtQuBGR07SKDeO7B7rSoW5VsvIcDPt8DS/P24pD/XBExA0o3IjIGYUH+fL53R24q2tdACb8sou7PllFTr7mwxGRik3hRkT+kbfdxlPXNuPNW1vh521j0bbD3PPpagUcEanQzivcJCQk8OeffxY+XrlyJQ8//DAffPBBqRUmIhVH31Y1+HRoBwJ87CzdkcKwz9doRmMRqbDOK9wMGDCAX375BYDExESuuuoqVq5cyRNPPMEzzzxTqgWKSMXQvm5VPh7crrAF577P1yrgiEiFdF7hZuPGjbRv3x6A6dOn06JFC3799VemTJnC5MmTS7M+EalAOtarxseD2uHrZePnrcmM+GIdeQVOq8sSESnivMJNfn4+vr6+ACxYsIDrrrsOgCZNmnDo0KHSq05EKpzODcL5cFBbfLxszN+cxANfriXfoYAjIhXHeYWb5s2b895777F06VLmz59Pz549ATh48CDVqlUr1QJFpOK5pGEEEwe2xcduY96mJB6auo4CBRwRqSDOK9y8+OKLvP/++1x++eXcdtttxMXFATB79uzC21Ui4tkuaxTB+3e2wdtu8OMfiTw8LV4BR0QqBMM0zfOalcvhcJCenk6VKlUK9+3du5eAgAAiIyNLrcDSlp6eTmhoKGlpaYSEhFhdjojbW7A5ifumrCHfYdK3VXVe698Ku82wuiwR8TAl+f4+r5ab48ePk5ubWxhs9u3bxxtvvMG2bdsqdLARkdLXvVkU4wdcjJfN4Nv4g/z7q/WayVhELHVe4aZv3758+umnAKSmptKhQwdeffVV+vXrx7vvvluqBYpIxdejeTRv39Yau83gm3UH+L9v/sCpgCMiFjmvcLN27VouueQSAGbMmEFUVBT79u3j008/5a233irVAkXEPfS6KIY3b22FzYBpqxMYPXsj53nXW0TkgpxXuMnOziY4OBiAn376iRtuuAGbzUbHjh3Zt29fqRYoIu7j2pbVebV/HIYBn6/Yz/++36KAIyLl7rzCTYMGDZg1axYJCQnMmzePq6++GoDk5GR10hWp5K5vXZNx118EwMfL9/DSvG0KOCJSrs4r3IwePZrHHnuMOnXq0L59ezp16gS4WnFat25dqgWKiPu5tX0tnunbHIB3F+3irYU7La5IRCqT8x4KnpiYyKFDh4iLi8Nmc2WklStXEhISQpMmTUq1yNKkoeAi5efDpbt59octAPy3ZxPuu7y+xRWJiLsqyfe31/m+SXR0NNHR0YWrg9esWVMT+IlIEXdfUo/cAicvz9vGi3O34uNl466uda0uS0Q83HndlnI6nTzzzDOEhoZSu3ZtateuTVhYGP/73/9wOjVDqYicMrxbAx68siEA//t+M5+t0KADESlb59Vy88QTT/DRRx/xwgsv0KVLFwCWLVvG008/TU5ODs8991ypFiki7u2R7g3JK3Dy3uJdPDVrI752G/3bxVpdloh4qPPqc1O9enXee++9wtXAT/r222+5//77OXDgQKkVWNrU50bEGqZp8sz3m5m0fC+GAU/0bspdXetiGFqqQUTOrsyXXzh69OgZOw03adKEo0ePns9LioiHMwyD0dc2486OtTFNePaHLTw8LZ7svAKrSxMRD3Ne4SYuLo7x48eftn/8+PG0bNnynF9nyZIl9OnTh+rVq2MYBrNmzTrrOYsWLeLiiy/G19eXBg0aMHny5BJULiJWMgyDZ/o25+k+zQrXorrhnV/ZdyTL6tJExIOcV5+bl156iWuuuYYFCxYUznHz22+/kZCQwI8//njOr5OVlUVcXBxDhw7lhhtuOOvxe/bs4ZprrmHYsGFMmTKFhQsXcvfddxMTE0OPHj3O56OISDkzDIPBXerSNCaE4V+sY2tiBn3eXsabt7WmW2MtvCsiF+6857k5ePAgEyZMYOvWrQA0bdqUe++9l2effZYPPvig5IUYBjNnzqRfv37/eMx///tffvjhBzZu3Fi479ZbbyU1NZW5c+ee0/uoz41IxZGYlsN9U9awbn8qhgGPdG/EiG4NsNnUD0dEiirJ9/d5h5szWb9+PRdffDEOh6PE555LuLn00ku5+OKLeeONNwr3TZo0iYcffpi0tLQznpObm0tubm7h4/T0dGJjYxVuRCqI3AIHz3y3mSm/7wege9MoXrsljhA/b4srE5GKpMw7FFslMTGRqKioIvuioqJIT0/n+PHjZzxn3LhxhIaGFm6xsRp+KlKR+HrZee76i3jpxpb4eNlYsCWJvuOXsz0pw+rSRMRNuVW4OR+PP/44aWlphVtCQoLVJYnIGfRvF8uMYZ2oHurHnpQs+k1YznfrD1pdloi4IbcKN9HR0SQlJRXZl5SUREhICP7+/mc8x9fXl5CQkCKbiFRMLWuG8d0DXelcvxrZeQ4e+HIdT83aSG5ByW91i0jlVaLRUmcb0ZSamnohtZxVp06dThuNNX/+/MIRWyLi/qoF+fLp0Pa8vmA7E37ZxWcr9hGfkMqEARdTq1qA1eWJiBsoUcvNX/uunGmrXbs2AwcOPOfXy8zMJD4+nvj4eMA11Ds+Pp79+10dCx9//PEirzds2DB2797Nf/7zH7Zu3co777zD9OnTeeSRR0ryMUSkgvOy2/h3jyZMGtKOKgHe/HEgjWveXsrcjYlWlyYibqBUR0uV1KJFi+jWrdtp+wcNGsTkyZMZPHgwe/fuZdGiRUXOeeSRR9i8eTM1a9bkqaeeYvDgwef8nhoKLuJeDqYe54Ev17Fm3zEAhnapy6heTfDxcqu76iJygSwbCu4OFG5E3E++w8nL87bxwZLdALSKDWP8gNbUrKLbVCKVhccOBReRysnbbuP/ejflw4FtCfX3Jj4hlWveWsbCLUlnP1lEKh2FGxFxG92bRfH9A12JqxlK2vF87vpkNc/9sFmjqUSkCIUbEXErsVUD+GpYZwZ3rgPAxKV7uO7t5Ww+mG5tYSJSYSjciIjb8fGy8fR1zflwYFvCg3zYlpRB3wnLeGfRThzOStWNUETOQOFGRNxW92ZRzHv4Uq5uFkW+w+Sludvo//5v7DuSZXVpImIhhRsRcWvVgnx5/842vHxTS4J8vViz7xi93lzKF7/vp5INBhWRExRuRMTtGYbBzW1jmfPQJbSvW5XsPAf/N/MP7vpkNckZOVaXJyLlTOFGRDxGbNUApt7TkSd6N8XHbuPnrcn0eH0Jc/44ZHVpIlKOFG5ExKPYbAb3XFqP2Q90oWlMCMey87lvyloemrqO1Ow8q8sTkXKgcCMiHqlJdAizhnfm/svrYzPg2/iDXP36En7eqon/RDydwo2IeCxfLzv/6dmEr+/rTL2IQJIzchk6eTX/mbGe9Jx8q8sTkTKicCMiHq91rSr8+OAl3NW1LoYB01f/Sc/Xl7BsR4rVpYlIGVC4EZFKwc/bzlPXNmPavZ2oVTWAg2k53PHR7zw1ayNZuQVWlycipUjhRkQqlfZ1qzLnoUu4s2NtAD5bsY9eby5l5Z6jFlcmIqVF4UZEKp1AXy/+168Fn9/Vgeqhfuw/ms2tH/zGtFX7rS5NREqBwo2IVFpdG4Yz95FLub51DZwm/PfrP/hw6W6ryxKRC6RwIyKVWoifN6/1j+Nfl9YD4NkftvD6/O1aukHEjSnciEilZxgGo3o14bGrGwHw5sId/O/7LQo4Im5K4UZEBFfAGXFFQ57u0wyAj5fvYdTXf+BwKuCIuBuFGxGRvxjcpS6v3ByHzYBpqxN4cOo68gqcVpclIiWgcCMi8jc3tanJO7dfjLfd4IcNh7j3s9Ucz3NYXZaInCOFGxGRM+jZIoaPBrXDz9vGom2HGTRpJRlaskHELSjciIj8g0sbRfDZXR0I9vVi5Z6jDJj4O0eztLK4SEWncCMiUox2dary5b0dqRrowx8H0rjp3V9JOJptdVkiUgyFGxGRs2hRI5Tp/+pEjTB/dqdkccO7v7LpYJrVZYnIP1C4ERE5Bw0ig/jm/s40iQ7mcEYut7y/guU7taq4SEWkcCMico6iQvyYPqwTHetVJTO3gMGTVjJ7/UGryxKRv1G4EREpgRA/bz4Z2p5rLooh32Hy4JfrtB6VSAWjcCMiUkK+Xnbevq01gzvXAVzrUT33w2acms1YpEJQuBEROQ82m8GYPs0Y1asJABOX7uGR6fGazVikAlC4ERE5T4ZhMOyy+rzWPw4vm8G38QcZOnkV6ZrsT8RSCjciIhfohotr8tHgdgT42Fm2M4UrXlnEZyv2ke9QK46IFRRuRERKwWWNIph6b0fqVAsgJTOPp2ZtpMfrS5i78RCmqb44IuXJMCvZ37r09HRCQ0NJS0sjJCTE6nJExMPkO5x8uXI/by7YwZETSzVcXCuMx3s3pV2dqhZXJ+K+SvL9rXAjIlIGMnLymbhkNxOX7uF4vmtF8e5NoxjVqzENIoMtrk7E/SjcFEPhRkTKU3J6Dm8s3MG0VQk4nCY2A25pF8sj3RsRGeJndXkibkPhphgKNyJihZ3Jmbw0dys/bU4CINDHzsPdGzG4Sx287er+KHI2CjfFULgRESut3nuUZ3/YQnxCKgANI4MY27c5neuHW1uYSAWncFMMhRsRsZrTaTJj7Z+8MGcrR090Ou4TV50nejclOlS3qkTOpCTf3xWiLXTChAnUqVMHPz8/OnTowMqVK4s9/o033qBx48b4+/sTGxvLI488Qk5OTjlVKyJyYWw2g/5tY/nl0csZ2Kk2NgO+W3+QK19dxPuLd2mWY5ELZHm4mTZtGiNHjmTMmDGsXbuWuLg4evToQXJy8hmP/+KLLxg1ahRjxoxhy5YtfPTRR0ybNo3/+7//K+fKRUQuTGiAN8/0bcHsEV25uFYYWXkOxs3ZSq83l7B8Z4rV5Ym4LctvS3Xo0IF27doxfvx4AJxOJ7GxsTzwwAOMGjXqtONHjBjBli1bWLhwYeG+Rx99lN9//51ly5ad9f10W0pEKiKn0+TrE7eqjvzlVtUz1zWnSqCPxdWJWM9tbkvl5eWxZs0aunfvXrjPZrPRvXt3fvvttzOe07lzZ9asWVN462r37t38+OOP9O7du1xqFhEpCzabwc1tY/n5scsZ9JdbVT3eWMKS7YetLk/ErXhZ+eYpKSk4HA6ioqKK7I+KimLr1q1nPGfAgAGkpKTQtWtXTNOkoKCAYcOG/eNtqdzcXHJzcwsfp6enl94HEBEpZaH+3ozt24Ib29TkkWnx7DqcxcCPVzK4cx1G9WqCn7fd6hJFKjzL+9yU1KJFi3j++ed55513WLt2Ld988w0//PAD//vf/854/Lhx4wgNDS3cYmNjy7liEZGSa1kzjO8fuISBnWoDMPnXvVz79jI2HkizuDKRis/SPjd5eXkEBAQwY8YM+vXrV7h/0KBBpKam8u233552ziWXXELHjh15+eWXC/d9/vnn3HvvvWRmZmKzFc1rZ2q5iY2NVZ8bEXEbi7Yl8+8ZGzickYu33eCRqxrxr0vrY7cZVpcmUm7cps+Nj48Pbdq0KdI52Ol0snDhQjp16nTGc7Kzs08LMHa7q5n2TDnN19eXkJCQIpuIiDu5vHEk8x6+lJ7No8l3mLw0dxu3fbCChKPZVpcmUiFZfltq5MiRTJw4kU8++YQtW7Zw3333kZWVxZAhQwAYOHAgjz/+eOHxffr04d1332Xq1Kns2bOH+fPn89RTT9GnT5/CkCMi4mmqBvrw7h0X8/JNLQn0sbNy71F6vbmUr9f8aXVpIhWOpR2KAW655RYOHz7M6NGjSUxMpFWrVsydO7ewk/H+/fuLtNQ8+eSTGIbBk08+yYEDB4iIiKBPnz4899xzVn0EEZFyYRiuEVUd6lZj5PR4Vu87xqNfrSfP4eS29rWsLk+kwrB8npvypnluRMQTOJwmL83dyvtLduNtN/jino60q1PV6rJEyozb9LkREZHzY7cZjOrVhGsuiiHfYTLsszUcSD1udVkiFYLCjYiImzIMg5dvbkmzmBCOZOVx76erOZ7nsLosEcsp3IiIuLEAHy8+GNiGaoE+bDqYzr9nrD/jyFGRykThRkTEzdWsEsC7d7TBy2bw/YZDvLNol9UliVhK4UZExAO0r1uVZ/q2AOCVn7axYHOSxRWJWEfhRkTEQwzoUIs7O9bGNOHhafHsSMqwuiQRSyjciIh4kNF9mtGxXlUycwu4+9PVpGbnWV2SSLlTuBER8SDedhvv3N6GmlX82XckmxFfrKPA4bS6LJFypXAjIuJhqgb6MHFgWwJ87CzbmcLzP261uiSRcqVwIyLigZrGhPBa/zgAPl6+h9fmb8fp1BBxqRwUbkREPFTPFjE8dnUjAN5auIN7P1tDRk6+xVWJlD2FGxERDzbiioa8fFNLfLxsLNiSRL8Jy9l1ONPqskTKlMKNiIiHu7ltLF/9qxPRIX7sOpxFv/HLNQ+OeDSFGxGRSiAuNozvHuhK+zpVyTgxTPzNBTvUD0c8ksKNiEglERHsy+d3d2Bgp9oAvL5gO8M+Vz8c8TwKNyIilYiPl41n+rbgpZta4mO38dNm9cMRz6NwIyJSCfVvG8v0YUX74Xy/4aBWFBePoHAjIlJJtYoNY/YDXWhXpwoZuQWM+GId//psDUnpOVaXJnJBFG5ERCqxyGA/ptzdkQevbIiXzeCnzUl0f20xU1fuVyuOuC2FGxGRSs7Hy8bIqxrx/YNdiasZSkZOAaO++YPbP/ydfUeyrC5PpMQUbkREBIAm0SF8c38XnrymKX7eNn7ddYQebyxh4pLdWnxT3IrCjYiIFLLbDO6+pB7zHr6UzvWrkZPv5Lkft3Dju7+yNTHd6vJEzonCjYiInKZ2tUCm3N2BF2+8iGA/L9b/mca1by1j3I9bSM3Os7o8kWIZZiXrMZaenk5oaChpaWmEhIRYXY6ISIWXlJ7DU7M28tOJJRuCfb2465K6DO1alxA/b4urk8qiJN/fCjciInJWpmny89ZkXvlpO1sOuW5PhQV4c++l9RjcuQ4BPl4WVyieTuGmGAo3IiLnz+k0mbMxkdcXbGdnsmtW4/AgH+67vAG3d6iFn7fd4grFUyncFEPhRkTkwjmcJrPXH+D1+TvYfzQbgOgQP0Zc0YD+bWPx8VKXTildCjfFULgRESk9+Q4nM9b8ydsLd3AwzTWzcYCPnTrVAqkbHkid8ABqn/i9drUAIoJ8MQzD4qrFHSncFEPhRkSk9OUWOJi6MoHxv+zkcEbuPx4X6GOndrVA2tetyn96NlZfHTlnCjfFULgRESk7+Q4n+45kszcli71HTmwp2ew9ksWB1OP89RtncOc6PH1dc+uKFbdSku9vRWYRESk13nYbDSKDaBAZdNpzuQUOEo4eZ8XuIzw5ayOf/LaXa1vG0LZOVQsqFU+mHl8iIlIufL3sNIgM4o6Otbm5TU1ME/7z9QZy8h1WlyYeRuFGRETK3ZPXNCMi2Jfdh7N4a+EOq8sRD6NwIyIi5S40wJv/9W0BwPtLdrPxQJrFFYknUbgRERFL9GwRzTUXxeBwmvxnxgbytfK4lBKFGxERsczT1zUnLMCbzYfSeX/xLqvLEQ+hcCMiIpaJCPZlTJ9mALy1cCc7kjIsrkg8gcKNiIhYql+rGnRrHEGew8l/v96Aw1mppl+TMqBwIyIiljIMg+euv4ggXy/W7k/lk1/3Wl2SuDmFGxERsVz1MH9G9WoCwMvztrH/SLbFFYk7qxDhZsKECdSpUwc/Pz86dOjAypUriz0+NTWV4cOHExMTg6+vL40aNeLHH38sp2pFRKQsDGhfiw51q3I838HjMzdQyVYHklJkebiZNm0aI0eOZMyYMaxdu5a4uDh69OhBcnLyGY/Py8vjqquuYu/evcyYMYNt27YxceJEatSoUc6Vi4hIabLZDF68sSV+3jaW7zzCtFUJVpckbsryhTM7dOhAu3btGD9+PABOp5PY2FgeeOABRo0addrx7733Hi+//DJbt27F29u7xO+nhTNFRCq2iUt289yPWwj29eLHhy4htmqA1SVJBVCS729LW27y8vJYs2YN3bt3L9xns9no3r07v/322xnPmT17Np06dWL48OFERUXRokULnn/+eRwOrU0iIuIJhnatS1xsGBm5BfSdsJylOw5bXZK4GUvDTUpKCg6Hg6ioqCL7o6KiSExMPOM5u3fvZsaMGTgcDn788UeeeuopXn31VZ599tkzHp+bm0t6enqRTUREKi67zWDCgNY0iwnhaFYeAz9eyZsLduDUEHE5R5b3uSkpp9NJZGQkH3zwAW3atOGWW27hiSee4L333jvj8ePGjSM0NLRwi42NLeeKRUSkpGpWCeCb+ztza7tYTBNeX7CdwZNXcTQrz+rSxA1YGm7Cw8Ox2+0kJSUV2Z+UlER0dPQZz4mJiaFRo0bY7fbCfU2bNiUxMZG8vNP/0D/++OOkpaUVbgkJ6qAmIuIO/LztvHBjS16+qSW+XjaWbD/MtW8tZd3+Y1aXJhWcpeHGx8eHNm3asHDhwsJ9TqeThQsX0qlTpzOe06VLF3bu3InTeWqBte3btxMTE4OPj89px/v6+hISElJkExER93Fz21hmDe9C3fBADqbl0P/93/jk170aKi7/yPLbUiNHjmTixIl88sknbNmyhfvuu4+srCyGDBkCwMCBA3n88ccLj7/vvvs4evQoDz30ENu3b+eHH37g+eefZ/jw4VZ9BBERKWNNY0KYPaILvS+KJt9hMmb2Jh6cGk9mboHVpUkF5GV1AbfccguHDx9m9OjRJCYm0qpVK+bOnVvYyXj//v3YbKcyWGxsLPPmzeORRx6hZcuW1KhRg4ceeoj//ve/Vn0EEREpB8F+3kwYcDEfL9/LuB+38N36g2w6mMb42y6mWXW1ysspls9zU940z42IiPtbs+8ow6esIzE9By+bwbDL6jPiigb4edvPfrK4JbeZ50ZEROR8tKldlR8e7EqP5lEUOE3G/7KT3m8u5ffdR6wuTSoAhRsREXFL1YJ8ef/Otrx3x8VEBvuyOyWLWz5Ywf/N/IP0nHyryxMLKdyIiIhb69kihvkjL+O29rUA+OL3/XR/dTHzNp15MljxfAo3IiLi9kL9vRl3w0VMvbcjdcMDSc7I5V+freG+z9eQnJ5jdXlSzhRuRETEY3SsV405D13C8G718bIZzNmYyJWvLebT3/aS73Ce/QXEI2i0lIiIeKTNB9MZ9c0GNvyZBkDtagGMvKoRfVpWx2YzLK5OSqok398KNyIi4rEcTpMpv+/jrYU7ScnMBVwTAv67RyO6NY7EMBRy3IXCTTEUbkREKp+s3AImLd/D+4t3k3FiVuO2tavwn55NaF+3qsXVyblQuCmGwo2ISOWVmp3Hu4t3MXn5XnILXH1wujWO4LEejWlePdTi6qQ4CjfFULgREZHEtBze+nkH01Yl4HC6vgb7tqrOf3s2oXqYv8XVyZko3BRD4UZERE7am5LFa/O3M3v9QQD8ve3cf3l97rm0npZyqGAUboqhcCMiIn+38UAaY7/bxKq9xwCoWcWfJ69pSo/m0ep0XEEo3BRD4UZERM7ENE2+23CI53/YQuKJif8616/GmD7NaRwdbHF1onBTDIUbEREpTnZeAe8u2sX7S3aTV+DEbjO4o0MtHrmqEWEBPlaXV2kp3BRD4UZERM5FwtFsnv1hM/M2JQFQJcCbR69uzG3ta2HXJIDlTuGmGAo3IiJSEst3pjD2u01sT8oEoFlMCGP7NqddHc2PU54UboqhcCMiIiVV4HDy2Yp9vD5/O+k5rkkA+7WqzqheTYkO9bO4uspB4aYYCjciInK+jmTm8vK8bUxbnYBpQoCPnQeuaMjQrnXw9dLQ8bKkcFMMhRsREblQG/5MZczsTazbnwpA3fBARvdpRrfGkdYW5sEUboqhcCMiIqXB6TT5Zt0BXpiztXBRziubRPLUtc2oEx5ocXWeR+GmGAo3IiJSmtJz8nl74Q4mLd9LgdPEx26jT1x17uhYi1axYZoEsJQo3BRD4UZERMrCzuQMxn63maU7Ugr3tagRwh0danNdq+oE+HhZWJ37U7gphsKNiIiUFdM0WZeQyucr9vH9hkPknVh5PNjPixsvrskdHWvRIFKzHZ8PhZtiKNyIiEh5OJaVx1drEvh8xX72H80u3N+xXlXu7FiHq5pF4eNls7BC96JwUwyFGxERKU9Op8nSnSl89ts+ft6ahPPEt26ovze9L4rmurgatK9bVbMen4XCTTEUbkRExCoHUo8zdeV+pq1KIDkjt3B/VIgvfVpW57pW1bmoRqg6IZ+Bwk0xFG5ERMRqDqfJit1HmB1/kB83HiLjxKzH4Jozp09cda6Lq06DyCALq6xYFG6KoXAjIiIVSW6Bg8XbDjN7/UEWbEkiJ99Z+FyXBtUYe10LhRwUboqlcCMiIhVVZm4BCzYn8W38AZbuSCmcN2fYZfW4v1sD/Lwr7xIPCjfFULgRERF3kHA0m9HfbuSXbYcB1+2qZ/u1oEuDcIsrs0ZJvr81Bk1ERKQCiq0awMeD2/HO7RcTGezLnpQsbv/wdx6euq5wuQc5M4UbERGRCsowDHpfFMOCRy9jUKfaGAbMij/IFa8s4ovf9+N0VqqbL+dMt6VERETcxPqEVP5v5h9sOpgOQJvaVXj++otoHO35sx7rtpSIiIgHiosN49vhXXjq2mYE+thZs+8Yfd5exodLd6sV5y8UbkRERNyIl93GXV3rsuDRy+jeNJI8h5Nnf9jCkMmrOJyhvjigcCMiIuKWYkL9mTiwLc/2a4Gvl43F2w/T682lLN5+2OrSLKdwIyIi4qYMw+COjrWZPaIrjaOCScnMZdDHK3nuh82FK5JXRgo3IiIibq5xdDDfjujCwE61AZi4dA83vLuc3YczLa7MGgo3IiIiHsDP284zfVswcWBbqgR4s/FAOte+vYyvVidQyQZGK9yIiIh4kquaRTHnoUvpVK8a2XkO/j1jAw9OjSctO9/q0spNhQg3EyZMoE6dOvj5+dGhQwdWrlx5TudNnToVwzDo169f2RYoIiLiRqJD/fj87g78u0dj7DaD79Yf5Oo3FvPz1iSrSysXloebadOmMXLkSMaMGcPatWuJi4ujR48eJCcnF3ve3r17eeyxx7jkkkvKqVIRERH3YbcZDO/WgBnDOlEvIpCk9FyGTl7Nv79aT3qOZ7fiWB5uXnvtNe655x6GDBlCs2bNeO+99wgICODjjz/+x3McDge33347Y8eOpV69euVYrYiIiHtpXasKPz54CXd3rYthwFdr/qTH60tY4sFDxi0NN3l5eaxZs4bu3bsX7rPZbHTv3p3ffvvtH8975plniIyM5K677jrre+Tm5pKenl5kExERqUz8vO08eW0zpv+rE3WqBXAoLYeBH6/k8W/+IDO3wOrySp2l4SYlJQWHw0FUVFSR/VFRUSQmJp7xnGXLlvHRRx8xceLEc3qPcePGERoaWrjFxsZecN0iIiLuqF2dqvz40CUM7lwHgC9X7qfH60v4dWdKseeZpkna8XwSjmZT4Kj48+d4WV1ASWRkZHDnnXcyceJEwsPDz+mcxx9/nJEjRxY+Tk9PV8AREZFKK8DHi6eva06P5tH85+v1JBw9zoAPf+fWdrFUD/PnaFYeR7LyOJqVy5HMPI5m5XEsO498h2s4ebVAH65uHk3vi6LpWK8a3nbLe7icxtJwEx4ejt1uJympaO/tpKQkoqOjTzt+165d7N27lz59+hTuczpdCdLLy4tt27ZRv379Iuf4+vri6+tbBtWLiIi4r071qzH3oUsZN2cLn6/Yz9RVCWc9x24zOJKVx5cr9/Plyv2EBXhzdbMoel8UQ+f64fh4VYygY5gWz+zToUMH2rdvz9tvvw24wkqtWrUYMWIEo0aNKnJsTk4OO3fuLLLvySefJCMjgzfffJNGjRrh4+NT7PuVZMl0ERGRymD5zhRmrPkTXy8bVQN9qBroQ7UgH6oG+lLtxOOqgT7YbQYrdh/hxz8SmbcpkaNZeYWvEeLnxVXNormmZTRdGoTj62Uv1RpL8v1tebiZNm0agwYN4v3336d9+/a88cYbTJ8+na1btxIVFcXAgQOpUaMG48aNO+P5gwcPJjU1lVmzZp3T+ynciIiIXLgCh5OVe4/y4x+HmLsxiZTMUyuSh/h58evjVxLkW3o3iEry/W15n5tbbrmFw4cPM3r0aBITE2nVqhVz584t7GS8f/9+bLaK0cwlIiIiLl52G53rh9O5fjhjr2vB6hNBZ87GRGpXCyjVYFNSlrfclDe13IiIiJQdp9PkSFYeEcGl29+1JN/fahIRERGRUmOzGaUebEpcg6XvLiIiIlLKFG5ERETEoyjciIiIiEdRuBERERGPonAjIiIiHkXhRkRERDyKwo2IiIh4FIUbERER8SgKNyIiIuJRFG5ERETEoyjciIiIiEdRuBERERGPonAjIiIiHsXL6gLKm2magGvpdBEREXEPJ7+3T36PF6fShZuMjAwAYmNjLa5ERERESiojI4PQ0NBijzHMc4lAHsTpdHLw4EGCg4MxDKNUXzs9PZ3Y2FgSEhIICQkp1deW0+l6ly9d7/Kl612+dL3L1/lcb9M0ycjIoHr16thsxfeqqXQtNzabjZo1a5bpe4SEhOgvRznS9S5fut7lS9e7fOl6l6+SXu+ztdicpA7FIiIi4lEUbkRERMSjKNyUIl9fX8aMGYOvr6/VpVQKut7lS9e7fOl6ly9d7/JV1te70nUoFhEREc+mlhsRERHxKAo3IiIi4lEUbkRERMSjKNyIiIiIR1G4KSUTJkygTp06+Pn50aFDB1auXGl1SR5jyZIl9OnTh+rVq2MYBrNmzSryvGmajB49mpiYGPz9/enevTs7duywplg3N27cONq1a0dwcDCRkZH069ePbdu2FTkmJyeH4cOHU61aNYKCgrjxxhtJSkqyqGL39u6779KyZcvCicw6derEnDlzCp/XtS5bL7zwAoZh8PDDDxfu0zUvPU8//TSGYRTZmjRpUvh8WV5rhZtSMG3aNEaOHMmYMWNYu3YtcXFx9OjRg+TkZKtL8whZWVnExcUxYcKEMz7/0ksv8dZbb/Hee+/x+++/ExgYSI8ePcjJySnnSt3f4sWLGT58OCtWrGD+/Pnk5+dz9dVXk5WVVXjMI488wnfffcdXX33F4sWLOXjwIDfccIOFVbuvmjVr8sILL7BmzRpWr17NFVdcQd++fdm0aROga12WVq1axfvvv0/Lli2L7Nc1L13Nmzfn0KFDhduyZcsKnyvTa23KBWvfvr05fPjwwscOh8OsXr26OW7cOAur8kyAOXPmzMLHTqfTjI6ONl9++eXCfampqaavr6/55ZdfWlChZ0lOTjYBc/HixaZpuq6tt7e3+dVXXxUes2XLFhMwf/vtN6vK9ChVqlQxP/zwQ13rMpSRkWE2bNjQnD9/vnnZZZeZDz30kGma+vNd2saMGWPGxcWd8bmyvtZqublAeXl5rFmzhu7duxfus9lsdO/end9++83CyiqHPXv2kJiYWOT6h4aG0qFDB13/UpCWlgZA1apVAVizZg35+flFrneTJk2oVauWrvcFcjgcTJ06laysLDp16qRrXYaGDx/ONddcU+Tagv58l4UdO3ZQvXp16tWrx+23387+/fuBsr/WlW7hzNKWkpKCw+EgKiqqyP6oqCi2bt1qUVWVR2JiIsAZr//J5+T8OJ1OHn74Ybp06UKLFi0A1/X28fEhLCysyLG63ufvjz/+oFOnTuTk5BAUFMTMmTNp1qwZ8fHxutZlYOrUqaxdu5ZVq1ad9pz+fJeuDh06MHnyZBo3bsyhQ4cYO3Ysl1xyCRs3bizza61wIyJnNHz4cDZu3FjkHrmUvsaNGxMfH09aWhozZsxg0KBBLF682OqyPFJCQgIPPfQQ8+fPx8/Pz+pyPF6vXr0Kf2/ZsiUdOnSgdu3aTJ8+HX9//zJ9b92WukDh4eHY7fbTengnJSURHR1tUVWVx8lrrOtfukaMGMH333/PL7/8Qs2aNQv3R0dHk5eXR2pqapHjdb3Pn4+PDw0aNKBNmzaMGzeOuLg43nzzTV3rMrBmzRqSk5O5+OKL8fLywsvLi8WLF/PWW2/h5eVFVFSUrnkZCgsLo1GjRuzcubPM/3wr3FwgHx8f2rRpw8KFCwv3OZ1OFi5cSKdOnSysrHKoW7cu0dHRRa5/eno6v//+u67/eTBNkxEjRjBz5kx+/vln6tatW+T5Nm3a4O3tXeR6b9u2jf379+t6lxKn00lubq6udRm48sor+eOPP4iPjy/c2rZty+233174u6552cnMzGTXrl3ExMSU/Z/vC+6SLObUqVNNX19fc/LkyebmzZvNe++91wwLCzMTExOtLs0jZGRkmOvWrTPXrVtnAuZrr71mrlu3zty3b59pmqb5wgsvmGFhYea3335rbtiwwezbt69Zt25d8/jx4xZX7n7uu+8+MzQ01Fy0aJF56NChwi07O7vwmGHDhpm1atUyf/75Z3P16tVmp06dzE6dOllYtfsaNWqUuXjxYnPPnj3mhg0bzFGjRpmGYZg//fSTaZq61uXhr6OlTFPXvDQ9+uij5qJFi8w9e/aYy5cvN7t3726Gh4ebycnJpmmW7bVWuCklb7/9tlmrVi3Tx8fHbN++vblixQqrS/IYv/zyiwmctg0aNMg0Tddw8KeeesqMiooyfX19zSuvvNLctm2btUW7qTNdZ8CcNGlS4THHjx8377//frNKlSpmQECAef3115uHDh2yrmg3NnToULN27dqmj4+PGRERYV555ZWFwcY0da3Lw9/Dja556bnlllvMmJgY08fHx6xRo4Z5yy23mDt37ix8viyvtWGapnnh7T8iIiIiFYP63IiIiIhHUbgRERERj6JwIyIiIh5F4UZEREQ8isKNiIiIeBSFGxEREfEoCjciIiLiURRuREQAwzCYNWuW1WWISClQuBERyw0ePBjDME7bevbsaXVpIuKGvKwuQEQEoGfPnkyaNKnIPl9fX4uqERF3ppYbEakQfH19iY6OLrJVqVIFcN0yevfdd+nVqxf+/v7Uq1ePGTNmFDn/jz/+4IorrsDf359q1apx7733kpmZWeSYjz/+mObNm+Pr60tMTAwjRowo8nxKSgrXX389AQEBNGzYkNmzZ5fthxaRMqFwIyJu4amnnuLGG29k/fr13H777dx6661s2bIFgKysLHr06EGVKlVYtWoVX331FQsWLCgSXt59912GDx/Ovffeyx9//MHs2bNp0KBBkfcYO3Ys/fv3Z8OGDfTu3Zvbb7+do0ePluvnFJFSUCrLb4qIXIBBgwaZdrvdDAwMLLI999xzpmm6VisfNmxYkXM6dOhg3nfffaZpmuYHH3xgVqlSxczMzCx8/ocffjBtNpuZmJhomqZpVq9e3XziiSf+sQbAfPLJJwsfZ2ZmmoA5Z86cUvucIlI+1OdGRCqEbt268e677xbZV7Vq1cLfO3XqVOS5Tp06ER8fD8CWLVuIi4sjMDCw8PkuXbrgdDrZtm0bhmFw8OBBrrzyymJraNmyZeHvgYGBhISEkJycfL4fSUQsonAjIhVCYGDgabeJSou/v/85Heft7V3ksWEYOJ3OsihJRMqQ+tyIiFtYsWLFaY+bNm0KQNOmTVm/fj1ZWVmFzy9fvhybzUbjxo0JDg6mTp06LFy4sFxrFhFrqOVGRCqE3NxcEhMTi+zz8vIiPDwcgK+++oq2bdvStWtXpkyZwsqVK/noo48AuP322xkzZgyDBg3i6aef5vDhwzzwwAPceeedREVFAfD0008zbNgwIiMj6dWrFxkZGSxfvpwHHnigfD+oiJQ5hRsRqRDmzp1LTExMkX2NGzdm69atgGsk09SpU7n//vuJiYnhyy+/pFmzZgAEBAQwb948HnroIdq1a0dAQAA33ngjr732WuFrDRo0iJycHF5//XUee+wxwsPDuemmm8rvA4pIuTFM0zStLkJEpDiGYTBz5kz69etndSki4gbU50ZEREQ8isKNiIiIeBT1uRGRCk93z0WkJNRyIyIiIh5F4UZEREQ8isKNiIiIeBSFGxEREfEoCjciIiLiURRuRERExKMo3IiIiIhHUbgRERERj6JwIyIiIh7l/wH++lvG8p/7KQAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Draw accuracy values for training & validation\n",
    "plt.plot(history.history['accuracy'])\n",
    "plt.plot(history.history['val_accuracy'])\n",
    "plt.title('Model accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Test'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "# Draw loss for training & validation\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.title('Model loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Test'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 单机模式小结\n",
    "以上我们按照官方教程在数据集Flower 上成功微调了 InceptionV3 模型，分别是**微调顶部分类器**和**冻结底层网络层微调顶层网络层**。接下来我们将展示如何将单机模式下的微调拓展到联邦学习模式下进行微调。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 联邦学习模式进行微调\n",
    "\n",
    "### 环境设置\n",
    "首先我们初始化各个参与方。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The version of SecretFlow: 1.2.0.dev20230926\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-10-11 07:16:09,152\tINFO worker.py:1538 -- Started a local Ray instance.\n"
     ]
    }
   ],
   "source": [
    "import secretflow as sf\n",
    "\n",
    "# Check the version of your SecretFlow\n",
    "print('The version of SecretFlow: {}'.format(sf.__version__))\n",
    "\n",
