06_gan_simple_pattern.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "DpII-qgno6EX"
   },
   "source": [
    "# Simple 1010 Pattern\n",
    "\n",
    "Make Your First GAN With MindSpore, 2022"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "7bkrfSHHAPUB"
   },
   "outputs": [],
   "source": [
    "# import libraries\n",
    "\n",
    "import mindspore\n",
    "import mindspore.nn as nn\n",
    "\n",
    "import pandas\n",
    "import matplotlib.pyplot as plt\n",
    "import random\n",
    "import numpy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "o91c-ALVmw0n"
   },
   "source": [
    "## Data Functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "hS9SJ5JI5rIW"
   },
   "outputs": [],
   "source": [
    "# function to generate real data\n",
    "\n",
    "def generate_real():\n",
    "    real_data = mindspore.Tensor(\n",
    "        [random.uniform(0.8, 1.0),\n",
    "         random.uniform(0.0, 0.2),\n",
    "         random.uniform(0.8, 1.0),\n",
    "         random.uniform(0.0, 0.2)], mindspore.float32)\n",
    "    return real_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Tensor(shape=[4], dtype=Float32, value= [ 8.59873950e-01,  5.64498231e-02,  9.21470821e-01,  6.21404089e-02])"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "generate_real()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "px_g8nYPBD3A"
   },
   "outputs": [],
   "source": [
    "# function to generate uniform random data\n",
    "\n",
    "def generate_random(size):\n",
    "    random_data = mindspore.Tensor(numpy.random.rand(size), mindspore.float32)\n",
    "    return random_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Tensor(shape=[4], dtype=Float32, value= [ 5.87776899e-01,  8.54252815e-01,  1.67392999e-01,  6.90094829e-01])"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "generate_random(4)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "AJ_E9wgYQybx"
   },
   "source": [
    "## Discriminator Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "MrgKlTsRAfSk"
   },
   "outputs": [],
   "source": [
    "# discriminator class\n",
    "\n",
    "class Discriminator(nn.Cell):\n",
    "    \n",
    "    def __init__(self):\n",
    "        # initialise parent cell class\n",
    "        super().__init__()\n",
    "        \n",
    "        # define neural network layers\n",
    "        self.model = nn.SequentialCell(\n",
    "            nn.Dense(4, 3),\n",
    "            nn.Sigmoid(),\n",
    "            nn.Dense(3, 1),\n",
    "            nn.Sigmoid()\n",
    "        )\n",
    "        \n",
    "        # create loss function\n",
    "        self.loss_function = nn.MSELoss()\n",
    "\n",
    "        # create optimizer, simple stochastic gradient descent\n",
    "        self.optimizer = nn.SGD(self.trainable_params(), learning_rate=0.05)\n",
    "        \n",
    "        self.model_with_loss = nn.WithLossCell(self.model, self.loss_function)\n",
    "        self.model_train = nn.TrainOneStepCell(self.model_with_loss, self.optimizer)\n",
    "\n",
    "        # counter and accumulator for progress\n",
    "        self.counter = 0;\n",
    "        self.progress = []\n",
    "\n",
    "    def construct(self, inputs):\n",
    "        if inputs.ndim == 1:\n",
