ChromaGAN with GP.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "ChromaGAN.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/KunalVaidya99/Image-Colorization-Using-Deep-Learning/blob/main/ChromaGAN%20with%20GP.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zA8bcFbU0-VF",
        "outputId": "2fdb244e-20c4-44a9-8c51-4812d7a730df",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "%tensorflow_version 1.x"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "TensorFlow 1.x selected.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CswEHXhU1YDR",
        "outputId": "8fca5711-bb5f-4c52-ba6f-d828cf6055ce",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount(\"/content/gdrive/\")"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /content/gdrive/\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zaoB1Kbu1iXX",
        "outputId": "4c886a6e-9ee2-4763-eb0f-ef860c11833c",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "import tensorflow as tf\n",
        "import keras\n",
        "import keras.backend as K\n",
        "from keras.layers import (Conv2D,Conv2DTranspose,UpSampling2D,MaxPooling2D,Input\n",
        "                          ,Activation,Concatenate,BatchNormalization,Flatten,Concatenate\n",
        "                          ,Dense,RepeatVector,Reshape)\n",
        "from keras.optimizers import Adam\n",
        "from keras.layers.advanced_activations import LeakyReLU\n",
        "from keras.utils import plot_model\n",
        "from keras.models import Model,Sequential\n",
        "import numpy as np\n",
        "from keras.preprocessing.image import load_img\n",
        "from PIL import Image\n",
        "import os\n",
        "import matplotlib.pyplot as plt\n",
        "import pickle\n",
        "from functools import partial\n",
        "import random\n",
        "import progressbar\n",
        "from skimage.color import gray2rgb\n",
        "import matplotlib.pyplot as plt\n",
        "from keras.callbacks import ModelCheckpoint\n",
        "from keras.applications import VGG16\n",
        "from keras.preprocessing.image import load_img,img_to_array\n",
        "from keras.layers.merge import _Merge"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Using TensorFlow backend.\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JWGloSyGgyav"
      },
      "source": [
        "Extracting Features for all images using VGG16"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eKF4y_Cc2pe0",
        "outputId": "0e762439-a7b4-4e0a-cd40-06cd5e52b8d6",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "def extract_features(directory):\n",
        "\n",
        "  model = VGG16(include_top=True,weights=\"imagenet\",classes=1000)\n",
        "  features = dict()\n",
        "  count = 0\n",
        "  for filename in os.listdir(directory):\n",
        "   \n",
        "    image = load_img(directory+filename,target_size=(224,224))\n",
        "    img_array = img_to_array(image)\n",
        "    img_array = img_array.reshape((1,224,224,3))\n",
        "    feature = model.predict(img_array)\n",
        "    features[filename] = feature\n",
        "    count = count + 1\n",
        "\n",
        "    if (count%1000==0):\n",
        "      print(\"Completed Feature Extraction Of \" + f\"{count}\") \n",
        "\n",
        "  return features\n",
        "\n",
        "directory = \"/content/gdrive/My Drive/val_256/\"\n",
        "features = extract_features(directory)       "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:From /tensorflow-1.15.2/python3.6/keras/backend/tensorflow_backend.py:422: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n",
            "\n",
            "Completed Feature Extraction Of 1000\n",
            "Completed Feature Extraction Of 2000\n",
            "Completed Feature Extraction Of 3000\n",
            "Completed Feature Extraction Of 4000\n",
            "Completed Feature Extraction Of 5000\n",
            "Completed Feature Extraction Of 6000\n",
            "Completed Feature Extraction Of 7000\n",
            "Completed Feature Extraction Of 8000\n",
            "Completed Feature Extraction Of 9000\n",
            "Completed Feature Extraction Of 10000\n",
            "Completed Feature Extraction Of 11000\n",
            "Completed Feature Extraction Of 12000\n",
            "Completed Feature Extraction Of 13000\n",
            "Completed Feature Extraction Of 14000\n",
            "Completed Feature Extraction Of 15000\n",
            "Completed Feature Extraction Of 16000\n",
            "Completed Feature Extraction Of 17000\n",
            "Completed Feature Extraction Of 18000\n",
            "Completed Feature Extraction Of 19000\n",
            "Completed Feature Extraction Of 20000\n",
            "Completed Feature Extraction Of 21000\n",
            "Completed Feature Extraction Of 22000\n",
            "Completed Feature Extraction Of 23000\n",
            "Completed Feature Extraction Of 24000\n",
            "Completed Feature Extraction Of 25000\n",
            "Completed Feature Extraction Of 26000\n",
            "Completed Feature Extraction Of 27000\n",
            "Completed Feature Extraction Of 28000\n",
            "Completed Feature Extraction Of 29000\n",
            "Completed Feature Extraction Of 30000\n",
            "Completed Feature Extraction Of 31000\n",
            "Completed Feature Extraction Of 32000\n",
            "Completed Feature Extraction Of 33000\n",
            "Completed Feature Extraction Of 34000\n",
            "Completed Feature Extraction Of 35000\n",
            "Completed Feature Extraction Of 36000\n",
            "Completed Feature Extraction Of 37000\n",
            "Completed Feature Extraction Of 38000\n",
            "Completed Feature Extraction Of 39000\n",
            "Completed Feature Extraction Of 40000\n",
            "Completed Feature Extraction Of 41000\n",
            "Completed Feature Extraction Of 42000\n",
            "Completed Feature Extraction Of 43000\n",
            "Completed Feature Extraction Of 44000\n",
            "Completed Feature Extraction Of 45000\n",
            "Completed Feature Extraction Of 46000\n",
            "Completed Feature Extraction Of 47000\n",
            "Completed Feature Extraction Of 48000\n",
            "Completed Feature Extraction Of 49000\n",
            "Completed Feature Extraction Of 50000\n",
            "Completed Feature Extraction Of 51000\n",
            "Completed Feature Extraction Of 52000\n",
            "Completed Feature Extraction Of 53000\n",
            "Completed Feature Extraction Of 54000\n",
            "Completed Feature Extraction Of 55000\n",
            "Completed Feature Extraction Of 56000\n",
            "Completed Feature Extraction Of 57000\n",
            "Completed Feature Extraction Of 58000\n",
            "Completed Feature Extraction Of 59000\n",
            "Completed Feature Extraction Of 60000\n",
            "Completed Feature Extraction Of 61000\n",
            "Completed Feature Extraction Of 62000\n",
            "Completed Feature Extraction Of 63000\n",
            "Completed Feature Extraction Of 64000\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-6GuKhL-5-aT"
      },
      "source": [
        "import pickle\n",
        "f = open(\"/content/gdrive/My Drive/ImageColorization /chromaVGG16.pkl\",\"wb\")\n",
        "pickle.dump(features,f)\n",
        "f.close()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CoHf2gb8yFA_"
      },
      "source": [
        "f_new = open(\"/content/gdrive/My Drive/ImageColorization /chromaVGG16.pkl\",\"rb\")\n",
        "features_new = pickle.load(f_new)\n",
        "f_new.close()"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NYtomIhCJfZf"
      },
      "source": [
        "class DataGenerator():\n",
        "\n",
        "  def __init__(self,features,directory,dim=(224,224),batch_size=32,n_channels=3):\n",
        "    \n",
        "    self.directory = directory\n",
        "    self.dim = dim\n",
        "    self.batch_size=batch_size\n",
        "    self.n_channels = n_channels\n",
        "    self.features = features\n",
        "    \n",
        "\n",
        "  def epoch_end(self,filenames):\n",
        "    np.random.shuffle(filenames)\n",
        "\n",
        "  def generate_batch(self,step_number,filenames):\n",
        "    \n",
        "\n",
        "    batch_filenames = filenames[step_number*self.batch_size:(step_number+1)*self.batch_size]\n",
        "\n",
        "    gray_image = np.empty((self.batch_size,*self.dim,self.n_channels))\n",
        "    color_image = np.empty((self.batch_size,*self.dim,self.n_channels))\n",
        "    VGG16_target = np.empty((self.batch_size,1000))\n",
        "    count = 0\n",
        "\n",
        "\n",
        "    for filename in batch_filenames:\n",
        "      image = Image.open(self.directory +\"/\" + filename)\n",
        "      image = image.resize((224,224))\n",
        "\n",
        "\n",
        "      if (np.array(image).shape==(224,224,3)):\n",
        "        \n",
        "        gray = (np.array(image.convert(\"L\")))\n",
        "        gray_image[count] = gray2rgb(gray)/255\n",
        "        color_image[count] = np.array(image)/255\n",
        "        VGG16_target[count] = self.features[filename]\n",
        "        count = count + 1\n",
        "\n",
        "    return gray_image,color_image,VGG16_target    "
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ACE0LabGvNW5"
      },
      "source": [
        "def patch_discriminator(img_shape):\n",
        "\n",
        "  input_img = Input(shape=img_shape)\n",
        "\n",
        "  X = Conv2D(filters=64,kernel_size=(4,4),strides=(2,2),padding='same')(input_img)\n",
        "  #X = BatchNormalization(axis=-1)(X)\n",
        "  X = LeakyReLU(alpha=0.2)(X)\n",
        "\n",
        "  X = Conv2D(filters=128,kernel_size=(4,4),strides=(2,2),padding='same')(X)\n",
        "  #X = BatchNormalization(axis=-1)(X)\n",
        "  X = LeakyReLU(alpha=0.2)(X)\n",
        "\n",
        "  X = Conv2D(filters=256,kernel_size=(4,4),strides=(2,2),padding='same')(X)\n",
        "  #X = BatchNormalization(axis=-1)(X)\n",
        "  X = LeakyReLU(alpha=0.2)(X)\n",
        "\n",
        "  X = Conv2D(filters=512,kernel_size=(4,4),strides=(1,1),padding='same')(X)\n",
        "  #X = BatchNormalization(axis=-1)(X)\n",
        "  X = LeakyReLU(alpha=0.2)(X)\n",
        "\n",
        "  X = Conv2D(filters=1,kernel_size=(4,4),strides=(1,1),padding='same')(X)\n",
        " \n",
        " \n",
        "  model = Model(inputs=input_img,outputs=X)\n",
        "\n",
        "  return model\n"
      ],
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FgukDnq_8Uv0",
        "outputId": "741611d5-9adf-4c70-9218-c05e94ad0aae",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        " VGG16model = VGG16(include_top=False,weights=\"imagenet\",classes=1000)\n",
        " #for layer in VGG16model.layers:\n",
        "  # layer.trainable = False"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:From /tensorflow-1.15.2/python3.6/tensorflow_core/python/ops/resource_variable_ops.py:1630: calling BaseResourceVariable.__init__ (from tensorflow.python.ops.resource_variable_ops) with constraint is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "If using Keras pass *_constraint arguments to layers.\n",
            "WARNING:tensorflow:From /tensorflow-1.15.2/python3.6/keras/backend/tensorflow_backend.py:4070: The name tf.nn.max_pool is deprecated. Please use tf.nn.max_pool2d instead.\n",
            "\n",
            "Downloading data from https://github.com/fchollet/deep-learning-models/releases/download/v0.1/vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "58892288/58889256 [==============================] - 2s 0us/step\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MVp9NhQx8IZP"
      },
      "source": [
        "def generator(img_shape):\n",
        "\n",
        "  input_img = Input(shape=img_shape)\n",
        "  X = input_img\n",
        "\n",
        "  for layer in VGG16model.layers[1:-6]:\n",
        "    X = layer(X)\n",
        "    \n",
        "  VGG16_output = X \n",
        "  #print(VGG16_output.shape)\n",
        "  X = Conv2D(filters=512,kernel_size=(3,3),strides=(2,2),padding=\"same\")(X)\n",
        "  X = BatchNormalization(axis=-1)(X)\n",
        "  X = Activation('relu')(X)\n",
        "\n",
        "  X = Conv2D(filters=512,kernel_size=(3,3),strides=(1,1),padding=\"same\")(X)\n",
        "  X = BatchNormalization(axis=-1)(X)\n",
        "  X = Activation('relu')(X)\n",
        "\n",
        "  X = Conv2D(filters=512,kernel_size=(3,3),strides=(2,2),padding=\"same\")(X)\n",
        "  X = BatchNormalization(axis=-1)(X)\n",
        "  X = Activation('relu')(X)\n",
        "\n",
        "  X = Conv2D(filters=512,kernel_size=(3,3),strides=(1,1),padding=\"same\")(X)\n",
        "  X = BatchNormalization(axis=-1)(X)\n",
        "  X = Activation('relu')(X)\n",
        "\n",
        "  color_distribution = Flatten()(X)\n",
        "  color_distribution = Dense(units=4096)(color_distribution)\n",
        "  color_distribution = Dense(units=4096)(color_distribution)\n",
        "  color_distribution = Dense(units=1000,activation='softmax',name='colordistribution')(color_distribution)\n",
        "\n",
        "  fusion = Flatten()(X)\n",
        "  fusion = Dense(units=1024)(fusion)\n",
        "  fusion = Dense(units=512)(fusion)\n",
        "  fusion = Dense(units=256)(fusion)\n",
        "  fusion = RepeatVector(28*28)(fusion)\n",
        "  fusion = Reshape((28,28,256))(fusion)\n",
        "\n",
        "  encoder_output = Conv2D(filters=512,kernel_size=(3,3),strides=(1,1),padding=\"same\")(VGG16_output)\n",
        "  encoder_output = BatchNormalization(axis=-1)(encoder_output)\n",
        "  encoder_output = Activation('relu')(encoder_output)\n",
        "  encoder_output = Conv2D(filters=256,kernel_size=(3,3),strides=(1,1),padding=\"same\")(VGG16_output)\n",
        "  encoder_output = BatchNormalization(axis=-1)(encoder_output)\n",
        "  encoder_output = Activation('relu')(encoder_output)\n",
        "\n",
        "  encoder_output = Concatenate(axis=-1)([encoder_output,fusion])\n",
        "\n",
        "  decoder = Conv2D(filters=256,kernel_size=(1,1),strides=(1,1),activation='relu',padding='same')(encoder_output)\n",
        "  decoder = Conv2D(filters=128,kernel_size=(3,3),strides=(1,1),activation='relu',padding='same')(decoder)\n",
        "  decoder = UpSampling2D((2,2))(decoder)\n",
        "\n",
        "  decoder = Conv2D(filters=64,kernel_size=(3,3),strides=(1,1),padding='same',activation='relu')(decoder)\n",
        "  decoder = Conv2D(filters=64,kernel_size=(3,3),strides=(1,1),padding='same',activation='relu')(decoder)\n",
        "  decoder = UpSampling2D((2,2))(decoder)\n",
        "\n",
        "  decoder = Conv2D(filters=32,kernel_size=(3,3),strides=(1,1),padding='same',activation='relu')(decoder)\n",
        "  decoder = Conv2D(filters=3,kernel_size=(3,3),strides=(1,1),padding='same')(decoder)\n",
        "  decoder = UpSampling2D((2,2))(decoder)\n",
        "\n",
        "\n",
        "  model = Model(inputs=input_img,outputs=[decoder,color_distribution])\n",
        "\n",
        "  return model"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_YK3gdNdGsRz"
      },
      "source": [
        "def chromaGAN(img_shape,discriminator,generator):\n",
        "  X = Input(img_shape)\n",
        "  image,color_distribution = generator(X)\n",
        "  output = discriminator(image)\n",
        "  model = Model(inputs=X,outputs=[image,color_distribution,output])\n",
        "  return model"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UhLxJUnFNhv-"
      },
      "source": [
        "chromaGAN = chromaGAN((224,224,3),patch_discriminator,generator)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Rl77My8-8rfy"
      },
      "source": [
        "def wasserstein_loss(y_true,y_pred):\n",
        "  return K.mean(y_true*y_pred)"
      ],
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6xaVoUKq91nq"
      },
      "source": [
        "class random_weighted_average(_Merge):\n",
        "   def _merge_function(self, inputs):\n",
        "      weights = K.random_uniform((32, 1, 1, 1))\n",
        "      return (weights * inputs[0]) + ((1 - weights) * inputs[1])\n"
      ],
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2_1jns4LB3d1"
      },
      "source": [
        "def gradient_penalty_loss(y_true,y_pred,interpolated,penalty_weight):\n",
        "  \n",
        "  gradients = K.gradients(y_pred,interpolated)[0]\n",
        "  gradients_sqr = K.square(gradients)\n",
        "  gradients_sqr_sum = K.sum(gradients_sqr,axis=np.arange(1,len(gradients_sqr.shape)))\n",
        "  gradient_l2_norm = K.sqrt(gradients_sqr_sum)\n",
        "  gradient_penalty = penalty_weight * K.square(1 - gradient_l2_norm)\n",
        "  return K.mean(gradient_penalty)  \n"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EMonNhxdLCoH"
      },
      "source": [
        "gradient_penalty_weight = 10\n",
        "opt = Adam(learning_rate=0.00002,beta_1=0.5,beta_2=0.999)\n",
        "patch_discriminator = patch_discriminator((224,224,3))\n",
        "generator = generator((224,224,3))\n",
        "patch_discriminator.load_weights(\"/content/gdrive/My Drive/ImageColorization /chromaGAN_GP/disc_epoch6_step no.1000.h5\")\n",
        "generator.load_weights(\"/content/gdrive/My Drive/ImageColorization /chromaGAN_GP/gen_epoch6_step no.1000.h5\")"
      ],
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nxpsgLffIYqi"
      },
      "source": [
        "input = Input(shape=(224,224,3))\n",
        "generator.trainable=False\n",
        "gen_pred,_ = generator(input)\n",
        "interpolated = random_weighted_average()([input,gen_pred])\n",
        "disc_output = patch_discriminator(input)\n",
        "interpolated_disc_output = patch_discriminator(interpolated)\n",
        "gp_loss = partial(gradient_penalty_loss,interpolated=interpolated,penalty_weight=gradient_penalty_weight)\n",
        "gp_loss.__name__ = 'gradient_penalty'\n",
        "\n",
        "discriminator_model = Model(inputs=input,outputs=[disc_output,interpolated_disc_output])\n",
        "\n",
        "\n",
        "discriminator_model.compile(optimizer=opt,loss=[wasserstein_loss,gp_loss],loss_weights=[1,1])\n",
        "\n",
        "\n"
      ],
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "16z-7oN3Mk20"
      },
      "source": [
        "generator.trainable = True\n",
        "patch_discriminator.trainable = False\n",
        "chromaGAN = chromaGAN((224,224,3),patch_discriminator,generator)\n",
        "chromaGAN.compile(optimizer=opt,loss=[\"mse\",keras.losses.kullback_leibler_divergence,wasserstein_loss],loss_weights=[1,0.003,0.1])"
      ],
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OMleCSntUjuE"
      },
      "source": [
        "def summarize_performance(step,generator,gray_images,color_images,name = \"Colorizer\"):\n",
        "  n_samples = 5\n",
        "  \n",
        "\n",
        "  images,_ = generator.predict(gray_images)\n",
        "  for i in range(n_samples):\n",
        "    plt.subplot(3, n_samples, 1 +i)\n",
        "    plt.axis('off')\n",
        "    plt.imshow(images[i])\n",
        "    plt.title(\"Predicted\")\n",
        "  \n",
        "  for i in range(n_samples):\n",
        "    plt.subplot(3, n_samples, 1 + n_samples + i)\n",
        "    plt.axis('off')\n",
        "    plt.imshow(gray_images[i])\n",
        "    plt.title(\"Gray\")\n",
        "\n",
        "  for i in range(n_samples):\n",
        "    plt.subplot(3, n_samples, 1 + 2*n_samples+i)\n",
        "    plt.axis('off')\n",
        "    plt.imshow(color_images[i])\n",
        "    plt.title(\"GT\")\n",
        "  \n",
        "\n",
        "  filename1 = '%s_generated_plot_%06d.png' % (name, (step+1))\n",
        "  plt.savefig(filename1)\n",
        "  plt.close()"
      ],
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XIyYJMjank_o"
      },
      "source": [
        "def train(generator,discriminator,discriminator_model,chromaGAN,batch_size,epochs):\n",
        "  \n",
        "  directory = \"/content/gdrive/My Drive/val_256/\"\n",
        "  n_steps = int(len(os.listdir(directory))/batch_size)\n",
        "  filenames = os.listdir(directory)\n",
        "  DataLoader = DataGenerator(features_new,directory,batch_size=batch_size)\n",
        "\n",
        "\n",
        "  for epoch in range(6,epochs):\n",
        "    DataLoader.epoch_end(filenames)\n",
        "    print(\"**********************EPOCH \"+f\"{epoch+1}\" + \" *********************\")\n",
        "    print(\"\\n\")\n",
        "\n",
        "    for step in progressbar.progressbar(range(n_steps)):\n",
        "\n",
        "      gray_batch,color_batch,VGG16batch = DataLoader.generate_batch(step,filenames)\n",
        "      \n",
        "      \n",
        "      gray_batch_gen,color_batch_gen,VGG16batch_gen = DataLoader.generate_batch(n_steps-step,filenames)\n",
        "\n",
        "      fake_images,_ = generator.predict(x=gray_batch)\n",
        "      real = np.ones((batch_size,28,28,1))\n",
        "      fake = -real\n",