    "# In case you have a running secretflow runtime already.\n",
    "sf.shutdown()\n",
    "sf.init(['alice', 'bob', 'charlie'], address=\"local\", log_to_driver=False)\n",
    "alice, bob, charlie = sf.PYU('alice'), sf.PYU('bob'), sf.PYU('charlie')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 定义Dataloader\n",
    "我们可以参考[TensorFlow下的DataBuilder教程](https://www.secretflow.org.cn/docs/secretflow/latest/zh-Hans/tutorial/CustomDataLoaderTF)定义我们自己的DataBuilder。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_dataset_builder(\n",
    "    batch_size=32,\n",
    "):\n",
    "    def dataset_builder(folder_path, stage=\"train\"):\n",
    "        import math\n",
    "\n",
    "        import tensorflow as tf\n",
    "\n",
    "        img_height = 180\n",
    "        img_width = 180\n",
    "        data_set = tf.keras.utils.image_dataset_from_directory(\n",
    "            folder_path,\n",
    "            validation_split=0.2,\n",
    "            subset=\"both\",\n",
    "            seed=123,\n",
    "            image_size=(img_height, img_width),\n",
    "            batch_size=batch_size,\n",
    "        )\n",
    "        if stage == \"train\":\n",
    "            train_dataset = data_set[0]\n",
    "            train_step_per_epoch = math.ceil(len(data_set[0].file_paths) / batch_size)\n",
    "            return train_dataset, train_step_per_epoch\n",
    "        elif stage == \"eval\":\n",
    "            eval_dataset = data_set[1]\n",
    "            eval_step_per_epoch = math.ceil(len(data_set[1].file_paths) / batch_size)\n",
    "            return eval_dataset, eval_step_per_epoch\n",
    "\n",
    "    return dataset_builder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_builder_dict = {\n",
    "    alice: create_dataset_builder(\n",
    "        batch_size=32,\n",
    "    ),\n",
    "    bob: create_dataset_builder(\n",
    "        batch_size=32,\n",
    "    ),\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 定义 SecureAggregator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:root:Create proxy actor <class 'secretflow.security.aggregation.secure_aggregator._Masker'> with party alice.\n",
      "INFO:root:Create proxy actor <class 'secretflow.security.aggregation.secure_aggregator._Masker'> with party bob.\n"
     ]
    }
   ],
   "source": [
    "from secretflow.ml.nn import FLModel\n",
    "from secretflow.security.aggregation import SecureAggregator\n",
    "\n",
    "device_list = [alice, bob]\n",
    "aggregator = SecureAggregator(charlie, [alice, bob])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 定义数据加载路径\n",
    "为了简便起见，我们在 单机模拟模式下直接加载同一处路径所对应的数据集"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = {\n",
    "    alice: path_to_flower_dataset,\n",
    "    bob: path_to_flower_dataset,\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 定义联邦学习训练参数"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "epochs = 50\n",
    "batch_size = 32\n",
    "aggregate_freq = 2\n",
    "sampler_method = \"batch\"\n",
    "random_seed = 1234\n",
    "dp_spent_step_freq = 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 微调顶部分类器\n",
    "我们只要参照教程里对模型的定义，在函数里完成我们对模型的定义即可；可以看到代码几乎不需要作任何修改，只需要进行适当的封装。\n",
    "为了方便作对比实验，我们额外添加是否加载权重的选项。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_inception_v3_model_classifier(num_classes, is_load_weight=True):\n",
    "    def create_model():\n",
    "        from tensorflow import keras\n",
    "\n",
    "        # Create model\n",
    "        # create the base pre-trained model\n",
    "        if is_load_weight:\n",
    "            base_model = InceptionV3(weights='imagenet', include_top=False)\n",
    "        else:\n",
    "            base_model = InceptionV3(weights=None, include_top=False)\n",
    "\n",
    "        # add a global spatial average pooling layer\n",
    "        x = base_model.output\n",
    "        x = GlobalAveragePooling2D()(x)\n",
    "        # let's add a fully-connected layer\n",
    "        x = Dense(1024, activation='relu')(x)\n",
    "        # and a logistic layer -- let's say we have 10 classes\n",
    "        predictions = Dense(num_classes, activation='softmax')(x)\n",
    "\n",
    "        # this is the model we will train\n",
    "        model = Model(inputs=base_model.input, outputs=predictions)\n",
    "\n",
    "        # first: train only the top layers (which were randomly initialized)\n",
    "        # i.e. freeze all convolutional InceptionV3 layers\n",
    "        for layer in base_model.layers:\n",
    "            layer.trainable = False\n",
    "\n",
    "        # Compile model\n",
    "        model.compile(\n",
    "            optimizer='rmsprop',\n",
    "            loss='sparse_categorical_crossentropy',\n",
    "            metrics=[\"accuracy\"],\n",
    "        )\n",
    "\n",
    "        return model\n",
    "\n",
    "    return create_model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 加载预训练模型权重并且微调"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:root:Create proxy actor <class 'secretflow.ml.nn.fl.backend.tensorflow.strategy.fed_avg_w.PYUFedAvgW'> with party alice.\n",
      "INFO:root:Create proxy actor <class 'secretflow.ml.nn.fl.backend.tensorflow.strategy.fed_avg_w.PYUFedAvgW'> with party bob.\n"
     ]
    }
   ],
   "source": [
    "# prepare model\n",
    "num_classes = 5\n",
    "\n",
    "# keras model\n",
    "weight_model = create_inception_v3_model_classifier(\n",
    "    num_classes=num_classes, is_load_weight=True\n",
    ")\n",
    "\n",
    "\n",
    "fed_model = FLModel(\n",
    "    device_list=device_list,\n",
    "    model=weight_model,\n",
    "    aggregator=aggregator,\n",
    "    backend=\"tensorflow\",\n",
    "    strategy=\"fed_avg_w\",\n",
    "    random_seed=1234,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:root:FL Train Params: {'self': <secretflow.ml.nn.fl.fl_model.FLModel object at 0x7fbd4bb1cb50>, 'x': {PYURuntime(alice): '/tmp/tmpc7wdf9us/datasets/flower_photos', PYURuntime(bob): '/tmp/tmpc7wdf9us/datasets/flower_photos'}, 'y': None, 'batch_size': 32, 'batch_sampling_rate': None, 'epochs': 50, 'verbose': 1, 'callbacks': None, 'validation_data': {PYURuntime(alice): '/tmp/tmpc7wdf9us/datasets/flower_photos', PYURuntime(bob): '/tmp/tmpc7wdf9us/datasets/flower_photos'}, 'shuffle': False, 'class_weight': None, 'sample_weight': None, 'validation_freq': 1, 'aggregate_freq': 2, 'label_decoder': None, 'max_batch_size': 20000, 'prefetch_buffer_size': None, 'sampler_method': 'batch', 'random_seed': 1234, 'dp_spent_step_freq': 1, 'audit_log_dir': None, 'dataset_builder': {PYURuntime(alice): <function create_dataset_builder.<locals>.dataset_builder at 0x7fbb907ee700>, PYURuntime(bob): <function create_dataset_builder.<locals>.dataset_builder at 0x7fbadc0d5040>}, 'wait_steps': 100}\n",
      "32it [01:44,  3.28s/it, epoch: 1/50 -  loss:102.29530334472656  accuracy:0.25494277477264404  val_loss:36.37749099731445  val_accuracy:0.25 ]\n",
      "32it [01:39,  3.10s/it, epoch: 2/50 -  loss:23.1259708404541  accuracy:0.27976685762405396  val_loss:10.871068954467773  val_accuracy:0.26249998807907104 ]\n",
      "32it [01:38,  3.08s/it, epoch: 3/50 -  loss:13.160760879516602  accuracy:0.2706078290939331  val_loss:15.88563346862793  val_accuracy:0.2708333432674408 ]\n",
      "32it [01:38,  3.06s/it, epoch: 4/50 -  loss:11.84687328338623  accuracy:0.2930890917778015  val_loss:27.221759796142578  val_accuracy:0.25 ]\n",
      "32it [01:37,  3.06s/it, epoch: 5/50 -  loss:12.541316032409668  accuracy:0.3080765902996063  val_loss:46.98291778564453  val_accuracy:0.1458333283662796 ]\n",
      "32it [01:38,  3.07s/it, epoch: 6/50 -  loss:15.415522575378418  accuracy:0.29142382740974426  val_loss:10.973555564880371  val_accuracy:0.17916665971279144 ]\n",
      "32it [01:38,  3.07s/it, epoch: 7/50 -  loss:6.303822994232178  accuracy:0.3255620300769806  val_loss:8.56163501739502  val_accuracy:0.20416666567325592 ]\n",
      "32it [01:54,  3.58s/it, epoch: 8/50 -  loss:5.007680892944336  accuracy:0.34637802839279175  val_loss:6.916962146759033  val_accuracy:0.3499999940395355 ]\n",
      "32it [01:55,  3.61s/it, epoch: 9/50 -  loss:3.740983247756958  accuracy:0.3855120837688446  val_loss:3.589289665222168  val_accuracy:0.2083333283662796 ]\n",
      "32it [01:55,  3.62s/it, epoch: 10/50 -  loss:2.5561084747314453  accuracy:0.35803496837615967  val_loss:5.7687087059021  val_accuracy:0.15833333134651184 ]\n",
      "32it [01:40,  3.13s/it, epoch: 11/50 -  loss:2.6316604614257812  accuracy:0.3921732008457184  val_loss:4.043315410614014  val_accuracy:0.25 ]\n",