    "            inputs = mindspore.ops.ExpandDims()(inputs, 0)\n",
    "            return self.model(inputs).squeeze(0)\n",
    "        else:    \n",
    "            return self.model(inputs)\n",
    "\n",
    "    def train(self, inputs, targets):\n",
    "        # train one step\n",
    "        loss = self.model_train(inputs, targets)\n",
    "\n",
    "        # increase counter and accumulate error every 10\n",
    "        self.counter += 1;\n",
    "        if (self.counter % 10 == 0):\n",
    "            self.progress.append(loss.asnumpy().item())\n",
    "        if (self.counter % 10000 == 0):\n",
    "            print(\"counter = \", self.counter)\n",
    "\n",
    "    def plot_progress(self):\n",
    "        df = pandas.DataFrame(self.progress, columns=['loss'])\n",
    "        df.plot(ylim=(0, 1.0), figsize=(16,8), alpha=0.1, marker='.', grid=True, yticks=(0, 0.25, 0.5))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "IFajrD3xCgsb"
   },
   "source": [
    "## Test Discriminator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "colab_type": "code",
    "id": "AXgCZ9DiCl3A",
    "outputId": "d055f70a-380c-46b8-97a8-1c12bdc817ca"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "counter =  10000\n",
      "counter =  20000\n"
     ]
    }
   ],
   "source": [
    "# test discriminator can separate real data from random noise\n",
    "\n",
    "D = Discriminator()\n",
    "\n",
    "for i in range(10000):\n",
    "    # real data\n",
    "    D.train(generate_real().reshape(1, -1), mindspore.Tensor([[1.0]], mindspore.float32))\n",
    "    # fake data\n",
    "    D.train(generate_random(4).reshape(1, -1), mindspore.Tensor([[0.0]], mindspore.float32))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 483
    },
    "colab_type": "code",
    "id": "aBh8AGJFDhcU",
    "outputId": "71b31b63-fa20-498f-b73d-f7cc4f032f38"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot discriminator loss\n",
    "\n",
    "D.plot_progress()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "colab_type": "code",
    "id": "MUMJn6efDkH5",
    "outputId": "2219d02b-9152-4187-d8e9-17cf11c6ff7c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9316080808639526\n",
      "0.014704041182994843\n"
     ]
    }
   ],
   "source": [
    "# manually run discriminator to check it can tell real data from fake\n",
    "\n",
    "print( D.construct( generate_real() ).asnumpy().item() )\n",
    "print( D.construct( generate_random(4) ).asnumpy().item() )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "SsE9-KrguowG"
   },
   "source": [
    "## Generator Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "j-JwJlKDh_Ej"
   },
   "outputs": [],
   "source": [
    "# generator class\n",
    "\n",
    "class Generator(nn.Cell):\n",
    "    \n",
    "    def __init__(self, D_with_loss):\n",
    "        # initialise parent cell class\n",
    "        super().__init__()\n",
    "        \n",
    "        # define neural network layers\n",
    "        self.model = nn.SequentialCell(\n",
    "            nn.Dense(1, 3),\n",
    "            nn.Sigmoid(),\n",
    "            nn.Dense(3, 4),\n",
    "            nn.Sigmoid()\n",
    "        )\n",
    "\n",
    "        self.loss_fn = D_with_loss\n",
    "        \n",