        "      dummy = np.zeros((batch_size,28,28,1))\n",
        "\n",
        "      d_loss1 = discriminator_model.train_on_batch(x=color_batch,y=[real,dummy])\n",
        "      d_loss2 = discriminator_model.train_on_batch(x=fake_images,y=[fake,dummy])\n",
        "      d_loss = [d_loss1,d_loss2]\n",
        "\n",
        "      g_loss = chromaGAN.train_on_batch(x=gray_batch_gen,y=[color_batch_gen,VGG16batch_gen,real])\n",
        "\n",
        "      if ((step+1)%50==0):\n",
        "        \n",
        "        idx1 = np.random.randint(0,n_steps)\n",
        "        gray_images,color_images,_ = DataLoader.generate_batch(idx1,filenames)\n",
        "        summarize_performance(step,generator,gray_images[:5],color_images[:5])\n",
        "        print(\"Step Number :\",step+1,sep=\"  \")\n",
        "        print(\"Generator Loss :\",g_loss,sep=\"  \")\n",
        "        print(\"Discriminator Loss :\",d_loss)\n",
        "        print(\"\\n\")\n",
        "\n",
        "      if ((step+1)%500==0):\n",
        "        filename_disc = \"/content/gdrive/My Drive/ImageColorization /chromaGAN_GP/\" + \"disc_\"+ \"epoch\" + f\"{epoch+1}\" +\"_\" + \"step no.\" + f\"{step+1}\" +\".h5\"\n",
        "        filename_gen = \"/content/gdrive/My Drive/ImageColorization /chromaGAN_GP/\" + \"gen_\"+ \"epoch\" + f\"{epoch+1}\" +\"_\" + \"step no.\" + f\"{step+1}\" +\".h5\"\n",
        "        discriminator.save_weights(filename_disc)\n",
        "        generator.save_weights(filename_gen)\n",
        "\n"
      ],
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xKDoqJLKy8pV",
        "outputId": "bac9b759-07a3-40be-a1d7-d603415422dd",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "train(generator,patch_discriminator,discriminator_model,chromaGAN,batch_size=32,epochs=15)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\r                                                                               \r\rN/A% (0 of 2030) |                       | Elapsed Time: 0:00:00 ETA:  --:--:--"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "**********************EPOCH 7 *********************\n",
            "\n",
            "\n",
            "WARNING:tensorflow:From /tensorflow-1.15.2/python3.6/keras/backend/tensorflow_backend.py:422: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n",
            "\n",
            "WARNING:tensorflow:From /tensorflow-1.15.2/python3.6/tensorflow_core/python/ops/math_grad.py:1424: where (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Use tf.where in 2.0, which has the same broadcast rule as np.where\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/tensorflow-1.15.2/python3.6/keras/engine/training.py:297: UserWarning: Discrepancy between trainable weights and collected trainable weights, did you set `model.trainable` without calling `model.compile` after ?\n",
            "  'Discrepancy between trainable weights and collected trainable'\n",
            "/tensorflow-1.15.2/python3.6/keras/engine/training.py:297: UserWarning: Discrepancy between trainable weights and collected trainable weights, did you set `model.trainable` without calling `model.compile` after ?\n",
            "  'Discrepancy between trainable weights and collected trainable'\n",
            "  0% (1 of 2030) |                 | Elapsed Time: 0:00:48 ETA:  1 day, 3:33:33/tensorflow-1.15.2/python3.6/keras/engine/training.py:297: UserWarning: Discrepancy between trainable weights and collected trainable weights, did you set `model.trainable` without calling `model.compile` after ?\n",
            "  'Discrepancy between trainable weights and collected trainable'\n",
            "  2% (49 of 2030) |                      | Elapsed Time: 0:22:33 ETA:  14:04:50Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "  2% (50 of 2030) |                      | Elapsed Time: 0:23:12 ETA:  21:44:54"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  50\n",
            "Generator Loss :  [0.018103562, 0.011777321, 1.408296, 0.021013537]\n",
            "Discriminator Loss : [[0.005090573, 0.0036048929, 0.0014856802], [-0.019092528, -0.019557646, 0.00046511728]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "  4% (99 of 2030) |#                     | Elapsed Time: 0:43:45 ETA:  12:14:09Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "  4% (100 of 2030) |#                    | Elapsed Time: 0:44:21 ETA:  19:05:58"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  100\n",
            "Generator Loss :  [0.016304038, 0.010473558, 1.2951388, 0.019450635]\n",
            "Discriminator Loss : [[0.0027100444, 0.0020506445, 0.0006593999], [-0.017794581, -0.017992223, 0.00019764106]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "  7% (149 of 2030) |#                    | Elapsed Time: 1:04:19 ETA:  12:28:57Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "  7% (150 of 2030) |#                    | Elapsed Time: 1:04:53 ETA:  17:58:29"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  150\n",
            "Generator Loss :  [0.019066911, 0.013656468, 1.3748174, 0.012859901]\n",
            "Discriminator Loss : [[0.0048796535, 0.0011764001, 0.0037032536], [0.0005929107, -0.01463304, 0.015225951]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "  9% (199 of 2030) |##                   | Elapsed Time: 1:24:20 ETA:  12:02:07Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "  9% (200 of 2030) |##                   | Elapsed Time: 1:24:56 ETA:  18:16:33"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  200\n",
            "Generator Loss :  [0.019604977, 0.014259327, 1.2669004, 0.0154494895]\n",
            "Discriminator Loss : [[0.0012230896, 0.00049171946, 0.0007313702], [-0.0157844, -0.016032586, 0.00024818658]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 12% (249 of 2030) |##                   | Elapsed Time: 1:44:04 ETA:  10:58:01Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 12% (250 of 2030) |##                   | Elapsed Time: 1:44:38 ETA:  16:56:20"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  250\n",
            "Generator Loss :  [0.01685161, 0.011552386, 1.2393074, 0.015813027]\n",
            "Discriminator Loss : [[0.004028389, 0.0023806598, 0.0016477293], [-0.013906965, -0.015075796, 0.0011688313]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 14% (299 of 2030) |###                  | Elapsed Time: 2:03:18 ETA:  10:40:54Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 14% (300 of 2030) |###                  | Elapsed Time: 2:03:51 ETA:  16:10:35"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  300\n",
            "Generator Loss :  [0.01706198, 0.0119727375, 1.2314322, 0.013949467]\n",
            "Discriminator Loss : [[0.004567315, 0.0031167895, 0.0014505255], [-0.013814607, -0.014336797, 0.00052219105]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 17% (349 of 2030) |###                  | Elapsed Time: 2:22:15 ETA:  10:43:34Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 17% (350 of 2030) |###                  | Elapsed Time: 2:22:50 ETA:  16:09:36"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  350\n",
            "Generator Loss :  [0.016045485, 0.010600494, 1.4055674, 0.012282883]\n",
            "Discriminator Loss : [[-0.0028106617, -0.0034550063, 0.0006443447], [-0.010285249, -0.010602154, 0.00031690468]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 19% (399 of 2030) |####                 | Elapsed Time: 2:41:17 ETA:  10:36:24Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 19% (400 of 2030) |####                 | Elapsed Time: 2:41:49 ETA:  14:32:03"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  400\n",
            "Generator Loss :  [0.017344322, 0.012641414, 1.2482837, 0.009580579]\n",
            "Discriminator Loss : [[0.0045710327, 0.0029826548, 0.0015883779], [-0.0070811566, -0.009848403, 0.0027672462]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 22% (449 of 2030) |####                 | Elapsed Time: 3:00:15 ETA:   9:46:13Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 22% (450 of 2030) |####                 | Elapsed Time: 3:00:49 ETA:  15:05:13"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  450\n",
            "Generator Loss :  [0.020557089, 0.014239589, 1.6000483, 0.015173554]\n",
            "Discriminator Loss : [[0.0072619338, 0.0048650806, 0.002396853], [-0.013923797, -0.014942415, 0.0010186181]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 24% (499 of 2030) |#####                | Elapsed Time: 3:19:25 ETA:   9:38:21Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  500\n",
            "Generator Loss :  [0.018664727, 0.013671089, 1.3661495, 0.0089518875]\n",
            "Discriminator Loss : [[0.0042618, 0.00010646112, 0.0041553387], [-0.008810848, -0.0114174085, 0.00260656]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 27% (549 of 2030) |#####                | Elapsed Time: 3:38:35 ETA:   9:27:02Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 27% (550 of 2030) |#####                | Elapsed Time: 3:38:58 ETA:   9:23:53"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  550\n",
            "Generator Loss :  [0.024494382, 0.018324219, 1.5878327, 0.014066643]\n",
            "Discriminator Loss : [[0.0031251214, 0.0011500112, 0.00197511], [-0.011540792, -0.012417813, 0.0008770205]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 29% (599 of 2030) |######               | Elapsed Time: 3:57:54 ETA:   9:20:25Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 29% (600 of 2030) |######               | Elapsed Time: 3:58:19 ETA:  10:11:50"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  600\n",
            "Generator Loss :  [0.018084496, 0.013024401, 1.1847408, 0.015058729]\n",
            "Discriminator Loss : [[0.005402577, 0.0035840601, 0.0018185169], [-0.01245562, -0.01356914, 0.0011135202]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 31% (649 of 2030) |######               | Elapsed Time: 4:17:15 ETA:   9:05:25Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 32% (650 of 2030) |######               | Elapsed Time: 4:17:40 ETA:   9:24:25"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  650\n",
            "Generator Loss :  [0.017634667, 0.012460583, 1.3052791, 0.01258247]\n",
            "Discriminator Loss : [[0.0042304597, 0.0026087847, 0.001621675], [-0.013365091, -0.014399017, 0.0010339264]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 34% (699 of 2030) |#######              | Elapsed Time: 4:37:04 ETA:   9:59:04Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 34% (700 of 2030) |#######              | Elapsed Time: 4:37:40 ETA:  13:15:37"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  700\n",
            "Generator Loss :  [0.020438617, 0.014327382, 1.5038991, 0.015995365]\n",
            "Discriminator Loss : [[0.00450008, 0.0035729173, 0.0009271627], [-0.013885514, -0.014104595, 0.0002190811]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 36% (749 of 2030) |#######              | Elapsed Time: 4:57:19 ETA:   8:25:19Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 36% (750 of 2030) |#######              | Elapsed Time: 4:57:45 ETA:   9:08:52"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  750\n",
            "Generator Loss :  [0.02037733, 0.014357638, 1.5295286, 0.014311068]\n",
            "Discriminator Loss : [[0.007077495, 0.005420484, 0.001657011], [-0.014595956, -0.01517677, 0.0005808143]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 39% (799 of 2030) |########             | Elapsed Time: 5:17:30 ETA:   7:45:24Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 39% (800 of 2030) |########             | Elapsed Time: 5:17:42 ETA:   4:21:01"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  800\n",
            "Generator Loss :  [0.018812994, 0.01319145, 1.3726413, 0.015036186]\n",
            "Discriminator Loss : [[0.00082502246, -0.000111463916, 0.00093648635], [-0.009031409, -0.009643039, 0.0006116297]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 41% (849 of 2030) |########             | Elapsed Time: 5:37:12 ETA:   7:44:30Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 41% (850 of 2030) |########             | Elapsed Time: 5:37:35 ETA:   7:28:44"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  850\n",
            "Generator Loss :  [0.01879562, 0.011614008, 1.7563088, 0.019126862]\n",
            "Discriminator Loss : [[0.011678328, 0.010274761, 0.0014035662], [-0.01943781, -0.020509511, 0.001071701]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 44% (899 of 2030) |#########            | Elapsed Time: 5:56:42 ETA:   7:52:42Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 44% (900 of 2030) |#########            | Elapsed Time: 5:57:05 ETA:   7:03:15"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  900\n",
            "Generator Loss :  [0.02371023, 0.017200064, 1.6157107, 0.016630342]\n",
            "Discriminator Loss : [[0.012504586, 0.011317691, 0.0011868952], [-0.020277835, -0.020666672, 0.00038883608]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 46% (949 of 2030) |#########            | Elapsed Time: 6:16:36 ETA:   7:20:29Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 46% (950 of 2030) |#########            | Elapsed Time: 6:17:01 ETA:   7:32:36"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  950\n",
            "Generator Loss :  [0.01646753, 0.010538495, 1.3430411, 0.018999124]\n",
            "Discriminator Loss : [[0.0088825645, 0.008166426, 0.00071613863], [-0.015514632, -0.01584393, 0.00032929773]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 49% (999 of 2030) |##########           | Elapsed Time: 6:36:48 ETA:   6:56:59Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1000\n",
            "Generator Loss :  [0.018936755, 0.012349555, 1.5644813, 0.018937573]\n",
            "Discriminator Loss : [[0.006904345, 0.00578574, 0.0011186048], [-0.016904064, -0.017818097, 0.00091403275]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 51% (1049 of 2030) |##########          | Elapsed Time: 6:44:14 ETA:   0:21:03Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 51% (1050 of 2030) |##########          | Elapsed Time: 6:44:16 ETA:   0:26:14"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1050\n",
            "Generator Loss :  [0.016190194, 0.010188282, 1.3629259, 0.019131338]\n",
            "Discriminator Loss : [[0.0036360065, 0.002550916, 0.0010850905], [-0.017974578, -0.018441202, 0.0004666238]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 54% (1099 of 2030) |##########          | Elapsed Time: 6:45:19 ETA:   0:20:10Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 54% (1100 of 2030) |##########          | Elapsed Time: 6:45:21 ETA:   0:25:05"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1100\n",
            "Generator Loss :  [0.018213017, 0.012547922, 1.3199415, 0.017052695]\n",
            "Discriminator Loss : [[0.0023066066, 0.0014204194, 0.00088618713], [-0.01012764, -0.010483362, 0.0003557211]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 56% (1149 of 2030) |###########         | Elapsed Time: 6:46:24 ETA:   0:19:04Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 56% (1150 of 2030) |###########         | Elapsed Time: 6:46:26 ETA:   0:23:43"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1150\n",
            "Generator Loss :  [0.016537618, 0.010660636, 1.3112004, 0.019433798]\n",
            "Discriminator Loss : [[0.0036254807, 0.003041327, 0.0005841537], [-0.015999474, -0.01629821, 0.00029873755]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 59% (1199 of 2030) |###########         | Elapsed Time: 6:47:30 ETA:   0:17:50Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 59% (1200 of 2030) |###########         | Elapsed Time: 6:47:32 ETA:   0:22:04"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1200\n",
            "Generator Loss :  [0.017475804, 0.011866205, 1.3321353, 0.016131928]\n",
            "Discriminator Loss : [[0.007602827, 0.0067968997, 0.0008059276], [-0.017121134, -0.017269682, 0.00014854866]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 61% (1249 of 2030) |############        | Elapsed Time: 6:48:35 ETA:   0:16:36Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 61% (1250 of 2030) |############        | Elapsed Time: 6:48:37 ETA:   0:29:40"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1250\n",
            "Generator Loss :  [0.01901546, 0.01270106, 1.5426764, 0.016863696]\n",
            "Discriminator Loss : [[0.003144619, 0.0022193077, 0.0009253113], [-0.012750708, -0.013083108, 0.00033240058]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 63% (1299 of 2030) |############        | Elapsed Time: 6:49:40 ETA:   0:15:32Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 64% (1300 of 2030) |############        | Elapsed Time: 6:49:42 ETA:   0:19:26"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1300\n",
            "Generator Loss :  [0.017535461, 0.011137863, 1.4189373, 0.021407858]\n",
            "Discriminator Loss : [[0.004325326, 0.0031896296, 0.0011356964], [-0.016772704, -0.017111516, 0.00033881055]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 66% (1349 of 2030) |#############       | Elapsed Time: 6:50:45 ETA:   0:14:34Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 66% (1350 of 2030) |#############       | Elapsed Time: 6:50:47 ETA:   0:18:03"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1350\n",
            "Generator Loss :  [0.018025827, 0.012057571, 1.426676, 0.016882272]\n",
            "Discriminator Loss : [[0.007031521, 0.0055290875, 0.0015024336], [-0.01701296, -0.017411675, 0.00039871552]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 68% (1399 of 2030) |#############       | Elapsed Time: 6:51:50 ETA:   0:13:32Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 68% (1400 of 2030) |#############       | Elapsed Time: 6:51:52 ETA:   0:17:00"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1400\n",
            "Generator Loss :  [0.01644141, 0.0103064105, 1.3583808, 0.020598575]\n",
            "Discriminator Loss : [[0.0011687361, 0.00016351105, 0.0010052251], [-0.013796424, -0.014215125, 0.00041870133]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 71% (1449 of 2030) |##############      | Elapsed Time: 6:52:55 ETA:   0:12:25Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 71% (1450 of 2030) |##############      | Elapsed Time: 6:52:57 ETA:   0:15:39"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1450\n",
            "Generator Loss :  [0.019193284, 0.013970384, 1.1885563, 0.01657231]\n",
            "Discriminator Loss : [[0.0056851786, 0.005138387, 0.00054679153], [-0.016839689, -0.0169616, 0.00012191222]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 73% (1499 of 2030) |##############      | Elapsed Time: 6:54:00 ETA:   0:11:19Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1500\n",
            "Generator Loss :  [0.019019851, 0.013857145, 1.3326006, 0.011649035]\n",
            "Discriminator Loss : [[0.004112022, 0.0025679907, 0.0015440313], [-0.011997858, -0.013196201, 0.0011983435]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 76% (1549 of 2030) |###############     | Elapsed Time: 6:55:09 ETA:   0:10:33Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 76% (1550 of 2030) |###############     | Elapsed Time: 6:55:11 ETA:   0:13:02"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1550\n",
            "Generator Loss :  [0.01943171, 0.013525481, 1.423246, 0.01636491]\n",
            "Discriminator Loss : [[0.0044696014, 0.0035788128, 0.0008907885], [-0.01587756, -0.01627851, 0.00040095116]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 78% (1599 of 2030) |###############     | Elapsed Time: 6:56:19 ETA:   0:15:31Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 78% (1600 of 2030) |###############     | Elapsed Time: 6:56:21 ETA:   0:14:21"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1600\n",
            "Generator Loss :  [0.017630009, 0.012298033, 1.2134135, 0.016917365]\n",
            "Discriminator Loss : [[0.0031933375, 0.001970216, 0.0012231215], [-0.013027768, -0.013718641, 0.00069087406]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 81% (1649 of 2030) |################    | Elapsed Time: 6:57:33 ETA:   0:08:37Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 81% (1650 of 2030) |################    | Elapsed Time: 6:57:35 ETA:   0:13:02"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1650\n",
            "Generator Loss :  [0.018938342, 0.012645009, 1.6564326, 0.01324036]\n",
            "Discriminator Loss : [[-0.0007238212, -0.0017243709, 0.0010005496], [-0.010561488, -0.011577959, 0.0010164715]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 83% (1699 of 2030) |################    | Elapsed Time: 6:58:41 ETA:   0:07:22Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 83% (1700 of 2030) |################    | Elapsed Time: 6:58:43 ETA:   0:09:14"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1700\n",
            "Generator Loss :  [0.016652975, 0.010969725, 1.4110658, 0.014500516]\n",