      "32it [01:38,  3.09s/it, epoch: 12/50 -  loss:2.2317721843719482  accuracy:0.4063280522823334  val_loss:2.3408310413360596  val_accuracy:0.27916666865348816 ]\n",
      "32it [01:39,  3.12s/it, epoch: 13/50 -  loss:1.6880862712860107  accuracy:0.41631972789764404  val_loss:4.694206237792969  val_accuracy:0.17083333432674408 ]\n",
      "32it [01:39,  3.11s/it, epoch: 14/50 -  loss:2.0930988788604736  accuracy:0.4113239049911499  val_loss:3.81128191947937  val_accuracy:0.17916665971279144 ]\n",
      "32it [01:50,  3.44s/it, epoch: 15/50 -  loss:1.8601850271224976  accuracy:0.42381349205970764  val_loss:2.3201189041137695  val_accuracy:0.28333333134651184 ]\n",
      "32it [01:47,  3.36s/it, epoch: 16/50 -  loss:1.5195019245147705  accuracy:0.43880099058151245  val_loss:1.9348057508468628  val_accuracy:0.3499999940395355 ]\n",
      "32it [02:12,  4.13s/it, epoch: 17/50 -  loss:1.4174845218658447  accuracy:0.47960034012794495  val_loss:2.867670774459839  val_accuracy:0.27916666865348816 ]\n",
      "32it [01:45,  3.29s/it, epoch: 18/50 -  loss:1.5716767311096191  accuracy:0.4995836913585663  val_loss:2.5769591331481934  val_accuracy:0.2874999940395355 ]\n",
      "32it [01:40,  3.14s/it, epoch: 19/50 -  loss:1.5218170881271362  accuracy:0.4787676930427551  val_loss:2.084118604660034  val_accuracy:0.27916666865348816 ]\n",
      "32it [01:38,  3.08s/it, epoch: 20/50 -  loss:1.3385194540023804  accuracy:0.4970857501029968  val_loss:2.267355442047119  val_accuracy:0.34166666865348816 ]\n",
      "32it [01:40,  3.13s/it, epoch: 21/50 -  loss:1.4143364429473877  accuracy:0.5145711898803711  val_loss:1.8168503046035767  val_accuracy:0.2916666567325592 ]\n",
      "32it [01:41,  3.17s/it, epoch: 22/50 -  loss:1.223915696144104  accuracy:0.5253955125808716  val_loss:2.421297550201416  val_accuracy:0.34166666865348816 ]\n",
      "32it [01:51,  3.48s/it, epoch: 23/50 -  loss:1.3997722864151  accuracy:0.5278934240341187  val_loss:2.1914479732513428  val_accuracy:0.38333332538604736 ]\n",
      "32it [01:39,  3.10s/it, epoch: 24/50 -  loss:1.2561317682266235  accuracy:0.5770191550254822  val_loss:2.868597984313965  val_accuracy:0.3291666805744171 ]\n",
      "32it [01:44,  3.28s/it, epoch: 25/50 -  loss:1.4317070245742798  accuracy:0.5611990094184875  val_loss:2.570885181427002  val_accuracy:0.3166666626930237 ]\n",
      "32it [01:45,  3.30s/it, epoch: 26/50 -  loss:1.3605040311813354  accuracy:0.5578684210777283  val_loss:1.9396979808807373  val_accuracy:0.3083333373069763 ]\n",
      "32it [01:39,  3.11s/it, epoch: 27/50 -  loss:1.1538910865783691  accuracy:0.5711906552314758  val_loss:2.737755060195923  val_accuracy:0.3333333432674408 ]\n",
      "32it [01:52,  3.53s/it, epoch: 28/50 -  loss:1.295885682106018  accuracy:0.5961698293685913  val_loss:2.398267984390259  val_accuracy:0.36666667461395264 ]\n",
      "32it [01:42,  3.20s/it, epoch: 29/50 -  loss:1.2216304540634155  accuracy:0.5845128893852234  val_loss:5.456529140472412  val_accuracy:0.3166666626930237 ]\n",
      "32it [01:42,  3.22s/it, epoch: 30/50 -  loss:1.8455870151519775  accuracy:0.6069941520690918  val_loss:2.622253894805908  val_accuracy:0.3499999940395355 ]\n",
      "32it [01:37,  3.06s/it, epoch: 31/50 -  loss:1.2096481323242188  accuracy:0.607826828956604  val_loss:2.6999380588531494  val_accuracy:0.3583333194255829 ]\n",
      "32it [01:52,  3.52s/it, epoch: 32/50 -  loss:1.25503671169281  accuracy:0.6153205633163452  val_loss:3.1902101039886475  val_accuracy:0.3916666805744171 ]\n",
      "32it [01:40,  3.13s/it, epoch: 33/50 -  loss:1.321710467338562  accuracy:0.6319733262062073  val_loss:2.9719605445861816  val_accuracy:0.3125 ]\n",
      "32it [02:03,  3.87s/it, epoch: 34/50 -  loss:1.2177612781524658  accuracy:0.6394671201705933  val_loss:2.9436466693878174  val_accuracy:0.3499999940395355 ]\n",
      "32it [01:38,  3.07s/it, epoch: 35/50 -  loss:1.2386927604675293  accuracy:0.6561198830604553  val_loss:4.200243949890137  val_accuracy:0.2666666805744171 ]\n",
      "32it [01:41,  3.19s/it, epoch: 36/50 -  loss:1.5145163536071777  accuracy:0.6344712972640991  val_loss:2.730128765106201  val_accuracy:0.375 ]\n",
      "32it [01:40,  3.16s/it, epoch: 37/50 -  loss:1.150541067123413  accuracy:0.6644462943077087  val_loss:3.111203670501709  val_accuracy:0.34166666865348816 ]\n",
      "32it [01:50,  3.45s/it, epoch: 38/50 -  loss:1.268009066581726  accuracy:0.6477935314178467  val_loss:3.186652183532715  val_accuracy:0.2958333194255829 ]\n",
      "32it [01:43,  3.23s/it, epoch: 39/50 -  loss:1.1637378931045532  accuracy:0.6727727055549622  val_loss:2.7717678546905518  val_accuracy:0.32499998807907104 ]\n",
      "32it [01:43,  3.25s/it, epoch: 40/50 -  loss:1.1945478916168213  accuracy:0.6594504714012146  val_loss:4.414767265319824  val_accuracy:0.3375000059604645 ]\n",
      "32it [01:41,  3.19s/it, epoch: 41/50 -  loss:1.4453009366989136  accuracy:0.6860949397087097  val_loss:3.308169364929199  val_accuracy:0.3166666626930237 ]\n",
      "32it [01:43,  3.22s/it, epoch: 42/50 -  loss:1.1969460248947144  accuracy:0.6835970282554626  val_loss:3.0412395000457764  val_accuracy:0.40833333134651184 ]\n",
      "32it [01:50,  3.44s/it, epoch: 43/50 -  loss:1.0718883275985718  accuracy:0.7085762023925781  val_loss:4.295853137969971  val_accuracy:0.32499998807907104 ]\n",
      "32it [01:48,  3.38s/it, epoch: 44/50 -  loss:1.3032809495925903  accuracy:0.7027477025985718  val_loss:3.5949547290802  val_accuracy:0.375 ]\n",
      "32it [01:47,  3.35s/it, epoch: 45/50 -  loss:1.3326811790466309  accuracy:0.6827643513679504  val_loss:3.4334940910339355  val_accuracy:0.3791666626930237 ]\n",
      "32it [01:43,  3.23s/it, epoch: 46/50 -  loss:1.1082518100738525  accuracy:0.7252289652824402  val_loss:2.774402141571045  val_accuracy:0.3499999940395355 ]\n",
      "32it [01:43,  3.22s/it, epoch: 47/50 -  loss:1.052221655845642  accuracy:0.6994171738624573  val_loss:2.918473958969116  val_accuracy:0.36250001192092896 ]\n",
      "32it [01:40,  3.14s/it, epoch: 48/50 -  loss:1.0545026063919067  accuracy:0.7152373194694519  val_loss:3.524212121963501  val_accuracy:0.375 ]\n",
      "32it [01:43,  3.22s/it, epoch: 49/50 -  loss:1.1607937812805176  accuracy:0.7243963479995728  val_loss:3.213914632797241  val_accuracy:0.3499999940395355 ]\n",
      "32it [01:38,  3.08s/it, epoch: 50/50 -  loss:1.0676075220108032  accuracy:0.7277268767356873  val_loss:4.0257086753845215  val_accuracy:0.34583333134651184 ]\n"
     ]
    }
   ],
   "source": [
    "history = fed_model.fit(\n",
    "    data,\n",
    "    None,\n",
    "    validation_data=data,\n",
    "    epochs=epochs,\n",
    "    batch_size=batch_size,\n",
    "    aggregate_freq=aggregate_freq,\n",
    "    sampler_method=sampler_method,\n",
    "    random_seed=random_seed,\n",
    "    dp_spent_step_freq=dp_spent_step_freq,\n",
    "    dataset_builder=data_builder_dict,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Draw accuracy values for training & validation\n",
    "plt.plot(history.global_history['accuracy'])\n",
    "plt.plot(history.global_history['val_accuracy'])\n",
    "plt.title('FLModel accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "# Draw loss for training & validation\n",
    "plt.plot(history.global_history['loss'])\n",
    "plt.plot(history.global_history['val_loss'])\n",
    "plt.title('FLModel loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 只加载网络结构同时随机初始化"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:root:Create proxy actor <class 'secretflow.ml.nn.fl.backend.tensorflow.strategy.fed_avg_w.PYUFedAvgW'> with party alice.\n",
      "INFO:root:Create proxy actor <class 'secretflow.ml.nn.fl.backend.tensorflow.strategy.fed_avg_w.PYUFedAvgW'> with party bob.\n"
     ]
    }
   ],
   "source": [
    "# keras model\n",
    "no_weight_model = create_inception_v3_model_classifier(\n",
    "    num_classes=num_classes, is_load_weight=False\n",
    ")\n",
    "\n",
    "\n",
    "fed_model = FLModel(\n",
    "    device_list=device_list,\n",
    "    model=no_weight_model,\n",
    "    aggregator=aggregator,\n",
    "    backend=\"tensorflow\",\n",
    "    strategy=\"fed_avg_w\",\n",