    "        # create optimizer, simple stochastic gradient descent\n",
    "        self.optimizer = nn.SGD(self.trainable_params(), learning_rate=0.05)\n",
    "        \n",
    "        self.model_train = mindspore.build_train_network(self.model, self.optimizer, loss_fn=self.loss_fn)\n",
    "\n",
    "        # counter and accumulator for progress\n",
    "        self.counter = 0;\n",
    "        self.progress = []\n",
    "    \n",
    "    def construct(self, inputs):        \n",
    "        # simply run model\n",
    "        if inputs.ndim == 1:\n",
    "            inputs = mindspore.ops.ExpandDims()(inputs, 0)\n",
    "            return self.model(inputs).squeeze(0)\n",
    "        else:    \n",
    "            return self.model(inputs)\n",
    "\n",
    "    def train(self, inputs, targets):\n",
    "        # calculate error\n",
    "        loss = self.model_train(inputs, targets)\n",
    "\n",
    "        # increase counter and accumulate error every 10\n",
    "        self.counter += 1;\n",
    "        if (self.counter % 10 == 0):\n",
    "            self.progress.append(loss.asnumpy().item())\n",
    "    \n",
    "    def plot_progress(self):\n",
    "        df = pandas.DataFrame(self.progress, columns=['loss'])\n",
    "        df.plot(ylim=(0, 1.0), figsize=(16,8), alpha=0.1, marker='.', grid=True, yticks=(0, 0.25, 0.5))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "SwjycPNXxF7G"
   },
   "source": [
    "## Test Generator Output"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "84vbfjGRwodN",
    "outputId": "77578817-8fca-46d4-b0e6-96fba89db611"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Tensor(shape=[4], dtype=Float32, value= [ 5.00697315e-01,  5.00057161e-01,  4.98958260e-01,  5.01136780e-01])"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# check the generator output is of the right type and shape\n",
    "\n",
    "G = Generator(D.model_with_loss)\n",
    "\n",
    "G.construct(mindspore.Tensor([0.5], mindspore.float32))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "5k9SRwIkxTLi"
   },
   "source": [
    "## Train GAN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 85
    },
    "colab_type": "code",
    "id": "radA76P9xWr1",
    "outputId": "8b5b1470-58a3-4551-9763-faf62ba0d8a9"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "counter =  10000\n",
      "counter =  20000\n",
      "CPU times: user 36.7 s, sys: 13.4 s, total: 50.1 s\n",
      "Wall time: 40.1 s\n"
     ]
    }
   ],
   "source": [
    "%%time \n",
    "\n",
    "# create Discriminator and Generator\n",
    "\n",
    "D = Discriminator()\n",
    "G = Generator(D.model_with_loss)\n",
    "\n",
    "image_list = []\n",
    "\n",
    "# train Discriminator and Generator\n",
    "\n",
    "for i in range(10000):\n",
    "    # batch_size 1\n",
    "    input_seed = mindspore.Tensor([[0.5]], mindspore.float32)\n",
    "    label_true = mindspore.Tensor([[1.0]], mindspore.float32)\n",
    "    label_false = mindspore.Tensor([[0.0]], mindspore.float32)\n",
    "    \n",
    "    # train discriminator on true\n",
    "    D.train(generate_real().reshape(1, -1), label_true)\n",
    "    \n",
    "    # train discriminator on false\n",
    "    D.train(G.construct(input_seed), label_false)\n",
    "    \n",