            "Discriminator Loss : [[-0.0019039803, -0.0032178205, 0.0013138403], [-0.013214053, -0.013688454, 0.00047440198]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 86% (1749 of 2030) |#################   | Elapsed Time: 6:59:49 ETA:   0:06:19Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 86% (1750 of 2030) |#################   | Elapsed Time: 6:59:51 ETA:   0:07:47"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1750\n",
            "Generator Loss :  [0.017331734, 0.012259262, 1.2409631, 0.013495839]\n",
            "Discriminator Loss : [[6.269931e-05, -0.0003759417, 0.000438641], [-0.010173794, -0.010542586, 0.00036879224]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 88% (1799 of 2030) |#################   | Elapsed Time: 7:00:58 ETA:   0:05:08Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 88% (1800 of 2030) |#################   | Elapsed Time: 7:01:00 ETA:   0:09:33"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1800\n",
            "Generator Loss :  [0.018587807, 0.012495332, 1.5806863, 0.013504157]\n",
            "Discriminator Loss : [[0.0032679383, 0.0014808689, 0.0017870695], [-0.013199043, -0.016057802, 0.002858759]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 91% (1849 of 2030) |##################  | Elapsed Time: 7:02:06 ETA:   0:03:59Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 91% (1850 of 2030) |##################  | Elapsed Time: 7:02:08 ETA:   0:04:58"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1850\n",
            "Generator Loss :  [0.019526633, 0.014034377, 1.3406644, 0.014702622]\n",
            "Discriminator Loss : [[0.0021315182, 0.0017025863, 0.00042893187], [-0.016287794, -0.016594484, 0.0003066908]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 93% (1899 of 2030) |##################  | Elapsed Time: 7:03:13 ETA:   0:02:59Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 93% (1900 of 2030) |##################  | Elapsed Time: 7:03:15 ETA:   0:03:36"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1900\n",
            "Generator Loss :  [0.0148288235, 0.009729986, 1.2221609, 0.014323544]\n",
            "Discriminator Loss : [[0.0032942551, -0.0014877146, 0.00478197], [-0.0017863279, -0.010649265, 0.008862937]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 96% (1949 of 2030) |################### | Elapsed Time: 7:04:22 ETA:   0:01:49Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 96% (1950 of 2030) |################### | Elapsed Time: 7:04:24 ETA:   0:02:12"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1950\n",
            "Generator Loss :  [0.017566428, 0.012674559, 1.2338152, 0.011904249]\n",
            "Discriminator Loss : [[-0.007392252, -0.008750152, 0.0013578999], [-0.007611705, -0.008406607, 0.00079490227]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 98% (1999 of 2030) |################### | Elapsed Time: 7:05:30 ETA:   0:00:41Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  2000\n",
            "Generator Loss :  [0.01706421, 0.011000944, 1.6193959, 0.012050789]\n",
            "Discriminator Loss : [[-0.00088879326, -0.0032148263, 0.002326033], [-0.009976217, -0.011323586, 0.0013473695]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "100% (2030 of 2030) |####################| Elapsed Time: 7:06:15 Time:  7:06:15\n",
            "N/A% (0 of 2030) |                       | Elapsed Time: 0:00:00 ETA:  --:--:--"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "**********************EPOCH 8 *********************\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "  2% (49 of 2030) |                      | Elapsed Time: 0:01:03 ETA:   0:42:28Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "  2% (50 of 2030) |                      | Elapsed Time: 0:01:05 ETA:   0:52:15"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  50\n",
            "Generator Loss :  [0.015794009, 0.011773065, 0.9563422, 0.011519181]\n",
            "Discriminator Loss : [[-0.0021224532, -0.0031770577, 0.0010546045], [-0.008820396, -0.009047481, 0.00022708552]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "  4% (99 of 2030) |#                     | Elapsed Time: 0:02:09 ETA:   0:41:32Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "  4% (100 of 2030) |#                    | Elapsed Time: 0:02:11 ETA:   0:51:45"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  100\n",
            "Generator Loss :  [0.015078667, 0.011483277, 0.8961225, 0.0090702195]\n",
            "Discriminator Loss : [[-0.0019617816, -0.0026511294, 0.0006893478], [-0.008636354, -0.008793176, 0.00015682218]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "  7% (149 of 2030) |#                    | Elapsed Time: 0:03:15 ETA:   0:40:35Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "  7% (150 of 2030) |#                    | Elapsed Time: 0:03:17 ETA:   0:49:59"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  150\n",
            "Generator Loss :  [0.01688582, 0.012238848, 1.2024455, 0.010396349]\n",
            "Discriminator Loss : [[-0.0064253826, -0.007201997, 0.00077661435], [-0.008100686, -0.008718287, 0.00061760074]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "  9% (199 of 2030) |##                   | Elapsed Time: 0:04:23 ETA:   0:39:11Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "  9% (200 of 2030) |##                   | Elapsed Time: 0:04:24 ETA:   0:48:36"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  200\n",
            "Generator Loss :  [0.015838765, 0.012257475, 1.0551226, 0.004159221]\n",
            "Discriminator Loss : [[-0.0074413936, -0.008168443, 0.0007270494], [-0.003383247, -0.0041320063, 0.00074875937]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 12% (249 of 2030) |##                   | Elapsed Time: 0:05:28 ETA:   0:38:03Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 12% (250 of 2030) |##                   | Elapsed Time: 0:05:30 ETA:   0:47:15"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  250\n",
            "Generator Loss :  [0.015499092, 0.011504232, 1.1200902, 0.0063458923]\n",
            "Discriminator Loss : [[-0.00464346, -0.0051615424, 0.0005180827], [-0.0071225557, -0.0074078357, 0.00028527988]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 14% (299 of 2030) |###                  | Elapsed Time: 0:06:33 ETA:   0:37:08Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 14% (300 of 2030) |###                  | Elapsed Time: 0:06:34 ETA:   0:46:04"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  300\n",
            "Generator Loss :  [0.014140613, 0.01009073, 1.0280329, 0.009657843]\n",
            "Discriminator Loss : [[-0.004728015, -0.006289018, 0.0015610028], [-0.007706568, -0.0086485315, 0.0009419632]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 17% (349 of 2030) |###                  | Elapsed Time: 0:07:39 ETA:   0:35:57Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 17% (350 of 2030) |###                  | Elapsed Time: 0:07:43 ETA:   1:50:38"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  350\n",
            "Generator Loss :  [0.015909852, 0.0122679025, 0.9552857, 0.0077609336]\n",
            "Discriminator Loss : [[0.003899109, -0.0031080307, 0.0070071397], [0.0016367016, -0.005386198, 0.0070228996]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 19% (399 of 2030) |####                 | Elapsed Time: 0:08:49 ETA:   0:35:09Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 19% (400 of 2030) |####                 | Elapsed Time: 0:08:51 ETA:   0:43:02"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  400\n",
            "Generator Loss :  [0.015260071, 0.011991873, 0.98434186, 0.0031517171]\n",
            "Discriminator Loss : [[-0.0071504163, -0.00833755, 0.0011871334], [-0.0008307863, -0.0026800353, 0.001849249]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 22% (449 of 2030) |####                 | Elapsed Time: 0:09:58 ETA:   0:33:30Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 22% (450 of 2030) |####                 | Elapsed Time: 0:10:00 ETA:   1:06:00"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  450\n",
            "Generator Loss :  [0.013460722, 0.009755336, 0.958849, 0.008288393]\n",
            "Discriminator Loss : [[-0.009098322, -0.011248046, 0.0021497235], [-0.0048929355, -0.005606993, 0.0007140574]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 24% (499 of 2030) |#####                | Elapsed Time: 0:11:09 ETA:   0:32:44Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  500\n",
            "Generator Loss :  [0.014450667, 0.0100692315, 1.1192172, 0.010237842]\n",
            "Discriminator Loss : [[-0.005400273, -0.0059672263, 0.0005669533], [-0.009260257, -0.00966455, 0.00040429353]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 27% (549 of 2030) |#####                | Elapsed Time: 0:12:21 ETA:   0:31:37Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 27% (550 of 2030) |#####                | Elapsed Time: 0:12:22 ETA:   0:38:55"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  550\n",
            "Generator Loss :  [0.014660558, 0.010298412, 1.1276633, 0.009791549]\n",
            "Discriminator Loss : [[-0.0036155626, -0.004715151, 0.0010995886], [-0.0068839495, -0.007404018, 0.00052006857]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 29% (599 of 2030) |######               | Elapsed Time: 0:13:40 ETA:   0:31:00Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 29% (600 of 2030) |######               | Elapsed Time: 0:13:42 ETA:   0:38:02"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  600\n",
            "Generator Loss :  [0.013860643, 0.010464171, 0.9788732, 0.004598526]\n",
            "Discriminator Loss : [[-0.007004404, -0.008029045, 0.0010246406], [-0.0028520864, -0.0031905808, 0.0003384943]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 31% (649 of 2030) |######               | Elapsed Time: 0:14:45 ETA:   0:29:21Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 32% (650 of 2030) |######               | Elapsed Time: 0:14:47 ETA:   0:36:33"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  650\n",
            "Generator Loss :  [0.016057549, 0.012144271, 1.0973768, 0.006211483]\n",
            "Discriminator Loss : [[-0.008084191, -0.008755243, 0.00067105214], [-0.0018975299, -0.0022540388, 0.00035650883]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 34% (699 of 2030) |#######              | Elapsed Time: 0:15:52 ETA:   0:28:24Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 34% (700 of 2030) |#######              | Elapsed Time: 0:15:54 ETA:   0:35:17"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  700\n",
            "Generator Loss :  [0.0145509485, 0.010877954, 0.9825053, 0.0072547914]\n",
            "Discriminator Loss : [[-0.009160289, -0.0125620365, 0.0034017477], [-7.449952e-05, -0.0029063954, 0.0028318958]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 36% (749 of 2030) |#######              | Elapsed Time: 0:17:09 ETA:   0:27:28Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 36% (750 of 2030) |#######              | Elapsed Time: 0:17:11 ETA:   0:33:31"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  750\n",
            "Generator Loss :  [0.01578312, 0.012261732, 1.0074515, 0.0049903435]\n",
            "Discriminator Loss : [[-0.007817646, -0.008514478, 0.0006968323], [-0.002572884, -0.002722499, 0.0001496149]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 39% (799 of 2030) |########             | Elapsed Time: 0:18:14 ETA:   0:26:06Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 39% (800 of 2030) |########             | Elapsed Time: 0:18:16 ETA:   0:32:18"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  800\n",
            "Generator Loss :  [0.013116525, 0.009132691, 1.1404543, 0.005624709]\n",
            "Discriminator Loss : [[-0.009557726, -0.010014368, 0.0004566423], [-0.0027058516, -0.0029145272, 0.0002086756]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 41% (849 of 2030) |########             | Elapsed Time: 0:19:27 ETA:   0:25:15Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 41% (850 of 2030) |########             | Elapsed Time: 0:19:29 ETA:   0:31:22"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  850\n",
            "Generator Loss :  [0.013692253, 0.009676362, 1.1208838, 0.006532403]\n",
            "Discriminator Loss : [[-0.011795166, -0.01252762, 0.0007324546], [-0.0016333306, -0.0022336922, 0.0006003616]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 44% (899 of 2030) |#########            | Elapsed Time: 0:20:32 ETA:   0:24:37Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 44% (900 of 2030) |#########            | Elapsed Time: 0:20:34 ETA:   0:30:22"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  900\n",
            "Generator Loss :  [0.014336908, 0.010625995, 1.0643094, 0.0051798527]\n",
            "Discriminator Loss : [[-0.0067180735, -0.010247039, 0.0035289652], [-0.002092551, -0.004680546, 0.002587995]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 46% (949 of 2030) |#########            | Elapsed Time: 0:21:47 ETA:   0:23:11Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 46% (950 of 2030) |#########            | Elapsed Time: 0:21:49 ETA:   0:28:42"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  950\n",
            "Generator Loss :  [0.01711533, 0.013309348, 1.1859204, 0.0024822159]\n",
            "Discriminator Loss : [[-0.007923056, -0.00882276, 0.00089970307], [-0.007151791, -0.0073107774, 0.00015898659]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 49% (999 of 2030) |##########           | Elapsed Time: 0:22:52 ETA:   0:22:14Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1000\n",
            "Generator Loss :  [0.0153103415, 0.0116619, 1.1731524, 0.0012898486]\n",
            "Discriminator Loss : [[-0.013585329, -0.016537443, 0.0029521142], [0.0021479744, 0.0009970856, 0.0011508887]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 51% (1049 of 2030) |##########          | Elapsed Time: 0:24:13 ETA:   0:20:54Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 51% (1050 of 2030) |##########          | Elapsed Time: 0:24:18 ETA:   1:15:28"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1050\n",
            "Generator Loss :  [0.013020811, 0.01059519, 0.7802633, 0.00084830227]\n",
            "Discriminator Loss : [[-0.012980648, -0.013764552, 0.00078390364], [0.0018975501, 0.0016901215, 0.00020742856]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 54% (1099 of 2030) |##########          | Elapsed Time: 0:25:21 ETA:   0:19:47Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 54% (1100 of 2030) |##########          | Elapsed Time: 0:25:23 ETA:   0:24:27"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1100\n",
            "Generator Loss :  [0.012538956, 0.009401049, 0.930327, 0.0034692518]\n",
            "Discriminator Loss : [[-0.004098662, -0.0070610843, 0.002962422], [-0.0014059781, -0.0036282577, 0.0022222796]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 56% (1149 of 2030) |###########         | Elapsed Time: 0:26:25 ETA:   0:18:49Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 56% (1150 of 2030) |###########         | Elapsed Time: 0:26:27 ETA:   0:22:54"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1150\n",
            "Generator Loss :  [0.012467812, 0.009503925, 0.84260786, 0.0043606316]\n",
            "Discriminator Loss : [[-0.013295073, -0.01431027, 0.0010151963], [-0.00081902975, -0.0018329122, 0.0010138825]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 59% (1199 of 2030) |###########         | Elapsed Time: 0:27:32 ETA:   0:17:28Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 59% (1200 of 2030) |###########         | Elapsed Time: 0:27:34 ETA:   0:21:36"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1200\n",
            "Generator Loss :  [0.013255793, 0.0103067905, 0.95127225, 0.00095186016]\n",
            "Discriminator Loss : [[-0.011298919, -0.012137053, 0.00083813345], [-0.00067323796, -0.00101031, 0.00033707207]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 61% (1249 of 2030) |############        | Elapsed Time: 0:28:36 ETA:   0:16:26Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 61% (1250 of 2030) |############        | Elapsed Time: 0:28:38 ETA:   0:20:21"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1250\n",
            "Generator Loss :  [0.013757766, 0.01044252, 1.0038896, 0.0030357752]\n",
            "Discriminator Loss : [[-0.011178757, -0.011932395, 0.0007536382], [0.00023520974, -0.00034822695, 0.0005834367]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 63% (1299 of 2030) |############        | Elapsed Time: 0:29:41 ETA:   0:15:27Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 64% (1300 of 2030) |############        | Elapsed Time: 0:29:44 ETA:   0:32:42"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1300\n",
            "Generator Loss :  [0.015496571, 0.011773865, 1.0492384, 0.0057499018]\n",
            "Discriminator Loss : [[-0.0128296595, -0.014051868, 0.0012222078], [0.00034425163, -0.0011408827, 0.0014851344]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 66% (1349 of 2030) |#############       | Elapsed Time: 0:30:46 ETA:   0:14:51Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 66% (1350 of 2030) |#############       | Elapsed Time: 0:30:48 ETA:   0:18:06"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1350\n",
            "Generator Loss :  [0.013907034, 0.010549505, 1.1178296, 4.0403662e-05]\n",
            "Discriminator Loss : [[-0.011536811, -0.012460942, 0.00092413183], [0.00061825226, 0.00031446543, 0.00030378683]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 68% (1399 of 2030) |#############       | Elapsed Time: 0:31:52 ETA:   0:13:14Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 68% (1400 of 2030) |#############       | Elapsed Time: 0:31:54 ETA:   0:16:22"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1400\n",
            "Generator Loss :  [0.01370839, 0.010199481, 1.1760082, -0.00019115141]\n",
            "Discriminator Loss : [[-0.010394315, -0.0128541505, 0.0024598353], [0.003059532, 0.002464441, 0.000595091]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 71% (1449 of 2030) |##############      | Elapsed Time: 0:32:56 ETA:   0:12:08Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 71% (1450 of 2030) |##############      | Elapsed Time: 0:32:58 ETA:   0:15:02"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1450\n",
            "Generator Loss :  [0.014862274, 0.011143044, 1.1715716, 0.0020451597]\n",
            "Discriminator Loss : [[-0.012481236, -0.013075236, 0.0005940007], [-0.0033952314, -0.0037314633, 0.000336232]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 73% (1499 of 2030) |##############      | Elapsed Time: 0:34:03 ETA:   0:11:20Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1500\n",
            "Generator Loss :  [0.0148041705, 0.011503644, 1.0596691, 0.0012151853]\n",
            "Discriminator Loss : [[-0.0130681405, -0.013834486, 0.0007663457], [0.00072415237, -0.00032324035, 0.0010473927]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 76% (1549 of 2030) |###############     | Elapsed Time: 0:35:21 ETA:   0:10:40Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 76% (1550 of 2030) |###############     | Elapsed Time: 0:35:23 ETA:   0:13:10"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1550\n",
            "Generator Loss :  [0.014454437, 0.011125426, 1.1088765, 2.382574e-05]\n",
            "Discriminator Loss : [[-0.010902334, -0.011937747, 0.0010354128], [0.0010378463, -4.812847e-06, 0.0010426592]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 78% (1599 of 2030) |###############     | Elapsed Time: 0:36:27 ETA:   0:09:24Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 78% (1600 of 2030) |###############     | Elapsed Time: 0:36:29 ETA:   0:11:34"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1600\n",
            "Generator Loss :  [0.015859265, 0.01288838, 0.98985445, 1.3197981e-05]\n",
            "Discriminator Loss : [[-0.01878025, -0.01965258, 0.0008723301], [0.0040223906, 0.0033900281, 0.00063236244]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 81% (1649 of 2030) |################    | Elapsed Time: 0:37:48 ETA:   0:08:21Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 81% (1650 of 2030) |################    | Elapsed Time: 0:37:50 ETA:   0:10:22"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1650\n",
            "Generator Loss :  [0.0150045585, 0.01123498, 1.266989, -0.00031388077]\n",
            "Discriminator Loss : [[-0.012478793, -0.014061361, 0.0015825676], [0.00079012243, 0.00032351175, 0.00046661068]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 83% (1699 of 2030) |################    | Elapsed Time: 0:38:57 ETA:   0:07:11Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 83% (1700 of 2030) |################    | Elapsed Time: 0:38:59 ETA:   0:09:02"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1700\n",
            "Generator Loss :  [0.015578728, 0.012205728, 1.1780655, -0.001611968]\n",