    "    random_seed=1234,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:root:FL Train Params: {'self': <secretflow.ml.nn.fl.fl_model.FLModel object at 0x7fbd4baefbe0>, 'x': {PYURuntime(alice): '/tmp/tmpc7wdf9us/datasets/flower_photos', PYURuntime(bob): '/tmp/tmpc7wdf9us/datasets/flower_photos'}, 'y': None, 'batch_size': 32, 'batch_sampling_rate': None, 'epochs': 50, 'verbose': 1, 'callbacks': None, 'validation_data': {PYURuntime(alice): '/tmp/tmpc7wdf9us/datasets/flower_photos', PYURuntime(bob): '/tmp/tmpc7wdf9us/datasets/flower_photos'}, 'shuffle': False, 'class_weight': None, 'sample_weight': None, 'validation_freq': 1, 'aggregate_freq': 2, 'label_decoder': None, 'max_batch_size': 20000, 'prefetch_buffer_size': None, 'sampler_method': 'batch', 'random_seed': 1234, 'dp_spent_step_freq': 1, 'audit_log_dir': None, 'dataset_builder': {PYURuntime(alice): <function create_dataset_builder.<locals>.dataset_builder at 0x7fbb907ee700>, PYURuntime(bob): <function create_dataset_builder.<locals>.dataset_builder at 0x7fbadc0d5040>}, 'wait_steps': 100}\n",
      "32it [01:49,  3.41s/it, epoch: 1/50 -  loss:1.6169096231460571  accuracy:0.2788761854171753  val_loss:2.0081751346588135  val_accuracy:0.25 ]\n",
      "32it [01:41,  3.18s/it, epoch: 2/50 -  loss:1.6776905059814453  accuracy:0.2805995047092438  val_loss:1.5821692943572998  val_accuracy:0.2083333283662796 ]\n",
      "32it [01:40,  3.15s/it, epoch: 3/50 -  loss:1.5732810497283936  accuracy:0.27560365200042725  val_loss:1.6988409757614136  val_accuracy:0.25 ]\n",
      "32it [01:43,  3.22s/it, epoch: 4/50 -  loss:1.5928469896316528  accuracy:0.29059118032455444  val_loss:1.6616723537445068  val_accuracy:0.25 ]\n",
      "32it [01:43,  3.25s/it, epoch: 5/50 -  loss:1.5776023864746094  accuracy:0.2972522974014282  val_loss:1.5816072225570679  val_accuracy:0.25 ]\n",
      "32it [01:39,  3.10s/it, epoch: 6/50 -  loss:1.5489486455917358  accuracy:0.3064113259315491  val_loss:1.5895153284072876  val_accuracy:0.25 ]\n",
      "32it [01:38,  3.09s/it, epoch: 7/50 -  loss:1.5415023565292358  accuracy:0.33888426423072815  val_loss:1.6028796434402466  val_accuracy:0.2874999940395355 ]\n",
      "32it [01:42,  3.21s/it, epoch: 8/50 -  loss:1.5391955375671387  accuracy:0.3488759398460388  val_loss:1.6169476509094238  val_accuracy:0.2666666805744171 ]\n",
      "32it [01:39,  3.11s/it, epoch: 9/50 -  loss:1.5235228538513184  accuracy:0.37552040815353394  val_loss:1.6155515909194946  val_accuracy:0.25 ]\n",
      "32it [01:42,  3.21s/it, epoch: 10/50 -  loss:1.5191270112991333  accuracy:0.3522064983844757  val_loss:1.600549340248108  val_accuracy:0.32499998807907104 ]\n",
      "32it [01:39,  3.09s/it, epoch: 11/50 -  loss:1.5050208568572998  accuracy:0.3705245554447174  val_loss:1.6656410694122314  val_accuracy:0.25 ]\n",
      "32it [01:45,  3.31s/it, epoch: 12/50 -  loss:1.5016016960144043  accuracy:0.3563697040081024  val_loss:1.5471644401550293  val_accuracy:0.2541666626930237 ]\n",
      "32it [01:36,  3.02s/it, epoch: 13/50 -  loss:1.465255618095398  accuracy:0.37968358397483826  val_loss:1.54298996925354  val_accuracy:0.3791666626930237 ]\n",
      "32it [01:42,  3.21s/it, epoch: 14/50 -  loss:1.4461535215377808  accuracy:0.40549543499946594  val_loss:1.5411031246185303  val_accuracy:0.3791666626930237 ]\n",
      "32it [01:38,  3.07s/it, epoch: 15/50 -  loss:1.4370651245117188  accuracy:0.40882596373558044  val_loss:1.490233063697815  val_accuracy:0.3291666805744171 ]\n",
      "32it [01:41,  3.17s/it, epoch: 16/50 -  loss:1.413556456565857  accuracy:0.4046627879142761  val_loss:1.590057373046875  val_accuracy:0.3333333432674408 ]\n",
      "32it [01:42,  3.19s/it, epoch: 17/50 -  loss:1.425291895866394  accuracy:0.40549543499946594  val_loss:1.5726838111877441  val_accuracy:0.26249998807907104 ]\n",
      "32it [01:46,  3.34s/it, epoch: 18/50 -  loss:1.411468267440796  accuracy:0.3930058181285858  val_loss:1.6937826871871948  val_accuracy:0.25 ]\n",
      "32it [01:39,  3.10s/it, epoch: 19/50 -  loss:1.4267196655273438  accuracy:0.4004995822906494  val_loss:1.7930010557174683  val_accuracy:0.25 ]\n",
      "32it [01:41,  3.16s/it, epoch: 20/50 -  loss:1.444821834564209  accuracy:0.39134055376052856  val_loss:1.909441351890564  val_accuracy:0.2708333432674408 ]\n",
      "32it [01:43,  3.22s/it, epoch: 21/50 -  loss:1.4686118364334106  accuracy:0.40216487646102905  val_loss:1.622828722000122  val_accuracy:0.3083333373069763 ]\n",
      "32it [01:39,  3.10s/it, epoch: 22/50 -  loss:1.4035924673080444  accuracy:0.4104912579059601  val_loss:1.453572392463684  val_accuracy:0.3708333373069763 ]\n",
      "32it [01:45,  3.31s/it, epoch: 23/50 -  loss:1.359868049621582  accuracy:0.41631972789764404  val_loss:1.6003872156143188  val_accuracy:0.3583333194255829 ]\n",
      "32it [01:39,  3.09s/it, epoch: 24/50 -  loss:1.3811829090118408  accuracy:0.42381349205970764  val_loss:1.5266224145889282  val_accuracy:0.3166666626930237 ]\n",
      "32it [01:42,  3.22s/it, epoch: 25/50 -  loss:1.3589435815811157  accuracy:0.40965861082077026  val_loss:2.0248773097991943  val_accuracy:0.2708333432674408 ]\n",
      "32it [01:39,  3.11s/it, epoch: 26/50 -  loss:1.4731003046035767  accuracy:0.4154870808124542  val_loss:1.4912035465240479  val_accuracy:0.34583333134651184 ]\n",
      "32it [01:42,  3.20s/it, epoch: 27/50 -  loss:1.3412293195724487  accuracy:0.41631972789764404  val_loss:1.3930459022521973  val_accuracy:0.3791666626930237 ]\n",
      "32it [01:42,  3.20s/it, epoch: 28/50 -  loss:1.314136266708374  accuracy:0.43130725622177124  val_loss:1.555684208869934  val_accuracy:0.3166666626930237 ]\n",
      "32it [01:52,  3.51s/it, epoch: 29/50 -  loss:1.3471096754074097  accuracy:0.42797669768333435  val_loss:1.6817519664764404  val_accuracy:0.2666666805744171 ]\n",
      "32it [01:38,  3.06s/it, epoch: 30/50 -  loss:1.3709834814071655  accuracy:0.407993346452713  val_loss:1.4343681335449219  val_accuracy:0.36666667461395264 ]\n",
      "32it [01:48,  3.41s/it, epoch: 31/50 -  loss:1.3120677471160889  accuracy:0.43463781476020813  val_loss:1.4698201417922974  val_accuracy:0.3499999940395355 ]\n",
      "32it [01:37,  3.05s/it, epoch: 32/50 -  loss:1.3166753053665161  accuracy:0.43130725622177124  val_loss:1.4647256135940552  val_accuracy:0.34166666865348816 ]\n",
      "32it [01:40,  3.15s/it, epoch: 33/50 -  loss:1.316184401512146  accuracy:0.4263114035129547  val_loss:1.7285205125808716  val_accuracy:0.25833332538604736 ]\n",
      "32it [01:41,  3.17s/it, epoch: 34/50 -  loss:1.3727495670318604  accuracy:0.4154870808124542  val_loss:1.3565254211425781  val_accuracy:0.3916666805744171 ]\n",
      "32it [01:44,  3.26s/it, epoch: 35/50 -  loss:1.280211329460144  accuracy:0.45045796036720276  val_loss:1.469690203666687  val_accuracy:0.30000001192092896 ]\n",
      "32it [01:44,  3.27s/it, epoch: 36/50 -  loss:1.3061801195144653  accuracy:0.41715237498283386  val_loss:1.4475810527801514  val_accuracy:0.38749998807907104 ]\n",
      "32it [01:39,  3.11s/it, epoch: 37/50 -  loss:1.2898013591766357  accuracy:0.443796843290329  val_loss:1.6531250476837158  val_accuracy:0.28333333134651184 ]\n",
      "32it [01:43,  3.22s/it, epoch: 38/50 -  loss:1.3455705642700195  accuracy:0.42714405059814453  val_loss:1.5225203037261963  val_accuracy:0.38749998807907104 ]\n",
      "32it [01:39,  3.12s/it, epoch: 39/50 -  loss:1.30066978931427  accuracy:0.4371357262134552  val_loss:1.4351023435592651  val_accuracy:0.38749998807907104 ]\n",
      "32it [01:42,  3.21s/it, epoch: 40/50 -  loss:1.2890324592590332  accuracy:0.45295587182044983  val_loss:1.563452959060669  val_accuracy:0.3125 ]\n",
      "32it [01:37,  3.03s/it, epoch: 41/50 -  loss:1.3143178224563599  accuracy:0.43963363766670227  val_loss:1.3913853168487549  val_accuracy:0.4124999940395355 ]\n",
      "32it [01:46,  3.34s/it, epoch: 42/50 -  loss:1.2762908935546875  accuracy:0.45711907744407654  val_loss:1.3839548826217651  val_accuracy:0.3916666805744171 ]\n",
      "32it [01:36,  3.00s/it, epoch: 43/50 -  loss:1.2671079635620117  accuracy:0.4587843418121338  val_loss:1.4633164405822754  val_accuracy:0.38749998807907104 ]\n",
      "32it [01:42,  3.20s/it, epoch: 44/50 -  loss:1.2803606986999512  accuracy:0.45711907744407654  val_loss:1.5430619716644287  val_accuracy:0.30000001192092896 ]\n",
      "32it [01:39,  3.11s/it, epoch: 45/50 -  loss:1.3015990257263184  accuracy:0.437968373298645  val_loss:2.1071672439575195  val_accuracy:0.2958333194255829 ]\n",
      "32it [01:43,  3.24s/it, epoch: 46/50 -  loss:1.4335447549819946  accuracy:0.4371357262134552  val_loss:1.4610460996627808  val_accuracy:0.3583333194255829 ]\n",
      "32it [01:42,  3.21s/it, epoch: 47/50 -  loss:1.276992678642273  accuracy:0.46128225326538086  val_loss:1.4364838600158691  val_accuracy:0.375 ]\n",
      "32it [01:44,  3.25s/it, epoch: 48/50 -  loss:1.2780916690826416  accuracy:0.4662781059741974  val_loss:1.4993922710418701  val_accuracy:0.3916666805744171 ]\n",
      "32it [01:42,  3.22s/it, epoch: 49/50 -  loss:1.2837833166122437  accuracy:0.443796843290329  val_loss:1.5112351179122925  val_accuracy:0.3333333432674408 ]\n",