    "    # train generator\n",
    "    G.train(input_seed, label_true)\n",
    "    \n",
    "    # add image to list every 1000\n",
    "    if (i % 1000 == 0):\n",
    "      image_list.append( G.construct(input_seed)[0].asnumpy() )\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 483
    },
    "colab_type": "code",
    "id": "WfGwR21QxYtV",
    "outputId": "1cbff4da-8f15-4306-e1c7-341f5e63200e"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot discriminator error\n",
    "\n",
    "D.plot_progress()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 483
    },
    "colab_type": "code",
    "id": "SIx0tJBixbGM",
    "outputId": "1c6f7723-bc0f-4a30-a599-38c3e36c696f"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA6gAAAHSCAYAAADhZ+amAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8+yak3AAAACXBIWXMAAAsTAAALEwEAmpwYAAAwpElEQVR4nO3de4zl12Ef9u/53cc8dmYfJEVKIiWRTtTErmxLqaSkTSITedhKG1gxnDY2GlRyrago4iRN2xR2DSSpGyONFdRtUTeJkciVgsZW6hqtWrg2jCSEbNQPybIU21IUybIlLk1S5HJf87yv0z/undmZ4ZKc3R1yzy4/H2Cxc37397v3d2fP/u75/s7jllprAAAA4HbrbvcJAAAAQCKgAgAA0AgBFQAAgCYIqAAAADRBQAUAAKAJAioAAABN6N/uEzjqvvvuqw8//PDtPo0Xtbm5mVOnTt3u0wB1kSaoh7RCXaQF6iEtaL0e/uqv/uqztdbXXO+x5gLqww8/nE9+8pO3+zRe1GOPPZZHH330dp8GqIs0QT2kFeoiLVAPaUHr9bCU8uUXeswQXwAAAJogoAIAANAEARUAAIAmNDcHFQAA4NVkPB7n/Pnz2dnZOZHnO3PmTD73uc+dyHPdiuXl5Tz00EMZDAbHPkZABQAAuI3Onz+f9fX1PPzwwyml3PLzXb16Nevr6ydwZjev1poLFy7k/PnzeeSRR459nCG+AAAAt9HOzk7uvffeEwmnrSil5N57773hXmEBFQAA4Da7m8Lpnpt5TwIqAADAq9za2trtPoUkAioAAACNEFABAADuMKPJLJu7k4wmsxN93lpr/tpf+2t5y1vekq//+q/PRz/60STJk08+mXe9611561vfmre85S35+Z//+Uyn07zvfe/b3/eHf/iHb/n1reILAADQiCs740ym9UX3GU9n+d1L26k1KSV5/dmVDHrX+h43tsYZd6P9cr9Xcnr5eF/18lM/9VP59Kc/nc985jN59tln8453vCPvete78k/+yT/Jt3zLt+T7v//7M51Os7W1lU9/+tN54okn8hu/8RtJkkuXLt34Gz5CDyoAAMAdZDydpdbk1FIvtc7LJ+UXfuEX8p3f+Z3p9Xp54IEH8k3f9E35xCc+kXe84x35sR/7sfzNv/k38+u//utZX1/P13zN1+RLX/pS/tJf+kv5mZ/5mZw+ffqWX19ABQAAaMTp5UHuOTV80T/3ry/n3Oogg16Xc6uD3L++fOjxc6uHn+O4vacv5l3velc+/vGP58EHH8z73ve+fOQjH8m5c+fymc98Jo8++mj+/t//+3n/+99/y68joAIAANxBhv0uD55bzQOnl/PgudUM+ycX6/7oH/2j+ehHP5rpdJpnnnkmH//4x/POd74zX/7yl/PAAw/kL/yFv5D3v//9+dSnPpVnn302s9ks3/7t356/9bf+Vj71qU/d8uubgwoAAHCHGfa7Ew2me77t274tv/iLv5hv/MZvTCklP/RDP5TXvva1+fCHP5wPfvCDGQwGWVtby0c+8pE88cQT+a7v+q7MZvMhxn/7b//tW359ARUAAOBVbmNjI0lSSskHP/jBfPCDHzz0+Hvf+968973vfd5xJ9FrepAhvgAAADRBQAUAAKAJAioAAABNEFABAABus1rr7T6FE3cz70lABQAAuI2Wl5dz4cKFuyqk1lpz4cKFLC8v39BxVvEFAAC4jR566KGcP38+zzzzzIk8387Ozg0Hw5fD8vJyHnrooRs6RkAFAAC4jQaDQR555JETe77HHnssb3vb207s+V5JhvgCAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE04VkAtpby7lPL5UsoXSynfe53H31dKeaaU8unFn/cfeOy9pZQvLP689yRPHgAAgLtH/6V2KKX0kvxIkj+Z5HyST5RSPlZr/eyRXT9aa/2eI8fek+RvJHl7kprkVxfHXjyRswcAAOCucZwe1Hcm+WKt9Uu11lGSn0jynmM+/7ck+bla63OLUPpzSd59c6cKAADA3ewle1CTPJjk8QPl80n+4HX2+/ZSyruS/Oskf7XW+vgLHPvg0QNLKR9I8oEkeeCBB/LYY48d6+Rvl42NjebPkVcHdZEWqIe0Ql2kBeohLbiT6+FxAupx/N9JfrzWultK+U+SfDjJHzvuwbXWH03yo0ny9re/vT766KMndFovj8ceeyytnyOvDuoiLVAPaYW6SAvUQ1pwJ9fD4wzxfSLJGw6UH1ps21drvVBr3V0U/2GSf+u4xwIAAEByvID6iSRvLqU8UkoZJvmOJB87uEMp5XUHit+a5HOLn382yTeXUs6VUs4l+ebFNgAAADjkJYf41lonpZTvyTxY9pJ8qNb6m6WUH0jyyVrrx5L85VLKtyaZJHkuyfsWxz5XSvlvMw+5SfIDtdbnXob3AQAAwB3uWHNQa60/neSnj2z76wd+/r4k3/cCx34oyYdu4RwBAAB4FTjOEF8AAAB42QmoAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAnHCqillHeXUj5fSvliKeV7r/P4f15