            "Discriminator Loss : [[-0.015076231, -0.017342728, 0.002266496], [0.003155603, 0.0013630773, 0.0017925256]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 86% (1749 of 2030) |#################   | Elapsed Time: 0:40:12 ETA:   0:06:05Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 86% (1750 of 2030) |#################   | Elapsed Time: 0:40:14 ETA:   0:07:40"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1750\n",
            "Generator Loss :  [0.015416231, 0.012078322, 1.0968618, 0.00047324222]\n",
            "Discriminator Loss : [[-0.016921146, -0.017385937, 0.0004647922], [0.003487303, 0.0033679965, 0.00011930641]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 88% (1799 of 2030) |#################   | Elapsed Time: 0:41:19 ETA:   0:05:09Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 88% (1800 of 2030) |#################   | Elapsed Time: 0:41:21 ETA:   0:06:21"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1800\n",
            "Generator Loss :  [0.016260069, 0.012795195, 0.9947866, 0.004805144]\n",
            "Discriminator Loss : [[-0.016857594, -0.017354477, 0.0004968827], [-0.0014035958, -0.0016243032, 0.00022070744]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 91% (1849 of 2030) |##################  | Elapsed Time: 0:42:30 ETA:   0:04:02Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 91% (1850 of 2030) |##################  | Elapsed Time: 0:42:32 ETA:   0:04:57"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1850\n",
            "Generator Loss :  [0.011938552, 0.0094055645, 0.8225373, 0.0006537597]\n",
            "Discriminator Loss : [[-0.015486382, -0.017647244, 0.0021608623], [0.004585133, 0.0025481316, 0.0020370015]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 93% (1899 of 2030) |##################  | Elapsed Time: 0:43:37 ETA:   0:02:54Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 93% (1900 of 2030) |##################  | Elapsed Time: 0:43:39 ETA:   0:03:34"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1900\n",
            "Generator Loss :  [0.015601295, 0.012326781, 1.1150701, -0.00070695794]\n",
            "Discriminator Loss : [[-0.015542313, -0.016244985, 0.00070267124], [0.0009815439, 0.00036324098, 0.0006183029]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 96% (1949 of 2030) |################### | Elapsed Time: 0:44:44 ETA:   0:01:47Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 96% (1950 of 2030) |################### | Elapsed Time: 0:44:46 ETA:   0:02:09"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1950\n",
            "Generator Loss :  [0.014276576, 0.010952537, 1.0939641, 0.00042146732]\n",
            "Discriminator Loss : [[-0.012329374, -0.012974851, 0.00064547616], [-0.00036845318, -0.00063840696, 0.00026995377]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 98% (1999 of 2030) |################### | Elapsed Time: 0:46:05 ETA:   0:00:40Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  2000\n",
            "Generator Loss :  [0.013267821, 0.010319109, 1.0317501, -0.0014653928]\n",
            "Discriminator Loss : [[-0.001176862, -0.015342735, 0.014165873], [0.0086066155, 0.0022030724, 0.006403543]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "100% (2030 of 2030) |####################| Elapsed Time: 0:46:52 Time:  0:46:52\n",
            "N/A% (0 of 2030) |                       | Elapsed Time: 0:00:00 ETA:  --:--:--"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "**********************EPOCH 9 *********************\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "  2% (49 of 2030) |                      | Elapsed Time: 0:01:03 ETA:   0:42:51Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "  2% (50 of 2030) |                      | Elapsed Time: 0:01:05 ETA:   0:52:55"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  50\n",
            "Generator Loss :  [0.012658259, 0.010112614, 0.83256483, 0.0004795032]\n",
            "Discriminator Loss : [[-0.017738098, -0.018153979, 0.00041587994], [0.0017768696, 0.0015484773, 0.00022839227]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "  4% (99 of 2030) |#                     | Elapsed Time: 0:02:11 ETA:   0:41:24Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "  4% (100 of 2030) |#                    | Elapsed Time: 0:02:12 ETA:   0:51:16"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  100\n",
            "Generator Loss :  [0.011633372, 0.0098941885, 0.7864417, -0.006201418]\n",
            "Discriminator Loss : [[-0.015961183, -0.01672779, 0.00076660665], [0.0065033934, 0.00629861, 0.00020478354]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "  7% (149 of 2030) |#                    | Elapsed Time: 0:03:23 ETA:   0:39:53Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "  7% (150 of 2030) |#                    | Elapsed Time: 0:03:25 ETA:   0:50:00"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  150\n",
            "Generator Loss :  [0.013146125, 0.010609982, 0.9876104, -0.0042668823]\n",
            "Discriminator Loss : [[-0.019323949, -0.020562854, 0.0012389058], [0.004862674, 0.004385512, 0.0004771623]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "  9% (199 of 2030) |##                   | Elapsed Time: 0:04:37 ETA:   0:38:58Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "  9% (200 of 2030) |##                   | Elapsed Time: 0:04:39 ETA:   0:47:57"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  200\n",
            "Generator Loss :  [0.012914924, 0.010522812, 0.91680783, -0.0035831134]\n",
            "Discriminator Loss : [[-0.018249853, -0.018728405, 0.00047855254], [0.0023863986, 0.002083502, 0.00030289654]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 12% (249 of 2030) |##                   | Elapsed Time: 0:05:41 ETA:   0:37:49Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 12% (250 of 2030) |##                   | Elapsed Time: 0:05:43 ETA:   0:47:18"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  250\n",
            "Generator Loss :  [0.014308217, 0.011896909, 0.8550869, -0.0015395358]\n",
            "Discriminator Loss : [[-0.016130276, -0.017874569, 0.0017442924], [0.004711994, 0.003256329, 0.0014556646]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 14% (299 of 2030) |###                  | Elapsed Time: 0:06:48 ETA:   0:36:58Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 14% (300 of 2030) |###                  | Elapsed Time: 0:06:50 ETA:   0:45:22"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  300\n",
            "Generator Loss :  [0.013433167, 0.0108561795, 0.9249741, -0.001979346]\n",
            "Discriminator Loss : [[-0.015647799, -0.01634307, 0.00069527194], [0.00093784893, 0.0007411043, 0.00019674463]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 17% (349 of 2030) |###                  | Elapsed Time: 0:07:53 ETA:   0:37:26Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 17% (350 of 2030) |###                  | Elapsed Time: 0:07:56 ETA:   1:23:06"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  350\n",
            "Generator Loss :  [0.014455415, 0.012225442, 0.9265394, -0.0054964446]\n",
            "Discriminator Loss : [[-0.017174475, -0.021784801, 0.004610327], [0.011485396, 0.006881317, 0.0046040793]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 19% (399 of 2030) |####                 | Elapsed Time: 0:09:19 ETA:   0:35:04Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 19% (400 of 2030) |####                 | Elapsed Time: 0:09:20 ETA:   0:43:31"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  400\n",
            "Generator Loss :  [0.013861274, 0.011076402, 1.0246787, -0.002891635]\n",
            "Discriminator Loss : [[-0.018632371, -0.02175557, 0.0031231986], [0.00911564, 0.005595374, 0.0035202666]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 22% (449 of 2030) |####                 | Elapsed Time: 0:10:25 ETA:   0:33:22Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 22% (450 of 2030) |####                 | Elapsed Time: 0:10:27 ETA:   0:42:02"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  450\n",
            "Generator Loss :  [0.012759056, 0.011066515, 0.67758423, -0.0034021167]\n",
            "Discriminator Loss : [[-0.023578865, -0.024480637, 0.0009017721], [0.0070567424, 0.00650333, 0.0005534123]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 24% (499 of 2030) |#####                | Elapsed Time: 0:11:42 ETA:   0:32:33Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  500\n",
            "Generator Loss :  [0.0188234, 0.016421413, 0.8880396, -0.0026213268]\n",
            "Discriminator Loss : [[-0.013543297, -0.016961362, 0.003418066], [0.0039828164, 0.0026621502, 0.0013206662]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 27% (549 of 2030) |#####                | Elapsed Time: 0:12:50 ETA:   0:31:57Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 27% (550 of 2030) |#####                | Elapsed Time: 0:12:52 ETA:   0:40:31"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  550\n",
            "Generator Loss :  [0.0138328485, 0.011241185, 0.9827222, -0.003565028]\n",
            "Discriminator Loss : [[-0.011711799, -0.01813791, 0.006426111], [0.0162607, 0.00338247, 0.01287823]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 29% (599 of 2030) |######               | Elapsed Time: 0:13:58 ETA:   0:30:54Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 29% (600 of 2030) |######               | Elapsed Time: 0:14:00 ETA:   0:37:53"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  600\n",
            "Generator Loss :  [0.012554388, 0.01025254, 0.9128772, -0.0043678377]\n",
            "Discriminator Loss : [[-0.009078432, -0.013520743, 0.004442311], [0.0077110445, 0.004035196, 0.0036758487]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 31% (649 of 2030) |######               | Elapsed Time: 0:15:22 ETA:   0:29:49Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 32% (650 of 2030) |######               | Elapsed Time: 0:15:24 ETA:   0:36:47"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  650\n",
            "Generator Loss :  [0.012495273, 0.0101186605, 0.79664016, -0.00013307223]\n",
            "Discriminator Loss : [[-0.0102209095, -0.011846118, 0.0016252086], [-0.000113606395, -0.0009103957, 0.0007967893]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 34% (699 of 2030) |#######              | Elapsed Time: 0:16:28 ETA:   0:28:36Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 34% (700 of 2030) |#######              | Elapsed Time: 0:16:30 ETA:   0:35:19"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  700\n",
            "Generator Loss :  [0.013660238, 0.010874954, 0.8696823, 0.0017623728]\n",
            "Discriminator Loss : [[-0.012622222, -0.0138334595, 0.0012112367], [0.0029239594, 0.0024082076, 0.0005157517]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 36% (749 of 2030) |#######              | Elapsed Time: 0:17:46 ETA:   0:27:53Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 36% (750 of 2030) |#######              | Elapsed Time: 0:17:48 ETA:   0:34:20"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  750\n",
            "Generator Loss :  [0.012661389, 0.009691835, 0.8655935, 0.0037277343]\n",
            "Discriminator Loss : [[-0.007267432, -0.008299345, 0.0010319133], [-0.0028623806, -0.0036345252, 0.00077214465]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 39% (799 of 2030) |########             | Elapsed Time: 0:18:51 ETA:   0:26:37Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 39% (800 of 2030) |########             | Elapsed Time: 0:18:53 ETA:   0:33:08"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  800\n",
            "Generator Loss :  [0.014085182, 0.011042334, 0.8222717, 0.0057603344]\n",
            "Discriminator Loss : [[-0.010242359, -0.01124478, 0.0010024218], [0.0011512091, 0.0004621742, 0.0006890349]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 41% (849 of 2030) |########             | Elapsed Time: 0:20:03 ETA:   0:25:24Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 41% (850 of 2030) |########             | Elapsed Time: 0:20:05 ETA:   0:31:17"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  850\n",
            "Generator Loss :  [0.014537545, 0.011974891, 0.81763947, 0.0010973597]\n",
            "Discriminator Loss : [[-0.0067331227, -0.008031834, 0.001298711], [-0.0016125768, -0.0019725435, 0.00035996674]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 44% (899 of 2030) |#########            | Elapsed Time: 0:21:14 ETA:   0:24:15Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 44% (900 of 2030) |#########            | Elapsed Time: 0:21:16 ETA:   0:30:58"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  900\n",
            "Generator Loss :  [0.017131073, 0.014396079, 0.8926934, 0.0005691561]\n",
            "Discriminator Loss : [[-0.012996706, -0.013725949, 0.0007292435], [0.0030015549, 0.0025637716, 0.00043778325]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 46% (949 of 2030) |#########            | Elapsed Time: 0:22:20 ETA:   0:23:13Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 46% (950 of 2030) |#########            | Elapsed Time: 0:22:21 ETA:   0:28:26"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  950\n",
            "Generator Loss :  [0.012831524, 0.009776194, 1.0176475, 2.3876222e-05]\n",
            "Discriminator Loss : [[-0.012005934, -0.012976503, 0.00097056956], [0.0019515122, 0.0015257258, 0.0004257863]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 49% (999 of 2030) |##########           | Elapsed Time: 0:23:38 ETA:   0:22:31Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1000\n",
            "Generator Loss :  [0.013415623, 0.011102529, 0.71640193, 0.0016388881]\n",
            "Discriminator Loss : [[-0.009296252, -0.010022484, 0.00072623126], [-0.0012121644, -0.0013668218, 0.00015465743]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 51% (1049 of 2030) |##########          | Elapsed Time: 0:24:55 ETA:   0:20:43Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 51% (1050 of 2030) |##########          | Elapsed Time: 0:24:57 ETA:   0:25:45"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1050\n",
            "Generator Loss :  [0.013015114, 0.009642053, 0.9855446, 0.0041642636]\n",
            "Discriminator Loss : [[-0.009176671, -0.01093822, 0.0017615489], [-0.00387232, -0.0043590334, 0.00048671343]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 54% (1099 of 2030) |##########          | Elapsed Time: 0:26:03 ETA:   0:24:37Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 54% (1100 of 2030) |##########          | Elapsed Time: 0:26:10 ETA:   1:42:44"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1100\n",
            "Generator Loss :  [0.012083968, 0.008922463, 0.8546604, 0.005975238]\n",
            "Discriminator Loss : [[-0.006517751, -0.0073042447, 0.0007864934], [-0.00343031, -0.0036798758, 0.00024956584]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 56% (1149 of 2030) |###########         | Elapsed Time: 0:27:13 ETA:   0:18:37Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 56% (1150 of 2030) |###########         | Elapsed Time: 0:27:14 ETA:   0:22:58"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1150\n",
            "Generator Loss :  [0.013135011, 0.009947769, 0.8308852, 0.0069458704]\n",
            "Discriminator Loss : [[-0.009412443, -0.01048539, 0.0010729474], [-0.0014915362, -0.0032387923, 0.0017472561]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 59% (1199 of 2030) |###########         | Elapsed Time: 0:28:17 ETA:   0:17:34Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 59% (1200 of 2030) |###########         | Elapsed Time: 0:28:19 ETA:   0:21:54"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1200\n",
            "Generator Loss :  [0.012037373, 0.009315488, 0.76093066, 0.0043909345]\n",
            "Discriminator Loss : [[-0.0068142232, -0.0072416263, 0.0004274029], [-0.0034340934, -0.0036614062, 0.00022731282]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 61% (1249 of 2030) |############        | Elapsed Time: 0:29:21 ETA:   0:16:27Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 61% (1250 of 2030) |############        | Elapsed Time: 0:29:23 ETA:   0:20:22"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1250\n",
            "Generator Loss :  [0.013661795, 0.010589934, 0.99753344, 0.0007926106]\n",
            "Discriminator Loss : [[-0.0072505944, -0.008029885, 0.0007792903], [-0.0032274267, -0.0037715563, 0.00054412975]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 63% (1299 of 2030) |############        | Elapsed Time: 0:30:25 ETA:   0:15:20Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 64% (1300 of 2030) |############        | Elapsed Time: 0:30:27 ETA:   0:19:07"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1300\n",
            "Generator Loss :  [0.0131380595, 0.009969814, 0.9869485, 0.0020740028]\n",
            "Discriminator Loss : [[-0.0112208165, -0.011763998, 0.00054318213], [-0.0023033577, -0.0025323152, 0.00022895752]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 66% (1349 of 2030) |#############       | Elapsed Time: 0:31:35 ETA:   0:24:30Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 66% (1350 of 2030) |#############       | Elapsed Time: 0:31:38 ETA:   0:32:34"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1350\n",
            "Generator Loss :  [0.015103264, 0.012376726, 0.8416842, 0.0020148526]\n",
            "Discriminator Loss : [[-0.0065246094, -0.007264753, 0.0007401435], [0.000624939, 0.00036197033, 0.00026296865]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 68% (1399 of 2030) |#############       | Elapsed Time: 0:32:43 ETA:   0:13:38Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 68% (1400 of 2030) |#############       | Elapsed Time: 0:32:45 ETA:   0:17:04"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1400\n",
            "Generator Loss :  [0.014114481, 0.011662026, 0.8114699, 0.00018045839]\n",
            "Discriminator Loss : [[-0.010208441, -0.010995483, 0.0007870422], [0.00068898697, 9.2903116e-05, 0.0005960839]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 71% (1449 of 2030) |##############      | Elapsed Time: 0:33:59 ETA:   0:12:46Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 71% (1450 of 2030) |##############      | Elapsed Time: 0:34:01 ETA:   0:15:55"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1450\n",
            "Generator Loss :  [0.011999053, 0.009480938, 0.8237087, 0.0004698828]\n",
            "Discriminator Loss : [[-0.009745534, -0.010494976, 0.0007494425], [-0.0013956794, -0.0016648477, 0.0002691683]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 73% (1499 of 2030) |##############      | Elapsed Time: 0:35:05 ETA:   0:11:38Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1500\n",
            "Generator Loss :  [0.014230052, 0.011478123, 0.90389365, 0.00040247638]\n",
            "Discriminator Loss : [[-0.013175874, -0.013668434, 0.00049256], [0.0028920148, 0.002635782, 0.00025623298]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 76% (1549 of 2030) |###############     | Elapsed Time: 0:36:26 ETA:   0:10:41Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 76% (1550 of 2030) |###############     | Elapsed Time: 0:36:28 ETA:   0:16:10"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1550\n",
            "Generator Loss :  [0.013334282, 0.010908481, 0.7767186, 0.00095644925]\n",
            "Discriminator Loss : [[-0.0097018145, -0.011246323, 0.0015445086], [0.0016579798, 0.00051347044, 0.0011445093]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 78% (1599 of 2030) |###############     | Elapsed Time: 0:37:36 ETA:   0:09:31Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 78% (1600 of 2030) |###############     | Elapsed Time: 0:37:38 ETA:   0:14:29"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1600\n",
            "Generator Loss :  [0.013327208, 0.010089379, 0.95572454, 0.0037065612]\n",
            "Discriminator Loss : [[-0.0060849916, -0.0067668897, 0.00068189786], [-0.001325323, -0.0017234574, 0.0003981343]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 81% (1649 of 2030) |################    | Elapsed Time: 0:38:58 ETA:   0:08:31Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 81% (1650 of 2030) |################    | Elapsed Time: 0:39:00 ETA:   0:12:47"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1650\n",
            "Generator Loss :  [0.0127814105, 0.009968754, 0.81154823, 0.003780114]\n",
            "Discriminator Loss : [[-0.012422684, -0.012660061, 0.00023737781], [-0.0023449396, -0.0025196823, 0.0001747426]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 83% (1699 of 2030) |################    | Elapsed Time: 0:40:09 ETA:   0:07:20Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 83% (1700 of 2030) |################    | Elapsed Time: 0:40:12 ETA:   0:18:29"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1700\n",
            "Generator Loss :  [0.013035342, 0.010385593, 0.83399755, 0.0014775604]\n",
            "Discriminator Loss : [[-0.011175394, -0.013743666, 0.0025682715], [-0.00033849198, -0.0015245449, 0.0011860529]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 86% (1749 of 2030) |#################   | Elapsed Time: 0:41:26 ETA:   0:06:24Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 86% (1750 of 2030) |#################   | Elapsed Time: 0:41:28 ETA:   0:09:36"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1750\n",