      "32it [01:41,  3.17s/it, epoch: 50/50 -  loss:1.29127836227417  accuracy:0.45045796036720276  val_loss:1.6569411754608154  val_accuracy:0.34583333134651184 ]\n"
     ]
    }
   ],
   "source": [
    "history = fed_model.fit(\n",
    "    data,\n",
    "    None,\n",
    "    validation_data=data,\n",
    "    epochs=epochs,\n",
    "    batch_size=batch_size,\n",
    "    aggregate_freq=aggregate_freq,\n",
    "    sampler_method=sampler_method,\n",
    "    random_seed=random_seed,\n",
    "    dp_spent_step_freq=dp_spent_step_freq,\n",
    "    dataset_builder=data_builder_dict,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Draw accuracy values for training & validation\n",
    "plt.plot(history.global_history['accuracy'])\n",
    "plt.plot(history.global_history['val_accuracy'])\n",
    "plt.title('FLModel accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "# Draw loss for training & validation\n",
    "plt.plot(history.global_history['loss'])\n",
    "plt.plot(history.global_history['val_loss'])\n",
    "plt.title('FLModel loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 冻结底层微调顶层网络\n",
    "我们只要参照教程里对模型的定义，在函数里完成我们对模型的定义即可；可以看到代码几乎不需要作任何修改，只需要进行适当的封装。\n",
    "为了方便作对比实验，我们额外添加是否加载权重的选项。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_inception_v3_model_fine_tune(num_classes, is_load_weight=True):\n",
    "    def create_model():\n",
    "        from tensorflow import keras\n",
    "\n",
    "        # Create model\n",
    "        # create the base pre-trained model\n",
    "        if is_load_weight:\n",
    "            base_model = InceptionV3(weights='imagenet', include_top=False)\n",
    "        else:\n",
    "            base_model = InceptionV3(weights=None, include_top=False)\n",
    "\n",
    "        # add a global spatial average pooling layer\n",
    "        x = base_model.output\n",
    "        x = GlobalAveragePooling2D()(x)\n",
    "        # let's add a fully-connected layer\n",
    "        x = Dense(1024, activation='relu')(x)\n",
    "        # and a logistic layer -- let's say we have 10 classes\n",
    "        predictions = Dense(num_classes, activation='softmax')(x)\n",
    "\n",
    "        # this is the model we will train\n",
    "        model = Model(inputs=base_model.input, outputs=predictions)\n",
    "\n",
    "        for layer in model.layers[:249]:\n",
    "            layer.trainable = False\n",
    "        for layer in model.layers[249:]:\n",
    "            layer.trainable = True\n",
    "\n",
    "        # Compile model\n",
    "        model.compile(\n",
    "            optimizer=SGD(learning_rate=0.0001, momentum=0.9),\n",
    "            loss='sparse_categorical_crossentropy',\n",
    "            metrics=[\"accuracy\"],\n",
    "        )\n",
    "\n",
    "        return model\n",
    "\n",
    "    return create_model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 加载预训练模型权重并且微调"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:root:Create proxy actor <class 'secretflow.ml.nn.fl.backend.tensorflow.strategy.fed_avg_w.PYUFedAvgW'> with party alice.\n",
      "INFO:root:Create proxy actor <class 'secretflow.ml.nn.fl.backend.tensorflow.strategy.fed_avg_w.PYUFedAvgW'> with party bob.\n"
     ]
    }
   ],
   "source": [
    "# keras model\n",
    "weight_model = create_inception_v3_model_fine_tune(\n",
    "    num_classes=num_classes, is_load_weight=True\n",
    ")\n",
    "\n",
    "\n",
    "fed_model = FLModel(\n",
    "    device_list=device_list,\n",
    "    model=weight_model,\n",
    "    aggregator=aggregator,\n",
    "    backend=\"tensorflow\",\n",
    "    strategy=\"fed_avg_w\",\n",
    "    random_seed=1234,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:root:FL Train Params: {'self': <secretflow.ml.nn.fl.fl_model.FLModel object at 0x7fbc1c65c160>, 'x': {PYURuntime(alice): '/tmp/tmpc7wdf9us/datasets/flower_photos', PYURuntime(bob): '/tmp/tmpc7wdf9us/datasets/flower_photos'}, 'y': None, 'batch_size': 32, 'batch_sampling_rate': None, 'epochs': 50, 'verbose': 1, 'callbacks': None, 'validation_data': {PYURuntime(alice): '/tmp/tmpc7wdf9us/datasets/flower_photos', PYURuntime(bob): '/tmp/tmpc7wdf9us/datasets/flower_photos'}, 'shuffle': False, 'class_weight': None, 'sample_weight': None, 'validation_freq': 1, 'aggregate_freq': 2, 'label_decoder': None, 'max_batch_size': 20000, 'prefetch_buffer_size': None, 'sampler_method': 'batch', 'random_seed': 1234, 'dp_spent_step_freq': 1, 'audit_log_dir': None, 'dataset_builder': {PYURuntime(alice): <function create_dataset_builder.<locals>.dataset_builder at 0x7fbb907ee700>, PYURuntime(bob): <function create_dataset_builder.<locals>.dataset_builder at 0x7fbadc0d5040>}, 'wait_steps': 100}\n",
      "32it [01:50,  3.46s/it, epoch: 1/50 -  loss:1.650011658668518  accuracy:0.20915712416172028  val_loss:1.656758427619934  val_accuracy:0.24583333730697632 ]\n",
      "32it [01:46,  3.32s/it, epoch: 2/50 -  loss:1.5472569465637207  accuracy:0.32389676570892334  val_loss:1.58930242061615  val_accuracy:0.28333333134651184 ]\n",
      "32it [01:45,  3.29s/it, epoch: 3/50 -  loss:1.4915741682052612  accuracy:0.36552873253822327  val_loss:1.5358479022979736  val_accuracy:0.30000001192092896 ]\n",
      "32it [01:43,  3.25s/it, epoch: 4/50 -  loss:1.4296882152557373  accuracy:0.4004995822906494  val_loss:1.5131781101226807  val_accuracy:0.32499998807907104 ]\n",
      "32it [01:45,  3.30s/it, epoch: 5/50 -  loss:1.3764379024505615  accuracy:0.43963363766670227  val_loss:1.485734462738037  val_accuracy:0.36666667461395264 ]\n",
      "32it [01:41,  3.16s/it, epoch: 6/50 -  loss:1.334008812904358  accuracy:0.4845961630344391  val_loss:1.4631234407424927  val_accuracy:0.3541666567325592 ]\n",
      "32it [01:53,  3.56s/it, epoch: 7/50 -  loss:1.2897144556045532  accuracy:0.5104079842567444  val_loss:1.4368029832839966  val_accuracy:0.3499999940395355 ]\n",
      "32it [01:38,  3.08s/it, epoch: 8/50 -  loss:1.2628018856048584  accuracy:0.526228129863739  val_loss:1.4215635061264038  val_accuracy:0.36666667461395264 ]\n",
      "32it [01:44,  3.27s/it, epoch: 9/50 -  loss:1.2168503999710083  accuracy:0.5553705096244812  val_loss:1.4049427509307861  val_accuracy:0.3916666805744171 ]\n",
      "32it [01:36,  3.03s/it, epoch: 10/50 -  loss:1.187528371810913  accuracy:0.5611990094184875  val_loss:1.3869177103042603  val_accuracy:0.3916666805744171 ]\n",
      "32it [01:37,  3.04s/it, epoch: 11/50 -  loss:1.1617594957351685  accuracy:0.5903413891792297  val_loss:1.3873077630996704  val_accuracy:0.3958333432674408 ]\n",
      "32it [01:42,  3.19s/it, epoch: 12/50 -  loss:1.1248602867126465  accuracy:0.5936719179153442  val_loss:1.3818570375442505  val_accuracy:0.4000000059604645 ]\n",
      "32it [01:39,  3.11s/it, epoch: 13/50 -  loss:1.0991374254226685  accuracy:0.6369692087173462  val_loss:1.366493821144104  val_accuracy:0.42500001192092896 ]\n",
      "32it [01:45,  3.31s/it, epoch: 14/50 -  loss:1.0843088626861572  accuracy:0.6353039145469666  val_loss:1.3633487224578857  val_accuracy:0.42916667461395264 ]\n",
      "32it [01:41,  3.17s/it, epoch: 15/50 -  loss:1.059274673461914  accuracy:0.6636136770248413  val_loss:1.3560353517532349  val_accuracy:0.4375 ]\n",
      "32it [01:44,  3.26s/it, epoch: 16/50 -  loss:1.0273948907852173  accuracy:0.6644462943077087  val_loss:1.3572434186935425  val_accuracy:0.4333333373069763 ]\n",
      "32it [01:39,  3.12s/it, epoch: 17/50 -  loss:0.9911747574806213  accuracy:0.6994171738624573  val_loss:1.3419238328933716  val_accuracy:0.44999998807907104 ]\n",
      "32it [01:50,  3.44s/it, epoch: 18/50 -  loss:0.9911114573478699  accuracy:0.67194002866745  val_loss:1.3493646383285522  val_accuracy:0.42500001192092896 ]\n",
      "32it [01:40,  3.16s/it, epoch: 19/50 -  loss:0.9671102166175842  accuracy:0.6852622628211975  val_loss:1.3433830738067627  val_accuracy:0.42916667461395264 ]\n",
      "32it [01:46,  3.32s/it, epoch: 20/50 -  loss:0.937901496887207  accuracy:0.7052456140518188  val_loss:1.3399254083633423  val_accuracy:0.4416666626930237 ]\n",
      "32it [01:42,  3.21s/it, epoch: 21/50 -  loss:0.906328558921814  accuracy:0.7277268767356873  val_loss:1.328268051147461  val_accuracy:0.4541666805744171 ]\n",
      "32it [01:52,  3.53s/it, epoch: 22/50 -  loss:0.8943104147911072  accuracy:0.7402164936065674  val_loss:1.3352607488632202  val_accuracy:0.44583332538604736 ]\n",
      "32it [01:44,  3.26s/it, epoch: 23/50 -  loss:0.8718297481536865  accuracy:0.7452123165130615  val_loss:1.310567021369934  val_accuracy:0.4583333432674408 ]\n",