K+Wwp5V+WUv5ZKeVNBx6bllI+vfjzsZM8eQAAAO4e/ZfaoZTSS/IjSf5kkvNJPlFK+Vit9bMHdvu1JG+vtW6VUv7TJD+U5M8tHtuutb71ZE8bAACAu81xelDfmeSLtdYv1VpHSX4iyXsO7lBr/Re11q1F8ZeSPHSypwkAAMDd7jgB9cEkjx8on19seyHfneT/PVBeLqV8spTyS6WUP3PjpwgAAMCrwUsO8b0RpZQ/n+TtSb7pwOY31VqfKKV8TZJ/Xkr59Vrrbx057gNJPpAkDzzwQB577LGTPK0Tt7Gx0fw58uqgLtIC9ZBWqIu0QD2kBXdyPTxOQH0iyRsOlB9abDuklPInknx/km+qte7uba+1PrH4+0ullMeSvC3JoYBaa/3RJD+aJG9/+9vro48+ekNv4pX22GOPpfVz5NVBXaQF6iGtUBdpgXpIC+7kenicIb6fSPLmUsojpZRhku9Icmg13lLK25L8gyTfWmv96oHt50opS4uf70vyh5McXFwJAAAAkhyjB7XWOimlfE+Sn03SS/KhWutvllJ+IMkna60fS/LBJGtJ/vdSSpJ8pdb6rUm+Nsk/KKXMMg/D/92R1X8BAAAgyTHnoNZafzrJTx/Z9tcP/PwnXuC4/y/J19/KCQIAAPDqcJwhvgAAAPCyE1ABAABogoAKAABAEwRUAAAAmiCgAgAA0AQBFQAAgCYIqAAAADRBQAUAAKAJAioAAABNEFABAABogoB6g0aTWXYmNaPJ7HafCgAAwF1FQL0Bo8ksX3j6ap7anOWJi1tCKgAAwAkSUG/AeDrLeDpLr0vqogwAAMDJEFBvwKDXpetKtsdJWZQBAAA4GRLWDRj2uzx4diVnlpIHz61m2PfrAwAAOCkS1g1aGvQy7BXhFAAA4IRJWTeoLP6ezeptPQ8AAIC7jYB6g7oyj6jiKQAAwMkSUG/QIp9mVkVUAACAkySg3iABFQAA4OUhoN6g/SG+8ikAAMCJElBvkIAKAADw8hBQb9DeKr7VMkkAAAAnSkC9QV1XMpnVbOxMMprMbvfpAAAA3DUE1Bs0msxyabfmqSs7eeLilpAKAABwQgTUGzSezlJrycqgl7ooAwAAcOsE1Bs06HXpdcnGziRlUQYAAODW9W/3Cdxphv0u9y6X3Ls+zGtPr2TYF1ABAABOgnR1E4a9kuVBTzgFAAA4QRLWTShl/j2o1ZehAgAAnBgB9SZ0iy9DncmnAAAAJ0ZAvQmLfJqphAoAAHBiBNSb0JX518tc3Rn7HlQAAIATIqDehFmtefLyTp6+spMnLm4JqQAAACdAQL0Js1pSas1Sv0vNvDcVAACAWyOg3oR+l5Su5OrOJCXJoOfXCAAAcKv6t/sE7kT9ruSN96xmZzzNA6dXfB8qAADACZCsblJXilV8AQAATpAe1JswmdU8fWUnm7uTTGc1b7jnlF5UAACAWyRV3YTJLOm6ZGXYz2RWLZIEAABwAgTUm9DvkkHXZXs070G1SBIAAMCtM8T3JvS7kjfccyr9XpdTw77hvQAAACdAsrpJw36X1WEvW+NJRhNDfAEAAG6VHtSbNJrM8tUru9kaTTKezCyUBAAAcIskqps0ns7S75WsDPsZTWcWSgIAALhFelBv0qDXZdAreW5jN72uS/WVqAAAALdED+pNGva73L++kq4rmaXmmas75qICAADcAj2ot6CU5PTKIDvjWS5tj3J2dZhhf3i7TwsymsxyaWuUrdEkg67LYDE/erwYjr637aXKL9cxyXwUwqklq2ADAHCNgHoLBr0uXUkev7CZJOlSMuh1WVv2a+XWjCazbO5O9uc2v1AIfHpjkiee20q/V7I7mWV3Mk1q8uTl7fzrr27k0tY4dTbLmdVhZjW5tDVKndWkJKeXB7myM06dJbUk6yv9bGxPUuv85svZU4OklvkxNUlqTq0Msrk9TlJS9o7ZuXbMPavDzDJ/ndT5857Ze52a9Lrk3Kml9Lsu51YHee2Z5bzxnlPHCr57v4szK8Pn/R8bTRb79jqBFwDgDiZJ3YK9Yb7PbYxycWuUp65sZzKt+X2vOy2ksu+4YXM0mabXdZnVmscvbObpqzu5tD3JZFpzcXN3HhJLcnZ1ERw3d/Plr+zmX9ffytrSIBu74yTJrNYs93qZ1GTQL9mZTNPfmaSWmtF0mmG/y2g6y+ZokmmtGQ66jKc1u6NpapLBoGQ0rRlPZiml7G8bT5PJZJrSlfR78312RtNMa81gEZC3RtOk1Ixnsyz1u+xOF+99rzyZ5crWOLvTSZ68nPzLx2vOnhqm13W5tLWbvanc51aHuboIvinJ+nI/26Npzq0u5b61Qd76xnNZGfQzrTVbo0kubIyy1O9lMpvm7Mowq8P+S/burg76OXtqKNACADREirpFp5b6Ob0yyJMXt3N1NMloWtPvlXzt689o+N6FTjRsJjm9OkhqDoWzlUE/m6NJ+r358w27ciDwXQuO6UqGvS6DrktSM+x6Wep3maZmbbmfS5vz3tHVQT/rq8OUvS7NJEtdL+vLg1zdGc8DaD/PK59eHqaWmum0vuA+B8vDYS9nV+fHpM6D7fDI6/QHvaws9TPZmqXX6zLo1Qx6XaY16XddlnpdtqfzoDuZ1Sz1S3YXQXg8rbm0s5vzFzfyW89sZjjocnlrlFmt6dLl3vVhnt3YTa01/dJlfXWQja1xpqnpUnLv2lJqkis7kwy6knOnhvn9D5zO684uJyUvOhw5uTYkOUk2dydJYogyAMAJE1Bv0bDf5U33ruXZjd08/dRuuq3dTKfTnFkd5PVnVzVeG3MwYB53LuVoMk1/P2xu5emr27m0Pc1kNsvFzVFmiyGzZ1YHqbPjhc3lfpfRLJnNZiml2w9n42nN8qBLVwYpSSazWdaXB+mPukXgOxwcS1ez1O9lfWmQq3WcWeZB6w1nTuUbHhzkvrWl/d7Evfdzu+egpiRfeW4rX728ncs7817fyaxmMpmlJlnpHwnP3TwIX9oaZTqtOb08zOmVYUaT+fOtDnrZGdds7Iwz7PWy3HXZnE4zGk2TrmSl18vudJadyTSzWrM7maQb9vLkpc08fXU3u5NJ+ulSSrKy1M/W4nWT5MypYZZ6Xbquyz0rg9x7apiuSy5ujVJKct/ach46t5quK7f0e1td6gu7AAARUE/E2nI/X/vaM7lwdZQL27t58tJmZrOSJ89u53VnVnLf+rIhvzfoxYJkcvzGf78r++FyczTJ+QububQ9yYXNUS5s7qbOkqRm/ch8zHnPZs3lrdGLhs3RdLrfszmdzFK6lwibK8