            "Generator Loss :  [0.015464947, 0.01282095, 0.87090176, 0.0003129147]\n",
            "Discriminator Loss : [[-0.015089774, -0.015740734, 0.0006509599], [0.005298137, 0.005038904, 0.00025923306]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 88% (1799 of 2030) |#################   | Elapsed Time: 0:42:39 ETA:   0:05:15Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 88% (1800 of 2030) |#################   | Elapsed Time: 0:42:41 ETA:   0:07:53"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1800\n",
            "Generator Loss :  [0.013341426, 0.0106408885, 0.84511334, 0.0016519689]\n",
            "Discriminator Loss : [[-0.01075167, -0.011868993, 0.0011173226], [-0.00094026874, -0.0014039659, 0.00046369713]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 91% (1849 of 2030) |##################  | Elapsed Time: 0:43:47 ETA:   0:04:07Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 91% (1850 of 2030) |##################  | Elapsed Time: 0:43:49 ETA:   0:06:11"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1850\n",
            "Generator Loss :  [0.013380312, 0.011120022, 0.7433156, 0.00030343197]\n",
            "Discriminator Loss : [[-0.0132282255, -0.013691873, 0.0004636474], [0.003017532, 0.0028354737, 0.00018205821]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 93% (1899 of 2030) |##################  | Elapsed Time: 0:45:20 ETA:   0:02:58Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 93% (1900 of 2030) |##################  | Elapsed Time: 0:45:22 ETA:   0:04:27"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1900\n",
            "Generator Loss :  [0.013987302, 0.011750968, 0.75405353, -0.00025826952]\n",
            "Discriminator Loss : [[-0.012618747, -0.013671247, 0.0010525004], [0.0010188431, -0.0004926067, 0.0015114499]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 96% (1949 of 2030) |################### | Elapsed Time: 0:46:28 ETA:   0:01:49Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            " 96% (1950 of 2030) |################### | Elapsed Time: 0:46:30 ETA:   0:02:42"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  1950\n",
            "Generator Loss :  [0.010777302, 0.008445001, 0.7539912, 0.00070326973]\n",
            "Discriminator Loss : [[-0.02007364, -0.02092741, 0.00085377216], [0.0038315565, 0.0035213449, 0.0003102118]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            " 98% (1999 of 2030) |################### | Elapsed Time: 0:47:40 ETA:   0:00:41Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  2000\n",
            "Generator Loss :  [0.013018031, 0.009973064, 0.97684705, 0.0011442531]\n",
            "Discriminator Loss : [[-0.013550676, -0.014001279, 0.00045060285], [0.0032439486, 0.0022264447, 0.0010175039]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "100% (2030 of 2030) |####################| Elapsed Time: 0:48:28 Time:  0:48:28\n",
            "N/A% (0 of 2030) |                       | Elapsed Time: 0:00:00 ETA:  --:--:--"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "**********************EPOCH 10 *********************\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "  2% (49 of 2030) |                      | Elapsed Time: 0:01:05 ETA:   0:43:49Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "  2% (50 of 2030) |                      | Elapsed Time: 0:01:07 ETA:   1:07:00"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  50\n",
            "Generator Loss :  [0.012008508, 0.00991034, 0.6556629, 0.001311784]\n",
            "Discriminator Loss : [[-0.012536815, -0.013390928, 0.0008541121], [-1.663604e-05, -0.0001805663, 0.00016393026]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "  4% (99 of 2030) |#                     | Elapsed Time: 0:02:12 ETA:   0:42:42Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "  4% (100 of 2030) |#                    | Elapsed Time: 0:02:14 ETA:   1:04:59"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  100\n",
            "Generator Loss :  [0.011734763, 0.010121571, 0.67020375, -0.0039741923]\n",
            "Discriminator Loss : [[-0.014598838, -0.0151987, 0.0005998621], [0.0014962709, 0.0013166823, 0.00017958855]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "  7% (149 of 2030) |#                    | Elapsed Time: 0:03:33 ETA:   0:41:34Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "  7% (150 of 2030) |#                    | Elapsed Time: 0:03:35 ETA:   1:04:15"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Step Number :  150\n",
            "Generator Loss :  [0.011645895, 0.0100544365, 0.687848, -0.004720843]\n",
            "Discriminator Loss : [[-0.014349607, -0.015714953, 0.0013653457], [0.0055661565, 0.0048090583, 0.0007570984]]\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "  7% (155 of 2030) |#                    | Elapsed Time: 0:03:41 ETA:   0:41:22"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vcneD1RfzVYT"
      },
      "source": [
        "directory = \"/content/gdrive/My Drive/val_256/\"\n",
        "filenames = os.listdir(directory)\n",
        "data = DataGenerator(features_new,directory)\n",
        "gray,color,_ = data.generate_batch(3,filenames)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4XfnTzNWcEVI",
        "outputId": "566dd359-dda2-41b6-8301-853e8405e5c1",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "p,_ = generator.predict(gray)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:From /tensorflow-1.15.2/python3.6/keras/backend/tensorflow_backend.py:422: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TKrhu1tJdjuM",
        "outputId": "fd31e496-ed4f-4a22-cda2-4f120a00d1da",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        }
      },
      "source": [
        "plt.imshow(p[8])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7f90cc4f48d0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 30
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAQEAAAD8CAYAAAB3lxGOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOy9S8xty3bf9RtVc61vP859276OE0NIRIxAUYywQiNpEFlB0CGkEyUSARGUpJMGEg2MGwjhToQSIjpEOCISSDwliIiiCIgiaNBBJlEEIU9j+Ypc29e+17nnnnv23t+as8agMcaoqjnX+r69z8vexrvO+fZaaz5q1qzHGP/xLDEz3pV35V359VvKr3YD3pV35V351S3viMC78q78Oi/viMC78q78Oi/viMC78q78Oi/viMC78q78Oi/viMC78q78Oi+fGREQkX9BRP6uiPy0iPzYZ/Wcd+VdeVc+WZHPwk9ARCrw94DfC/wD4KeAP2Rmf+tTf9i78q68K5+ofFZI4HcCP21mP2NmF+C/Bn7fZ/Ssd+VdeVc+QVk+o3p/I/D/Tr//AfDPPnTx6clTe/Le53miTxDgc3cLiPGqvkKARRbMQK2AGHJaEYWlCSagIhhgAAKS3xEkPxPw5BfzM9c4yLh14hZekukJu4P9q+x+I3K49Pg7n+KfDtLGk9VuVhOvItPxwwV5c6/PUDMMEBvP2r9jnN+1iIe65/qCR8rrrrDXdL492Ij94au2P1BEsq+PPWdXR66fdOMpt78ebh3zb1fD/DgD1MdIDGQRKFBKRaSgJhiCFQWMrSmGcXqlYLDJBQTKvVf3Sy/e/6aZfe+xKZ8VEXhtEZE/BvwxgLvnn+Of+Zf/FX7bd3+IKvCjv/WLNGn8/S/+HaoUvq98iWaFD9bnyPkV5Xu+wdN74Xver9zXyoenShNjE0WqIEUwCkbhZBUxoVgQB4kvzUFQy15vPlxqLVoYi0P9lx5WoEhBgNM0aiJCKYAUTIRaKrVOYKv6IGIFQziVGgTL6y4xE8VWAC5tA6zDtVebf5Z4GSGIYPFnL6cYTimYWScaGLApXJSiG2KNl9vKRRvFFMG4XJwwZGkoNrVJD5+tL0R//7YdCFj8W0iyM0r+7pPfYhGKd3ZT9afnI1oSpHi2KZgh8WyNQcpnqCQLyPonopp0eSbWnQg4WRbLfwyZSLUgqI5GC4Z2NrOfLxad389HJdrPy3ReYh4IJsHCxJDV4N443RvLxShfKZT3Cu89/RKn0x3vb2c2K+jTl1CUX3z/A5oq3//3XiDN+Mb55wDj+dd8lv/HP/WXvsaN8lkRga8DPzj9/k1xrBcz+0ngJwF+4Ad/k/2ef/J7+eZXV8SEv/VzZ9Qa3/juFzmb8eXtJa/OZ772/U+xu0L7ynM+9+LEZXuKrYq9bFQ2KhulCqWCngwt2lncGoOxlj1GsJi8GqNj5iutL4gkAjFBS3aZFoqABTEoGKUUymkBM4pWiigiglMgQ6SADI69LAURYcUJgTVBTFjj2S+KE4Nz0KWtvgKgUgHhJAURKNXf5q44kbHafDKViilog1YaGyuNBrZxaStr25z4GVxkxVSdSAnYFouqT3r/ksihxfFciFaCQpkwrd9+e646y0vEgpP5IjXMnyHGlteRi7xFDV7bphYEKshjH7s9UeglUM+g4zu4NrUtiAATmij71lsgqljDGBPTEGD1Z6v48ZaEvXejn6/qFElLdLhUhGBWYlQUOxU4F+yLT6Dc8Z1y5p6Fb38gLNvGt97/Jpd15YdO3+WJKF//rV9hPVW2f+o7FAz58LdjCO996+d4rHxWOoGfAv5xEfnHROQM/EHgLz50cZ8eGn85BNX/jJwk/ic6UH2WPbcR9qd3mLxPrnHusd+36hnnfUITpH4634/Nt0v8zVXM107f++ScntXbPi2oXof/zceMQ5umyT4+j3/0unf3XbX9kbbJ4W8vI1314a5/ZhFK5uO7Cw/v9PoxTH57Jb71K2+151ZLB+fe3T0PnczXxTOn936sC7J00aCM1xlTI5hKVhEPFXMiLZQgzup/hUdX+meCBMxsE5E/AfxP+FL+82b2fz90vbSNJ9/5ZdZfUkB4/+mXsbOhv/c528uNb/0v34S18j1/v9Hu4MX7laUW7p8IG4WLGMKCmHDHwlkWBIe5i7o8pVYxxGGu4BwROJUGGNYqMCh4wrUcKC3eNk4lDgsNQcy5e6uClMp6WpBSkVI5UVgotAKtCOcCZZopxZz6t+KLt56Di6zn/mkwkAEXAO6aD9v92T/LWXE8cgeimKwucmwLxaCqgK0YK80UVeXCyqU0FnORoBUXABZAdIgDi/mkXWu+d3C4JBMhRm0nP1469LYxYWEnqHdxwBwkaci0mHZCb+Z1OVTfAn4nuFa/N2Z2C3RX4myNh2m0Lbm5qXd9yTG1/J2IJIamTHJDvP+gOwn7E+6nCBdIKdqyBXqxQJBsgWLM51GRPRKwIHRqBSuGlkJ5dqK+d+IkT7grd9RXK219xXnZqFtDXpxY7Uz9ylexpfCVX/4azVZ+YPuNFIRvfeFvYBgvfvgHvA1/lZvlM9MJmNlfBv7ym1wrOKUoJwdkp6pYMaoKVYWTFcwKLmtPhM0mvh4TJ35MDcmfE/+3iQjbfpJKv+QG1BDbIRAJ3Lt7pI36U/3Qv1vwAZFxDX1e9WeXaFben3L18fr+XvHiVwvODm3AYpFacAwb72DBWUIJlf3S26YW6yKfkef38vp8fNcYO/LOuCbaaHPfWZ4bbdhZsu2Bz+MFU78a19ce0eRV/VNb+wDcasetex86dzx+9duRbnAuJ7IlbtS8XjExDHVW15y6iYJQ2LQEQfQ5YVdIal9+1RSDczmr8IMvF17+7nuKCf/Ed7+OSuHrP/+bOL9q/LZXL1jLiV98/n2004WX5duYCdt9YynKk7LxYjU+vDeWZXMioi4zLNuCWOFkLmlKaRjmspoNrmEp5AZFb+2KKgBCWQL+lZDHRREp1FIoBUqFWmGpcBY4iaGX4PSn4lwpWNOpOFR8GoN0Up9lqypmcL7ko2MxrP5Zm+sKRF37y31BpLiiUaAszl2aQRWjloZuG9u2uYJAG3W755WuiAb33RqqxrIpYkbTQAL4fFyC6ybH63rA4Ihb10Lq9O+8iBI2z1DeiZGGqKdd3NNuvQDQTYNY+SqwbsqIOpPAJNEo3tayjd/C4PQ12pDIQLoY5Ry6t/Gm7sCirgPlSt1IIqGWTQqEGboEizFu2eT4LIFySsq6ReG7K+2XCuXcOJ0aW2s0NdYvfkB5duH9X3Jl8Re/tiJWOX/+PawYP/3iBzArLF/42yDK+6czj5W3gggYxoZSrDg8bA5GRRQ25VJhFbggKMVRoxCTwlEDloNhzA5QGlylM/wYu673s6CWnVP5BLLUMOdgdVmuBFst0/Hks/1iN2kyEW9GE3NC5fFOZwJp2HzQTwzZPxsv0+8ud0+TNB6YuhTri8cXU5/E2cD87F05GmC7/hnv2N+pP2t02BiCwZVzsY06ZLpXeh1DZzHG0ab2cBjDh5hr3pL9kaZUiwNdlyMy3SPZ2Lml168e4/ggGJBDf4XSwAmWgGR/lek669flmEpvfYs/RduGtYZRoLgIqwrWFszElYoCqj5XX6f4eyuIwCqNXzh9m+evfgtixoe/8A2MypPL59Fi/M3vX9jaiW9+9zl3+oqvfLggtlH0QziB3QFNqapoq6xUVAoqhRXvz7KpT37xxZqcahGn1O2w+CwGp9R6+O1d5uvGMJrLlBIyqlVU3ZSkxdhC5iwC0gCTPhFMfIiX6tO+hbmum5GiMWso3tvlzJgQUEuNlygIxdsmgXZUwe4xU7ZNWbeVy7ZS2kZpG3bZYG00FcygteiDkIFNHSPdh6msaC4K7weN34kIEjnlwrUuI/lqEYtFkGsiFqFhmJbQIfjJFkioxf0JVpxSCa3p1AuD+yYtrDFWWmMZpbyNo7e02hpuqclF4r8NxCZ9nE2Ee8Lz8R4akyafUezkp2NlpZ5C9IQBJawHrvAGrf4lkYo1RUQpRZFNKE0odsHkQ4QXVC68/8svuWjj1flztLuF7778Jq8M7Ju/HezMl7/3Q0SMn/nuFzCMr24Ha8mhvBVEwAT0DGV9CWZsesIoWN1cSbJcUKDKPUXuKfIKoVFlQ8Wcg4pRJL0DzHF5LCMRQWRz6it7RU+flBqHQmuU1qeSv4NAl5pTTzAMNR2EuxgWNnub/4qhIkj1+n2OClp8EmqCCxyKa8mFFMql+E1Ju1xzDrH4jbJEa6r3gxXvE0VBDSlKK0oTxeJPxZ+TKN4kJ3iouuJZ0pUCB9bfDd8cjnee2j93WvmDeJqabr8/nok7u5RoixQNNBemwpKIab8gs0klOHRfgMlVU6Mu0z3xeyA6i3eeeHwnANNnZ/TzcRtt6LfmtxCT+lgOIoS42G/4mKTyUoKIUf1PrbLJQtU7TtI48RRhgcVF3YU1UJL3k0rxOrd67PZdeTuIwAku39+wb/wtsMIvv/gRpBjPvud9Wmm8+Nz7yOXE8w8a5/KKZ6efRyhUKltp3MvGSQpI5W45cSoVKyegdi4ioYGWU9plvfdrDFJ2hCwh+5aKr7Mk6XFcUk3nHOiybj5oJWBlLe7BWASVwiYFqfFXKog4uhBhSyQQRKCEN8N2aWCg4rL/ugYxWILjLSHbPn/mE2g54cii4kv/wtaUV5eVZPP3ZeVeLsi6UmRjVXeqsQKY0EJXImGlHyJKfuYRhyWi0R8h6xZNwjUD5PF9Z5v3I10gSlFFQiYp6kRB1cU4W9060DR0IU1D3Mr6gxLFeluC1WvqIZLopk9H/E4GL6WzC/pijr9ptXtb0/Myp8W0iGEgy3L4RH1M2ynqcuUQ2rG6v+/WdEJM8eVuoZ0qH9rCS5QvXM7ctcKz7XvZ7A6+/DOo3vOV+i0qwrc/cA/b++UpJnD5+vOHlaC8JURATKiXhUU/h688Cw5ZfRKuhjXQKrRS2OpCoSJ2prE5pNVKsQVYYlRPBLkNBhccXEtMnZT9kyhEW3JUJPAaTgzSw3DYmwwzxTa3WDh+FFCHhikipDNPN2kIaLW8NKoekh9AC0VizPn+mY5NOU11df5lGp+xiAy8v1Zcf9KgbbCtUDbQTWibucOPOttUlZC7vS1NOhiO947PUHRB3hdtvEkEsrUzETgSg1joDCSQDjUaegs1CaLg6CsdcIaHYzYijpdJZBHCEVGwsC6J7nUARceyTxWh7YjAgA6mB3frwRN87MKRq19P3fVb65jD56FphY48og8s+6RgVpGlIItim6BNaVbZpNCKoTS24tdf5BVVhO20YgaLPHExVTYeK28FESit8PwffoEnpy8BYM9cdv/w/BTVDf2OS9+XZ0KThfPp81Q7c2rP2NYL969ecuLMnZ0dcotAqy5/bw6JLSCBqk8Es6DEMVip9VcNCh6fiQD65E13Y1XUlLb6Ct0CCooU9/gS8xli7sHonoOuji4B+Yg2pPZaQim1rQeX5ku0LTXOd+pLdbVAEg6XpQ1noVXhsplrk138537zBYC65cCtfgVB0M0VrkXCHbbEZO5a7+yDtH9Hm6bz++V/Az7bXizIxZ/LjQmG+wJJ+t187DZHK9vRrJsegiV8GHQ/ZinSmLiHppUkcEF8u57IV7SYhQnOOsfOsUrlahdFuuwfn2n5SQVpIEdJr8a4r3b65URAaO45qfmpbFtl2xbOorTFWL8D20t48XnhcgeX04rJyrfvlbYZS/lFqigv3vsahvK8/g6gMHuu3ypvBREQKdTTE87bUwDMXmEC26vqDiynM6YLdikspVLt5I42toAotpyAM3DnclQxp7zmC88nQcrXsVCkBJ1P+J/QMIhAKt26ljjlSuK+5B2avDf+d+Ve8CTnLVaCYyd3SA11KBvJtqTLsj86xYGWnE50tEXotv3SNf2pSms0bTRdB7GylWYXsIawsWGomDs7JRAXqKmHyPfti3ZAZRhiQlcJXI3qUTcgdI14H/iuRuz1976boJEEpyMVfV2RkdWkIJ4iW6KVbLr3bie23fEp/gvUU8hAs0ADKSskMTQYFoz4t+T8oNfpbZmRj5B4XCdHakcnPl9c/6GotVB0Rp9IQ1GPl2BSMF8M6uZTvDlT23TBMLQ+x1BaWwBBn2tv363ydhCBUjk/+xKf/8Uv47LR1zEpnNsd21J58f3PsXXh9MGJWgun8gRsAT27HHxesPIEypNQEKpPCMOVa2qUGmwlIXlwn2KJBKItoRNIIpCdV8hJ4FwXcWuDyMWVeaaDo2GIuf+AUw2H2p3zaegPovba15nL41tA9K1uIB0IQPgmNDn5O2h4MUbbijYwxdjY2sbl8pJmyqbKRVfu2wWhgTU2jIY5MTSwEqilS0au5a5J/vrCSt/4GSEwqMJxbJl5ux+ZHYlc4TpFQOaaiTvEzMVDAlgRaAa6157YpMfJsbNECNJFNQk/ihHUFWMvZYhCNp7la7ebEhg2Xjoa0d1iH685pMZ8e783jQOpfHRCbCy6+cK1DTELRWHDysJmG601MPdAlVebM7/qTkN1E8Qq9/YUwdiefB+IsX5w5239anv7iYABTRofPPdFUNXtodR75+RbQZpwJxeqNE4lFF4sGMqGsrBSRKiSUXcB39hcTjUNE2ws5qTZSXQDandkIBm0khM2hL/NoaltF69zXX2A1RLxkiDXV1TpKCJf1kKJOL+/PyyDdGJyBqxN6Jjidi3erqLDNVUA3YI76UprG3q/oaFg02bhwGJ95XY+L2BhIxWZXXXpXDLblq3tMv4BAgzP+OTWe2/9bnLLazoBDCQgiUIIWR1XPBjhnWjUkMsTMvWA8UQCHUoFEQhCXIKylBQP4s4iTsC7QxPmwDEJFDJiWlIKCBTTTaf9kdrHaLQBUEd5ZVt7kzAwVcdhbQVVrA00qFZoWhBZQTfssrjZUBoF4yTuYqTm+OLupff3cl4wMU4vfOzq07Ro3S5vBREAo9WVb33pFWA8byti8OzVh+6bs1ZKg6fygiLGua4Ofc05tJqwyIUqGwvuguwKmUITp7DphVMCxCcCSG2vpIdg6gSmtvmHC7HWXFOtlxUzhXZxN9c2OF7+J7VArZPjTIgFyeryCQFBB7cNuXEJNFJzIvnHKR2ZIuQ47zBVzBTd3DrQ7i80FVpzhZKqTu7ChULpdnEnAkYprp1O5dzRe66UsYB6v0xtzuvTnbh75uQ1lm7e7Ex14xHDLcsyIjGUA53pzp5fDPHIYuwkZDZtKYe7cjcXf2EggW7ZMbo33+DbMlzUsz+yCaH47UQsCE/qiHo/xX1lc2ZRw+lDxPULPp8UaRcfv7Z1lNjM2AxkvUBdKbogWimmVIPangKV5c51GKdvnxAV9MkZCtTvOI7bvjDmya3ylhABQdoCzwqY0e6fIgiX9566n9RJ3Y7GCamKPLl3Z5bt5JzOQJdCrWlZNrCKBSU1c+WgE3gfte4VHF/aYXIPBpcLLgfNHVtaQMjWlphAcaWlUGBQq5sDA3nk3z6KcVo4IX/W8GCTUzorJIGK6+N4j0kPBtWaK86sGq0pjXC0cSkhXjphb/UpnsEF6aAUisEeQFX2be927t47DxCB5NZ9veaCtelO2RGC0lFUKOx6WLL166Mydh2S6O2Umvj4C3MbGa4bplkJvVB3/hBCXMuOCGWldUznzhy4Q5OLC+GKfAwoynmUgURbQkz/XNMC4jyFJu5IVcSZ2RptFYTNQnyoBZZKaQuUiprHgBS5A61ocd+Q8mQBhVbckzCt2VsEyT1U3g4iYFDaCZ44JGvbM0SEy5On5NwVKnJ+5lqQ00s3ba0VT57R0KVitaLhDIN60JGqw6VUXyv+NVFaam1z8HTPdAcMbuEctIb5rkTiDj05Mgi4KqFhd//vBSndA6GzvnSBzU8JK2Re2xfS+RpaA11v0dLyEwRojSQkqk4QVDS+m/vJt9lxx2MqhiifyzPIaJjN0jjShaz8fWxa/MyIvLkjDcIPYL7SwpQKw1U2UUIQwd1rJwEbLR2egukgEJ+JflMjny7etYaVRvpzAh76M3UiAjbarcTYmqO7tAINeZIee9KJZJxbu/eQI9dLd+TxuhvF6XAR1AprWqIRNjUPPlyq60VKBV3YzEdj4Q6RE1IbCGxPnWiu1SeUiyTmuqVHyltBBKQIp7vKU/U8PYutOEd0zv6eGRXjiTXENpZ2wZpg94KJ+0dV3VhKiUAMBfOY6q1tASdzinvJgI6RaSKJQMh0HdzHMGfgRwTvaGtBgLqPbyc0RYwiUM0o5mG+3fmDicPFBEkhYNHAKTHBZJsWnjEcdMKKkGNrLU1XIaq0jaINbSuqnsCiqbtVF2neP7TeXu+X2XWni+FDLEruG9GEhxXa12j3eEhT2o5I2A7K+2KSrkRLj74u5fcxG2OBDGA00Ecs8rxbk0julZgSodsZQOQiSepAwKxg4Z8fzR+LP2ZCTXExhP4RMBQEZ92indH2NYKutwtmxraFw1O2ePPnlXbxOekcz7tYHVGJeTYo0eKqsvR8DHGwsiEYp1NIvXoPYujdE3/P7deATkBEWJZ09jGq1QStGPAUdwU+0xBrFF1dybW6XLYJ1OJcpYTnmWj4q7fmRCBKB5IRJWcZc558MCLVkgi0rjvIyR2x7d2hZRCBFAeKQC3e5jpN9l40iQA+qZM3p2JQEhHkhIt+6vHxMUnT/2lN2BqEafNJ09rWUU9RRUyphMm0E8ZwDsp8dbHkeg7CcLTpaz5dew9iwYgS7B3sh7t2nEEoCQ/BMpR1PvGjr3K1x1ho3hp1dxt7D/IKsSgchUtC980XlMaqSq+QmnZe6O7m3VlKmvdBdo8meXTFctEQKUKU6NnoQkmra3J6f0BNb8/1ggF1Xf11swnq86m0jaKum+n8wswtW6IIDdEWqMFCgmsgQmF1RWHEoGz2oc+Ds0+yoa+6XT42ERCRHwT+c+Cr+DN+0sz+IxH594A/CvxSXPrjkVvg0WLRoZ4sIlWzoRAyjzB0OdNYmzkjw1NzbXgYahGcYlojdYEbzd1jdUxGh7vdur8rGoNZk5KHp1e3/Iub7Zol/Iv2R2U5wEUqRQpFCjVgby6EEc3m77OQiEB2zy5BFLpDk7mMmy66rdM2r2cTF/4bK2rKJUJwGk53moQMeuS4uK7EgUEgnv3okN5zff0d9Bp5osTkld6q6XSYw4yI98h6zIlFSgVJV5PwNNmP0xGddEV8/K4Z7kzM/vTtDRCTeSFTRFnEYztcclZMGk0CJEq4m8frSpgbS4hVSTRLUIPLzqpk8dsQuccMtkjD1lwR4W2MmWx4QFZKSJsZqxklEr+4coeeL9N/C6I+x1UDh9VkCd6JB++Mq/JJkMAG/Ftm9tdF5HPAXxORvxLn/oyZ/amPXGMq10j7d1D6DseSOHpoq9Nn63Cxs8egmBhsbHGNnx/ATqdJ3x9Fs5kIDCXeIALOxTRMRJaaZXxxq7lcXIq7hrrZMohAcLyORqaOdA7lvzOCbTZlQYgoRk+oku6zSdpaOC6pKQ2lSShNGYRAxdsnu2UsXbmYuGgHwckFaSM8t5+4ltMlfs9Etk/DWXrItRkXZnvytfIZQ6INYniY1N0jkLm/jG133YgOqOaBNRaMQ8XdeZvhAVYoW/RZEoH+nsnhJZ+YyNFHc80Iz3iJFqpoofn8CCKhFig2wuGzx1N3W8yZ25ZKwHyWCTXmmETHqTq6S+uI1QwgG+9+pNlz+dhEwMx+Hvj5+P6BiPxtPNX4R68LV+615lk0fCEKhJsotromfg2+ppuP3FZQcacXF1WNTJpR1CGsxkKe4/Sd24RCJsNOc/ofXE6LLr2N/m8NhXoQBR1TzzmCK98yGQykC+rEGZsNYgakhj7bV8LuX1MHIMn5gyAFm25rKvEkkICvqLZ5OurVH+99YNY90RR3krLOUQgolisS0ptxmPxy9cbbJhTPMUzuXeJdd+7GSfBsWvFpBRgcOnUUFgFC2SGp2x79cCACS1KmbKN/rpcQAzJdWPHFcxJXKBuekFUivFtwLb2Iy+LVwCMXnQiIgEU2kBq6Ed1i3FuaAg8+uu3s/awLIygJao02LBErICumg6GZ2BBF6ga1oVZRSs82nRS01A0RRcQD0LQEpF5fgQn6mlX+qegEROQ3A/808L8Dvwv4EyLyrwL/B44W/uGjFRiYacjuU+rrEMwsYI9/Njp7s4wm85xzEolKxVxDLkr4hfd1Nv5JR5PU8uasSyLQNchlNJKZA04oJcOZzYlJcvK+ss167DyGiyJhUfBHhown6XAUEysN40Mgj4+wVGTG2iBEbvkQt4gAHoAyrg9NV69qOE4ZXUFxi3WHs83uRNZj+34ZXFxG1Yy6rLchTX4zi0o9Qio6JdDbIJB+mY0mz2MytcGQkf7bxkUC3dXXPLomnuMeAUkIavRXmgkliMUwSo7zQPdilHTdToLU2fHez4AeZehIID0PRTQIbMiz4s9234EMlA/Tp6To4DkIusY0RQCNdaMHwnQon5gIiMh7wH8H/Jtm9h0R+bPAT+BD8hPAnwb+yI37+r4DX/7yl2gmrM0VgxIaUodmxtoaYiXWre2SgaZOQNRhkytPGjWQgPvlC3snYIZCKiZ+2Z2YOGDPJ5BiwTTzZCitkni1WNAiscFHV4zFb0mRgojvz4kgHULWntMglkmf7flon1DN2lggODHs8zkl7n6vY8zUhueNjgakL4x8oURO6cAzWP6obtejZT4+eUJ0AJGL0MYE9o7qQtec+Qi1rtBtkQmp5EKa+nR+RloBlsjYYS0jroj5FOMqtRNEKUKtEXlKpZbiSuYWupdYUOnDkQrlpfuV+O9qGWKdacJi/oSj150sfl/vuJr8AUPRVlA17tXRr+IihphSigX3LzQktq+wYJxuKhaENfaxWNkcGXbCNASDW+UTEQEROeEE4L8ws/8ewMy+MZ3/c8BfunXvvO/Ab/5H/xHDJvm7Q0jrHZKBstiA7n5NTrBccDr5pvsNsRycH87zZp6wfZF7KYdJ3plNt2kf3ieu7o4twvCam65JkSCRd0yx6cxYQrnABxbx8yW+jTiawd33CzZln1j1kz2+N2xERI2XmqH+9J7zK9utg3ORiZA8eJlcdzAEhyFFKIcAACAASURBVJ3ErFQmzubF+Y2zI7pLYWprp8lvOTbZF/khAfU9T6OUCDSy7rIUhLPmo73lCT86N85nJxPJlsjUtDRPSr8gI1pVCk3cKpUObxnnKJMOp/dl71AL0Ss/53kR8TEzUb5RPol1QID/FPjbZvYfTsd/Q+gLAH4/8DdfV5cP8D2qd2BGfbm6/HYOLaetiBXKVsGTjLvKePPI3lIk7KmGWJh4pKBFuhIpofiQG/0jnV+OTjA9BVVysG6GClZZCDl7qi6SYbid2a8cO+lYnJYhBjAWvDC8GVNll6m7hjbcP/ouQB2SR9v6q+1t/j7RpC+sYMAkSjax4fHYSc+4Ng/b1IYOjMNisSdb5EXZif2cdP2HhfydQTvB+TUVZdZ9MSQ5eijdUul7jLVAPOipnEI+D+++Uj2QS+tClcL5dOqr6bRUnjw5YVJRWajVPU+tmwy1E0UE9N7jVnJe3S8XMKiXQDnRUemL0CSRgLd9UX+whSeputcQVjwcuIRTkTZlVU8JL7qBbEgTtiacaFRR2upouNQVQdHYrIbi0YXb6oPc6kRsb5RPggR+F/CHgf9LRP5GHPtx4A+JyA/HMP0s8MffrDobk8VyhQQ1tL3DRvJOGOMzT96OZ49FHvg+H5bj+QP+HZgguMm0VGRggblNRzQwPoMd2WDaOZV2zRuXXTV7V3dwfTu2q1+ZiGOqduLYV4v46gly+/z8qLnym2+8xy1OkAbn60hqHsuDFWQ4KOXvSW8zzffxmSgzlH5JBYXg/H6wiFBqoVTfRs40HJlkIAoprlBNw3EVIgTb25iSWCKFBF8l9TwBkbreJ2QzDd2SeziMukoXkYZ1I60CrhsTMv4ju9pPG6mzkOMwHsonsQ78b1zPSXjDvQZ2deHh0ZcY5EjV2L3Y7uMdW3SUpMu7eIquFj7f7gTjwEnxEBmxtOhGKa4ITKtBSX+9CZ750+N3OsFYWhHycpdTPR1UTF01tIXHIo1m1j0QHa75y1oLT5SSE9+Ht2XWo0g6KGHyycSZPagnRrRljHs+ocupo2P7UjJn2slG02txmPbGwro5X5KwHBZ3X/Oy/yIBvxN9aBAbdQiF4YjIrThRV4CsLfQBm3m7XjUwCxkZG+Hf8XYl8yEWx1I1ksEui6/Aung+NjufqKVwPj11hy4RTqeFp+c7KAtWToEEIheF5YdhEce8bR7okyZB6sVHMHb9LHyI4dmigO4h6NBR2DIeutQJUWQ6k0Yt7uqtobspkkreQGvNwklIUQ2P0MjDLYvnDxAJx6RQIGaSmofKW+ExOE+omaF0fmDzVYMfSb+O7vo5qNINjso4MBR8D1ywa9k1D5xxx/5Bwc3mhncOxjiefzJz6On6XvGoZ06lvj8+GnhrAduuLfMbXPfRIwzjNhDYS1cPXra/ZlxxBA55dO7f0c9HArX/1d8mqh/Ky4Dg4t9HwJb0v4QI7uAVyGGicBlnEQJGb7P0f6/xhzdx/p0ixmjV8Z5HUPtV6QLbsKeObpHDsUfKW0EExNxv/i7SSC2XBkVYW2peq3N09ZTamQAjdyVy/61wrZQtuEWEvNgWVgIvZjW6LsFVcJEMQJsJjjFpqMM5qW/MGc4d3T3WOZn/9h99N3B8jqU/vjbncK2I26lzsoYNug9dj5dvu+PJOTPt940enV4iXzz8AmICjpUl4a8wI59puWbbetz8wR044x/6gpt0BTb6STWUpaFL6Rl2AjoL5tp89RyC2U+eD+GCmbJuFzAPuvFbvQ3VfBrXJQhb9cSrS6C+2nG04z7ZLpQqLLWyUCgtiEGJQPPWKKUgi3g7zXouwcYHmDXW0FNcWuSTaC/AYDX3ddFLnr8HA9VXGHBJJPDKmYCnvVNEV1Bom49HU2NTZdOG1g2tjXWtaKtYu0dotM33GbDyCnc2cgvFvd778L6K7eyWx5f5W0EEgGuiZfOXa+6c8jPz586enWdu8ZrU+3Yc8fr29OOD+/b7OyyfOPdtpk23PccCnJv34Cvv6trhhZuX3SINV3VP4OTmY960HIjGnqXPLT18nRs/oZRH23OFhPJjf1evegJ7RwRxi3cfv/d7dxUMbNKxzBhQrt75YM7sqcOm4xK/xyPyvA18eESDMtUaiHLeu+E4uI+N7VtBBJyjQl3d6ecciKtvDxWpuk7F7ewsYTJprsDZiitxTKcKSTlUen8IYaNnmFP6qPco1NHJBqzq3ovZ95uObbKNEUOQcyMDkIZTzgH2JRM2GyGueydFz0hrjJTZcVwPQ5nTJlQhvYqeaafPYP8xlI/xvjlPbAIGUUsioF1dDKLanVJDA58ehCXevwcYqfN86V5ukCmQsx9S067W5mmPNUdTunoshDRHFy26uEYClNTbZZJYxdHFqTiHr3duJEvUcm/GglEWsApycq18kbW3s1SP/UgfBQ2/xRdyoZWNdfPReLndxxr0DeZfFUcCryJy76Xee3+oxwZc1nBwixTk1rYI7NoQhKIZoq4jK5Q0lOaJYVqh4Wni1rWiWmi1IWIZccN95M9MU+HyeE4RHtcY/EqWR0jVjlLL7eOfdjkut/zXbpwdVz10/NMut1o3f/+IT/wUOvLTf8dHKv448OXwjle3yZtVZg/28H6OHOfKHg9M99h0X0c2h1e8ItLzgU/e828FEvD3UFYuIJEUVKSTsi0Ya4YH+5YAvoNPA9fAK2EodxiRgC9z15fg5mlx6DLwwd+8p6eLz/TXT1G/tRREQlcQwU4S46kd7nlTbHL06EBR2HGm3gk91DD57RyieOyv8dmdoKZJ5HD4IBpdLZ5pWu01c6TCydgfH/46I67Bhy9QSWfLCV/Dqy3eK0OF8zKLYKtS3FZjHWP4fgBiUJeNYoqsIU9Hiq5ulguFxLnGfWGTL7FZi2ZkWrtg6vZzO3kCV6Gyyp1vJhvmPwzP9V8Ei5T1a9swVV5eXqGm3N/7tZf74Oi4dWANhLNdYu5GPIw1D2e6aIR3X9LX1BPBFNkQCksEsY8/c2/OWj1zUqnuN6GKnUInUP1Kre4xeL+6T8jJ3Fqi9XEo8PYggal8JNr2mbGgX+nyWeKaX6nyaQ3GJ+yLN2jG8ZLXg4uHeP8b3rtj5W/XpH0rkIBhrLbRdHFWs15AhNI8j/q9XdxE0yqUDdqKaEE0/QuIwCFF2PDc+qELSAc/C51Aj25Lbzr/3V3ndZLxgVQ0ZCJJ2TK+PHl1Bsrc0Mgj4adAl51d3p7TWw2QKF3bn3XtJ8tNpZXN148TNukxZhSyD7aZ1Ka9imyLTb96NwydQIQyttwVSVsgn5jvze/PEOoWzl5LpmpXl7ELRj0LtYyoyiKeAUhqRGk2N9m1UNGnnmeJFOypVEmdwBq5Be0+MgxdPHOyWkNVWS8rpcJy8bFQrd0/3/doMKSKT4rLBVHF9BWGctncUvFqdX2FRA7xUyCBzDjUXnp/bXSFhfdTRB1ySD6iYuGzchcORupJV4q4fiw2ubUIknPLglvIRCqIUu+d8y+XyCMR4YOTr9PN8vYggY9DHF97z8Mc5eFbU6R47B45HHvoObnIPyJnOwqFN54L11LCg9c98P0zKw/Kqnux5Y3KY137UQHDLU48izn2AHdPsWgSt65fIZnLzZM4+Xx4XI717K+5Zce4daskbf9I5a1AAphi673LSQb6ynUCdhfRUJkWyoy+518TuAibRFIOC0tAnx0R0Ns9/pKDSaogAPq+ddlviQSS86X/fkYwtuCQ6XtvMx01ejoU/x3uK8mKd2hhhKWOsGSZf45Jk793rsuD08+Tk3zeuCyPsv92mK199mT2H9k9o1d5CCE+msBaRNNlLkaz2MDd3JOymmYApo9xKbQmnlVp8dRdp2Vxd9rYXq0smy/E6hzdSmYncI5X+1KJHZs0dQbpSy+YGu3ed2O6tHtYBe49RqBp6XkotEiolnzRanDdpm6hcC+UyrqG7B55DM+R0+BUTtlRwIhsLKdwCLbM9pSaD0ctZSk0Fba2IIFKclfr0wmWs/sVqggsmyOFzfNo+ma7JRKUeN1m1j1O9cSj5e0gAlH2Hn9wmy7f+v3QsceO377u9Vc/TL/TD38c+HjS7W7tftyrHn2Rx9/ytvT7OvX6uGyOANydvEIIN6nV7my36It/s8nD8o3LrIGfHzc1afjb4wRgJnIWEYw3FbTjujQwXIlfMolRXIOY4/cHRzVOTvFMV70wJy55eP3sy1tDBDwtdk4e5/wSXMV0C4q5BsXe6LvLdFbr0WeediezwQxjeBKYsR+gl4dccY++yokkMuVXet5dhQuJ7I7MENBmCjFRPJuu9uv2TehL5dDWK9KYVd4c8z2wvLqo+/zne+2vKXK8LRGBf2+Z8DBjKtYtvCgTzfmuOQvNA3XAg3dwuVapnltSCksp1FJBK6awLGcM4/7sGvZ1c416xhZaZPM5RcflnhJrxi80b9O6NkcCl0ugNsOasd03j7lX91S1TO/ce8L8PcJnACBT+We3nCP76TnyB5wyLHU5TdeNWJLcxqx5gGPsRVhomSAnok0Vz2K0hCOQ1syE5LtzmxaoHktwwn0Vzi9XJ8ShK7kvj0OBt4MIxESanFfp0dy3fN5nAif9g6FooysGZ9vXji7ufLr70UfK67jPHm4fl1w/Pns0zqjhxvW/8uXNOMdHr+942GZN5O7qqx7I+Ps5QMJPvIGofOP5B6R2XUFmntg/RiKzj5tjhziYVx6jT2+83v5Rc1YWObxPv7PLkXumZDGbj8hq+uo7HL3ZfHoriIDLjM0puxl1vTjraeH5FH7aDfV0y0J4CHoG/g3X0nt0We74UwlHdYC+Z1xq00cizYPcncqAg/fh3NrxkbqHWQ8wT2bZj5Gxy+eR5Sitv+kaPJKv6zVxC77uz497++zy3xkjUPa1S9lbT/rvSQPtBNnjQNrqXNuqojHO6SkpMfFNhdKEung4r6t2zOP6TTymQgpb0AKt591r1TrJ1+D5IwVyf0jNnIXpR1JOePai5gvFCkilFuv7RGbyGFnct+EUgR8riXDC9yBs8Kfi1pHcQ0IOm556wkJBltwstfTDnuFIPSw4RBL3pVDPqi2NJg1bW3jJBupdvU5dLphkWjthiRyKW4nYAfs1EjsQoAsI5Vr42M/ZhfK/DtFS9kz5zh6KQ5+AeZ/nx6Uzw9vp93R+XlKjjsEPDA7XDGp+Cxnc6oOPdNyOV9yo/8hAH7jaGe2OBz7ahvH7RuWzi5t57PxQv8enxHUZb3XtFtdl32OWpzQ7PkYrD9Lc+HbrpsgPOZjxIKe5cYxOGor54p4BIF3Abb8RymAy+1l3C3ONJDVOzjT6Rc06cUjxy5891RLoJEGFN9Pb2bMgPVDeCiJgwGrKlhrVtvkLxUYNL2WN/HKr26eL6wZKw/MJ4IqcYobjgqCUImMi5T4D+YwcpNz+JglQtyIQv/c6hL3J9QbvnSBYJ2wdZcTkSR+Fw71HQnQ9dAfUcrhiVqzOrR5NmqZ4opL+Q/raHDfZiMcIRNC3Be8OAV6jWe5oFBsfNo/408yZtzVEItdChOxmWi8zsKX2dE51FWpGigI1ovu0+CLZfN9u1uyH1GdoEOMICe07odnGWFX0fRIzwYcWXzBueHIk4Is/QouL9HTumUA1NnFmDS67BVG4bP7su/QjmXQtBp4fU9y/xQ8Q9MQzT7W+h6PQNqNtylIaVRqsFVoJj1qNjNaC6AUR5VTcqrKeQy+xhaWmznPsunwaiUZ/FvgA18ZtZvYjIvJl4L8BfjOeXegPvD7jcFDjpHpdDkyqH+6k05rtSYdJKjqobc7tvhyCgg7nmsG9gW6eG8z1NQtSso4c5HQamqn8VOdMtK849o3rbpRrbnw9uHvGf5QzD9fIdN1DDwz7Zk+ukqdsvigJUmSKthmvxV/cGE7WjDw5XsTSZTcSsqpNitRsSrQhP8e2I3HcaxyBVqMnbH7clEDG25att85FB3HMfvR70jks8r6Q0pAEOhih1Nkv+dAMB0+F9ui3nlptniMqTkxVMW3ujKUNVLyOoqiWaZ5PyCkO9R32dN/Xx/JpIYHfY2bfnH7/GPBXzexPisiPxe9/+8G7zbBtc2duM1jDT2Bzf6tmkX01EgPYEqt/g1WMtTjhqKpJMWL+OxLwiSGxUGeYJlewsiOEGCQ9+FP1a3dEQHYX7CwC0Dnanj8/tqjz2f1hhxts9/MYEHkgZ0xk0aFvYoiktYyFtW+4dDGhE7u+K5KNT3OfeYAtEUHx7LkrBdDIf+ha7Urmf/TEm5n/oqj76W/rBqKxizDdilAWRw21xV4QmhA8t/vOjT3yFaKNGZOxxHvpAqa+Z4V4fKCa+xCkHrIWTy0mpQZx8RN9t6gaOw+1NnSWGGt4DmrfHirECHV/gMGVozGm4SnZBlppuGdhU2xrmDRHVOsG2wmtG1IUlSdA5YRn26ohvpwjBqO1QFL2q2Md+H3APxff/zPgf+UxIpBltyJtYsPWZUgPAx52WxPvt9KRgP+NfHXgk+9aLp851A1Wv29aX4fHpTZChWdDht041n/fWm83n/0ILDhUMHOE20R/rHhjcA0bh2/iiiQOw0qTn3OgkEXG35RdxxjlXhKd4+UvdQWg9joU9zwWtDrR6Bl/LLh8Knmjll0S2f4yY6/CbOnACyNzf//PMquvjfyD0YeZBCXNwcJ8vkQ/ll2HWuYMzN7KNuEoNE3Iuz0a2c8Xtdg5atIHWOhPUuyV1LVMc0RDNWil+rtJ9tBnLA5EK/5ncTz0n0Qq8a9OGYd/Ad+vcFfmfQe+8IXPYWi3oXpM+pDnXUTwLcdNm8f0K6AbDYmt452q+rLUWBjpajJ1QsC2FA/00EHHZD3H/Qa6LT84ZXZAn4dMqrUDKjiO+Bi+xwfp6v5Dubp7gpq3KshUbGBTNvIUZ7LFR/NIfHZBO7dB8r62kD97Jp7WIkPO5n0ubUxIzUXs8raH1itihaUKy6lgpbp3YRHKqXrSTZzrFnF7eMZz5P6RiVJKV875vxKUz/dlMNzOpKzWAoq7nF4wEpbUYpF9yFUVFqJVKcVzW8Y2h7G3bG9DzsCWxCCcFpaOnHKIBtE0zD1T1bwv1P1mtPlu0lszz4bdDNuciPmmNwrF91IsGFtZEIR2dgSVW9eX17gMfhpE4Heb2ddF5PuAvyIif2c+aWYmN9ST874DP/ADXzWzMMtYdxVy5wjALJKOWv4VPGikdn7vk3qYBMETjsgxY0f+PuR+u1qZvcW35fehRz9C/8fve/2xG+ffEBD0d8+FnbfONDChQP/M6wecmFHBoMNzxyTBpS9omc9P3DXz7udGKuCJQou5g0y1grvSSjj1CCoFLQ0RDxLTzVOEa8uM/F76hsD5GtGGNr1BtsIJXmz9JRNxjoXkU88iGKd4TrIinhBUythcpZZABJFtUHNPyzxfe3fCIEzDnyDmW4mNYYLolHQxL9F/uXUaQqnqqcjrAraEmdzNmlAo4m7GfY/JvqmNBdF8fAJ9YiJgZl+Pz18Ukb8A/E7gG7n/gIj8BuAXH68EVEv3sx497nBmC66hVoPhR0oYFnz7SJ3msZJbSjnOzeSOh0Xf18t+e8sbvnL9rpjf/UqZJ1Oeh05g9vTgoQU/vA6vyEafXHa4Yz5yY4AnatRzDRyvD9gu01ZDzpzSL39PDbsJT3I/9NwFKolAwH6LPM5xfZFI272kOtAo6taaYrCYOsdr7uHmbvyNJuKKrwJN7rAinoUYusIwPQM7xuhEINudiCCiDIVQArocYhJa9tJ6lxCcXpYgBBJbkUcC4iQCNeZmTaq0uVhQI41P7llI7EI8ZlFudHFkHtazZql4f5ZiaFHKSamLonbCykIRz6lp5Q4o7l9R1KMqEdayhhdiPLssPMZwPukORM+BEhuSPgf+eeDfB/4i8K8BfzI+/4fX15aT1UjabrHlk2/k4WYoM3XopVA0fTdtLNCMHfYWYhHKe1wIw1S4n+zXINgOv6dP4yrFXt8NZnfd8aKZSFgfnyv9wbGHJkY/t22+dv+uUysmUNFlzPwjum2SuzvsJ3+GQjCJRNtfV3pP+o7QJfcSiO20ewCXRKqvaIuap2jbzCixUevFmjPi6ok9ttNGQWJzrUEEMlVGpnyrsfwTgteU5Yv098aMYu64VMJ/IYN6BCIcvfgi7Dn905yJKyklN0zxoDYz8+3Qzd2id/00jQsTMhpp29IfJpR4QmwjLzQsEq4ItXgSHYv9J9Pd2hgMr6kTAQ09hVImjvEwGvikSOCrwF8I5dEC/Jdm9j+KyE8B/62I/BvA14A/8GgtnTn71EivPYvNHVVb6JjdVKKt+WajExGIHZ4ZRkOvOGX3vUJvDrHYD9bRk7Cf73kIRrPt8Hs+P+o/EIDp9/HeI/y/8gPIqw7PGN9tdz4h/lFBmq6vmPbFn/4M/XxEz2VdPboy+yPDK2PBlbQaxC6+VTTqalNbCpnt2RhZnzY11uYoTg2s+XjXIAJ36prvSwC4fNf78PFYI2Lx1Nvm/XgWRRDqKSC8zkpKH+tunsOJi+d/KI5WcCLg0LuEl1JsJJpKyjDtJUNKK4n2eRXXBZlJVFJCoBz5I+nrQMVjCpoJzYLwVNd9dVHGHBlLRGQYyqaeh0Gb93HPl7mn51flExEBM/sZ4HfcOP4t4EffvKLk4nTGngvYmXtuRFJwIufUTzPIYrpP0pIr3mO5+cjQ4icRSG60h+7H/tIOKaOdnWewuz/W0FTXQSSYjvXXvtboMWsbrsbuQBw6EZllvisUMDOCmVRMaCXbLpHgq1O3IUrkpG6WYDtcuXUsoEQcbopd3JW1QOz04SJZWSimvniawqaBJjy4B7XYbEViazmgVbBCrUN56R2eMN8DgroVYNqDMLdYG93kDkrgFsjcJTl9BnKWFBFXCUilJLc16du15XU5P/o8kSXalM+fiIGAZm77qKVM84ZAGsTsFHOxSRb1v1aDALj20MLkXYNYS/ggDNtHKGuP5t9DeWs8Bgc8hb5QLSFj6QORCkCzgloJ+675gicmQvcHyD0JJkvDQU7s7p7Zlg5bc5D3i3woyOQ603het1MTPDwARyCQZGMQk9tX8ND5G4jPusJv3NNFhmnxdw5riVUCCXRR5QEiYHsxIW36i1XANey++KvL20ulWKPaSpNAGItrum1rNFU09qAq7j8OWsGqZ9kRyN2Ja4zhhTjevRk10I0PkJoT1pLKZyILcehBJIlAmE3M3CLgcnUQAQvIHSl9+/ywNFiGvookAjmamXEpSa/3iyZKizHrZkgpIX2EBUAZRECjUUoQJCdMxXBLSdEQB1I/k6jk8dxBbwURkOnPf0untrmw3Wd7D+VzwLIStyLYyNeZl011HTn0kPmPx/MRtjswy9s9Qcl8Ty4u2dc33vRw/XzFwWnhKi36cZHfoAFjAUdjY8UnbZX+EsbxpcQYpr9wvMr3HanVA6WEM0yLtGISyrUintptKc61pLpijfwUR2rV6vDu3ho07Yu14YqsiBvyJJulenixQKkWiz6geb1EN4WbrF2cM9Z4/fSYk3DtCuVfnwuSvNfH29SJV6YuM5Phlx/oyHUH4p82xKWcESkupHiZO0mX3YSFFNvTIalUTzkusbhNBGpDqmBL7C+srkS1rYBKbOxayD0dTX1DlEyI8lA8SJa3ggjMxSdpLtqxcNNjK/uOuM5fO33GE4rOPHImL6PMUJhJbp4JzNSi8XOGb1g3a44jB+Qw1XG7DZOzUT9xwAQHynR9/b7OHflJrXM/NEGAnUkxCUGc7zsrJUGKRR/1DkVhLv6xCIq4x53JIN5SYpPPIAKFQrGYqFo8YlQSXfmugyIaXnvFo/JK7M0YiFojTl5KTGONXgu9Qy68nrFJvGtL8TlRMm6kSO8ODbu9OAyYXG7HPIi1CwznKCwN2/sRl0mBs5sFsv/ihMAXtFu0UnktUM3zLbaC086KmKIR7VhCaZkPyRwOlltgifLQHIS3igiU7k2VRLqV2JLJXIXjgUIDBrmLRDiAWGKGdBSadmXPyTgN5s7OP1bOjTLjB1+Y85GjOD57Jo5xPgD3ee2Nn1MTDov/UI4ecnuqMFOrWxqh7lnBzn14KtKPDYonlh5y8d4p9IbrrkTQioXyiqV0rpsnxKkDRX2zTCuGifomoFQcKntCkSJCPTs3Pz95zlIrtYRbbw9v9rqXUASWWLBbDEq1tFykVqiBecJOCdPljIkGgoq/dOOdrDmmblnIZLFbOElZpEE/pqfrdsyS8rqnSHMfAVwZ2V9FmO1TErCfzZOGsFbYCkJs2Q4gguoIzAKJ5DyG6n7MHipvERFIOd56LLejxTn2PylxUMuM+857Lf52y3C/+HtJpHygyMcl+RAS73flenS02enJzuXGyDicHb15UOa/xeKPjZ/uv+GKtacjO73A/p5OPw/05NbjBUbm2jynsu/emMwS+yxSYqW1RHIlNO7SXXJLcd0NugAFKb77T1lcHKjLyfcNrJGbIN8h/PAzdD/1Fz37U/R2mcQif+FCDyrpfX4gxzsxUAaAiqhIImJRczukXPyx6C33XUgpaib3BpabG/acjYmasgU+MG718n0GUEEaoWz0824qnNBdPDuRxC0ifyxvBREwwk86B1fd46m1TERhQcc3DPUkllpBBS3pLJQbV6Tb8CwJyYBWOR/ijO4mwcxH8/deHBifD5yf36srGY8nxjNug5CcGI8TgSsbY7Bw6Vx80tz3RR4tNgXzfp7bLulKm7qBrDpTrgf89223nPOJwLZFgIw4F79EkE9NccncQ66aUFRYtCIaG83UwskqwWN5ejaWKujJtxQ/l8IilbrcgcBlSzNb9u8r/22+5Zemsrcrd1054ABS2MR1DqmZtwnS+1cflY6IDHI32UukKsv+zFRniQR6ss/s0+jG0tPh9ZU6xiYAQJBC7995zJoixV3m06/B0aWvC2kXwKiyuPqlOaEqLSMef40ggeTyyUUtKFwOkuVL9zxvAeDSlJUdMnHJhd5G+AAAIABJREFUWaael+ysKhkbPA7It/vsgxa6hsNi33/O3nnzM9hTnul4MppdbV0Oz6qON+ZLHdj3jCCy8qMocjhvTDqDeOjw15gq740J2X+iin0hEUvHhJbWGpsWk7rsWsx3FzLAikfmCRLRosppgdMibJFxt9t4ZInmRH5hiUr6NL4wxwyPsfE2tHBUalK6t2daQSwI1RVhtvnPs02rjdHVTjj2VpKWIoQm3D9OnD3T6Y/KfpzyUWbAUDyITNLS0YUZoEE4ZOh1rGNrHitvDREguXdSOhO0FdRcJCrYBP8jCWWrmCitWfiEhiYXw+XL0BOIuOZ56oucst2GeoOT+++cUEekwKPHj11vx9Ger7h66IEIHFvUz+db7KsoSfVycucrhshSuugU80qiLTZQcvd8zM0xYhK3BAgp69cUm8L2XjzY1jfNHItJUUcCq1DVTWSXrXHZNrQoKuq5uYqg9YSWwrO7O2oR3ltOLFLR6qp0vbzEgPuLh+2+Wj/wZ7/yHT/uVz++hCKxi2xNHQlgIZIkYwHfEMWCNho5b9zaZtAcpXTX5bCObKs7CVlLm3y+bzw0r7/a1CY+uwzjA6ThjDRYiIT3omKr+QY8qBP3SIVGSb8Cx7U9uOkS5spzmUbiurw9RGCWyZKDdS4z2bHJvAD+PTPcOHLIW2PhdjOQhLw+S+N7bjHaER8H8fx6kTO16fb5W9fPv5KhH8dHpmtufFyVYeG4CTjGvXb4bkzWDev/3XqffrzP2Ql+hGwt0ANrdvWYC2oYqMZY2NjtJ0XBvRm3UItnGKph048HkK5j3Q6eIkpG8iXU7sFi8QbxXAvi1nMjJKfNYKCYfwhdw26TyTQ17764A42mDuCQlnnsrLgnAsPaJF05JZbIwSYRMtsVz1TrE8TyvWbYkmvF6HqLx3xV4C0iAiGSAZM8FamnMhFkIoF0DNYclFSg9A62afIRgwljmQw62/Xnx37qsQX7333sDgtzP+mnao7Vxs3zVBlWpP0M2e9FP9c1OH0/3wlByvxThx5akG+W7sNelU6Nme6NyV3LdHe3gIDVThX89qad4JqBNndeuW8rEE46apzU2HTl0i4sZ6Oe8OCcUih3i1sGnrl/gFJplLAOwcvVX+9lbFbz8t4X/6IbmPEqkoyshNvwFPZswBY6Ave7sfA4VQ+CitfvMY/hkppEIBzVHX0CbfN7eoLaQEQJmDKKtTOVGDOdJ17qA80oPRlJzgEoRSlFoRWkCZKtqAYiVHE9jmr1z7YwoiWHPueh8vYQgcOX0ezHqdjVIrMb63m66Ch+X1/8Ro+9UWIlPtLfD6GCx4foRg0zK795yYOYoX/u1Qz2wH3XkQ+Dqk7H5HierojHLHSQwVnTCSaQgO9QlNUETC825JIyntkJ90QDO+fGGMraQZCNkXwkz8wMIpnJkLlz5ThV7clr8nN+VZsIdX/1/QQ2bnzemibRmHQyHrtA7en9IPyTWNiHwDoynvvrdeXtIAIGRM4AMoWYwFo8NmDFqGahbIm/yClgpp3qFnLiJWmdbPUHb7zds2+2yU90jvAwlbqu7E16fzeh9jf0SXxbKTAaY4cp1j+m1SHTdTM3iqi3XkMmuEzz2sFjcCiqsl/CcYcUQCOST7bw948Zvfpnu6TOZwFpFGms1rjHkFOlSkUqlAWMJzQ7I+WOUgtNN4zGurmeZ90+xAxevHiFAS9fhHXg3i0W95tz+fOdy8i5nbzvXJ7E2omTWqIhPybTkvIL4vIACd3yEBE/0rJHcuPZXVcPhDDHDEz0W66orOfEGCnSQn/jmIYeiqR4nIIJViIXt7nVY5NwFoqw75FM93Z5O4gA7DrizTnjY/UJOw6VFX9UDj+tn8fb9RFafUtL+LGqe91Ftn/nj9Sxb3bxUQIeRHciVJ1tD4tPXwk217KveR/3MnH5K84/Ljly2nx0X6C7/rgm2v09bs3H3fyZfQM/ytgfD9y6/6FjD/+2m9c8+NBdeSuIgBGOQfHunpZZaNZQ8y2pDfcbsKIgLRKxFrYSGttQ7BRLa4zSw387NeWKWQ/RLI8cOvfAjYeOwI/rzDjmF8Kuuv7qfH/kfjIftf/9Nom3yeM9nexUj01AMJJbzu3LCHRUuynJZOJ0Lnm7GcDoKdg9MpNIo2XJCl0mxlPGC9Ai5Tarh3w9MVfUiW1gG8U2FmmcyhpJSTd8O+4TWs5oXVjD0WjdBFXhg1eeMtzaCgbfef99zIwX3/0uYKyXb/vr3r+MfqqeAmyN1OMtApLCXbhFTjDvLvdDcdSjZNBR79akVXFNOgdreCemm3EfgwximpyA5sFNa1TrKtD8jEjAWbFHRLtGH1YtHkRkOu7uFrPMR5guDd0flIPLx1V5HCe8LWVeuJ3Kf2p8+WOXKw5EMj7bnfvY7emwsc/ERyt6UHo4PN+mb0fO+dr23LjyMM9vVnalg9i16tjCQ/0HUDGQRZr3om84ilg3ZsqRsMa/NvXFrgbLWm+1bzfSH7+8ITpNwvBG8v5HaNbHRgIi8kP43gJZfgvw7wJfBP4o8Etx/MfN7C+/tsIwgXjrfXsobZ5kYrPCkh0gAs3jB2xzajcSjYrv9mTZUUklp7E/cO2bziFzs/Ks5bQaq9punL+aZDZqv7X56X7iTbN9vv7Q9gFj9kggVUnDOmA33jd6Rie1k+SzLPIG+HnM9lpsQGLKdA9CiXRiGbseyUZMNpoa633DrKH6Ckx9S7m6cLZnnBdjOcOz5Tl38oT65CnyZOF0d6bUwlI92k8XRwLryxVTeLFeMDNeBup4uT0DoOZW3PUOgHOYPwueo7BI+Pmn33708dAqjeUu+4GJLrCerSeTDKd7cLLT7i7cxy6VBGPwjDBzhzUnCdiVwjXnvBkZytyP9UZEhKNIeB8K6dutHbdWHisfmwiY2d8FftjfTyrwdeAvAP868GfM7E993LqvHzb68iPc8ileNzS1n0UbftXLnkJ8+nVfQYXHyiMNmBbmm/dtksVbd3xEef5j3P9ZdOmnXT4tncCPAv+PmX3tahOLNyqRZMEV0qjODhCCaMo+kTUWidx0fq2qx1+72Ul20NBuLWC5npeD2d6Ci0wc/kDp91cxIEFOkRtY28Z1AlzttJNyeHKVg3zZ25gcZAdh494JCVx7Jhrs3K+Hp1tfMoG9W/RiZuPJ2IGayMDAENY1dAPhYdfW5kjl1T3uk3CJdAJGWYzlmcC5wqnw7NlTnt09pz59QjktICewwqmdEBFevPwQU2V7cY+ZsV08im5tYRVonk+gZEy/rQhQw2vUNLxIS2QEiaCmvjdCoqJ4597/5gbG3jHslc3dqWcak/77gN4kPC51GqyMiEzdPwZzMoyMh7MiYTErvgZ69FcgXiFyIeZY78f2dQzs09IJ/EHgv5p+/wkR+T9F5M+LyJc+pWe8K/+/KXL99TWK7Yf0HW/8nI913ZsztIea92sBDX4aexGegX8J+Hfi0J8FfgJ//58A/jTwR27c1zcf+dzn3os0AJEWqQQ9k2E12FuLIiJLDavJ7Sclkcqgllca9uScM4UHDpT8Sk7PX0dZsTPZG+z+wZlxaNt0newO7K8bfWApkO6qHbaQ/XnJo1M4o7vD6pR3b5juTIxIFNSTapTFIvtN1hm6gVKccza3WcvmOR3qdsEUts3cDf9uQUqhLhV9ckd7/pTlaeX0tHJ3fsLT00J9YsiyeeSdFeSyggrb5QO0KetLzyXY7h0J6Ob9ftanACz1gmv/PdCohrdeS1fe3q37jMfdBzVy/qOD8btZc8CkrGL49VnH/AN/zugtXJUOep5hhfDfPd6jjePZPquF3HizD7kEEiGtGkorkQItjreIQ+hxHg+UTwMJ/IvAXzezb8RLfsPMmrmXxJ/D9yG4Kmb2k2b2I2b2I8+ePh3HX/OwW1rc29d9OmWAqo9Y6SM04HXP++gXDcHndTe+aZtuXbenp5Ptcfd9PnIgZrdKX5/XqdpsT/aiDX1pPF7va4vd+PYplo/YvDdGEtPk/1QsE3w6OoE/xCQK5KYj8fP3A3/zdRWEFBNpl30baKfo4ScQsl3Tholi3U+goqq0zKUWmVhyZ1yiXqfWk9AbZ8a/dI+4kTcoKfe+rV2m2zmqTP8eKUYSA5m+k8k+DyjjiDCOeohsa37rmX7nSqblZFOi0N4H072WeRcmRFJaWA9ig9F0Czh5RUXOACzxmWm+n1huunEBlLLde8xAvUdOIM8Fq0ZboDwT5AsFrQv3y5nv2hPW+zOn01NqPXH/QmgbPNsuFDNeaqGhtPWVZ/aRFxjG1tudHC/7NiLxamx7njb+zr59IJRxzohkpzvinVYf6/fkDsbdoy+/HKwCo0/LzvpyA/xFv84uvwcyZ75XgqkhWhDxTEMinmJ8M6WgvlmqCL6PgvU9Gdtrco5/GpuP/F7gj0+H/wMR+eFo/88ezr1B2U/46x+Hq+3hyz4i4/54V3Xl22vusP2b7WjFUfTcUZOHW3abS1wfupJsd30WVzwEdw4/b0vJtm/tkQv2DDjSowxl+q/AdDwq6Is34gtSiWd2RZgfQoY2vdNxPt3s1fkl5HBcrl/r8TJV0OubrQkHy4KxP35owxxj+XAZc2vX1tc0+pPuO/Ah8JXDsT/8sepqFuGDycU9cssjBJ1SK0ABtfSiCgKtGQQSduFpkHWeAHNnRIDI0enumvsmae4S2GhzDpbd6nC79RFjO3GNabD79DjqDObnzVd2GNNtAvEztQD7B3cpMzst3QT99aJOP1BabFQS3nW2umKm1Myz7xr5NTwIt/IKDO4318w/tYpRUHy/wdN2RzkX6t2ZJ+czz8szTqfK+W7hrGdOulAERyCXb3O5v3D/YsWa8e31fZopl29/gDXjuzXfJKavhX4ik55GP6waXo+5aWr2Ue5r0BecRWdOliXzTrFAl93hNMPT+wTLMbw9ZmNmHc6nr4J21YNfMqUj6+JQaZ6jcBPf7iysHNLVWo7+tKxgQyegbWiJHitvhdvwRyo32d9Nfvf4fck+Xkfa+yK+UcFDdT764OPxqZ4HX+G1OOOmeHGTo71p2RHMqbsOeLYnVjVuEK24N9adFN/Qw/ML+vZaRUpkDRoNTeLeg3siCcnw3pte+sFGX7/DFRz4CNr/X41y1eRduT5zYE/TiccH/q0hAmoTElDdTQppMVyh5U1ImPu4e0Bc8Gl12cr6biyZhDQGvWPOJO0DQnmnHybyTQ/AGcZdT/wZ0nJ1drr9Jq6dvlzBlPjZF2Imz8jjeXm20YYve4KAHceysTtPlf2sM+cgFum4M4tuypk1PQbv3Qvemm8bvkU6sRWwAu2klFPl7nNnllPlyfMzT57e8d7n7zgtTzgtT5FLQVZha0Zr8OFLePXKuLy4oK3xnftXnnG6rZj9f9S9W6ht25qY9f2t9T7GnHOttfc+u07qGJMUiWIeTJACg/FJA3kwBKUQJMSXXPTBQHwTNIWCIgQCKiIEhIihLBAlENAgPhgFiS+FeHvwQqTKC1alKmWds/dea83LGL239vvw/39rrfcx5lpr77NPXLsv5hr90nrr7frfL0qZfM58PUwp4h5GivKBLtJBlz/KR2yAG3SLVWDEpLTlspG3yFBJDNMoG2UYvwsWT9ptjcZI/Ki3eahkpC521fUpUhDfO2KhzxBtlEC4ULfI0M8cHw0QiGNH8Pb7V6H+fkOOJyND1bd4Q2sXIXrfBz21LZjrLbxoxOX1bpPvgU7fg7tvXCys8RvvgPI7ymSs1dSI23c3GRAGgCAMAKaNW9CtkTTWSfSga8NHtiXztJDjkdbPbV1jp27qCDlArdVYwupRoyKKlFPust+ByuUmvdid7z+uIc5xKDcw/H3LQK7fvijTKKlnNuzV78Ra7m0eo3A9B4/2x8cBBJTBhntw14hEDC2kdEj6ezxCk+xWes4Bh+LB4zVwvt1eDV83HfxuhsfGbe4/s0G3tQ7Pxg3f1nx0+pr8iVi6+8hCm0SjbTeM+EOGDRtleuUa9hMSY+hRgNTyekQCC/usRetv8TF9z2cxLYBUi+Onq/muWyJOYV1vSKLUSdEs6OEGmSem4w3zUTm8UPIhwXSwzEIZalqpUljOphVYnk6sT4+c7k+UtbKuFnuvrhnUtEKKcLyZfWDDqMGe62SjvXo+gvD0bNujyoYP76IRC2/f7oNbserlZn4GDl+ggGvqgGFaQjcT66JpG9pd3wOjYHCz8YVaTLy6eg9Ly6dgdU3fDXYgtr1sGr/3PY+ow1bCAUCsdQ02oEcUtjoYh7nVtlXxcTl7cXsHuq/Dgmsa28vJv8TJMDatLQzdNkevf/QC1F+UazRtPAuMC6FKDAy7Ub97FeM4dWMXf+5+35HOK7kTTkoFESVPq4f4Vjd3VYp4SHIya7XcArUkCy1foK4VXRUtq2XRqeZ8pLp6wyzApngWo5GcZ5znUNeN9zcA2VZJxDpsrKQDgE6R9HU0zte4cqLspVzQn+8pE7Vv79XB+5iW/b632du6WcoBmxoE8d5dzNl3gB1QnL/36+qxAMJOwDKsCBWzE6iiTirqoD7CSchOCVgGAldNbaaORja2Sbwa1vtye9c2y89t8n6j8aS7ajf3BywzAgHi2dWat4s6Jj/Kb1JkxRtBUvv4hW9ByFPwJDcxLskBgngsvfBIC4kzZ5PBJ8+ld5zMim+dH0hJOb46UQQeJVGmwknuyGlGp1syt5yWW+5KJqWJ81NlPSl6/wTnleXhLcv5kWV9NMm+p9USsXgCcx77CWuNMN++YVZPzd1kJmGfP8bvcwpRtc2pOlCsDgjsXdntTOnzNk5Rm7LtDd2tk6Bz9xszoURuxgA9BkDFbAS0mg/NYDHYYhU1xOgfUsvKFfENa933YXt8FECgHSPp1PgkLtf6uCkCtTk2u+AAGwaUDcTd7+F4d4dcr+317bd3d0cKfqtOHOt6Brty/fnzQMJ/hnyIctEXH5tGBfiAqsexG0KP2/s7BmVIOaT0VOSxoSyMtwGXjk9BdCaJcJszcz5wkw6kNHNIB89DuCK1UrSw1hNLfULXJ1hW1rKy1mJZy1VYNQ8CWTMGCmrP+nud8b7I8hgLqJUbVIFte8az2Ox0Kum6dG776Ln1cmW9xSSJDN8YpLxtzQZA3jWj19WBgHXHVYS9qnfKBT4eIFDNh90WEN7bCpqanUCLgJNoZH6DAWqLwQWk7QgKKmZrLyjqlMAG3iCt/ABwiDka2YI95N8Ck3E1XIEZV59dLKgLIBEnsrkZY5eG9hnVUTvbFP4ZMd4+buGgMXks+0ipjcepKx4v4KmGLt69Bs/GiU6z2Q2YJEbI9RWTJF7eHpjzzCfTp6RpYp5uzMqzPJKKshTl8XTP0/KIPj2hjwv3pwdO68pjsShAS518DCKr0cGSd0YSUJd+N8qneOITCfYyIieP0FQbG+QiJiB8/Id1U4cNO86Qf2pvi7cPebZFOrpdT5vFYotYq8u+AmGFGKfie0Q6i+B1VU/Wt/paXuMrkYuxfkdiDHZJq3ZUhtJnYQBrjABgRIvv2HEXO7A/kM3z/Y67XlFz0tm9Jc/Vs7l9ubs3gGNEM1fr2D2Poemxz7oQzJ9fAKbNYu1SE9XqRlfaht8+6Qsu9pMHD1nLighMkQRUhJyF6SaRcmK+mZllZp4nUp6YJLMqLKugBWRVljOczlhyjdXCxRU1VVdLdShggky13IKqpDS2vgOBbtjD5mSUDTU1II0Cp3U79rg+N9S6v3E5Y+3FDgWGYbdy+0lxpGPL3ykBJ+22lJ2yBSh+HcCsRXZ+zzry4yMCAtoEOm0BR8y8arkGI/x0I3wG0nYE3yHx9poZIXcXyQQpuR+o60BgMwnxOxQdAXo835NtHdhsgcB+ueyfj/VsAVYLUO0bNnhffy/SdKmVkV1ljaoRmtFPDOXkjnWyxua3b51jvCJ+wHpCBG4c0szTRM5w82JCpszxxZFZZm7TDTInkszUCuupIGtFFuHpSXk4K/VUqUvlXCqrVs88icXVVwbMvwJC8s29tpmJfu8nxUdEQgA6gMhYdr7ZGvjUuN5O0KUMYLgYN1778bbpto2tXp+6FhyrdqrUNrZpaGoDfvHRlrnT664erTvAgxLJWd7NDHwsQMBHpAZE2+6vYQ/vNn2byvdAPN2db3jsuil0NcOv/2z5/B0KYOTZhpv6Lk315SZn8/51INGFnNfeH9/tX7RAoUL0N4CJWa9G6m5IEqE07UgSwjXrXV2cJHf2wCT/oJPnGbyZSDlTD0fylMk3B7JMlj0wK5IKSEF0QddCPVfKeWU9r9RzoZ4LZVVKjXZFyvGW8cwCnGxJJ29jAIGYGwf5gUkdQdQdAG/D1QaF9x8XFOI45OP87zfgJchvbR8Ajwyvqo55ES6/rQOgU6Ul5wkg+Z7cIx8JEMDVNJ5GqO/vYIp6uQ4EnE0YWQV/ZoMSZMNWnKLb/wZsu99w2+fbqdtClR0yGNqy7eUGSLTHev3dsW3PAPIRWFiXpC1k+4bTh979poAdSNVoU/L6Jim9HQIycL2KUNYZFKZk+QCrb0idzdAg32ZSntDDLTon8t2RTCJVTOWQC8KK6JmyFspTYT2dWc8L5ckAwVo8oWf2+HmeuizN1trkloI9/mHjWbzdV+ZMx40mvXwbnz1E2Y+2Xjn1gd28/wxCGjbuft7HMk2d6RPTVd77NmiXY/g+KQ5AIoNXFwy+Gwp8FEDAlyYtQZbTRKpmLVbVBB/qYZM6ueML2je+No+ivjUv/Dt2G6o+BwT25Fu8v2l1/70+zpcLq4OjfqOV2pMEz6ynLbDoZI0Ov20d+UIXH4TGSjm6iFx7dVIQyO5+IxathezJLg8uIKyygMAxmbHQmROIkA43ltbrMEGaqPOM5ETVA2sRns7Z/AWmxPlJeLwHfarU80JZF1QXVE9oXRBZSVIpYkZBtSbD4B5Us2ylbZTduIVJc7e7iUFxcn9gfdic97Frs3+JYzYfu5AZ7LH0yH4R62kbO6EDZ9waUrswMpk6PNWK1tQFGDKsc2+kuCxAqrMKEZr8u2AnADF4TcRhZI26b4CqAQAZ7QJiw6uTSx3Sd3VSTHg3oLjk6bZ2Ag2ZtOf+G9f7nXkNQaDsh1035cZ3R03D9qRD8OdIyjiXzQJteuNNUzvAUaHzl8U0BzWpeWjW7G75GVRI2fpycNfM4pv/4L+JxST10xEE6jRBzgYEUqLWGSo8nYSchKkmlifh9KhwrnAu1Go5CbSeoZ6QXJCkqM42dyX7Pp76hqXbfvQwAW0nbsfc+98SoQ5zYn/Sxi3MbttaHIHAbh7Hb16sl92NUWuwpR3tXk/Npg1ggwHppCPCi4bTCD9t/bP7AVRaHsTvAiUAtBkR6LNaA2P1SW2hnnFgAX0DbMF4f2dAt217SnxiNyHPXO+2/uYD14a4vbVfNAxbelf3lap3X7U3N3EmRwDR8vqFRiVUgx0dRZhrQzjaKgsEU0LAlIznL45OC2dUlKdq9w9YNKi7aUYEXuhLEOF484qUEneTkBLMk2kP0mxjUVellHtK+Yp1WSmnlYd64kHPLOeVuhRqUAAyGUUoZhfQgps2AyYfvxLptrajdQkUbK7G5RXuwhrYly4c3M5BG8HddT/rEpfxxS0DWbcr4GKtVh1kGuP31Oarq88dkEsgOKV47RGhLDk5kd4dU+TDgICI/GXgHwd+U1V/v9/7HMs78Lux4CF/TFW/ELMj/XeAPwo8AH9KVf+H936kbd5h47TMkx2DjZtdx/f6WHUcMQDO/pk9qB7qYZiTZzb/tSjD4/v7Z9f6ualZeV529Cx02Mo5pEm2e9TcTiZuq9FhUEbNSGPFMEyirn5bStSx+v8hKLSsPoecLO8fNwjC7XyDJGF2V+EpOUDKalRHsRwEpT6ylsKyVM66cNaVcymWZUonILt5sLRfzw/UfCPaXIZwIPIE7iYzetAAugTO6axBG5b9FFwB4uOEj3DahjUg9Laezft7QMBWPK3Yhu85M4IBHt/RXs7fqu1sXPd60Yf98aGUwC8AfxH4xeHenwP+K1X9CyLy5/z6X8JiDv59/vcHscCjf/CdtasPXu0dMmm2OwGVauou0xOhbTn0srH4gyQiZNwd2Y2foycarVtYvZ3DtqC21N0WfG+nJt6/HPn9q3tbg7HOze+u8vheUwWGteTgetzkAfFeE/nHotdBTiKU4gst+cj4XMjTamOYFkBbwI7p6K24MyBw8+KAJOH2YGHCZzHLwEM9UotyWs7UtbCeKvdPwpsTLOfCeVl4rCdOulDX1fjf4iHOEJBkwkkVSw+O0pJ7NiAX5LzdDurlUo5j55HjpkUpGgBzrKv95u6X2xt9z9tJ3X90kPBvANcgvLRTD202pIjXqpGf1P+0QzCvowGeABSibr2Jy9Ks7ncdHwQEVPVviMjv3t3+OeAP+fl/APzXGBD4OeAX1Xr9SyLy2S7u4PVvRGfAvQXH2Hex2cd9ob4B+6CP0vLmhHMFq3aZeK/w2qaFUUAY744t3iyXof4oohf39cpz6V3YfPMatrnWhnY6NlLH27Gh2A6KwpjHLjCLwQpfOKttPAkHIbfYT5EQZwIRIc8m+Juz+QpnMQehrAfQYhRAUZZVWFZYCpyLcqqFtVaLHzlqfBp7qMNudiWy7InsbfcvgENgxF3ZzVIajNCC0mrgf3y+h+StCt1db9tyOZfbtgcSiwsNQOD7IBzkIn8kbGU+Opz0NnbK4JnlDfx4MoEfDBv7N4Af+PnvAP6fodyv+r13AAFtQhEYombhxh3hzeLmkdTIB13bfRel9oF01Dfsh93hpPTOTmBo0ubksh51wLPpxYbq2Av7+70rkKm1drxfN/cv+xGqseR+AH7ZUF2XDkuuTBK4yIJUMiSyKMVDrXkQkSQnQCl1BVXKwdkAMTaAG/udXlnyz8PNhCSYbxZrsWQDGItSl8JyuqecC+ubhfLwQL1/pJzOrKczVQuiYhqJJCRmRBOZiYTaBu2JAAAgAElEQVSg4vMcUzUZMFij+7IbtxbAYwddd1u1G5W1XTTM84BIxjr2MqPdxOzZAXaAqO3yHfnYtFCxrkJGoVca0bGF9dKvi8s2gosON/C/IypCVVW5cMN79zHmHXjx4kWAPho0azPwjHpjB2H7+R4d9ve3G3n//n5DXt7dEIpXBnasf6QsrgKQq9/QZ56PfbpyXMgphnHwxSeEwMlW/kamKAZoRKEneeiAJLDShsJw90GZLF9gcp1+ymGboAPAdtfgUiz+wFrsr1Sqayca0I5gI3jYsSs0mtou6fd3w3IhYtkP35WV2udWrs7t5Yt7APPcG9fX1/i6jPe2S/bigW6uxzqCYhjXqP/8BIHA3w4yX0R+O/Cbfv/XgN81lPudfm9zqOpfAv4SwPe//1PaA4KqhZKC7kqsGahULShlCEBaqcWEWE3Q7WBw1O838KKX3NE+glefsv7+tftDRzbPhwfPL5ABgG1ZnN3zPYLb3w6z6vhcAUQp4mnHnb9sGEVwtdGEBedMZKzsHHLrUowlKAuosqhpA1K2iJjnZI5C0ytrxeICwbcnRZJAfYFW4fz2SAY+BZa1cloK9XRiefuGcn4L5UtqWSlldXdxSJNYohLB6vLxjlAzbsFMWmw7lIYlHWgGC9PiCfhI1djiW+1AvBrWkC2YyIjZ9+Bbdythh6X32qSdMImuBE+b90BNDuMecApG2GL6/1oVqdWtJ6OMglR3BKst1ZmUGIeo+SeXfOSvAX/Sz/8k8J8O9/+E2PEPA1+9Tx6wPy6EL+3/4UzH4esYblN+LDMChutf9HqD/+rfuPoX5cb64h29/MiG1d2V33RUr7x00drN2r98OIxB8JPDkNClhMEyiWNdUyuZ15p65mL/E4+DJ9X+sv2FwV2pUKrliChk1pIoJaEFzxFhCzkChVjmaQfqUcjZu0YMbLvk/Ravb5ifNky7ARwm48qUtPEYVc1Xjz7p7ca1urY3Y03u7m8mb7vIdXOim1fao2ERXqtm/0AYbAyeOT5URfgfYULA74vIrwL/KvAXgL8iIv8s8H8Df8yL/+eYevCXMRXhn/6Qb6DVTb8sdLgKZv+oNMheGDeSQc2gAsa+j5swSKT2mR0gqKPTNX2eL0hx3Y11r3B3v2+/ixeG2dSLe2M52TSiazLG7QA60jXDIlL3ttMSDkRiwVaqkD3Sb822EasWUGXNC5ory7miFW6KaeUXx653y4QAL+5MFnDDSxD4/t0dILyQW5MNIFQqZflNqMoXD4VzXfji9CXl8cz5qzecdOWtwqITS5qpmlDJZDGB48xE0sRaPR3qUi2o0GQjUZ11bNRC2Dq0ybN+lxhEbXGIrBydymjCw800DRtQYyJ0++yZPXWJaHT3ozRho5cIQIv28Lj27Q7AzBhwdBALTQmoFNBCUXMibgFgwiHv+f0PfLh24J9+5tEfvlJWgT/7IfVuX2z/Ddf9XPfnu44pDIIcH2i4xLSXr7Z3n2/XFhvBCCS2hZ3qvt6PTcuuNOTdTbg8ubCF3i69K7FUNx8fS5uxTCeZm2eibusKFWGEFcue5y5j6Du1phhwqbX4X6XWQtFC0epcm6CebdrOO2JrfyPJr+ZLf8Hz+6+MszNoQvZLoM/lNQBw/fy5Y78ePqz0O+q2bobMr6kGLwTQrZbo6AcupivHx2Ex6JqB8IdXtw9QMd6fUmiRadWg4GifNerbW9gsh+ABGEbWbDOgunW/vdjkTYV40egAxJtbUYMMZcbHEo1o3xmpAul9sJNNa2qEXg8KIe2+OQrLVCxJi+BZusyGojhfmXw407maBd66oFopq2kJlmL4cslnQDncvkAE7m4t/dgnn5rF4KtXnwJweLgz/PQI6yr86CugKuubJ5a68OXp0fwEns6UVKlTceFiZgJEKlkrqSq5WBxBdd52qTrMoRD4MoanWQ7GwMT9gJEbfnyUCyk6ANAGawZgu4tYecGdvc/C9N0ygd06GEha3WkHVNVSrLt/QOjFjCK2GkKZW/296lRcVuXaCo7j4wACcVwFqwMqfi+4/ebQ8LKqd9W1Ux/tyl/il+vHNbn31zsCXXAxNtdq0lgMop16GLB9LLZW2ciAugghePXIC9HzQ0iLMh4bST1yUUscolvqw1670tIPQsdjO1sT+uV+fp6t2uq5LLuF3lsK5H01f8ij7bpu1InSx1HeWetlde/cH8/X9PEAAe3xBCyAhbikP2ijipGYxYJKDCpFaSPQeecWURYwrtoGIeDvBc8f2HQ3VnvSvZNiAb33/bh44bKuDbrprbzQerZIsyHl9keeeFM91JdEaDBP016qW9ip9daIJ6ME8lTJye0tMAmbVuVcC6sq5/MdtSaONxYsRI8W1vtwd4sg3H1ubUmHGwAOR3Pm0PODNfEk6LqSHl5T1sLTFw+sunJenqhroZwqVRIl3yI5Mc8ZoZrkO3vw0sn4/WUNs15LoyEeLj20ABKGS+E12LZ1HzdDinW4vyWs+7iP5/Rxa5cDt38NbgWm31MGw/tj27bvOh5X74V6X93+o6WMG4CzBGBWCD6qVqORqquCq1MR79Pefxupyb/VY78Pni/xTZ8/f3wYX7f9zvV33tOGb5Fg+doVXxTRazc7TB3F9I0K2F5fr67P5EYwq+PjfQWX37ra5G/l+NBaf2KT9RM6vn57PwpKQDHvtYCii/tDn8XsBJaayODST8tMQxWoHq8+dWzXmaWR19ZtHInh2007MBCHMEpoh6d6STnsqMb2XyPpaC/2coMKwopFXSET6O22YltKYQylbZ3YfqPiY1GMghJ1vwsxqiGtyakEyKJIhjsslfjD4Q1VlVevLHT4Ui2k6Hl+BSLc334OwFk/RxRenG9B4XT/BlTJrx9Yysr5y3tqKSyPbyhUhBOiBUlnppxhnilSqZrRohZvMCVzFlqTJd/0iEIRJCNkBLLB7AMWRnv/6R57YT7QxJ4q21l8Vve/XRUyjvl2iob1Uoc3NiU3bRyZRlsD1VWfJotRUTTIVlWS2xBodVlYUMdIoxRElSSKuFdljF9loCKuHB8dJfDe48cEzB9CR+wp+n7+NT5+8e43a8/XP77tGq/Vd1VM+oFN+P8Ls/4kv/tdoxa2x0dBCYCalZM7q2h1zB3SN/dwC8uxhvWLEwQtg07oVDbAfQPtu1OJY42GbeM6XttSAq2l1wSGu3dafVepBOh2AFusEM8jkWTUF9l9Ot/p1zXuGyxfo7bkJqQpjHCKxWFIHh+0mHsuCOvZrdHkBFJY1BKLpvoZWRM/88kBRPht0w9AhLsXnwFwKm/sm8trUOXN+oVZ8Lx5w1oL9+VL1qK8XgsFeGSmamZVy0A0yUTYUkhNSBGmrJZQYzWZRqghq1jPgucvLe5hzEm4ENtRBizby8k75nQsTWvXOHeN/b54eXjnyvM+1eOc7kXHbvHRtFp+t6pbhdpfpJzf/Am+MwqV1cVobgHhocbH5L7Xju8eJfCTOjYkwNd8b3ty9fLv5PGuT++J0j2B08Wr/UI2Nz/sY8+Rw1btdxtzfu3jI+/ux0EJqGHgVBsXZ3xbhMwuZlBeHSpWD6CZRl7HeSc7d0g+ZJFxBX3TEzT7gYjH1tqynzHd/XT+cAQA+sz1RjQwVhkGC4Piv0WIjqOFhNlK4qrD7pANRA+CgghLSyk9vkJZLYw3gIogtSAVyukJLYWbFws5K9M0A8Krzw7knLn73CIGfXJ8ZXEC8hGAt6+9sS7B/60vJrMzeDCC4PW5UqrymuiU2flOupJrZlottVy3ZxCPYQDCZBhO+lirtxtgbf4hXccDA3W3vT0k+dxThZtiXT5zhYLbvjE831GS+7qvUwraPR/9ubixVsiyQuPVVKuhZrWQUK1OC53nf/SIw/YVm+/UpCTXj4+KEri2Wd4PRJ/ZtBeXIyay6xaU4+q7H0oa/Fjkw49xXNYhm3PdgY54yzbexUIfK5HxLSPLRXZlYnxiUbeVp8MifP54D4V6vVPPHBrN+EiOb7UpX6Oyb/rdj4MSYKvjLW4VVWJyxXXaagEpKatDTIdhKhuJugUezVgi0kGqa2LXVg46QNArwAJGST2MNgJ7CqEhH90Zo2q3A9i8GtRJsw6Blvyygebg/e15Dht5j6cQiUHCmDKjIEqSlWYnEOijCGtN1PVALTOH6ZGcztT4ynFGDjAfJ+MhP7mDnLn5qReICEf/RjpYgNGp3AOwvF6oKOc3X0FVzj+8p6pyenDcJxMRJUhEkQlyzuQ524cjVIBa2yt4FCnHbG0CBwl8pB0bxvxifOnYd3Nb++9AIvbyF4hna13w4TKBWM87amW3buJOj5rdw+kooQTQiER6pSMDReO2NjWiQvmDJO+mBD4aIABbUmzvobfH0dsw7z5Nm/np92wx7spuv/rOVu2//56i/QsbwPKcUn37eNPPHYAZX9OxQOgaLz41GEtJGEoNnoE+wuYhyFDX8OdgtEXkbfMjnRBoDVeaJ2Az7fZvqjarO73WNxngInSjxq9DjMn+xq5Ia3yvOFSGDmWuv/8Mgri811esXn2+b+s4eNdW+nMr5/o3392254+PAggoHuXWedazS73XsqIV1lJJVAoLqDmjSBVS8VSMhU1kYnVLQ+0KYlpQjNDK+viEVVUMWI8vsJvMATixfbJ7Hhf9eac2djvdZzgmufhrxamJ2kjtoHisPyUohAitU82lt8xmQydpwpTMM1kg50oulQOVtS6s9cSx3jPriQddWUSZpwPTIcNxhpSpL+6QnFkxS8E3b2dEhRs9IMDr3zJKYH4yz7WkT2ityPoVgjIfnigqlJKhCmURZBLSZJkNtGZSNQRXiAi55nQkRZuuHFzWISYz2g+fjfle49MHXoepGDHmZuoaRLogAxrweG5TX1CQz8obvG0jFNBdSQ1NjdfT4gtW/7MlENaCkgqiCalmJ5A0gLpbksZ4vCcF0UcBBNqm2QxKnF7jbMfnu1ubCd++s92Wur+5uXyfMPwa0vjm8NihwSDYtK4P2IrLhdV/O77QXfkq3dXK07f08mrGOUkr2fIEISlDyiSh/Yn/jia77thtsQGxlOVa1TPjKtXFURopwySSivnfHqg+h2R3Qr5nhw+GqM1boPDBx9W5/Hp1fPAa+IBqN/TBlQo22+MKRRDb6X3yl48ECGB2AjUswpxQrclY5hpJspJHUpUG4XX3Z0KvLhaThiV2ACHytMWGCai5Aw57SL+PRLRdOAMmiTx4MpTf7F5t5VMg+sAingos7ADQ0AbY85KsbGhPQgpcZGokv6ZCTbQAHMgC05mlwKrKxMzMgbu8gFRe8QlTnTi++Iw0Z14cDqQkvLpVkkDWxRZUebRxqH8LFL56fUZRflTeorXy9umMirAc70yuMxklkEpCk5KTWjay1SXeFaPcIqe8DMvZYXXdhTbr2RU20peNcmicpc41hdBSd/M2nF0gheFeYOjx+Xsogf1XVPYfGO6rgc/YuKEVqGrSr5FqbWwVvg6oFAIAO/wcfE7edXw0QEDHWY979I5uZHpDyQ+C0zYifrFHLUEz7FfQMyTCu2D9bpNfvm5lLmyEdA/QZS/A35YPLKDDOUC1VN617R5tRlVhahr7yZEzJHPWkUlIOZFzQlIiSSKJkJh8QZqjUPGcX0v8uorV8gV0028Jd196fMMEHs24C/ZsLp+ZyZESuF7C3txP7dXy+/nZFxzXnW7no533Df4+SvHZ4x0FYuQY53RoU4zSNVHJHhB+YGuADwACzyQe+TeAfwI4A78C/GlV/dLDkv9vwN/0139JVf/Me1vhvbPOBcanbYTgacokLjVWF3TFxpXdaqDnHxg3dQM0fSNK8JkNw+9WynNopmWsuezLSIZuZMuVrj3Q/jQBKhZnD/WIOQqlyQJ8eeS+8QC02KBVjyOgayEBJVUfn4KsSn4qSC6QV3Q1zDwdknkd3iaYE4dPbznMB17c3ZBz5ma+Q0S4yZ8iIjysln/g4d5kAT+8t2++eTyhwNvzI6hSi6Ulm0lAskzCIqRkAshcC0ImSXbMF9oc052r91fbwPShDbvAGtSdz6nu5uhiC+2mks38bB5cwokGO3Sz+a5vum17r1z26x0uCkvYWLaC2TsExespNwYZR3hCiAH92DDjt2LdfAtehL8A/JHdvb8O/H5V/QeA/x34+eHZr6jqz/rf+wFAHG02/E/Vw1AHNrGcbA4i6AT/8M5FhTZ6l5KB4a+hxUDR+7rG67F9w6s48AjBzFC2PWN8dqUdo4ByUBeN8Q438f5QkvPZgoeXLvjKiQWhROw+RY09EDdUSQJTQmb7yzmRciLlSpoKLZagu29rWah1oS4rdbGIwawrpRRKKd3blYTiNsoIvYXBEvnoiJt8E82VPg9sTjfnl2vm2lhup2w7o5e7VPf3tdc16qgCQO34z6/5N3zjgurcN3zfJ18XF30Y2roBFFd7fHG8lxK4lnhEVf+L4fKXgH/qffW86xBMx716ZFtqIYJfoljEGSnMwSMFqSm1Uwy+kCIxcQihIumGSFgjDliEQWYQjNNutYW99pDGctv4YSHaIo96Yx/2RV2dYhG/UZ17C3uG2CC1Guke0v8AHdNkdWWK92m2lns+vXo2Sol5QmVFpxOoR+VNFZ0rVZOVP07IMZNeJTgIxxcHjtPEzYuVNEGSA5CQ8oAgrKd7VJXltdkJ1LdGETzdPwE9B0CK/IEcEWBO2davmIWgem4BkrjFu82bJsdpoi546TKdrYBvf93n4JLfvpwy20t7IL+FJ1vZzjPl2m3dPX9my40aCO/H0KL2XK/tYIewBgM8ApcJCZw88CpqUFb+6p59fOb4NmQC/wyWkzCO3yMi/yPwGvhXVPW/ufbSmHfg7tbcUZsWbz9Aor7JvZMS+NUEhC0j+Tiem14HPO87UFv12j+1f43+/Mp6ujx2hbSv5bE3ERNkU5du3nOIXrdAIJ5LsD9iOqRQGZkkXzxisCKrRQzuA1LRVNFcyHMiH/Hkw4m1QqqJhZksCV0PKMKdywSKRwnS5extOPvwLd7m1YR6qWKpw/ZDM8QHVCMyikcgHsdAwi36ApX1pbyfgzbsz2ga9gY7cR5j3relbuZhKDXc2F7rxcn1tl28HgOhXd4VBkM4GxAhxnq/tI9FA1YxVnsLWL1OPV05fiwgICL/Mua89h/6rV8HfkZVfygi/yDwn4jI71PV1/t3x7wDP/XZZ0pVpljk0Z1kHm1VIIlS3cfaQ6qZjYBCLe4zrgqrU7K4jEFMEFVD3TYI44BmV173mz0YwMHUQBm0A17PZqD1clG0gJ10QBWwoWWK2cG8+GgAgdCa6BrcWyboJwSy9vFKClOJCEzN75LkSjtJC8yF+UXl8DJzkiMqcKqJuiayvCRJ5unpDlXhbrZU5efzV0YP3b+1fpc3gFL1ETSoF9DZ4xcmHyTXkpgdgHSp1hqS75BfYOrHgXKKs61tRre36JufzbGnHPbQuY6TpONTvVrfuJUvZQDburcg7RLAXAMqCiTH9GNqcq0JTdW0AzUiBSmpOp+vvtCbnYBH32xs6TPU6+74xkBARP4UJjD8wx5hGFU9ASc//+9F5FeA3wv8d++qy/hUbRs0Nkfs63COCMhpIZgEmgMNLvnWJiMMiiI1Tlo7HS597nqOud2k7RBRX1jRDjbvRcmGSbwRWyFh3I57nR2wws8Fy4jS3iuHQNUz/lYxnr+kBVGzqhatzDWjHpZNqkBNZEmQE1OamWRimo+QMi+mmSnN3IhpBlZ9QCucn6zHD8sbVOF0fosqPLmzzyrm7LM6S5OCLRMPD15ts+OxYo189UlS8TxDuhvw7gjT+t8m9YoP4sWkyeZhAI0GDvaCwStUx6Z6vfbVbR0XAuV96ZjLrR06DbChjRXo63cQmnrQnaAMmtC8vdFtQEaKaU+RXTu+ERAQkT8C/IvAP6qqD8P93wb8SFWLiPw9WGbi/+ND6jShlTW8OHSM3y36rqj7vBMDE8MhtcnFICINW1qNJjht6au9yh12uMhrt7syEl/YKF8vFpWfKQPwGQFCkBdbL8GWi3EXOSh+i/WKHOMkJrG3jE2Vkk9Y5i9hUkFq9vcLUhJThWnOTHPmkI8c0szt8Y5pmrhLN2SZOCRLJHouX1JL5eHJrPi+evwhCjw8/AiAtw4EzsmiDy/lgAhMyeQ54elYa7WGLwLJfAdQRavLfCSiBw/AYRhL9U1syLFN0mbuLqZhS3O3WdnL/mMPxpVwrc7rGH5XS39vB5D60nim3MYTVBmpHmN9TXaUNLxnccCAZx0Wly1VqvhZyGc2rX0eHHyIivBa4pGfB47AX3esFKrAfwT410VkwVb6n1HVH73vG2AQsoQZSGwGd4lVLW7ttqJq6au1JrNQw6zUIuJSLbTQ3LZhPDiFOyXZ3MjAd/v3lf30W/93AL7xtA3rbye7z/kA+dVZHCcFQhLeuLgdAKrbKnCVPLl2rNFao5AWy+i71hWtQqnZeO5SUSmorEw5k48TaZ6QaeJ4c+Bwc+Bwd8s0H5jWI0knWDNVhccHi/SuDyasfXtfUFVOiy3E6n7P4vt2zjbok64ml/B9nYqa0HKUa4SwNbwTHXB3TB0AgI75hnl5nhLYzUFsvB0lMBJvFxv5Yh04JeATf/HuM425XA+9T9v3pcl0mkjAhQRV+prx5dsMh3zJO3C4rL/JDNq4XDS0HR+iHbiWeOTff6bsXwX+6vvqvHhPDBIWz0hoUk5pPJCyAkrRs+UhrBZnsJZqSKaGAEyoRamFFlffEmL4xaAu1B0y3ptXN7i5AwLjDQG27EA81iYUNLsA6QtMaTtng3ui/PCF4merA4ewnY8Cyax4yOeVqkotBUVYiwlO1jUEgSuaE+lGmKdMzgcON0cOdwcOL+6Y5yPT4w2pJGoRqPDwAGtR6tsVVXjz1gSA69kJ0FSiEaYFcHv1CfNgzCYEQIph8qq9c9oSDg62AU1wKC1kYvPvGMb1Yqiv3NBx7Q83epLb3UwO5bebVIc6dfvOM5vqGgDZf0s3jfOpHzeyA4LY6OJIZ6PFlm4LFqzEpj8DEFD2avLt8VFYDAogSUkuyDp4QtJHt5aQVT3cNO5UMWNNnxGUCR8drSBWh6kYIUI5S9IQB/gRQ2bfSs0UJZZkx3TQM+/UhlUq2ykfBA3QZ2sjl5IOCMDYmqEpktyuoMaGipqtQLZ9SD54GwN1UECUOim1CutTtvHKFbKicyLNmXmeOR4z81GZbjJpnpmnzDwn8vlswMqzvstyj6yF5f4eVFken+z3ycb3nC0dWZrtN3to+JSyz6f31dkGzT4mTaW15dvbRh2BZQDTBjy3WPZyZV8HEq24b5RLwK0dW8Z32wf0qkxgH4i2GeQ8xz60goE8+reNGOpU8GW3tP95G0X7fQ1T+eCm6hbQ6HvAwEcBBMA2QG4x8zxdtdGRpFVtE2d1HnJCdQKyW9spFpbHOqtS3aBIEJc+Xw0goqCUxiLYEfYCLWKf/e/We/392q5tfUobam+Fk7h9QdkkOysi3s8B24QFpGgAMp8eXxhptQ/moe22CAwI6KTUBdaSLbfAXEwFOAkyZ6Y8Mx8Tx1uYDhmZJvKUmLNA9lwOqy+y5QlZV8rjA6pKeTqhKMtiUGLJFntQDhMoZAnbjeQpyq2fdfU9n3zlpmramyESzjgf3t2G3eJ6e7I9Lu/q1Qeb3I2bkr6Z+tVFm7wXz3994OWvPQ6DsW7F15gTL1gHJ8ZunShBGzVAEK+N59vP6chuER6z18cOPhYgoAqLkhxLaDoambNaKKs8VZJkbtILp40KSELrhGYfXIzkXFRY1e0KVZpJZKMkAuM7I1Ui1bNv8lAlxpuNn2yGPGHBaO/VzQR0EBB3Yr7A4VSoPESh1s0kbZychvMUObld2JkWs+OvLiyYyKgqL8rJeMb0lpu58OndwiLKQ4acJxPKzUfq8YZ1+gTSHVVv0TKxnB7RpcDjGaqyvl1Z1pVyOtvCrSdflScDREYAsFizWpTTlL1tEdDEsVKzePAkKCkk3RFI00bXxzyAXF8eSo9utBe27c1lt1Hk99L23b7e0Avb78btUaB4sZVGCuZ6iU25tjLkoiEGfEcPtZCVVPGUc2YoJKYLbJSghAbGLUlFwwmp9+ldx8cBBABKN3pRJjeesJVmQS8ykxzMEUUMJbbIOVSzgiP7IIQra484JG3CfbNHRhdn3puxbxPY900PMMb+1wYi5GLBhm/DiIw0Xqzi+38QKja5x7YOe9m/HcRJ+NcX9650u4GcEkrlWIWilSWdOaTK7XxCEE4IKVWLQpxndLqhpBtEbql6g2qmrGfKquSTzUU5FcpaqatTCHV1GYvt9pQn39xOntRYte7x6WqaHIAQJVR/ZgM/Ashxg/h4aqeh1Odpg6mvLmx/foXc1+euW2V9Bhp3MDzuoP09O+q5x/v716iaaEJjIUOAPMgCfL1JrDFvYKz7ce21jnwXgIAtJqEUS3lVjtnW3cl6+VSfmFLmRZrRpCxpgSzUKSHZBH9FKoVIz1zA/eOrE2LNSSLCU7cxNHWjpMBVbgEXobwaKWE/ocpJ1C6B3Ql5bJaGhaWYAEyBKtSWedVI6xLvpWnzLa0hBPBZnIxdqqstjjTNoPB0dgupfLJkogKHPCPTHXdz4uVh5uXdp3w6/xSHuxfMr27NGUlOHKaFlIR1+pJVz5zWt7AWHvWHrBS+WhdUlR9V2/SrWCLSNN1621xomUxqMruXiy4e+FtD1RW9rC27XMxBk91rUEaDwLQBgGF0G1bdkwb737iyeRjDdm1nS595hzbHnXjfzXe05UJwuD2aPcr+SWxidwuvg9xKxZyDhOrvV7eataOKsbrFVYQtE30Iu52yzqn7b1w7Pgog0DgmNa5Ys2MSWz5U7yayulHR6g4onrFGEtWBgPHJpat0nNPeA8UG4RuPFpO05Rs7dvbnvoGrlgalN1PbQlVtsYvBgn28otARS7tqUb2gAQtpAGm0ABMmU529mV8AACAASURBVLqz1IpIRZLl+qkixg+mA1OeOB4O3M633KZbDvmGeb5BZUFlJadiwtb0hKbFjYQKhROFwrnapj05Fi9iS2ZKZoOQk0utJWgjE5gGWdtNPHxMtI9LHx0XAA+4vqtPt+wVrfw46jL8v5m1DabX/bPNarjYtq0lm6srwshticuaL8sOh+imtO5Ka7eld0Bg7KMpmPo3o4dd6Nxb9L68Ax8FEBAqsyw8udT7dLo3SuDRFsJ5sc23ns9orqwvzhZZ4zQhWaiH5CrDgtZi5raC215bEJKW49THo/hsFldRhenuujNJLcEOuCS/loj55CRCOLu4eks2OMOgdzNuqR6rr7iDUDHqoIi5Ts8eWiqpYVHxtFvJ+e1lKd4mMxsuxcxF68MDkpWbTxYSmRs5cnszc/fZDdPtgeMnL3g5f8qr+ROmuyNTnnk6J5YyUZwVevv6gdOSeLo/UJeVL8+ZtcJXHqpi8f5kp5RuI/ipGyS5XJBUfJwiUgq+8YNt8oAoMSHBDkgDBB1n7+0oOjyNE9lC9iYND9Zt3JhXQISO5wP1oZ0+uYjMNWyubWVbqqQ1eQSC2xeG8trI+1p7vIXwIo1Au7VUUsSgU/O9CJlSGFR1Q1ptZud1/83d8VEAAXBsl3zbrMU9YI2vrDXZ6eoBsopaQD73p1cPaEEMWmzAC3VNTEZfXB4ro09aCADbgtrdN2nWxcK8BmvtG/HEAYB2YqE6Tdxkk/GSk79BYrZgsX4dso7q4ZhrLZaDzjFtFiFlMSehOZMPE3mayPPElMx0WDR5ZieTL5Q1sa6Jpdh4ryqsVVgbKBupECU5zZmbajXat3Ncie5XGknc+xr9dItAn5+2iZ9DYOOm3yP9q8dISfSXntmWl9ebjb7D7RpgK969BoAuK21gZijQNYAOCKQNaVcJ+uDF+tbNe+N6fYYyunJ8HEBAhbpM3KVPUeB0+MIMIV5NUBW5Nwn1SRdKKTyd76Fmkk5ImRCZqWuhFjeaUaVOFU1i7raSHCpqx9w+RKVd2qiVWOyO4Wokzwg2ILyTSkyyvSfJtAZ5WLnm9ozHRHDQJDRAFeLFcABKS6TmzkbuNYGEsR41z0YZ3d8iKDMPgHKSRM5weDEhkpinifnuSP70Uw53B15+dssdE7espFsl3a6k8wnqwvK4QlEeX/+Qp+XM/W/eU86V+69Mc3KWG9uM3v98cMx/cPuAkg0rufVfc+5xOGmsCQyi1baAg3Lq6kA3ixUzegpM3zdr2GbEGO8ENrqzON0usv6zwcoDOSCXz3Sovm3U/qTVO8oQhq9d2X4da9ThhQFWbsbJ9V5mOuzAN/lYFreG1TCFd6q2tkZEjKlL1eh4fBxAAFydZ/npJIJSptQ2WUyAopRaHXu4DkDNrrq6p1UlfseoKTEwu2nZYeFBHMUIW0OlEzEChP6erXF7GBTq2N6tUGa38FozBk3G3n+h8YQDW4EY5aTqwiPQlOzdLJAT5IxEsBAHRKZpqQiWrbhqgVKNjSqFulYDqKtR7HVykt0FgNE2Cf6pWIf3bKfNm3QBoEgbsy0+3aNyV5U2qmEg03cYvFkZtKLbDbmpd4t22eHNbfmY6yvf3JcPmP+hbruXNezuyR6Lj2Bv0LRsSnTtVntnV8e7jo8CCOSc+PyzO94U27KfywtUhZt6y7pWfrX47quZtQpfFluTc5pcUPhESQslLeiirm7MSE0ccyFJbuSr1hlUmFI2KOpBTG88zPVTKP/qFnrWYktuLaFitPttwVVImkjqfotiSVSqDqRiqANdACruLhnm0dlNcKeTYdtlOYDCzWQ1TKsF9HjkLYJyx1dUUb6aT+Q5c3fzEvIBbl6R7+6YPv2EPM3ADUs58LAcOKSJWTKPrzNv35yYF+PjpVQShSxvEVl5dfdAVeXBU5OX1dqUkhsIlBsgwojBKYc/26lpQYCWGMXGpZqalODlg3S2Ae1bUrpgS5QWls0HsqoB2At37n2i0t06k9DNXwHA7YWQB+huo23gxWACtqsr7b/dZAJ+MsTyUgxwyYCYevIa8cJh+6pErvKWFkIt8laoXJN0/l+ks7TynvhhHwUQEEzFFFRLErP3yyTviLQ/1Z7WIuazimN/OqmZtJfZzPxIfmmXnDbCsiHf7SSrL9owGoplIEGvxoAzBtbewueRb2x6Xhriac9blNj2Ldm8L674TCGJx7G7f1jFgoWKJF8BqY2dySXMkEqr/20Wso2VhTCrINL60ot1dkrCg2hs4RUUJN6fLRVw7ZBxtmztB3/9DDHXYMC+4k4ybK83Bd6HL6+0+B2vdYC/u3GlKc9/+UNw+AC4GNfG16/qowACpJl099N89uIVoKTlt1AR8s98gq4r+Tdeo4vw9BsTlMrd6cBU4bYsVBHKmllXoaxQJqgHMUEZhbmezWfdef6zQ/haLCzWk6vi6xSur4FuQhaQWhtB0OLuuS6Es6AYYlhck5HWYs8jnLbx1I79pC+qcIwpYtJJmQ1oLIun4r6xT7892bUuFtLr7ZMBhrvpMxLw0zcLOU98r35Omg/cfP/7HA+ZT49wmCu3NyvzUjhMJ2ClPBaW0484n9/y1Rcr60n59dNbnkrh9VcP1KXwdHu0EGDZ5RbFlsrRZSTTg41fw3CLRxbKq4MlK5+S9zfsBNz9Ncysw/W7NpYiJPSYqTFYtCKl7XKt25WvJQB7UHu+roJ92kGB5kMwLMGRhbN+0d5pEZHiwfDylegGw7c6ixcUQs+WGlSQVxZyjNZ5M0TTpNZfGRCM2LKMcQwUNpU+dqDoZHOQ5fBMG+34OIAAANJ5Yad3zMJNSNndgwfec8+DyvB3Cf12HF1jmC65vgtNTjMjfgaDtcX5YRzYsyWfBeUjJWoLsFMSaVgcRja2UOGDm+6GmGp6eqedPEx4yFNCmKmNjpErm2Vo1cbUeU95fShX+nyfL1DdpsyW0rqYug0pcIUPuCAVtvdb7a3iLS10Revf3gqgsq/7gigZgMmzTj7vJp12tT/Xp+ePjwIITJL5PH9K+t4rAI7rLZXEr8v30HJmfnFPneBxTpynEymvpjUoKyKJlDILM0u9ociCamEVsWi2i6IFF2QJui4Wn7EUV7OZXltng6aRvvvGscuajZROs2PnkOi7zKCg7j8/OUozAx6P+kCLArPBdLbNwhCwxRh5siksgck8nl/ykF6TPKIIB5lJAi9e3kIC+ezIPM989oNXTPPMJy9uSVPmmGYmZqZ6gz4Vnu5XzsvKean8rd9a+OLtmafXr1lPZ754eMtSVs5vHtBauUtHJAmLGjkyOSVwd/JxWey6WTivS+uM4mbZ4r4EggNK13eruKw73GHF/eVto4aVZ8OOdZAR0DF9QwShqdlj8VHAqgNQGDZVfMdYqs562G2bx3F7ygZj0y1LvdLA+PE87wBKbt/sbTAgbQ5EOcmgXnQWzimopOZkljVZ2SKklKiSwWNLqnpsBwV1YW4IdZ873htyXET+soj8poj8z8O9f01Efk1E/if/+6PDs58XkV8Wkb8pIv/Y++ofh2IbnHqUqQdK6xx3YLnneK7Lq13BcSZ0+GvPR7LvEhRfDqu+8/IdN7eP9gzv1YJ7nLc/5NrQNKpT/TPxt613W/9IBwRdcEkC65XzK227VuxbOC6mbl/9N/jWt9W8D6tnW+r9OPxbHDw+jBL4BeAvAr+4u/9vq+q/Od4Qkb8f+OPA7wP+buC/FJHfq+phaJ45csp8/uKO9WDSz4NMVIXb9YxK4fhTn1HOgp6PzJyZZEKKks+VOmXKPFPVrAUl0piv9vP6YWFZizm/VeV1OlGpTGfTsacvntwA0JOgPjmGdxPdxc0CJFlwjXMxV1nJbsfvGy7XFUqirgvglopOCdSE2Sw4tmk2B64NqJgqLRJvlsWFdcnDe4th2frSqJHyZIlGp89fIhnufnBkzpnv3U7kY+bl91Yim2XWhXw+sa5nVj2xPKyc36wsP3zL+vqR+8fKaRUeyy2rKnI4gCrlMCMCBzWtdEuV5tFgq2OXeydbzhOA2rgolNXHhQlU0WKBYZSKijRLTRCLQ1gw77jwCfMxQQZQtAsr1sjxpjbwNjZUMvLII4enw7V0DmkEgm670Sy1NJikUbg5bFe36Qg5RrOKEGcKXBZUs0HhZtW341/tzC3YUjKB+VQ8pqwikdg1GfssYnEaUWFWG5Ozy63EfVGmcPl85ngvJaCqfwP4oBBhwM8B/7GqnlT1/wR+GfiHPuzVdzDFV0rtUd0e5zc1ynN1v4NIuHp8DeD7IUW/FVh+IQD5gI5cpTR29745K/+O7/4E6vzOHd98YN8pR3rX+QccP45M4J8XkT+BRRL+F1T1C+B3YMlI4vhVv3dxjHkHPn/1KYc6MT/dAMqdWMz7+XCLSOHudqWe4f4k1OnI+U6QIshp4jRPPNweTCq/KtMB8gT1EeoZvnrzhvOycHoyu+wvHt5QauH8+BZV5enNV6gqORsGT/erD4zztj5CNZt57dNqEX1nv17U/A5mSagWqq4WyKSudIxl2MYE5dLNZ9MECFkdkvvYlNkk6ZPP5unWCak7g/Cn+UDOwt1PHw1bz0dymlmnlzBN1OkT1gUe74WbJLzMYsE96gldn+D8lgNfcDc/8LAsrFo4SGUC3PePOZsdxdFzCCzWETNCAtan6JtbEjY3ZwUi5gIIBa2KrE5SpR0WFfPHqCukwJIYjxsudeG/0TE/Xi5kBrGq6uZHIoqPM+jiFoVj3AYjACJoR1gkasfgcR2tDvmONyKF0ZRTACk0GtHUHN+J/kZ560jSXq/EbaeELElLgklgUgujAcxOCcz5QJIJVXP3Vl9zs1Mlaz54G969zT8kDdm1498F/l7gZ7FcA//W161AVf+Sqv4BVf0DL+88WMgOXYyTrh2tbzDgBtvL8DieSZBNtN8gFRq1IF0K0TQQEaB0PMaKr/dq97s/9u/JO+oaScSt9YF4/6WNw6Wj6IZQ2pHDl23ppTulJYSNwDiwsvt7Z1fG+XondvoAWckHFv/6x4dU9iGotc/P9TndDsbzzr3Pk7iyv/EtHd+IElDVvx3nIvLvAf+ZX/4a8LuGor/T773zkKQcbhbOdwbCn57Mq+5hBk3wG3c35INw94MjIoXDdEMiMdfMDRO3HNqGuEmFQyrITQYVHl59yroWnk4rpSqvlwNFV56+vEGr8unLAwqc778ElNe3Nsg3yyMKHOejZfOpFj/g7WxmygeK+e4vnjqLhSpKSWZ+q7VQJZldt9pfACT1ABxzPpKSB0JFmIthorNL3o+rZWY6ZouwdJxNUi+fvyQl4ZPbW1ISPv3sJVkynx9ekHLiZhXWdWU6P1A18aCZkz7ypGeetPBUhTe3n/J4fInewLTC7aNAgem2IKIcjgZkpvMZUVjPRiEdDoY3bkNEn25QUZ7EUpbrYtqU1U3AqxSoRkmBIqlSwmgJ3fLibs5cqjk2mdm4uO2GEJFzGpbdAeUA2koIcgLLOiVV3JZ+2PeiQzAOap+jwMZN+u8ZMJ0Xb8bgKfhv1xq1kBDWljRHI0MvYH+1SWYNvadEC8EoeBDZlNGcYa7o5O7ik5KzyZYmMe2DLD62c7YxCqrELU3J7/Yd+EaUgIj89uHynwRCc/DXgD8uIkcR+T1Y3oH/9sMrtp+tfFrb/yPUHpCgY/IRs9t56MtNiDLq0D0OgQteUqMWUvdbaEIXv4+0Z+2XvhCvQXbZXXSKg5Fw2eP54d8VamYQCI1/AUi27di1SPt9CWrTv5HG9smuPcK2Ta1du7H3dzvhMFAQQ7uepX02ekA773hRh+dbqvGS0NBhp+/1Bh94yFD+CpJv6+9rHdsxGOva3h8o1uHNPeW1+b1+64OOb5p34A+JyM9io/R/Af8cgKr+LyLyV4D/FZPP/9n3aQbA+KVzyiwen+4xG9Y5ywNVlLdSmFLidlrJKBMrc068mBOahJIzU8rMKXOclDkrE2bHv/DSePm3lboqb09Hqha+ePmIqnL8rR9SVfnytZU7VtPJJzHJ/Avn+cp5RVW5nSzqLnVBtXI+rW5ws7KqBeEgu1Wh7yzx+OdZqm8S053Ps+l5J49lPDnkro5ODmeXODt18uLv+gQQbvIniAif370kp8T3Pn1F0sSL8x0C5BPImnhbKlIrUiy4J6pGQR0Tn4kFH7mfYSlCPiakwjR5QtHkbZwsUuh5MnuK2e0lymzYZXU99sPqWZBmMz46rQa41+ragWTuXCpqav1a216VtSJrBIRTtFh0oZTNp65Y08luX1EjgIlbea4lKAFrY/DZYUlYw4NUI/ROUCE7dJO0IYK+6f1cQ0Zg/+UwJPUI0TnZ06Nj/uzlktufRN7IoDBKM2rowD0p5ujliEpTMtIiVzQrOlksyZStugMzmUyqC9RqwXgAnT35rq/D+izLace3mnfAy/954M+/r97NOzgJ14J61jZPKGipUDBPN1GUFdVsakEt5kGoFjikVHWT4RDnmIV91drj12ubV2hYz7Gjb/oIMd6gtYfQjl8QqC43cBJfECS527J79KkLd2IxGYS3uPuh5mnf8vFoUXqa0Cna4i1t1IC91GMnROZaBmclhr8wuulofC8vsbppIQM7yb3DftG2OmL5nbxCcdI52u7j3fbViGkD9Wk7V7/dckQ08dqeLvTn2suxKbET1sULF3vDa98VkGh39DGe7imDWFhDqS327n1sEa7HdYE0t/NOCeguavW+X7Ggh/Oh9Xq9o5vjo7AYVFYe64+YF+d5z/cIcLvcULTy5f/7SMUyDRWp1MOJlCfu51tEbiC9BE9ocUjKLEo6mBvtcZpISTir2W6l9UwuluARVW5uLWR2Pd+gKIfPbLBntaF5KysKlHVB8TB/qnCaKbXycFrRqizLimBJUy3US4UklvTDN/yULLvBPNm98GScXfrt0b7RegSEl+G/cPRoPi/urI83s228eTJpe32iqHDSxSo4Cee1sjycWYuwrJag5VQn1nJkrRmdjmiuHO8yhyrM00qqSj6Y81DjWUPn3DIOeciziKy0WNtvHt3RwVPEZbe0WzxGQD0IJrYvrGvlFFGfqrJINf26WialSDQba3d1DJ6SaXAiqlPkZ1jW2IJuuxF8vL/fAs00m4+A4x7JaSDIazLqLPTvktxrLzz6sm300NxIBFdxKi5H+vjJas0hE4g2+85cwt/B7U1ywiIJF0voOqUARormguZCTZZqzJK9Kols7UqPiBYitFueDBAmV00sz0ZnseMjAQJmMBNegGGCqVpZtVJ8AZ499HItFjkwZ/eLr8VHUIxvFCUVSKLkapNZHBOaZ5wJT0TVBpug4IV1DrNgM2vNxCawts1YEA3NShJhzpYQlJo6dpNq2Csn8EWVxJyMbPMPQABhyuKOIL54fcGF2k3DUkc8sYdPbm1upLZiqg2kud6G1iswiJgAM/5MChWfNApG1MK+S5g8D26uPe+C/cYCi6Zlp4/VU5jnXJvJq01JYHcbj5yqmQMnSFWGAJuditiukUs8yMUd3V76JurCxIAsQjiUyUDi2NgOlBH07FXtvR3VtvkdqYTdHb8OQrJlombUSPXvW7ulUSWjNGETo2JDavh62ZBsG3rm6vFRAIEKPABSnwBl4YQCr9eFWitfPX4JZH7EJ6Ar6BNpykxHJdeFqZwMGEjhkGxjH+6O5Dlze7yzmPvFSLBXs5BUmMVI589vbHBuHRo8fe/OlsrZfPnvPAsSq/Gb52rRd/XhRK2V09NC1cpSCkXhXM2zsGg1jJJcEEnPizjnmZQSM4bRZ8c2Z9dTn862weZkvgML5rsv00tbKIc7ROFcrC/LeoSqPD0uSFWmUlnWytPZItKsukIqxs+XxXwubjMchLpmak3MMiEFbu+M3VhXG59UT7Z83COtuA/BjYeDL+4ysM6OpVejRm6rUQTFAZH5CFRqWVlLZZkKazWLzPN5ZVkrtaQG+A2euVdgcR8MB1iL75hgo4obCiQiMKtv9wAGzavQMb/7g1jGp2Qh0pJvwOxAyreZUQWgng4+sPMUruoOJLMEcLTfg2sNZufLk0O27FqS8+J1NZbPXOkl4RGEHHwFuymVVSyqdrAdc8QYChsHy1RPcSpi9nwVEQLuueOjAAJGA6vRw4As9pu0Qq2uNlEn/+wvFUirLXrzR7X7ms2Csy4WxKLIAsnJf2Alk9VkB2DURrRBwRyT6MBffMDDc0/EFldNtghzTog2J1Zy1RbsQbIDAYf22Q1v8uRaCyPqXYahLZfBlHGS2scnZALJBGWTU0ZVqy2uKoa564JWpayFUiq1Lu4r6FlaMfK7qgvPKqARajRwRfL8LubQImEuW4MN2GHdwNotEKLR3BJURu3FErY4jT1K7gyj5JSoycatqlFMil2jSknaja0IbNqxYSccOgmwxXsda3YJUA9v1tFwJ/sbVk4dq/uCse/0wI9d2EA/7b/BqkSbaqe+otywzrZ8fVBPxka5bzpmGGSh7hp95ZSuMz7XyKVnj48ECCi6rLCaBHV6bZjg5rhQVVlUTZp8/whpgektSYV0n0FNX6B1RevKkjIlZfRmJk+JKvckyeTJBV75xqD17GR+cWm3xxSc3Jc/LL2K28YvGcI+GwE9JpREPmSqwqzmUTgHFlNTIUuGHgQkkRCmOSNJyFWMKl9tIcWmPzqfOEeGIY/CXGajDLLb4Z+5t+w09wdUlfNpMcvJhxMr8GTyVIetZqtQU6UmJS0ZUWGSlQwkim3K4irThwJVW96FabJFmiI/wzQAAQHJZ4zvXn0DxQCa7jtZdYjHQ5zmiXWtnNdKTpV5ShSnBHIuFpWpGFWR1TBuqrGBnPUIGUBQADVSx03eMNfSxPbzNgkJSc7He8QqGXWmyV61mCzhN2EdtTYpImZbqb5+YtPHOlo9MUxy2ZJkk0mFViE7W1nD+8QloIpb/6EWM1AE1WLBdEuFVanFhIXnLMaWFYsFV5nNL8O1IUcHXmWAYdeOjwMIYCGjpBqUm2tIRG1sJrHr0iC5gkrje0cNUKS8oyjFMUaiMrkufcrVg3/iENjDYvlARaz/zo8HT+wkQCCCrI1cE/92YFLXxtkCynbPuWHD/MmAQAspH/G6w7jFsWiKuH5BCWzMXbsRTcI22LK6eq0oBWVRXKAKlrlpwDJaaaGWXYhXUIqaLKA6Rm9x+3wOapjFOpeuTns3kUbTiNn9bgQThQxjJQVNwpTE7WisfFUXkomzEn6tdahzh23DFakRvcFTt7YHP94mvdlG4GtsK+UPWBAzNhy7bzeXYpfqx/MSz4nfkccPMt/1HdLnVJ1qFGcHqgpjyvqQmYUGqHqMTUJq4rYUMR9B0b7r+CiAgAILFXkyH//bpRjZeGMTfJtWqmYeXGiSqi32YoiPpRggKKqhfSNXSJMys5I0cVMtDHedAgvHCBk2afzlHDyetS0it0bWYkvv0XPxWR5AOBRxQEVTV4YRkvF0+KYTUs5YSu9I8WWLW9yTMRIRRbrv1piIb4gFYD2vRrouVShFeXtfqLVwPlmWgBXTViyizDlxyLmr59QiOcfCW3FS3RO0lmzJTlMyjLd6GxzRtXTptZolnIe/86ZKy0konl4uZQfaoXvEiLElCaskSqqcS2JVs6fQqpwlGc+9Jouw41Z3ZXVVZ8TVD8Ad0MFVxCkHUO07N6I+mUFikP7BnnViPCNt/kR7CDZ1QNVCPpTYbTFHDrBHgN4AjNo3VSjZCfkGqJwo6QnuWqwFxTRNkcqy1urCc4unkWsxi1YpgOfYQBrS+E4kHwHDcpFaMFxHLY2SIHogSeKQD85vHlBJ5DqxJjPNrdXJxZSRlMnJiO+k4ipU0+XXEjramFX7Kx70sxlxOEm+1u3kVm+dRhbe1ANn+rQDGvZATdJrPGBoMPCN76Rf+62dRAHKuJDA2ELwdN8edkoVLZVaPQ1VBa2JilLGgBg6cMGBfBSCx63VNl6pFquw1kLLGiR0p5sdChRPhlHd5DlkI11Kou0bIj5uEjYVZviSPERWFqNOzAbEhHOqSj7ggkEDAtnHOhKyRjbnttfDlNeFcs3uQtw4ODb3FMtAjRprZL9PXADuAZ02+4tYODpgYBsQGx5fN0VXkyFVEzDWkvs3YhyDiqpO1qoLQnFgpxYBqhaPAOWfM7lUAld/N/uPoARint8T+eojAgKJNQcZ5BFus/VmqncIiXl+gcpMmU5ISSQ5cFZbBHVNlAXzmErZLKmQ5himq/GB69k3put7jWH2/HcKq1MGT9n4b63/X3tvF2vZlp0HfWPMtfaput23f67tNG27nbaDMQpIOE4UIhEsBAiwhdTAQzAPYKNIKFIiEQkkDOYhjwEpkYKEIhnFwkFRHIQD8QNIhAgUkLAhNnbbwTi2Qxu75XS37W73/ak6e805Bg/fN+Za+9Q5VXV9bZ9z1TVLp/bea6+91lzzZ/yPb0h369VP+cvboKW/AnjMJxeZtqVMLnDbyUBtSIq+ZyRowEOCmAY5BTucpcO2Js7WWXswhuwOW9UmJLLxNhIYtPgPUAIwtxlDHgCaGVrbRVfmRCTG6IhIwGgoTBGBUUFSVhiLN+TgTdw4qyyxJKn6PJcnA73QqBa5L3BTnTwXPsFYSISb8/JNmAztNDl/IjGEBjU6MR6WRe5c3TPKIi9O7+Wukz2DjJL3zZFIGzAPVfs1niAikOmYCMlQHzL1vIlZOq8qA3lFUoowVbk2Xc832iu8xD3FOpgbx1T2oQwiLI00hAfSA9GZKdu1njwbEg0tF1j6rD+wLziuukS9v709DCKQgI+EVwAKKLY2p1+8LQ6Ew+MEeDK01RtssHpxB1Cx/2kLYDTAeTI8kwWItTOjaYyo84YW8RBFDxl6rODEur6vkmDi1q5kk7nAFKy0x7TtkFgzCkxn7hZ2utXCCrqM186KWpzxAeK6IkCzoq+u4kZjyNJYq9G2M7mOJxNbnONhYZLXCeQBgKXCsWMMDnH0HOJIer5WVn+J3j45REr0xQAAIABJREFUnYxyY+wczQAIol01shS/cfAMWJvusDQAjbEZIxJLyJbR2KelS/8dUpc6x2uTinJWwJLJilngqOWa1bCjSnVXpXc6VkIqngwPU0HfwUb2GINEaea7Ho9JuCn11Hri9728UhLRy87j9cOZ5izuXj/MmuekYbAMYFNOhu51UGJylyoordZaBC6wIG+0B0EEDLQeuyymkOC6tkVim8PCscQVJ/N0AnoDxgo4nQrRV8QWCDQkHB6EHR8Z4qzC+h9V3acDlhhnWpzPGtg80/FtqggcWmCiFWgzyiyZkARM8RbGCEErWO/J00sKqIko7e7MDaioxHRy+FHcQypJOQ8hTle6Xl19sYRZ4LRuyBHoyzWaAbZQj05bYNmAnE5J1W7MaT8fQftHhOL3+6WtpIH6pkMcXoQsyp4hUSmFlpRRzy83VorbOS3aS1sQllgMFMthWIN7YqScYtr8Z6YvYGwkmlcbsRwKQedanL/AHzavAJwat/mfxHkROuoGCicnIaD4TQJg+75Dbf4KNhN73b0CE+l4dznDAFM04+ZlTNnXEYkA1aOmbg3Vv3CNQaBMgV0uW6kDxjnYAdp3AjSRk9TNjLsJAPBAiABATh4V838lq7smMWLAIjHAyMDS/bxRt6N4S05FKzRLayEktnuAiSHkZIZatEk9FFAaL9DLGHeg0ACmIXGCYnF0kWPMrY3y2WOnwFn6gRWYpgw9uUNVtbGQ05Uuq8q/tQFLPqpNXwy3KP1mQKSj+4pEYCyJcEOu7It7g4fDo2FtjqUZ+qjCKCbioOeQF66yvsqoNGapq8prUNTirNip45I0Svap+g80txuw+PyObjlMIupSPYqhwbjJTgrsGgq4SacxNqcdQsu4ctUqlbYmpmIcQF27Cp1kBCXAhdWYWHRW/SpJII6FYhKVv6HaWGjqY62TMggWuMgkRDIS1WhFESLFiNT9ajwTJIajZAyre5RnAbAZNkwnbyHXFlRemTWsSufd0R4IEZAoX7h9J3LVwmUbY0icZHx0pnijVwiqjDdON6MpYo8ZV7LAl9huVfG3KyBHK06cb9TyL6oqKJym+PXdZcilXr8zYecdBztLlDRJJxCXofyoCELHhNWvSr9LYQxpQYmyF7e5dH0lOuSZcObw9wXMKFubwGkMbTQsvWH1xKLMtwBU0oz7JuqWOYXeGeg0SrTUom6VoVbqQyk7TZs8jm7RJB10qiMmiaQ2uUlVM6lxNIwB0GZ3bYUmi+DwwgeQNDITVSU56XsaGjEJtSet6ixgC4UqU/pBDGAMzGAh0DYwDaKDnNaaEsPmUHHO+1T2ighIbarYBK/YBmBXHhKzkI6euQhzJqWActlOd4T6lpLoLGkI5xgLOVvdryCiIkh3tQdCBBTTrtTQrZexqSg2DR8ejSmpbUGaInGckXkeQA6Dx6AYN3Y3UiJmMskW4nga9LIszyKOalEid+l0k4SLnoi4poIODGCa7IEOmIiVgymqVQ+wTnBfOFkSoX2l+lPIFLWxJEji7BUcI9E7Bna5QJzPAV9oB4AWbHOmVbfmWD3RwGKt0wQzCdNu9a5FXv7tIkS1FvfY9xLLRIgbwTfcTWK1rNxNqkJrCDSqJ8AU2206YRValQzAQtKaT8m5NqQIUe39qDkrdlzjQQI+oz2zxGufQWgXBLU2tzh7TtHIZAtJcVXbJcTpfqsNWJtfw4Jjl0wenUP8QBa3TuSgkdfk+485D8qNCdp1SqqtuCbPyo+IHbA2K08EM6HqrvYgiEANgsm3XEi0U9LMJl+tHtuEuGK+F98cRxrJRBjqcz4XM0BDVhGBfaKlK87e7BuwjtcCm7+RhX+vdsNqOhWdRUmBfWnGTdSyXGO8DfPX9+DXZSG2HkVtm7pcLbQuG9CMqZ/GnqP4azIG2uTKzRsRjIJuuGZGYrVrLrMd/eGT203icGPSDuM3I24qJC522wA4DZA7gPOYjCIqIuDaWKZNT8monk4EXYNbHovKxhlLza3URxn4pn9cIo5ps0QaLCrWYf8Lt3mPrDmcRoFQwJXPRUBJqYhjeR6K11+qciXlzfVWz5Y8O4wMJ0cWmaKskKkKTCrCWyHpGqsKLKq1hkyUNaYeP2eI8+3tZUBFfgDAvwzg85n5j+vYXwPwLTrlIwC+lJnfamafBPCzAH5O3/1oZv6JF90DRl1yPRowDOjMosDWrmHhsFyA6PBO/uQ5aAi04l2BsA4Dy5OxLPYgJYXKYEkfLzWgFaXxEiX5ubhvHMR/AEi562otWFBUDfnX3bRppqE5yckMKN9hZeQtNC8SFgpGSQA2JSLZJvd7T/GDn8pAlUrBK4HTmsRvJOgcGIIxo6sOGbQ4Y2BAGAQSuWds+pTCZO3uJeZKDfAK6ihxt3RlvbooVupIEYlJ0KVO8WqT2KBeU+OVDI4hwSnVS8+bHOOl1KSLVOLc4ytclsUcNLiJ+67qV3hjDEmSOA7f7TWTYVipNVZci4Sggi6lCJT9rxbQorDhgh07Fi1JqT4oImep46m+sD/wgLVABUhYqz5QMqn1nXONc+4r5uW3wzD4X+JG3YHM/NfrvZn9OQC/eTj/FzPzW1/iuhfNalIPzcEJDfmuoxIoRgJG/a5Uhkm6EUgLReub2EmARqObbpbd3eIlxs0Q3b1ffMODcTQ6aTEFLTqkykr+waxBWJ2T+xmi5kW7TZxbEkBdA5h06WClq81Rr3P05vOT1hSHlaXZtIi1OTKpIgEKQkH1MTENGbVYtWNCCf5lQsk43BeAVcp1yTXaJBV/M7dUHnTuw315S4nc8xm1QUrXn+65uletk+K+efHZjuOltVFONeQuTjfzaR9Io61gqm5Iif8HVc64csxscn4o9sIqpdrLyi93ZUEMxqVMNQu6TuKAaTcKkJHZxH7j5k/b55f9CxHQIoZsMYnqeyQCmfm3xeGfaUZ5648B+GdfdJ3nNTOBKIjz1wYZoGiDa1nOeof5gNmmBFADs+AGIjpGdmCcyUJ7Qw5Dy42WXwbG7uK3DDW2KEdfnG6sRQQ0Kwq0qe9DBp+5IU3nBnd/gsdS8NllNit9FwaK++ZYWqPPXLaB7os2LRd5ZToutchdUF3okn5iSkx1XwDI8o8nOfMSFH8ZvMQS5D35tyVjBLYgQlMqYk3OSI0xppJbm3qtzyJYBaqBlfderSmqTeNR1JT1sye33zdESR+pTWA70VIAjbniKLTRwmRMO0prtd9R3pXcOSjoiiPYa8pZwQShGI4YjiUY6FPMvmi4iTjmUYeyUj2AXo8nm25x/BXl7uXnPnKnt7oHJQ3q8y0mrZj5C1lqVKGQcmrRdc4icaTbelgXOdOhy+5xV3uvNoF/GsDnMvPnD8e+0cz+LwBfBvAfZ+b/+jIXMhyjNg9JG7kzcxNqhk2qR+w8C+NGjwCCC9kUjE0CsCdZXNyxuDAw48snpNfkcPpceqUXfzgQAxxWy+Q8u5Fpmttq7cjzkYe/CjRKk9XXDpR89tv26x6+KcPPTggMR+JfmkkBrmBKUHwtiaqOHR6/6Nw+avv6B3AwgE1/VLliFbteQ3IMXDn8PwWlfW9UfBElGuwEo4x7sDKYHa85H/0wdxoL3d41B81InJpEtTaTf6hNt7BpLympwSbRujEAapXeMV2DkmKmm7e+n2tFzGEGm2FKjsexT90z9jvzu5IYElOyjMN3uDE2z2vvlQj8GwD+6uHzrwL4hsz8dTP7gwD+OzP7xzLzyzd/eCw+8uEPvY7RGjB15XFY0ExnNXEsAELu4kRxizsiHSMckI2A4WiGPqRTFrhnTUJF4VVGjHysBWJZwXo1eSWROwThJYCp1uj/byP2SUsF3wTDP8toM2SnaDOLkFxxNeICnGJIUqHxcqh8+pCx6rw9IZfernFkec2wL1SOEP3qkkMGPwBDNgEQfj2S30VQUA4LhMa+i3SVAaswDMomUFGNJeq3sncsHJerSdS42MM6yheeTptEZXyGQnEh8JAot4WeL5TdNMV+PedUl6IiKRV2PSrvo4yUNXe1oTgPo1H/T1f8wEomMhmGpDw7iizaqYVDUTa3ghvzWWW2oha1ySeR0HrquceQIKG0iElMG8jhA468CuRJVaQrL9uSSXIJtE6j56JQZkYp7qzCXrDNf8tEwBjR8q8B+IN1LDOvAVzr/Y+b2S8C+EfAKkUXLTO/H8D3A8DXffxjNa24yXsuKNnkGHtI7vybn6U7TUpoEsYlWkodmHNa1Gbest7X5OXepdy5/iEcZkYP8lJJ2CwRmgn8kLsEMGHN1ajySfXJg06rjKFyjWXIDZVFJA+8wXZxGNp40+VXrDVKgonLsZv/jhJGofsUN77hkahNYLW1arxKBDUZ0eV+rD7NQc8pnZRV++LvcN4hs4xXOLLKm+tDv7vtvP15S5qSddeTz5OUwvZn2+e8RnpmQV4QhZ1Lz7Tvw3jwc06JZPajPpSMbwdJVKKBw6tshVSrmMbJss8c7Vs170ciMMfzjvZeJIF/HsD/k5m/UgfM7GsA/EZmDjP7JrDuwN9/0YUMNF4VQANiAwxYeiMkVZxh6doMHdavwWTPgEsXR3QgOpRAO6elZbBks3FD3cR0L0NgYeS1UwXPlBogih61XQQSId/+onJkrRdJKNRj5huMTmNmzgADwFXiq1x4qQT6kC6/ySL/9sFangDe7rT+VspxW29wulJVjOt4L3ARdD2NnO6i2phEagYKoLCWTgj1NCV6bxVZWZ4LuzluzMlwGcjcXPuHEk+FD4f0/TJNwkoNiQktPqbxTGNO0/mUAEbFkRRAawVxaPOee3FxWWSSG2Wp6wnLoWVMAjTVIUt6OWWcM/NJzGDk1Amj9RC5l1/TVCwlxExlRF2bWAqak6YNO6i/W8buKkVFTTQYFuQykMvAWJgt240ORBsrGPOwALmPX9c8TnzIF3gH/LnfAlDdgf8dwLeY2a+Y2R/XV9+FS1UAAL4dwKfN7CcB/DcA/kRmvmQx07zxd3y7G41w5BrznF3kngdrAI7Xu8k6JtE9uIMw6fItfxWTrtdjIRBngAxxBX0i1u5/Nz8X9a8/9dZujkTZxMtohjkON9zc81IXks1hqPbPd7HQI7eeQ7sP+S2zdXHvC/lmP1aSEy6+t/2kGzNU9z3OHOazH/qQl+ddfD6KArl/Pv7+cvhuGANtn/PLg0cvwcWg65r6d5D4at3Ma9rhvna8b1n9yUzmn9lUf+c4752YDC8PcsHFHL2g/VbrDiAzv+eWYz8M4Idf4r6XvwOYcKJgoTRKAl4LcCvRkj7RNHJ7aVPIMHLdUCCQ0e+baRjJ6LBKGba1fPUAzNDL762+lG++Bq/ZcYATOcTxoI2vqEVf2pQsTdLlNEZlcRtOUGUxe3G086CYKP90CuJsXCu78Cn7c/2kEogAGHAqk7NzoTUwr5wLN2RbCUQMmp3bwW6hh2zWGG0ZDDk2uZrjREBVH3F4fgCn8qoov+GK5nBfuEJ92d2SAJCNbsgxOo27RbCzfNq7wmZJzEEmVe1LOBUQUFz3XBmdBeBRmItBWPjKBI3pYpVNZunHuwGKHOxRShKmcU4yAgGjYZMolgG5FXeVN2WqAUXbbkgEU6JsigOUuFah8dNx6pQuWqUVZgBrIk+A96aU944BxpNYMrnOgnD8aQB6xWGovUAdeKEk8LvWiq08X3LR97c9FAft9se9yeO19HI/fFMOuUlFuXme7VwerrmziPpozzzS7MlN1nVxp3qe+VgXAtBkcs887M5ubnLlZ4bh4nXnLFMmOkorhysl8Owc3SZGXfwiDw995FX8fv9pXhyzZ8695d437vTMDB6kxsuQ6AOR0Vw8KyHhljG+5YSbE/y8z8eHOIoEOjzlonn88vo3zED7qz1nYF7QHkjYsKG1hi7/eC6uoA3DgOMswejkVwhzdCxo5hiVP2ALhpOSuzO8MwMKzJBvuQxLtt8V2PXJJkpemaAV3DGm4ae4Er8PifxYxCHA82MGqACFy8cSU77PZTJ1OIxW3pENCMNZICFDLOysgie9gEXzHfY1KTauol5NHuFlIujQHuCyRaTwBrAYIqmrO2iXCIHYhnTcTYgV18EEq1Bpq1Da7tVCnnUSvtoCSkBrsnyZG49XvYY02h2M7JXltDwY+F9z4iG7XAAhOwVyuhijEoyk41eRlpqMriSubaNo/FRo1TNYdmM8ReUa9CJ1nkJkqiCkoIrnmJvKWpuBVmlQHcvKFNDEJ4QTYDPGo+Imqk7DjC+ZWYKl3mpBJecw0Vhnwisl3eDLYLyKEutiODwMmQ2ZBNtldzT/kkZL1YhbmNexPQxJ4DbqeQsJnYZpiFfY5TnPMMdbON8zhP0uxSlvvrn9xBcyijvbgRMdr3W4b16ce/zNTemB7W4mtHOTIw28OPfm9/PYrs3v//aL3768bkgQd67BGxzxznbnUz0jrT37RU5J6njO5SmHMb00Kty88u39tGe/uBir4/E7ODaXsuEujn774dtEsnffHoQkAADWDE2gb0u52xZGCF8tBkTCtg2JDcvocAw0G0BbGRKQHZ4DHsyeqqizimirJAsvdiDVwauMVC2SotiFL7CVjiurtgvn0BaGJyejEJfmcxNnUB/t/YwxzoSIzsS5M399jA5kwNYF5oY1rmgMSnqbN63JKqrxVKWnnw5y2ZPT9uEbpYuTrMoFEVLhsEu6osm1EUbu6c7u+3uNjWViWUkp1yCAxxgELa103eVKEsH6iPeWDeCqEfClOGHpDRGCyIqVkFkqzjKt5QZJazGz9irYpbIHTYkYLpfpxBApV95CKDhTJZS2Fp/uh/PoGck0RKpQyha0to89s3Q37FICyIWem1yVlDW4IV06fuzyhu5Vy4dvFqV7NG3QGCapTTYY28cBqfDsyJkvgTRg7Wgx4HmCoeFkDeEVTcKlmgk0D6QlvHG9lYcrjkTslvYwJIH33O5+yOfRxmf0zWeueZuO+LK8/wbrue27O396e69v16TfTbtFNLr4+gYHn2ftx+e7W6W359/12aPP3uuFF7s47Y45svruOVe8Iell3nGt+2rP7cqz4/NbkwHYHoQkYAAWa4iVivjJrgiS0RiZ1WwlNU5HItBBTJVRIBRmE27JnWi0w5jXXkafmRCUpSdx0iOGXkVVRymOFAtL31xX6cYCo1w2RvdlrEylD+raIW43euD6fMb1+YwUwMnWB8ZInFToY/ErGByrUqQnxmIVQHnC8ui41ut4SvuD4LjEbGGjcAj84M1KtCUnqEQ6Iy8XjVfzpiQo6p1VHWnMxSeors77dEkl7YrHKyZ+OSleooAzp12Cfn0Tcs7qFd2ZjLIbO5mN4oC6d9O5FWdQJ87Am6PlPfcoxioow3gTK8zOPb/bVRqNJhahNNM7wDRhCN0q0Vx1GVuqKCklJVd/Ct68siyL61YE4U5mS0/n+afKuZCnaDO+zpSCUBTnDPaWBEtxAgxo4ryusj85iEG4yc7gykZ0pdePF5CIBygJvCxN2/XC93KVix+8qx/9TnCN937N269wl9fk4bXf7n6+cEpfVuh4F+e+39qDkARgBm8rgBMAwJeNAB0rc7xtZUbggCOsIXyFmWG4w5cVtlwBrcHWhTj2Luu0uVBzgUKGIexTYoj0Vlnrcgt4+b+vuRy3EhdVB5D/p8KJg4VKDUhLIUYbogPjrJLgqggUEdiiI5JxDm6BEfR6tKBtQKDDLMUNsI4idt23WkkxY1zDDNi2zoCSznTmrbDmrIAzoSyiLEgDWDbFtQvAZIKXln88y3SARGItRt+kh1/46G2iDqPzegU4HgLJKK6c0TDClbFIMXx06rXRgQxDb1EBeQAMbUKYyYq/I5noWfhpLW+BnebXAICrMvwJstw6cxQgyacwIgDhLjja4iwnvzhh0cu2dAPzIpQPMXMEFsVVyKZUQCdVRzIrV0WR7C0kCQmy3EP5DzRYERbdAuGDEonmwxLIxniabAOwMVM6o4rYYKMtrOIl7mgPRxK4qZ7WmxuG2edpq3bLiTdi2J652W7ttjt14jvZwB2q9btqJQUnXsgF88brxbFpgnj2QqnjOd8/e53bpIXdjXXb+c/+Ni8+7QeOZpTLMy97+Oyxm3d9wQhp6o6zeLE+DtF+c+rs5s9via94zu1e2O7s8ot+nc8sr98pQeTBSAI41Oa7Yi4HHiUZ2GPp/L0BaawE3IxltXxd4VdXKLScdSH1pgO/VUI/KgasvAN9sLhGCCq7ixKfpY+nuImpCIm7dONuFzXrA64SVkxGWRCK3Bpw63DvOOfAhsR17+hjoDUwTvxMlNguffQk7a0iBp8O3vup3Fab8hdCpH0L6slbBhyGLakv20IvQBiLuMRhPzHA0bAYfd6LfMnLovRlcKzXRtvzUyEnzXRKmeabYhlOAhup+ApTctNkPpMCxORwjNOHPidGDtlmCCRrGbM0t7J42HexrNWr3DevbYpiDGEPPtbOHoW1JL27shVHV2GVHrOcPASwsoeCczxaawrbrRLltAmUTdKtH5CAgZQXZUoAFQc+wd3LtcHnChVcrDqcs7hNJqMBocrW7sShaAuAzuzUWocMa9zjUraUtFBSSMEd3d4eBhG4qxkuA+rMLkEdbrOSFrUXuS96ulN8ww6yYLjpt52fb77O68yr7/3KS455uBMq8m4PM38ePc+d/t/0QhiOd3zmd/vbF5wzV2yi3KRlxJsssr4+3Pr5V83Z812tOM7OgdMfRJc8/J6HbrnL4djlNXcJ6oVNjzznzg5GRzDwKvMQ6z/na5/rOy67v9aHvPxyX3bsfc6HOEo6t8tml/cpOeX4QPsZd64bPH/+gAdCBBhz74iFu76vKc5vCBfaTzPE1pAIwXA5WluYv+4LDANuwNqIrW++wqzRVRwJJeruoA+y9hc3adLhlqhSVrTIb/0Jv1ca73om1Jn3aySYm5AwjLHIB+2qUNyw+BWusMI9cMpEs40+6QUwS2zyKIzsl8BQwgs4R6EL1wRXjDzPPgnHz4Q6bFZlsiskmryQMFjMI6jowC1ooQ7b8w4MRqQlA2Kht6Cy4iaaT3EXJVn0is6rVGO5F8a04MtRXpGU8mlXWHwGMLbO0mLihtEA2F7dSemGgJd9QbYSiRPn8ofX+Y12ifCVzyAFPIwSgHnMvIiWybm3wEQSLiKAgoU/KBayKZU5f8LSaZtV5OkoyegGEtV5L2NM0pmbJNI42G6MEYO2AFiIp7sCiCsgFwAbMoP7xYARiwZS0pYqNDWNw0Vo+y3tQRABtudrPjvHONC0A7U9cv8LnQ/AruvnM5Fxx/x35A5JlTe5gNnues7Lfh6p7PG+0nKYXgzMjDBh3dyQJm5Q6xuP+XyGd5d0tKMbwSoqwmb/p63QdjnGbvYDl+Nwl8aceXi1A0M6iMrHx5oGyxv3u+D7aRePc/kcmNLXTRtEfdiX/3xyve6fa56On3E8Pr/bJZzZbB+vO6WFmYCWR3F0f8RnH/yir8+835e4xvh5kuXz91S1B0MEmKNN8Ag3WofD6FfuOYA0ZXkVukoCDmXymWrbGdrSaNllnt8FICigKjdGHpkJLEoEX8TBslN/6uXHrSKXsmRnL05gU/f1o00jc2bKXS2GbA1bD/RInH1Bj4ZugvR6h7HyXR3cujAPOvtbhmQImGQR/poV0Ka4eMmjhoYdMAOoYpkDhhGDFodabJ34AksrrD3G96+65rKSAxY3XpqiGUvPFvft5RS4VkakxPxR1nQRXhd+ZAKKqEyUeX6MgRFj1nhI7f/WmD+QEp/LjnCum+r8qhPZ5JNYhIDsSqVbVdK8uO8mOPoBRtYNidYGm2ngZUeotN1QJSJmp+ZUKScQjGDPC+KhqODimoOSVrb6frdPIJOQeABtEY7p5XIH1kYbDlTGDd0RAXTN93CuTa+qWTZHnrdqz7f/PxgiAJVfPub/7wFhO5m8pMe7XElDCC2KVfACYPgwqTUvljhyu53H2rxizhtd0s+5tchDKzPNQDG7kGkCwAG9Z15crKVsAyWmTwTrMpbV3W6NGhQp01d7tV1DVRmEgcQRmJumcOinzn0Yj8kpqwS3cqAPWqbunBevO4eshXeYs8Mj85krPJaurbADl1V/kUrcqRzYKaHNKTpIZYfpwLMSgWZ5Lp3jUCawox0Bh1HAfDcvfZv95i4RZvZtspvZ110O3R9ipigcLy02f1PamgjP6vA+dHqKQ6Ykb6GxnqzvmcV80V6m7sAnQLjxj6kf35+Zf8HM3gDw1wB8EsBnAPyxzPyiEIj/AoDvBPAOgO/JzJ94/l0Cnu9gDObDR5yRZvCNUE/lnXeiigJg2VGi/FY8NrPlRtDfPkDLe0OVvOKQL8rmCxMuQGV9FWS2dLgUYNEs4TQYleW4hmUgxzWjGmWs3JQJaN3gNtAs9gwvZzXaxRe0JIdKAGMl1p8NSjo9VJi0CmDIxF457GI2LMcF4KQIQZdl/1QUpaWiF8fUvXsM9BgwWaNHDgQSTdFylC6MuigMJrzD1ki4Cj8/k/kLFS9QeH61kupjK1G+MQNvWSQJJPEXuwOF/50FChuYWAzgT2GWiEEJpsZhC4XCbRBxk4dnOsRJBBzV531DJmxWHx+GWQC19PDWGpr7xBUofAOLAQtjROlu3RNROaI6aR1pjtYKWRDf8FlWPhAIdBWwUJ4kYE78Rs1DC2IrNhFoQst1ADHxKzA2IAYzOM0m3uGSJ5jtBW7vai8jCXQA/15m/oSZvQ7gx83sbwL4HgB/KzP/rJl9L4DvBfAfAPgOEFbsmwH8kwD+ol6f256xms5Y7pzsZeesAoss7pc7Qxn76bMdpYcSjOL4WmHHes+zDRcWWIOSkvbf0f3GjsfkDEceg/ku1dcEZoES3y9NFUXcrwyBVX24fGM+zYemNUuC5kItchG2lNsPBckNmkV9WuREQA9BSBPWQipEqj5AhV7PjTTVgUvJYIJqVLLP9BK4ip5QwkDmHtoMhrdWKPfEyrOLfQYJBdQ6AAAgAElEQVSrhDI9/+6ByMOv9iGPOQ8Kmy6pzXdpsFyVMxZCm323JBS3PiIa5UR+qucu9ErVXpnEoBYUg632pV1hx7Q5+TQ6lghPH+8uKh0lQoVvYV85tcYuvWblyqxrPtchhZdDFvpVEEUYmfmmmf0sgK8D8CkA/4xO+0EA/wtIBD4F4C8nV8mPmtlHzOzjus4d9wD6sB3xN4QRGBuAgDVWix0CdU9zloPCgoyGDJ9GuJ5EYS2DH6HaU8FUicFKX9xoSUnAjPgAwJ4HPjlc49IYnZ6KLbg5w0hwngYtxN1S7IUTX0U/ueCVG65S5i2FeeBlnNMsKf5+BjHOOdbi1yZo8mVfLcxobCIC67JwEUtHnEY6AD07LfqRjGYLco8S40cuAAzDGsyBrszGopZjVMig5mCCcapPwtuvKkFVWTCkDiwmVCO3GURAv3ZM9OG52yxUMKVosrAYda8xqmA323XxjaqWNEXyAYMhOsvQm5CVql5BDxHJEcwRmPJ2xe3vwDMO+uJDQKOuAJSVwzYrDJ1Ut6JrEte1qhrz9eqauSotFtq7pFpmwUU5N3Q0oQZlm4QnckWkkBKFRcExVI1k6Ru2VBAdcxcLDv2u9q5sAipC8gcA/BiAjx029j8A1QWABOKXDz/7FR27kwgAADImsAUXSe5AiaKAjibuTCJAfH7fOUMe9HDplhMlN8RhgqLfjBWIPXWVn3XPPEgOSaBuR7kUS9ywucmm3zsheHRwk6kmnwsckncirDUBP8RdDrEGKaPFbrMvPnTpvz5MjBaKgmKqX3KpZWjh237PQH3G/jwCBEUCkdpAlTJ7Yf4/yDsa+6ENPN2Zc2y0mTTGKXRisWIOaiRH10IDLqOc1+98roNEqtjp3qcdan0Ox5y7iWaMPYmsblnoTpVGbQlWtT4o0KbNpWxkxKCKVa7BcktaOzw39vgTMxW3Ud1Eb1T9XBbr8iDmNKbQNkIEa9YkJDNLEKbdENFhEaxdGLUygNJh2jRqTpkGwI01c2gvTQTM7IMgfuCfzswvHxdiZqa9qMzJs9ebdQc++qHXYX2guybqmqKjr8qGckb/uT+CWVcUnwOQzg3HwGA5co2F5aAlV9R+nbnVZC+zbiZ7w0q+2IlGmQIqHj1ysL6AaWMZ5sYBGHkYwfz5Qq0NBCIDSyPqT5VGs1RZNaMNoKtWQuHqEXEIs5bc7GWJxcIUrA3XhGozIEQmLbD0gRyqL4DBZ4hARKA7M+hqg/gRgQlA7+K+YIZiZeo1FPqRNoEwfM8zh0DehTbdA8Q6sIUcqy3k2DaACNpDwE3GjZI7elOh9ZQqNTZd22lb2ISqWxISUvdTFKTKyZOQQ8VrOGVOownHQJWqHQkrGwmEI5ApPz438TaUd1DLvySkel5hQrRZv0K5KFIcllMZlR2RwEkxHaG+o/F5uLoDixlWJJYM+EhEXxDnDSMGYjNV5ONztaRdx4Rt0SbTEeO6o70UETCzFSQAfyUz/7oOf67EfDP7OIDP6/hnAXzi8POv17GLdqw78ImPfywjSOkAZUwiWV8eiZOlSjAlTEEdpadmBjc7GHo6idNEjExxaVq/M8bl940cZtaME+W3AhvRxmwgXFcR/B3wQhtXzG3o8wDdiiylZnRDoaTdJLHRj6yCYZQ8Ytpgk8NVgNMBo+/A82Hz//pdRQBWZE5lqRwWgtZFcn5FEMSRDKiaecUt998W66rPQ3fnhvQpt7GxBoyIiLEOA+TmtSTBYrltY4JMSmUwEj0As0JQ2e4XTV8ld1WZv/IoTK+JPB6Tsx9dDRpJbnhZa0qv3mn7DHEOMZMxVJGh7C4VDJV7GnepgQCw2DxLUmINXd3LlNe133vv186/ZQPXNFblKOzgqKjJ0jqxHSjnWV/PZXsZ74AB+EsAfjYz//zhqx8B8N0A/qxe/8bh+J8ysx8CDYK/+Tx7AB8wEZ3ZdUiisaQDwxgF+KgNjrUrsou1dIkkrPz9jA5P5pFTopWLpIxgoK5VhGZ6cJariwFrpcuiCJD0SLkhq7LOpsqjrkjCWigjMKWEHvRBW4m9fFpgDBKZoToCQxMu7IKKy28HNyT7XKah2uyKFEzsMQOwGW8RqHoDO8GpGgZFBaZ3Sd+32mqKnoP6hh5ao5fqQNkEiNy3L2aHqjQrK29diAUxAFn5UwbLve/uWq7u2o8SbwcugpiqsNB2INA8zoHyqj0YezwFx3FXH7I2cspwWQQijxF2hh4k7KNznLahakleJsTKEZDUIXtCk0frSvB//Npov0iQaCSmPciFf4EoVTWlfko6StWyyETPku7IbIr4VEVtayUJaY1H4Rrd3l5GEvinAPybAH5a9QQA4D8CN/9/rToEvwQWJgWA/x50D/4C6CL8t198i0S3DeeVnT9vj5AGPA1BW4V0thAXAYlCasdZBjy4sdydAbC18csaPmpxa0EUsKgCTwaGiIHEXEFVIYa4Ibmqa7GkcXGt2K3ngVSEYGJRvxBBN4+zXBl1dBrBQgsq56qWXaGycWrDiQu7x0G1Y9KOocBWbFqi5Xdj2XSNUwYJAoa4uthIWbnzUCiTdIQEJmyjajZtAWyVKlwdGq3Gq3TkBXDDsjbAAlHxx/A5lz6SbtBtA3qfgB4uY5sQPyfSyeiEBctBZhEimjZ2iQAw+KhNUJJTBRdx/AnCmmhKKNp9Ajn7WHjlUQR0cB6G1IFS5q24epWRF0y8nZZ5b0vApU4shySmyED2AoArz8dePToMCNfa7QwtjqTUFoNdjEw0FY3DNYmKvc5osy6gmMxHwHPIwMt4B/6351zhn7vl/ATwJ1903YvfINHtjOsrWtGvn1whLfGOwqu8a6JEBCwHF0+X0QUJE8ZgS6DlMsXY0ObdAy60mJuQdITw6xI96/vozOArER3Jyaw6BBAPOEnXy2CNnWbJ+nAi3dkDPiAiQGkiK4NRly4lODcRFLskAjsEso4GuUzIm0RcAptciH731CrBXuNvxIHz40AEkrEJAeSSEqZklxBfrPoMFZVWRKC0gu6b+qbF0k7MyLtakcrPN4gwZlAPHyQ+tp2RfdvHf6VxrCSCwoQYvUuFEaFWjHx5GxQVsBMBOxCBLGJKTppScWjfCUkk2oySjCxyJwKCJRyjEz35gB5sAF0UBpgwIMpC75IQFkUaFhHoG3Emc6PnIlw6SEleSIQ7siViI9EZEcJrTIxB1WRkwtFhGLAnCYTBP9QBz+llmqrvHe1BRAxmAuM68KXHFOEfq+Bid4rcYzhgDlsZDGH+QbnuGswD1jg4Ix1tWZBLk0hMg0uptJY7AGRZBkdTNMeiIJhZapqLftP5ZcTDwvLgvnUgEqelS1dkJeQYEjdHILaOwGBYpzNgJwqFFMHjKMEPCOKsTpfQ0GLf4jLYo0+b2178EzAspmCrNC1u2lIqniGk+/M9DXBVECUUnru10mcpGZwkbRQ3LSKwlW1kGhLLNci+rDbgMFwNh/sOVFrWAyY08d9TdGzo81oeDqSkDzMsgybVMpSaU9cfsziqiGjF7K41XiUbFB3PyfkzU8lOiQI5NZfAF5IkjZy/JwFlEYltMNx8uqCjQpR577O2lCt0eaHRCYp3QncaNbs7RgKjueSAMlGXO1jrwOjoazCqlyPosSlhJIFmC+0nj4l/n1fse2pO8vx+CBtOYJwTbyY30oddxiZj1uA5GESdywKYs/LrAD1Ky0BbhyikIdcVaIwdmKouIJw5w+jkhqYsuSHZ08rRW75mxSTMMHWXGWohYTLf4JlYlo3ccID63DC6HXsgszNP3gPhiY6OEYF0EpBJBCpRXvj3Uf5wDU+hH5U3IMq1MWhQSxm/uhXhK0LCsLxsMlxqMYbsFjm1AxOeI7C5iSNy4Or5cZMISP8sTMLe2bvmFGfXDGLjB/EUU/YLoAhSEYHA0yTmwhTNh2oYUOfCKiJQUlrle4QqVqFiOxZt/hM9FihduKL5RjlpJQWUSG+yd5jRhpJJuwUZKwaCRCwgVKTde+TyxW/ituciPCICJ32u8vFDHp7uHZGOIaJQW78MrNP0a4bFKN31IPdHkMl5UMJtbeHvrkhwsLKDWYjF2/tBEgCDJh6fOyyBD53fAmB49PSDyDS8kwMBw3kzhDuur07AZsgnDctVYPnAwIKOxTquzgtO7kg0ZBqiyZ1XfuzsEst4bxNxOIHWFdeGW4obV3KKdN0tRYVV/dGifLJca4tBATAd3hpa09RmUN8DCFElj0UejG0MkyCABLBPm2mR14asclz05QPoWlhKimKFblO4RbkS/eAokFjkIPRZacUJwoSJMAAHo5I6s0tKNYDaVnFZBDO2AXdD88FkLo1r+foDQXF2BM7RsaWIqTicmU+DqPYoAV0NEwhGsY0IcduU66YKvZTBrOWB2BrFfnjSq5fyDihpZwwmP1XxGjO5MXMjzBeumcLspV/SGN1s1bPxc5VI8+t6bq50r0Aua1T7TtTPYgQs9pBwM+r3vjl8MfjqU8VtLeCe6AvTkK9ltby62kjAlZLtCvHGep6a5W3tgRABiu1r73AYHvWnMDg++PSD3BgL8foxDL0B12jIrWG8s9AKezXgxuyvFY5lOP3ScISL+iuOvgn5pizvpTgVc/WtDFsSPc8S2NO0WcoGUEEvFanNJdCUATmMYrA78xly6mVyWCUOobbSy1GnaAHVlasykzjXrPMno1MWuo/EgOgmIuBTrISIl9YRqhOFjVjxAaUthfpLe0PN0mSqO4DvJAL6ffHaCLgb+iC3bJ11fxwkdmlE9ukj0G1Q9RERQJafQ9cMeQ+WkmBKdJ6USa+4HDfUOPKyhcQzk+omoTMiLq05XXuhQBzLpP7fBkvD20B4zDyGPTxNsR4pHi4CPYmfc37diuurE6oRkCFUo1BIuCUZ1PCLgCcAXFci+4mBLjvD1bJRAhAhtwoF992Me1t7EETALNEenfGxDwLIxK9nh6Xhqb1FQ/frJ2A0LG8KeOLtM6wDy9mwtMDV00CeRN3RcIIjoyPh6NvgQu/icKPTkKjxQa/FT0PgmAuDwvi1jE8FET2UrOLiRm3Q5mBDum4GYgQTRbaB7AOuUlcEGo0dX7qYqRZem2pzLZxp2QEALIf8mMwEzvu3iVQZdAB9J/zcTBVPcDhOHFa0QX03aabAEPEbknymNaIKk5YcXJvfSt3i3A3BfY9h8JY4ZxDKNAcroQnZKIwqE3oi2yCQZlbykiLlqiKtc4P1C4gzSjcwwJTcdBKSySqiWYZBFoYxhuDK3abwQI5LhevLmZ8WjOUY3D8+OlFSkgR9BC0aMBI4wHAKSgKrDDaPpJqcNIBDjszFrsjwnNLQmldSPxpaGK68cX1awOwE95OMpMDZE+94whrDvd9qV+iWWKPTiPn2m5QUcAUA+NhHvhoG4M3lbTyvPQgiAEvY0vHa2pAJvNkYuum5AW5YH1/BekN7hwE72DYiqEaiDSLDRFBndiRaypOejORLyx11OLqoJFuSLc0IsAkmUlhx2uxMrKFBy8ryngr0SUj8086OAEagxUCMoai0lIEi9ryd6sdkaDmP87BpkevzjNORBXmQpZf+ngcisF+W4qYZxecpYnPY0YIhzEWYpJHsrr6ZPFDXUt/LU1GdFzFlHxRuHEZjJ3KqGRl6IB8stdST3LAVt7M5ptbygsvLKYLyppSE5BM/oLq2SwAU/+UqCxKNPXy7JK093IrRevI+ZMJiYRRhLsweTaqZVdtSRSCA5CZ33aNVjsGUrCiPuPy4lRLlSvRuaGhmWLAo2Cfg1uC2zN4OC2wGZRkGniyNSWe+0Rh9fgLLwKLxWfwxn8jexPPagyACiRXn+Bge/fIXEWZ456NfDYNhOb2BtIG32+fQLPGhD78DnDte21iY0x+x5NLIgNuJxTxKx3ZaDnNItF1NRKEEby2cwW1U5chmKS8Zkdp1cX6J3/1t8Apnrk+5jgrSq8KGo4JHEmjyy59GSPTWMlyUElsewGn1rw3YpHHIJlA7cPD4tpDw2DamkQ8JRlRCoBm16c3gaNQKDFyMblNflh1UsfCG8li4CM4eFM4LFkyWKZjlCZ5o/GhopQifWDolgIaOwk9IulO0STniLRq8NZg7Fnd4Gn3eASyCPDO5+Ma1xHr1NewpAKBJdalt44L5tvL8KJfCMxVejYPYLKMc8U6RZ7rm+tPAuSf6mYVAeywMVdvOXC0L18d5pVRxNnLhxU+aswIelbF3uZ6Rg47Elew3DQ0tgRN9xLBccLKG0+p0n7418PSdBbE53nmc2JbAmoHmwFe99RbWHngS7yAReCQbyjLe5Px8YdsDPG9pD4QIGEZewZ8wG268cQUzR7TXkNbRpUPGSr3MneJQJfNQy2ItAjNxsjLchCLAJA+XTlyDUhLBUe8FMBNKfFeSQXl5E8faxHXFjaXLElyv/M3iMBUaqpuTieS8bwkGE0ho9qGU3NLtDzNpyQi5i8kVUQAuEoTSFLoLmwlLhahbY1La0cQVQIXskkqUJF4ItlOztfIq0PhU2XBkfIkWFVasgBxT7yYn3WM42CdD/Su3WyvHY4lQMsY2hRdXFOgetlsiUwBmsFG5HiGJoIIZatYpoVGKA6UeBVblSETHrImQMrSmlkvhTRRhL6zFcncW5iI052mV70Bi6EWERYpmnkPaRIVGAjkCy0gsnf1L53h4AqfRceoDZ0XSLlFJR4qgvZ4unlvbgyACawN+z0cDT/0tAMA3Xz+hePVFBmX4R1lS6kud7pQ8vY1TNrzWF5wb8HQF1lPidJKulsncVnNsjUE8IViwTVGIxflLDaiY91HGOqMB8c1rqg8jOgKBN+OaHM4LGJIbi1UkFBlmAbMBKBQ0mqL7UuXT0AELLCuTcyp3v50I9jm0UJZWsQviyrJTZFewS4F8nmX53wAglaAE1O4Kh3IwfUoCrS0EbC3XupNLhwTV1WnwdKkJayVYacCuN3E2wVvnRqI4+mu0WKt0WmZDJHAejsVY1GREsqxZUUZLBlmtDe701LQ02GBnr65OZA5aL0Ph00XIz7OkPbnvOMuecSXKr+Se0Uig48gWM4HochMGg6YGAUqGJ0aegbbBbIONxBpP4Ahc5xmBxFPp/vmI9/71VZz/8WUCkQ1G8Q2lGgcDSmaUJ9VYwxINLRwtVlwNx6PRMPIaIzd85HrF43R8eADdEuMpvQJ5ehNx6lhfY+zH9ZfeATLxpbc/AsCwvV6ly29vD4IImAHLkkhZvh5tA4iG7bzBPCk15x68Bw9gyPZZnF/cf0/JlZTg8jDXdxIJQwpmGbrK3lrMt1x3IZG6i3t1GdkKort+neLAbSY5pZR4uQH9oN5KgrGWFzktvuhBSkfV56p6M6vf1HXqs2LlZ9kehe1xbBR9B0YsWWEzNheWPW9ZGAXVlglYwpstK12FIbfArFmo6DuPXY1IpltiDmjaDEaatoR0wIMGypJMmBC6Swbgb71RQox69PIZFsNX6mzIvRYSCabxV8aByi2I3CWmYulpzB6cblQX13YlOxmRq8xoM6hqSHXvYQEzQ6/4iVlegPcs2wAqtFvDUyaFUrlIiww+iJLpUY/JHIlVco5neTAS58ZQelslGanmYh9Kn/c9IPq29iCIwHVu+KX+BSyPvx4GwyevHlHEPFFE24yb/qufBDocT+INrKcF/uET7DphbwdGOq4lRi44JoGwsRwWo8zSUi7CRGFjT+AcsZvzkJtN8tkYjM9mTgHdMNOeFglcM088g9jChlAueUNIIpgurbYCBqx9hZkrcQZwhUH7Wpyf4vqOKCQQVAGkYOFkmwh9g4pqKD/ftcLMyFmZUwHYUJQZCM4K16Y3g8nyflrbvpGQWKwAWBWoc1Z47HaNBHAVJw7nRgKZT87kstc0jEYfQCSue6APYBsiCp0ZlsMSp55oDozGYKHWKClRqDX0dqIGhrP2LlWQs+wRCL4OGUZn2kyuoIREYj1swNKYqJZyqZUKNxwYju16w9gGzuOMcw6Mp8y9uO5nRBIOvqIyATC0F0Bfv0zbwCN24lrFayo7s9bTVkbPhWHDtG06uq1IG/A2cD0cWxiubeDaHfGRBfnhBV/qievoiHydZc0++mHkkrj6whN6Cb7wOWQmfvNbvgYwQ169+fCJADXsAGwR0eYmyfZ0nmCWU2SybORs3ijGlVh+8MUDJSKUTmZclOXezyOJuHEeoGvlxVfF9S44SP1frqciLmX9Kb1cf8d7HrPVSsek+WI/6zh7dvFavDIpOdT99AObl6t/R535wG2lG89/tr/uROAgnZRUMt2Y+++RmJF25rnHaks6guWuO2t++D7nGBYrrngM3kTBTLghfc2x3uPtOE18uAnNpUhMTLyJsg3Y4X4iAjIN0N0UIvZKNKr+iQOn+gZAuHZgvggwaxGmuEoxmbmUaq5aznU5V4oiNsvaVXJHSPSdKSGmCWo08iKFHBUNiEA4xaz0w1je0h4EEegwfD4e4ZvO3Oznr/0sYhg+98sfBcbA9rkv4HFL/KOvP0XEY3wovw5P7Qpf9NexPXY8feRYlg3rcsbTLK/TgGcoZruKbyTC20F43xd1K0MiHQ+40iQ+HrQNLMHCIk+juCsXgg1m6o3uXJ4BxrY35tC7N4QxVn9ZuHmyURohrDcmLt3iK4AdkKM4uStIKCRi+1YLSKLmwuUTKaDSpCTQTMkI4ESvoOgMMyynR/C28AsH1rbQLiI1oJ3kLliK5LAvj8r9+JRl0k/xQSSAc7/mPtsUYpuuMtsBG4mn/cwAmJUeBwbKMVBpEaH0E2BLwtcVzR3LKtg0u4LB8EjrpZ0FAy99PBQnQNDYGSs5VbsqmOIq1IlK/5buUQlEDcCWSgFfwQSeSCAaC6YOINsKZGLp7HOl6y7CdzzFY947CipOfRM46lL5DbK/2qkhzTBWlTrzFcohhp8M65UBY4FHYjsN9PYWbPlNwK6BX/si8gzg/AEAjjc+uGJJ4Asngtl8aPkSYIZHWgO/iNvbgyACxRmKS1EXs8kYRjC4xGS2mgCjVqCLFe3BTXJJ88QdRAymp+DoMD++3vKxeE9th+NPjgb7o+Rh84TDHzBtGHbL8Rk1dCkLXOjqmI9p9bji0pdyxjzx5lGzSQgmfJdhegtmtR0raeByvPYuqo/zHvI0WMqrsUsg5Ft1raMkdZBY5ljUeXsf6mrF68stuXflxnPXHNe0YB+fnKvD9pvPk2/jlkfx6XCuXZ6+u3XvWFCH85jrcdmP8ozMK8whuBTfTIZUGlRlG5kRlPK+uE7XWPvtXZntQRCB5me8/vpn8MbnTuRSb/0ewByf+ORrwDnQ3vqH4MuG8dEvAGlo49ewbAuufuNLePwBR/swAcOXZMTf6IKKzsRQ7sDVIOfpSriJshqV4WaTbUAROX1anShWjula+yCARNrG0NfCtfFN+d8MWHJLApSEwxst781d8fTMtT/5StedLO8ncbDiFpUjgC6uou9D9ojKMI4CPtHC8oWyjzvdmW5ByzNoZPPmOK1ntGXI6MZ7O4zSAYC2cuO6gEWLu1a2W5NvflPE5OONktbTJ9eICJyfkIKPc2fii/zrFmeK1yPgTu+PLYwPMMGb011G3szMR9A9uG1Ayr6ARJ+M/ZKgp+w8rcrNm0tdK1GdCdIKsSe+ABINgdE7C6FsieyA24aGwBoVWUlMy1gUiRrM1T/LlnJd0lnFWLniJwoT4qnsPI+59LxLnF9XOAynlvBwrNcNax9Y+xkxEmMkPvBOg6Mh+tfiaTq++CZw7sDrH3sDbV3w9vJZAB2Pf/EbYCMRH3iCBPCF5Wu1Mv7urfvvQRABAOIQN3j4gajbxYec55P60ddcpnbqp5Pvqx04R06h4Q56/a46jmk7uHH8Ti3shhByfP88PnJnX29wpf3kg43Abkgt9uz9JyDnkRvdce/5+UISAvaolLyQyvKGdPNMP+3CdHrHXW/c65bb332uXZ6UN/50MHHj2I1FMiWX2wbcbutDPvfjbe2ZpxbzP/ZpVjvOkh1sSlRzFBWinc8bRzwQInB+coX/79P/MP6J33gdacBnviFh5njtjdeRY+C8PIG3BVfL74VFYNk2NCQerYnrc+DLv3bG623Bh9qCEOQYKmVzKCTYD2XJAAFcJqKALoUkFGeOdN/o527bXn4KAF5r5HjXVLuxNENEwxaNFl7CDaAHdfqBprBVY6GRBCwZwxBJ8bsNvnYhDK2abC9OVkL1EPYBA/WxyDjVFdp8ZVSrtr0aKBKObo4wejosF2A0bGdKJWtb4OY4S29eG6Pz0MmxJtBqBxdcUALo8sWfVTDFOtnqMlg2+8MCyniCjmyJq8cdI4HrLOObIbHSVbgmck3mDqSxL2FYOt1u1xt9w1Oll6pQklHN6bSdyI3bpMIMrxBvSgLdVhgSzQUiF4zd2DCUOWjYWmAYMxTNHCOIZ9BO11hAD0wAeGRvwgAs4NyEf4CdkUTwlrZYUwyDa3wFRI5WeAQbycvT60TLwIpgqHJ8AH3Z0NeO14bhlIb1tQQ88FX+DnIbePLolzDawFd96UNow/FLv+91hBve6F+GIfHZt3/qubTHn/PdPTRx+6Oz/oIXHT/v7Xkhke+pvYvrvneJ4t3e9mXu+Cz/fnfXvlCsX67dKpHc+PLmNP52Dt6tl/ttXiDvpb+3SguHNkWRd9uN3CW9d9k/u7Um/O9yM7MvAHgbwK/dd1/eQ/tqvL/7D7z/n+H93n/gd/YZfm9mfs3Ngw+CCACAmf2dzPxD992P32p7v/cfeP8/w/u9/8D9PMMDUwdetVftVfvdbq+IwKv2qn2Ft4dEBL7/vjvwHtv7vf/A+/8Z3u/9B+7hGR6MTeBVe9VetftpD0kSeNVetVftHtq9EwEz+5fM7OfM7BfM7Hvvuz8v28zsM2b202b2k2b2d3TsDTP7m2b283r96H3389jM7AfM7PNm9jOHY7f22dj+M83Lp83s2+6v57Ovt/X/z5jZZzUPP2lm33n47j9U/9jNc8MAAALtSURBVH/OzP7F++n13szsE2b2P5vZ/21mf9fM/l0dv985qBTJ+/gDc6l+EcA3ATgB+CkAv/8++/Qu+v4ZAF9949h/CuB79f57Afwn993PG/37dgDfBuBnXtRnsJ7k/wCGnvwRAD/2QPv/ZwD8+7ec+/u1nq4AfKPWWbvn/n8cwLfp/esA/p76ea9zcN+SwB8G8AuZ+fcz8wzghwB86p779F7apwD8oN7/IIB/5R778kzLzL8N4DduHL6rz58C8JeT7UcBfEQl6O+t3dH/u9qnAPxQZl5n5v8LFsj9w79jnXuJlpm/mpk/ofdvAvhZAF+He56D+yYCXwfglw+ff0XH3g8tAfyPZvbjZvbv6NjHci/D/g8AfOx+uvau2l19fj/NzZ+SuPwDBxXsQfffzD4J4A8A+DHc8xzcNxF4P7c/mpnfBuA7APxJM/v245dJee595Xp5P/YZwF8E8PsAfCuAXwXw5+63Oy9uxnz0HwbwpzPzy8fv7mMO7psIfBbAJw6fv17HHnzLzM/q9fMA/ltQ1PxciWt6/fz99fCl2119fl/MTWZ+LjNHEijwv8Au8j/I/pvZChKAv5KZf12H73UO7psI/J8AvtnMvtHMTgC+C8CP3HOfXtjM7ANm9nq9B/AvAPgZsO/frdO+G8DfuJ8evqt2V59/BMC/JQv1HwHwmweR9cG0GzryvwrOA8D+f5eZXZnZNwL4ZgD/x+92/47NCIrxlwD8bGb++cNX9zsH92ktPVhA/x5ovf2+++7PS/b5m0DL80+BcC3fp+NfBeBvAfh5AP8TgDfuu683+v1XQZF5A/XLP35Xn0GL9H+ueflpAH/ogfb/v1L/Pq1N8/HD+d+n/v8cgO94AP3/o6Co/2kAP6m/77zvOXgVMfiqvWpf4e2+1YFX7VV71e65vSICr9qr9hXeXhGBV+1V+wpvr4jAq/aqfYW3V0TgVXvVvsLbKyLwqr1qX+HtFRF41V61r/D2igi8aq/aV3j7/wGNEG5Ab8oHXAAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Z0MmTtTWdoJq",
        "outputId": "ddc443e9-b8c2-4b85-e724-dc6be9b1f5ee",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        }
      },
      "source": [
        "plt.imshow(color[8])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7f90cc4445f8>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 32
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Hg_iThNIzIXU",
        "outputId": "ca7ca35d-3776-49fe-e359-7846d98a9fea",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "filenames[:5]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['Places365_test_00036857.jpg',\n",
              " 'Places365_test_00000428.jpg',\n",
              " 'Places365_test_00018729.jpg',\n",
              " 'Places365_test_00000427.jpg',\n",
              " 'Places365_test_00018726.jpg']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 27
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JtjCB0ewWHDo"
      },
      "source": [
        "data.epoch_end(filenames)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9Z_QCBTOWNBT",
        "outputId": "49e4105f-b49a-47e6-c4cf-42a48818ebd0",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "filenames[:5]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['Places365_test_00020536.jpg',\n",
              " 'Places365_test_00022029.jpg',\n",
              " 'Places365_val_00028755.jpg',\n",
              " 'Places365_val_00028968.jpg',\n",
              " 'Places365_test_00000949.jpg']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 29
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E659yJP6WOgL"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}