      "32it [01:53,  3.53s/it, epoch: 24/50 -  loss:0.8595983982086182  accuracy:0.746044933795929  val_loss:1.3058205842971802  val_accuracy:0.46666666865348816 ]\n",
      "32it [01:40,  3.14s/it, epoch: 25/50 -  loss:0.8207837343215942  accuracy:0.7618651390075684  val_loss:1.3195736408233643  val_accuracy:0.47083333134651184 ]\n",
      "32it [01:41,  3.18s/it, epoch: 26/50 -  loss:0.8151105642318726  accuracy:0.7718567848205566  val_loss:1.3052656650543213  val_accuracy:0.4625000059604645 ]\n",
      "32it [01:39,  3.11s/it, epoch: 27/50 -  loss:0.8001042008399963  accuracy:0.7718567848205566  val_loss:1.3049904108047485  val_accuracy:0.46666666865348816 ]\n",
      "32it [01:42,  3.20s/it, epoch: 28/50 -  loss:0.7668044567108154  accuracy:0.7901748418807983  val_loss:1.307594895362854  val_accuracy:0.4625000059604645 ]\n",
      "32it [01:40,  3.13s/it, epoch: 29/50 -  loss:0.7354434728622437  accuracy:0.8043297529220581  val_loss:1.306896686553955  val_accuracy:0.47083333134651184 ]\n",
      "32it [01:42,  3.21s/it, epoch: 30/50 -  loss:0.7426672577857971  accuracy:0.8018317818641663  val_loss:1.3118059635162354  val_accuracy:0.4833333194255829 ]\n",
      "32it [01:38,  3.07s/it, epoch: 31/50 -  loss:0.7213742733001709  accuracy:0.8126561045646667  val_loss:1.296970248222351  val_accuracy:0.4625000059604645 ]\n",
      "32it [01:37,  3.04s/it, epoch: 32/50 -  loss:0.700598955154419  accuracy:0.8184846043586731  val_loss:1.305118203163147  val_accuracy:0.4583333432674408 ]\n",
      "32it [01:37,  3.05s/it, epoch: 33/50 -  loss:0.6826915144920349  accuracy:0.8259783387184143  val_loss:1.3077011108398438  val_accuracy:0.4791666567325592 ]\n",
      "32it [01:37,  3.05s/it, epoch: 34/50 -  loss:0.6565826535224915  accuracy:0.8251457214355469  val_loss:1.301528811454773  val_accuracy:0.46666666865348816 ]\n",
      "32it [01:40,  3.13s/it, epoch: 35/50 -  loss:0.656404972076416  accuracy:0.8226478099822998  val_loss:1.306329369544983  val_accuracy:0.4749999940395355 ]\n",
      "32it [01:40,  3.13s/it, epoch: 36/50 -  loss:0.6250473856925964  accuracy:0.8384679555892944  val_loss:1.3014804124832153  val_accuracy:0.46666666865348816 ]\n",
      "32it [01:45,  3.29s/it, epoch: 37/50 -  loss:0.6090459227561951  accuracy:0.8426311612129211  val_loss:1.3036556243896484  val_accuracy:0.46666666865348816 ]\n",
      "32it [01:41,  3.17s/it, epoch: 38/50 -  loss:0.6007826924324036  accuracy:0.8409658670425415  val_loss:1.3221803903579712  val_accuracy:0.46666666865348816 ]\n",
      "32it [01:45,  3.29s/it, epoch: 39/50 -  loss:0.5874541997909546  accuracy:0.8467943668365479  val_loss:1.3001883029937744  val_accuracy:0.4583333432674408 ]\n",
      "32it [01:40,  3.14s/it, epoch: 40/50 -  loss:0.5784027576446533  accuracy:0.8459616899490356  val_loss:1.29843270778656  val_accuracy:0.4583333432674408 ]\n",
      "32it [01:45,  3.31s/it, epoch: 41/50 -  loss:0.5616000890731812  accuracy:0.8434637784957886  val_loss:1.3310949802398682  val_accuracy:0.4749999940395355 ]\n",
      "32it [01:44,  3.25s/it, epoch: 42/50 -  loss:0.564281165599823  accuracy:0.8559533953666687  val_loss:1.3145432472229004  val_accuracy:0.47083333134651184 ]\n",
      "32it [01:43,  3.25s/it, epoch: 43/50 -  loss:0.54606032371521  accuracy:0.8592839241027832  val_loss:1.311663269996643  val_accuracy:0.4625000059604645 ]\n",
      "32it [01:44,  3.25s/it, epoch: 44/50 -  loss:0.5349538922309875  accuracy:0.8601165413856506  val_loss:1.3167394399642944  val_accuracy:0.4625000059604645 ]\n",
      "32it [01:46,  3.32s/it, epoch: 45/50 -  loss:0.5256470441818237  accuracy:0.8667776584625244  val_loss:1.3188197612762451  val_accuracy:0.4791666567325592 ]\n",
      "32it [01:41,  3.16s/it, epoch: 46/50 -  loss:0.5097993016242981  accuracy:0.8676103353500366  val_loss:1.3174797296524048  val_accuracy:0.4791666567325592 ]\n",
      "32it [01:40,  3.15s/it, epoch: 47/50 -  loss:0.49822941422462463  accuracy:0.8742714524269104  val_loss:1.3001458644866943  val_accuracy:0.4583333432674408 ]\n",
      "32it [01:47,  3.35s/it, epoch: 48/50 -  loss:0.4908905625343323  accuracy:0.8667776584625244  val_loss:1.3209236860275269  val_accuracy:0.46666666865348816 ]\n",
      "32it [01:39,  3.10s/it, epoch: 49/50 -  loss:0.48140043020248413  accuracy:0.8726061582565308  val_loss:1.3325889110565186  val_accuracy:0.4749999940395355 ]\n",
      "32it [01:43,  3.23s/it, epoch: 50/50 -  loss:0.4809170365333557  accuracy:0.8792672753334045  val_loss:1.336763858795166  val_accuracy:0.4833333194255829 ]\n"
     ]
    }
   ],
   "source": [
    "history = fed_model.fit(\n",
    "    data,\n",
    "    None,\n",
    "    validation_data=data,\n",
    "    epochs=epochs,\n",
    "    batch_size=batch_size,\n",
    "    aggregate_freq=aggregate_freq,\n",
    "    sampler_method=sampler_method,\n",
    "    random_seed=random_seed,\n",
    "    dp_spent_step_freq=dp_spent_step_freq,\n",
    "    dataset_builder=data_builder_dict,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Draw accuracy values for training & validation\n",
    "plt.plot(history.global_history['accuracy'])\n",
    "plt.plot(history.global_history['val_accuracy'])\n",
    "plt.title('FLModel accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "# Draw loss for training & validation\n",
    "plt.plot(history.global_history['loss'])\n",
    "plt.plot(history.global_history['val_loss'])\n",
    "plt.title('FLModel loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 只加载网络结构同时随机初始化"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:root:Create proxy actor <class 'secretflow.ml.nn.fl.backend.tensorflow.strategy.fed_avg_w.PYUFedAvgW'> with party alice.\n",
      "INFO:root:Create proxy actor <class 'secretflow.ml.nn.fl.backend.tensorflow.strategy.fed_avg_w.PYUFedAvgW'> with party bob.\n"
     ]
    }
   ],
   "source": [
    "# keras model\n",
    "no_weight_model = create_inception_v3_model_fine_tune(\n",
    "    num_classes=num_classes, is_load_weight=False\n",
    ")\n",
    "\n",
    "\n",
    "fed_model = FLModel(\n",
    "    device_list=device_list,\n",
    "    model=no_weight_model,\n",
    "    aggregator=aggregator,\n",
    "    backend=\"tensorflow\",\n",
    "    strategy=\"fed_avg_w\",\n",
    "    random_seed=1234,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:root:FL Train Params: {'self': <secretflow.ml.nn.fl.fl_model.FLModel object at 0x7fbc24074700>, 'x': {PYURuntime(alice): '/tmp/tmpc7wdf9us/datasets/flower_photos', PYURuntime(bob): '/tmp/tmpc7wdf9us/datasets/flower_photos'}, 'y': None, 'batch_size': 32, 'batch_sampling_rate': None, 'epochs': 50, 'verbose': 1, 'callbacks': None, 'validation_data': {PYURuntime(alice): '/tmp/tmpc7wdf9us/datasets/flower_photos', PYURuntime(bob): '/tmp/tmpc7wdf9us/datasets/flower_photos'}, 'shuffle': False, 'class_weight': None, 'sample_weight': None, 'validation_freq': 1, 'aggregate_freq': 2, 'label_decoder': None, 'max_batch_size': 20000, 'prefetch_buffer_size': None, 'sampler_method': 'batch', 'random_seed': 1234, 'dp_spent_step_freq': 1, 'audit_log_dir': None, 'dataset_builder': {PYURuntime(alice): <function create_dataset_builder.<locals>.dataset_builder at 0x7fbb907ee700>, PYURuntime(bob): <function create_dataset_builder.<locals>.dataset_builder at 0x7fbadc0d5040>}, 'wait_steps': 100}\n",
      "32it [01:48,  3.38s/it, epoch: 1/50 -  loss:1.5617533922195435  accuracy:0.27783557772636414  val_loss:1.6185896396636963  val_accuracy:0.13750000298023224 ]\n",
      "32it [01:45,  3.30s/it, epoch: 2/50 -  loss:1.4592828750610352  accuracy:0.34970858693122864  val_loss:1.6081194877624512  val_accuracy:0.15833333134651184 ]\n",
      "32it [01:41,  3.16s/it, epoch: 3/50 -  loss:1.388456106185913  accuracy:0.4254787564277649  val_loss:1.5979021787643433  val_accuracy:0.2750000059604645 ]\n",
      "32it [01:45,  3.28s/it, epoch: 4/50 -  loss:1.3454346656799316  accuracy:0.43547043204307556  val_loss:1.5885674953460693  val_accuracy:0.2916666567325592 ]\n",
      "32it [01:43,  3.23s/it, epoch: 5/50 -  loss:1.3147145509719849  accuracy:0.46461281180381775  val_loss:1.57625150680542  val_accuracy:0.28333333134651184 ]\n",
      "32it [01:42,  3.21s/it, epoch: 6/50 -  loss:1.2754597663879395  accuracy:0.4854288101196289  val_loss:1.563430905342102  val_accuracy:0.30000001192092896 ]\n",
      "32it [01:38,  3.07s/it, epoch: 7/50 -  loss:1.2425297498703003  accuracy:0.5045795440673828  val_loss:1.5455553531646729  val_accuracy:0.2750000059604645 ]\n",
      "32it [01:45,  3.30s/it, epoch: 8/50 -  loss:1.2141278982162476  accuracy:0.5129058957099915  val_loss:1.5239124298095703  val_accuracy:0.2874999940395355 ]\n",