P0dyfXgtfSPGzuhbNBv+TUcJDTK8mZlUHuX5/P0UzJdX8HvzA+nz/yjjfecYHnNevLh3qik5f+N90aTXL+ua1c2R7n0vYovV4ymvRSS8l9ZwZZG/SzOZrmys4oq+VwyF3t93JmEezrrGQ8mSWly/qwn8xm6RZDlKfT+RDmpd58GHSdzrIzm+XqzjhPX052x7OsLvWTlCQ1u+OLObU0yO54mpKSdMm51UHqYu5uapJuPg/36u4kqUkpNevLw1zens/tXRp0+Zr71vLg2dXcf3rpRXtzDw5PXl3qZzydZWs0ObTP/Cuwrs3ZHfTm4dv8XADgTiA1nZCzp4b5PfevZfZUTUnJVzd38q+evpx7Vpdz//owb33juSwNeq/ICql3wjF7wbHmQC/lYrjq5micJ57bmgfJrVEubIwyX4Wn5uzKfJjmwTmZp5f7ubJY3CepWV0eZHPnWrA8szJMyvyY1JJ+v8vKoMvuZDEfcxGGJ7VmaTEfsx7p2RxNZ1kedCmLsDmeTrN2pGdzffnWwubNrHL7wFo/D96zemL1+JUy7Hc3teL168+uXneI9V5g2xpNDq1c/EJ1dGs0ybNXR3lucye/9dVZJrOawaA7FGrnPbnDTOos27uz9HslK8OS1UEvFzZH6XclK8N+lgbzubZ7C1WNjszd3Z3MsrE72Z//uzuZJdvj/WHb2+NpHr+4nd+5sJGru/Pe3HTzebdXt8bz1acW83A3d6fplZLTK/2cOTVMryZfeW474+kss9Q8cHo596wMsjurWR300uuS0aTmvrVhlvr9PHhu5dgLUt3I/+1LO9M8cXHruuE5mQ8zz3x0+f4waYtaAXezvcX79hYRvN717uACf0kOrb6/utR/3s3GozclV5f6SU22xvMboHdKG9Axr8y3JWyM5m2SO/FzVkA9IcN+l99z/3qS5PGLm/nys9vpd/Nen8+c38wXvrqR8XSampIuydm1YcpeL0syHyJ6YLXT65XvWR2mJtdWVT2yT1mUL9/AMcd9ndPL/WuL1mS+euvV7WuL2JxZHaRbrBK71yd2anmQrZ1xZououLY0yObu+HnB8fLe72Cxz5Wd0ZEgOdkPklu788V+Ds7J3D6wuM94mkyn8/mfS4uL+fz/ZZdTw0GWB71s7Eyy3OulDuYr0S4NrzMf80jP5rDfe9nDJi/tpYLtjYxUeNO98xsTb33j5CU/GPaGJG+Opjmz0s/960tJko3RZDE8eTFPd9A9f+7uIvheWQxnHi7q28bOfEh26izDrkvtleyO5725O5NZdkbT1K5k0CU7i3o+mk6zMuzlmavbubg1zlK/y85kml4vmUznC1A9eXkzu+NZlgfDnF7p8uzGKEu9LtNac2Z1mKV+L89tzG/w1FpzdnV46P//6lI/W9vjTJN0peTeU9euB10p6UqytjzI5e1xulJSUvPUEzv5V7PfSb8rObMyyOnVQbokT1zcycqwn2SaB86spN+VLPd6WRr0srY8yGg6zT2r80Wtum6eYGe1Hvsm17DfS9JOg+BWj9kb+ZBk/0bMoNcdqwGaHJ4nffAGwMGG8v6xx3jeG3k/e4uOHTz34zrJ3/XeyuYv17/pnVKX7oZjRpP5WgSDrku/V1LrfCTS3ra9fUaT2fwG9+IzfzSZ7o8kKaXsP+/wwD57/wf6R8p7QXE8nWVyYNv+dWdRHk2m1z1mtNj21MY4n/qd5zKeTnNxc5LXnlnONLOcWRnMp9108+d88vJOak2ms2mm05rfvrCVi4s20ZvuWc2kztc16adkMquZpeYrz21nOqvpd8nrzi7n8tYkXSm5ujvJ6rCXqzvzG4Ml82t+cq2tWVKztjw4tM/R8sFj6v6on0Gu7BwcBXS4fHZ1mLr3OosOhdPL88+W1Fy3fG51af+zJYsOhvUD++yNNjpYPnvkmOvt80LHzDspFl8RuWgrv1D53IFvJLhe+/p65XuOHHMS7fqbaqOX+UiuYb+X556b5N94diMP37d2x7U/BdQTtLbcz9e+/kxed3Yl95y6ms8/dSUXNnZzdnmY4aCfqzs1/d68t2U6mSV7vSy9eQN0c7ToxeuVeXl3sr9QzM60Zmc8TZJD2zYOrpA6neXqkfL1jjn6vEdfd2eyOGaxbS8ETmvNsHdt0ZprobBmNpullpLalSwvehyns/lwydVeL6PpLLM6O9Qj2e+SWrv9xX1G01l63TykLvXnQXKp16UOBovX6s0vnKWm7C/C8xIL9/R7WRkM5mFz2mV9pZ833rOSb3jo3P4H4o18gCbC5t1iL+yeO/XSPbkHhyQfbbAkt9bATk2e2djNV69s57cm88bX+pHe3PVFPb+yM850VrM8GOT+08uL4FHTlZKlQS+9XpfepJ/Xnu3nytYk40nN6aVBkprxrCa1ZnP3Wu/teFb2V1leXtzwmUzmwXhlca3avx7Magb9ZHsyy2wrGc2ujUDYHiej6TS9Xi9PX9nKhc3+IjzPsjOdZHNnnO3daXZms2zujNMvvbzmzFKe3djNbFrT65UjN6xKulJyduXaB3O3aERc2hplurjNdWZ1KYNu3igqi2HXZ1aH++VSak6vDPcbX7XOcmbl+R/up5b62Vzss9dbvff1Sb0uhxpFs1mdn8vKvBFUMm+cXGuw1f3G195XO5XkQANtsjjTWdaX5kO9S0r6/eT33Lee16wtZTKb5cLmbjZ2Z1kbdnlmY/4cV3cnObXUP/Q6Z1eW0nXz39e55UHuXRtmadBLvzdvmN9zav6au+NpfvfydrqS7E5qVnpdntrYSS8lV3enWR32srE7Sa01tdb511DtTvbf396/z6WteSOuK/OG7cbOJL2uy7lTgzxy72pqkuc2d3Nxe7r368x0Vhfz7ZNer5c6m2U6q4ubEmU+p3v+L7Ro6O59JdbzG5xJPdSY3G+wlXk9+PLju/mdX/rt5zVS5/9ei99bas6uXGtM7tWd9eXDX4l1emX4og3Hew+87tFz22u0nj5yzLnVQQ42/q/3vHujgvbe37lT16YN1MVd4rXFV37t3fA9Wj5zarh/0/iF9tkrTxf/TnujMg7faO5nc2ecuvgt7d143nuO06vzUUUHp8OsLQ8WxyzK17k5XcqRYw6sRH+98gvd0G71mKefGOUL9al0XclsWvM7F+fXroPXu2mdr50w6PXmayF0XZKSQb9kNJnlty9s5fLObpb6/UUPbMmg67Iznt+U3BrXfOXCTjZ2d7O2PMh4lkwms+weuDbvreC/c+BG/2wrh/Y5Wj56zO5klnpgn+uVNxfHHOxQuLI1PvS8R8vXjrm2Ldvj/c+W3euU9zopjo5IOs4x49ls/3Wu7hz+doGj5et9I8HR9vXR8tZ4mvln7bUOlOe1t4/Rrr/RNvrOtM4XljzQZt/rNe1l3kYfT++8XlQB9YQN+13uP72cs6vDvOneU/ntZzZyZXucZzd35/PlMp8Pt36kl2X9SNC6XvnMymIOXc119zm6Yuqw673kMdd9nf25ejm0SM3RRWuu3+NYr7vQzfKR8sHgOJp088V9el1WB4Ns1cmLBsnk5u/WJtf/Hk14MTc7JPm4Xn9uNZu768fqzd0aTXJ5a5LlYZfptObMymB/3up4Nstnf/fyfNGm0TTnTi3l8s4oX72ym6Qs5u2+yA2eI+XDc3czX215EZ43dyeZzpLVQZfVYbI2HGQ+77qf+9dXMprMsjPeTa90ec3aSmpKJuNZXntmNTu7s2zsjLPU72V52GV7Mt2/YTXvLemyM53t34BbXoTdqzuT1FJyqt9bhPtZRpMcanhc2Rrvl3ePNooWN+3qgUbEaFozHk8zy+LrkhY34CZ7N+SONIr2GzRHGkH1SCNv8+hXO12ngZbZtXPdGdc8fnE7X3zmaq7uTLI06KXrSvql5Mr2tQbo9MhzbO1O0vWSjZ1xnrqc7J6fZX1lkGlNRpNJppN56NuazDKeTNPrerlvfZBnxtNc2DjwvNuzQ7+nqzuTQ+9vsHu4Abp9oGG7MuzlyUuj/O6lnexO5ueeJGvDXlKSK9uj1HnUzdqwl83x3mickpX+kekWRxqtRxucz2tcTueNyb3fdUrJ7vT5jdSjv/uDDdC9bYdu+B6oS8v97lpD8UDDcWtRlw7+Gx89t6ONzd3JbB7aF/Vrv0F65Ibw3g3g0aKxuTdtYDhY1NnFjaRhr1y3fPCm8QvtMz5wM2o0nc1vHJSSWkqW+tduPve6Ra/hgRvPezeay2JcWK/rMlxs2ztmr1wXN6f3yl2ZH7O3bf7ZPO8V3dvnYHm8uHmdxfXh6D57az6U5NAN70GvpM5Khl1v8Z4PH7N3/sOud+DGeUlqed4+h8vJsHftmEGvpNZyeFtXstSfT7OYdslkPMtyv5fllV52RtMMeiVLpTdfDyHz0Ht6eZBnN3Yzndb0S8lyv8tgdSmTWc1oWvOateVMp3X+O0yX0i16ZAe97ExmGXTTnFkZzr/PPPNr89rKIKUms2nZ37a+PJiPgHuh8sr82wWuHdPL+sogG9vXVuN/Xnl1mNT5Gg/X22ewKF89UD69OpzfEJuVQ/tsHPhmgKPfFHBmdZiaeu2bAoa9nD6wz9ERcUsHvl1g73Pv6D7Xe52X+kaC65XPrhw95tbb9cdpox/dZ7BYTHNp0OVqt7ix0buzwmkioL5shv0uD55bPdTz0vJQmtaOSQRJXj1upDc3OTxv6ehd0fvWlrM1mmR1OO/lv95wy5P+v/2JX/ndvOOdb3xeeF4dzufrLvV72RlP8uSlnWxPppkuz7Lc6+0varV06IbV/EbeqSM3tdaPNiL6vf0bY7NZMsu84bG66OGZ1nmDZ21lkKtb82PWhtdu2r1QI+J5oX04b0hl0ShKrjVoruyMM6vXb4wdPGZWDzfQZnXe6DtzoDHW1Zphr8uppWH6ZT63fWc8ycrKMMu91exOZ+m6WU6vLh16nTOrw0xms+ws5kmvDueNkYsb46wvDzMbJr1+0k1KXrO+nK3dWSbTmtXhMP31ecO262Y5u7qUy1vj5Mi0h72bEmdWh5kdaoB28x7U7Ukmk1m60uWetWG2drvM6nye/sqwnzpLxv2aYX8+kmbY6zKdZT9EnFrqp+x0+w3ho43Wg0Pir9u47OaNydnidz0oNWvDwfManGvLg1xZvL/rNUCv12g9daAn8HoNxb0G6Qs2fI80Yg82Lmcv0rh8qa/8Wu0fvZF0nfKRm8YvdczKgf9Pe8f0r7fPkfLacH7MeDEd5ujN6KPlpX4vpxbHjA5MoVlfGiT12j4Hy8N+L6tHrg9H9xn0S9aWBql7q9n3uiz3+4ubHd11X2fY7x0+pt9lZbB3TO95+1z3mIOvM752s31lUHLf+rxn9b6VYaazWS5tz693e8f0esmgNx9psdTr5f4zw7zlwTO5b20pg16XZzZ289zmbnZHk6R0ecM9p9KV7N+UTE0ubY9TU7M7muW1Z5b35/+31J5zzO2bg/orV750Rw7vTQTUl93L3fMCvPrMryvX/8BZW+4furHzSlx/vrDcy4Pn5ot1HQ3PbzpQHk1mubw92p9veXRRq6SdD/dX+pjU7DdILwx2c2qpn3tWh1kZDtLr5UUboOPp4XnSp4ZdloddunTZHk3yutPLubI8zmtOraQrNb/3gfVsjqYv2bA97vvZW3Tsys4oFzbm535wnv6Tl3f2h0bft7aUZzZ2k8zb2K87s3KsufzHPbcXWtncHFTHvJLHfOJXnsi/8w0PJsn+tJDrXe8OOjp9aD6yZrL/2PUWWnqxm5WwNrxz64WACsCJORqeD5aH/e5QeDZC4rC9Buleg/NGVjw+Ok869doCSHsrfY5ns6wO5zcwTrphu7fo2MFz33ve16wvH3qt159bfdka1XfqyubcXb6w3Mv9p5cPbbvR691xOjhe7GYl3Mm0DgCgAS/UID1OA/R6x57LCzduT7ph+2KN6Re7aQEAR/mEAAAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJgioAAAANEFABQAAoAkCKgAAAE0QUAEAAGiCgAoAAEATBFQAAACaIKACAADQBAEVAACAJhwroJZS3l1K+Xwp5YullO+9zuNLpZSPLh7/5VLKwwce+77F9s+XUr7lBM8dAACAu8hLBtRSSi/JjyT5U0m+Lsl3llK+7shu353kYq319yb54SR/Z3Hs1yX5jiT/ZpJ3J/lfFs8HAAAAhxynB/WdSb5Ya/1SrXWU5CeSvOfIPu9J8uHFzz+Z5I+XUspi+0/UWndrrb+d5IuL5wMAAIBDjhNQH0zy+IHy+cW26+5Ta50kuZzk3mMeCwAAAOnf7hNIklLKB5J8YFHcKKV8/naezzHcl+TZ230SEHWRNqiHtEJdpAXqIS1ovR6+6YUeOE5AfSLJGw6UH1psu94+50sp/SRnklw45rGptf5okh89xrk0oZTyyVrr22/3eYC6SAvUQ1qhLtIC9ZAW3Mn18DhDfD+R5M2llEdKKcPMFz362JF9PpbkvYuf/2ySf15rrYvt37FY5feRJG9O8isnc+oAAADcTV6yB7XWOimlfE+Sn03SS/KhWutvllJ+IMkna60fS/KPkvzjUsoXkzyXeYjNYr9/muSzSSZJ/mKtdfoyvRcAAADuYMeag1pr/ekkP31k218/8PNOkn//BY79wSQ/eAvn2KI7Zjgydz11kRaoh7RCXaQF6iEtuGPrYZmPxAUAAIDb6zhzUAEAAOBlJ6DeoFLKu0spny+lfLGU8r23+3y4e5VS3lBK+RellM+WUn6zlPJXFtvvKaX8XCnlC4u/zy22l1LK/7Som/+ylPIHbu874G5SSumVUn6tlPL/LMqPlFJ+eVHfPrpYRC+LRfE+utj+y6WUh2/riXNXKaWcLaX8ZCnlX5VSPldK+bddE3mllVL+6uJz+TdKKT9eSll2TeSVUEr5UCnlq6WU3ziw7YavgaWU9y72/0Ip5b3Xe63bSUC9AaWUXpIfSfKnknxdku8spXzd7T0r7mKTJP9FrfXrkvyhJH9xUd++N8k/q7W+Ock/W5STeb188+LPB5L8vVf+lLmL/ZUknztQ/jtJfrjW+nuTXEzy3Yvt353k4mL7Dy/2g5PyPyb5mVrr70/yjZnXSddEXjGllAeT/OUkb6+1viXzBUS/I66JvDL+1yTvPrLthq6BpZR7kvyNJH8wyTuT/I29UNsKAfXGvDPJF2utX6q1jpL8RJL33OZz4i5Va32y1vqpxc9XM2+IPZh5nfvwYrcPJ/kzi5/fk+Qjde6XkpwtpbzulT1r7kallIeS/HtJ/uGiXJL8sSQ/udjlaD3cq58/meSPL/aHW1JKOZPkXZl/c0BqraNa66W4JvLK6ydZKaX0k6wmeTKuibwCaq0fz/wbUw660WvgtyT5uVrrc7XWi0l+Ls8PvbeVgHpjHkzy+IHy+cU2eFkthgS9LckvJ3mg1vrk4qGnkjyw+Fn95OXyPyT5r5LMFuV7k1yqtU4W5YN1bb8eLh6/vNgfbtUjSZ5J8mOL4eb/sJRyKq6JvIJqrU8k+btJvpJ5ML2c5Ffjmsjtc6PXwOavjQIqNK6Uspbk/0jyn9Varxx8rM6X4bYUNy+bUsqfTvLVWuuv3u5z4VWvn+QPJPl7tda3JdnMtaFsSVwTefkthkK+J/MbJq9PciqN9T7x6nW3XAMF1BvzRJI3HCg/tNgGL4tSyiDzcPq/1Vp/arH56b1haou/v7rYrn7ycvjDSb61lPI7mU9r+GOZzwM8uxjelhyua/v1cPH4mSQXXskT5q51Psn5WusvL8o/mXlgdU3klfQnkvx2rfWZWus4yU9lfp10TeR2udFrYPPXRgH1xnwiyZsXK7UNM58U/7HbfE7cpRZzVP5Rks/VWv/7Aw99LMneimvvTfJ/Hdj+Hy1WbftDSS4fGPIBN6XW+n211odqrQ9nfs3757XW/zDJv0jyZxe7Ha2He/Xzzy72v+Pv5nL71VqfSvJ4KeX3LTb98SSfjWsir6yvJPlDpZTVxef0Xj10TeR2udFr4M8m+eZSyrnFiIBvXmxrRvF/5MaUUv7dzOdj9ZJ8qNb6g7f3jLhblVL+SJKfT/LruTb377/OfB7qP03yxiRfTvIf1FqfW3xQ/s+ZDzXaSvJdtdZPvuInzl2rlPJokv+y1vqnSylfk3mP6j1Jfi3Jn6+17pZSlpP848znTD+X5DtqrV+6TafMXaaU8tbMF+saJvlSku/K/Ga7ayKvmFLKf5Pkz2W+2v6vJXl/5nP4XBN5WZVSfjzJo0nuS/J05qvx/p+5wWtgKeU/zrxNmSQ/WGv9sVfwbbwkARUAAIAmGOILAABAEwRUAAAAmiCgAgAA0AQBFQAAgCYIqAAAADRBQAUAAKAJAioAAABNEFABAABowv8PKzLmvVx9Fa4AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1152x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot generator error\n",
    "\n",
    "G.plot_progress()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "vddYtD84xkmI"
   },
   "source": [
    "## Manually Run Generator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "mePLPy3Yxj9A",
    "outputId": "350aedcf-49c2-48bd-8e09-2bd2ebcaf844"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Tensor(shape=[4], dtype=Float32, value= [ 9.69303131e-01,  2.38597877e-02,  9.69235599e-01,  2.38773953e-02])"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# manually run generator to see it's outputs\n",
    "\n",
    "G.construct(mindspore.Tensor([0.5], mindspore.float32))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "0gZ4BG6O53SZ"
   },
   "source": [
    "## Visualise Pattern During Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 426
    },
    "colab_type": "code",
    "id": "jxHrZmUyxm-z",
    "outputId": "678d51ca-14f5-4579-a24f-296314913284"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0xffff21a6e890>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot images collected during training\n",
    "\n",
    "plt.figure(figsize = (16,8))\n",
    "\n",
    "plt.imshow(numpy.array(image_list).T, interpolation='none', cmap='Blues')"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "name": "06_gan_simple_pattern.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}