      "32it [01:51,  3.48s/it, epoch: 9/50 -  loss:1.1969578266143799  accuracy:0.5237302184104919  val_loss:1.4922692775726318  val_accuracy:0.38333332538604736 ]\n",
      "32it [01:37,  3.05s/it, epoch: 10/50 -  loss:1.1684715747833252  accuracy:0.5653622150421143  val_loss:1.4653935432434082  val_accuracy:0.36666667461395264 ]\n",
      "32it [01:38,  3.09s/it, epoch: 11/50 -  loss:1.1470826864242554  accuracy:0.5678601264953613  val_loss:1.439568281173706  val_accuracy:0.40833333134651184 ]\n",
      "32it [01:42,  3.19s/it, epoch: 12/50 -  loss:1.1196565628051758  accuracy:0.572023332118988  val_loss:1.399143934249878  val_accuracy:0.44583332538604736 ]\n",
      "32it [01:47,  3.36s/it, epoch: 13/50 -  loss:1.1069118976593018  accuracy:0.6036636233329773  val_loss:1.3591387271881104  val_accuracy:0.47083333134651184 ]\n",
      "32it [01:48,  3.40s/it, epoch: 14/50 -  loss:1.0912777185440063  accuracy:0.6036636233329773  val_loss:1.327014684677124  val_accuracy:0.48750001192092896 ]\n",
      "32it [01:50,  3.45s/it, epoch: 15/50 -  loss:1.0678472518920898  accuracy:0.6094920635223389  val_loss:1.2993195056915283  val_accuracy:0.5083333253860474 ]\n",
      "32it [01:41,  3.18s/it, epoch: 16/50 -  loss:1.0440828800201416  accuracy:0.6111573576927185  val_loss:1.2653034925460815  val_accuracy:0.5166666507720947 ]\n",
      "32it [01:48,  3.38s/it, epoch: 17/50 -  loss:1.0225656032562256  accuracy:0.6294754147529602  val_loss:1.231634497642517  val_accuracy:0.5249999761581421 ]\n",
      "32it [01:42,  3.20s/it, epoch: 18/50 -  loss:1.020599365234375  accuracy:0.6203163862228394  val_loss:1.2129393815994263  val_accuracy:0.5333333611488342 ]\n",
      "32it [01:45,  3.29s/it, epoch: 19/50 -  loss:0.9966413974761963  accuracy:0.6286427974700928  val_loss:1.1990076303482056  val_accuracy:0.5333333611488342 ]\n",
      "32it [01:42,  3.21s/it, epoch: 20/50 -  loss:0.9871851801872253  accuracy:0.6402997374534607  val_loss:1.176818609237671  val_accuracy:0.550000011920929 ]\n",
      "32it [01:51,  3.49s/it, epoch: 21/50 -  loss:0.9735381603240967  accuracy:0.6511240601539612  val_loss:1.164151906967163  val_accuracy:0.512499988079071 ]\n",
      "32it [01:44,  3.27s/it, epoch: 22/50 -  loss:0.9548952579498291  accuracy:0.6486261487007141  val_loss:1.1686121225357056  val_accuracy:0.5333333611488342 ]\n",
      "32it [01:43,  3.22s/it, epoch: 23/50 -  loss:0.9554542899131775  accuracy:0.6444629430770874  val_loss:1.1495128870010376  val_accuracy:0.5333333611488342 ]\n",
      "32it [01:55,  3.61s/it, epoch: 24/50 -  loss:0.9482444524765015  accuracy:0.6494587659835815  val_loss:1.136460542678833  val_accuracy:0.5375000238418579 ]\n",
      "32it [01:52,  3.53s/it, epoch: 25/50 -  loss:0.9186312556266785  accuracy:0.6702747941017151  val_loss:1.1416422128677368  val_accuracy:0.5416666865348816 ]\n",
      "32it [01:42,  3.21s/it, epoch: 26/50 -  loss:0.9188517928123474  accuracy:0.6736053228378296  val_loss:1.1264033317565918  val_accuracy:0.550000011920929 ]\n",
      "32it [01:41,  3.17s/it, epoch: 27/50 -  loss:0.9028259515762329  accuracy:0.681931734085083  val_loss:1.1303426027297974  val_accuracy:0.550000011920929 ]\n",
      "32it [01:45,  3.31s/it, epoch: 28/50 -  loss:0.8891870975494385  accuracy:0.6844296455383301  val_loss:1.1231117248535156  val_accuracy:0.5458333492279053 ]\n",
      "32it [01:43,  3.23s/it, epoch: 29/50 -  loss:0.8797008991241455  accuracy:0.6977518796920776  val_loss:1.1278127431869507  val_accuracy:0.5416666865348816 ]\n",
      "32it [01:47,  3.37s/it, epoch: 30/50 -  loss:0.878724217414856  accuracy:0.6752706170082092  val_loss:1.1188760995864868  val_accuracy:0.574999988079071 ]\n",
      "32it [01:40,  3.13s/it, epoch: 31/50 -  loss:0.8734909296035767  accuracy:0.703580379486084  val_loss:1.121822476387024  val_accuracy:0.550000011920929 ]\n",
      "32it [01:50,  3.44s/it, epoch: 32/50 -  loss:0.8486524820327759  accuracy:0.6985844969749451  val_loss:1.127307415008545  val_accuracy:0.5416666865348816 ]\n",
      "32it [01:45,  3.29s/it, epoch: 33/50 -  loss:0.8515317440032959  accuracy:0.6944212913513184  val_loss:1.1144821643829346  val_accuracy:0.5625 ]\n",
      "32it [01:44,  3.26s/it, epoch: 34/50 -  loss:0.8291710615158081  accuracy:0.7135720252990723  val_loss:1.1182698011398315  val_accuracy:0.5458333492279053 ]\n",
      "32it [01:49,  3.41s/it, epoch: 35/50 -  loss:0.8350727558135986  accuracy:0.7019150853157043  val_loss:1.1012609004974365  val_accuracy:0.5833333134651184 ]\n",
      "32it [01:48,  3.38s/it, epoch: 36/50 -  loss:0.8297075033187866  accuracy:0.7227310538291931  val_loss:1.1068880558013916  val_accuracy:0.5625 ]\n",
      "32it [01:49,  3.41s/it, epoch: 37/50 -  loss:0.809542715549469  accuracy:0.7293921709060669  val_loss:1.1091495752334595  val_accuracy:0.5458333492279053 ]\n",
      "32it [01:43,  3.24s/it, epoch: 38/50 -  loss:0.8065611720085144  accuracy:0.7185678482055664  val_loss:1.09843909740448  val_accuracy:0.5708333253860474 ]\n",
      "32it [01:46,  3.34s/it, epoch: 39/50 -  loss:0.7942371964454651  accuracy:0.7310574650764465  val_loss:1.105940580368042  val_accuracy:0.5708333253860474 ]\n",
      "32it [01:41,  3.16s/it, epoch: 40/50 -  loss:0.7934896349906921  accuracy:0.7227310538291931  val_loss:1.1048657894134521  val_accuracy:0.5666666626930237 ]\n",
      "32it [01:43,  3.23s/it, epoch: 41/50 -  loss:0.7780565023422241  accuracy:0.7327227592468262  val_loss:1.0963842868804932  val_accuracy:0.574999988079071 ]\n",
      "32it [01:43,  3.24s/it, epoch: 42/50 -  loss:0.7623825669288635  accuracy:0.7452123165130615  val_loss:1.0992350578308105  val_accuracy:0.5583333373069763 ]\n",
      "32it [01:44,  3.27s/it, epoch: 43/50 -  loss:0.7639929056167603  accuracy:0.7385511994361877  val_loss:1.0936287641525269  val_accuracy:0.5708333253860474 ]\n",
      "32it [01:40,  3.13s/it, epoch: 44/50 -  loss:0.7649388909339905  accuracy:0.7485429048538208  val_loss:1.0890998840332031  val_accuracy:0.5625 ]\n",
      "32it [01:44,  3.25s/it, epoch: 45/50 -  loss:0.7667314410209656  accuracy:0.7477102279663086  val_loss:1.1028246879577637  val_accuracy:0.5833333134651184 ]\n",
      "32it [01:39,  3.11s/it, epoch: 46/50 -  loss:0.7435023188591003  accuracy:0.7468776106834412  val_loss:1.0851361751556396  val_accuracy:0.5791666507720947 ]\n",
      "32it [01:38,  3.07s/it, epoch: 47/50 -  loss:0.7359048128128052  accuracy:0.7643630504608154  val_loss:1.0931051969528198  val_accuracy:0.5625 ]\n",
      "32it [01:37,  3.05s/it, epoch: 48/50 -  loss:0.729657769203186  accuracy:0.7626977562904358  val_loss:1.083522915840149  val_accuracy:0.5708333253860474 ]\n",
      "32it [01:44,  3.27s/it, epoch: 49/50 -  loss:0.7189591526985168  accuracy:0.7601998448371887  val_loss:1.0874546766281128  val_accuracy:0.574999988079071 ]\n",
      "32it [01:39,  3.10s/it, epoch: 50/50 -  loss:0.7207762598991394  accuracy:0.7760199904441833  val_loss:1.0822055339813232  val_accuracy:0.5708333253860474 ]\n"
     ]
    }
   ],
   "source": [
    "history = fed_model.fit(\n",
    "    data,\n",
    "    None,\n",
    "    validation_data=data,\n",
    "    epochs=epochs,\n",
    "    batch_size=batch_size,\n",
    "    aggregate_freq=aggregate_freq,\n",
    "    sampler_method=sampler_method,\n",
    "    random_seed=random_seed,\n",
    "    dp_spent_step_freq=dp_spent_step_freq,\n",
    "    dataset_builder=data_builder_dict,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Draw accuracy values for training & validation\n",
    "plt.plot(history.global_history['accuracy'])\n",
    "plt.plot(history.global_history['val_accuracy'])\n",
    "plt.title('FLModel accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "# Draw loss for training & validation\n",
    "plt.plot(history.global_history['loss'])\n",
    "plt.plot(history.global_history['val_loss'])\n",
    "plt.title('FLModel loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Valid'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 联邦学习小结\n",
    "可以看到，对照着 TensorFlow 的官方教程，隐语能够无缝地兼容所给出的微调方式；并且我们可以看到，通过对预训练模型的兼容，我们可以不需要自己再重新写出复杂网络的模型结构，InceptionV3 的网络结构源代码位于：[source code of Inception V3](https://github.com/keras-team/keras/blob/v2.13.1/keras/applications/inception_v3.py)，并且通过对比实验我们可以看出，加载预训练模型的权重，可以让我们的模型性能更优秀。\n",
    "\n",
    "## 总结\n",
    "本篇教程，我们以Inception V3为例介绍了如何在隐语的联邦学习模式下基于直接加载 TensorFlow.Keras 的 [预训练模型](https://keras.io/api/applications/)，通过直接加载预训练模型，我们能够获得：\n",
    "- 不需要再次编写复杂模型的结构代码\n",
    "- 基于预训练模型进行微调和迁移学习\n",
    "- 使用预训练权重模型能够使得联邦模型获得更好的性能"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "base",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.17"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}