{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"# REFERENCES:\nhttps://www.kaggle.com/amyjang/monet-cyclegan-tutorial\n\nhttps://www.kaggle.com/unfriendlyai/diffaugment-is-all-you-need\n\nhttps://www.kaggle.com/victorsullivan/i-m-something-of-a-painter-myself/notebook","metadata":{}},{"cell_type":"code","source":"# This Python 3 environment comes with many helpful analytics libraries installed\n# It is defined by the kaggle/python Docker image: https://github.com/kaggle/docker-python\n# For example, here's several helpful packages to load\n\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n\n# Input data files are available in the read-only \"../input/\" directory\n# For example, running this (by clicking run or pressing Shift+Enter) will list all files under the input directory\n\nimport os\n# for dirname, _, filenames in os.walk('/kaggle/input'):\n#     for filename in filenames:\n#         print(os.path.join(dirname, filename))\n\n# You can write up to 20GB to the current directory (/kaggle/working/) that gets preserved as output when you create a version using \"Save & Run All\" \n# You can also write temporary files to /kaggle/temp/, but they won't be saved outside of the current session","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:03.873674Z","iopub.execute_input":"2021-09-23T19:37:03.874031Z","iopub.status.idle":"2021-09-23T19:37:03.883925Z","shell.execute_reply.started":"2021-09-23T19:37:03.873955Z","shell.execute_reply":"2021-09-23T19:37:03.883076Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import tensorflow as tf\nfrom tensorflow.keras import layers\nimport tensorflow_addons as tfa\nfrom kaggle_datasets import KaggleDatasets\nimport matplotlib.pyplot as plt","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:03.88526Z","iopub.execute_input":"2021-09-23T19:37:03.885761Z","iopub.status.idle":"2021-09-23T19:37:08.621328Z","shell.execute_reply.started":"2021-09-23T19:37:03.885725Z","shell.execute_reply":"2021-09-23T19:37:08.62052Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"try:\n    tpu = tf.distribute.cluster_resolver.TPUClusterResolver()\n    print(\"Device: \", tpu.master())\n    tf.config.experimental_connect_to_cluster(tpu)\n    tf.tpu.experimental.initialize_tpu_system(tpu)\n    strategy = tf.distribute.experimental.TPUStrategy(tpu)\nexcept:\n    strategy = tf.distribute.get_strategy()\nprint('Number of replicas:', strategy.num_replicas_in_sync)\n\nAUTOTUNE = tf.data.experimental.AUTOTUNE","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:08.622898Z","iopub.execute_input":"2021-09-23T19:37:08.623203Z","iopub.status.idle":"2021-09-23T19:37:08.634283Z","shell.execute_reply.started":"2021-09-23T19:37:08.623171Z","shell.execute_reply":"2021-09-23T19:37:08.63352Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"GCS_PATH = KaggleDatasets().get_gcs_path()","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:08.636471Z","iopub.execute_input":"2021-09-23T19:37:08.636892Z","iopub.status.idle":"2021-09-23T19:37:09.083782Z","shell.execute_reply.started":"2021-09-23T19:37:08.636857Z","shell.execute_reply":"2021-09-23T19:37:09.082929Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"MONET_FILENAMES = tf.io.gfile.glob(str(GCS_PATH + '/monet_tfrec/*.tfrec'))\nprint('Monet TFRecord Files:', len(MONET_FILENAMES))\n\nPHOTO_FILENAMES = tf.io.gfile.glob(str(GCS_PATH + '/photo_tfrec/*.tfrec'))\nprint('Photo TFRecord Files:', len(PHOTO_FILENAMES))","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:09.085075Z","iopub.execute_input":"2021-09-23T19:37:09.085423Z","iopub.status.idle":"2021-09-23T19:37:09.737579Z","shell.execute_reply.started":"2021-09-23T19:37:09.085386Z","shell.execute_reply":"2021-09-23T19:37:09.736703Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"IMAGE_SIZE = [256,256]\n\ndef decode_image(image):\n    image = tf.image.decode_jpeg(image, channels=3)\n    image = (tf.cast(image, tf.float32)/127.5)-1\n    image = tf.reshape(image, [*IMAGE_SIZE, 3])\n    return image\n\ndef read_tfrecord(example):\n    tfrecord_format = {\n        \"image_name\": tf.io.FixedLenFeature([], tf.string),\n        \"image\": tf.io.FixedLenFeature([], tf.string),\n        \"target\": tf.io.FixedLenFeature([], tf.string)\n    }\n    example = tf.io.parse_single_example(example, tfrecord_format)\n    image = decode_image(example['image'])\n    return image","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:09.738833Z","iopub.execute_input":"2021-09-23T19:37:09.739359Z","iopub.status.idle":"2021-09-23T19:37:09.746558Z","shell.execute_reply.started":"2021-09-23T19:37:09.739318Z","shell.execute_reply":"2021-09-23T19:37:09.74569Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import re\ndef count_data_items(filenames):\n    n = [int(re.compile(r\"-([0-9]*)\\.\").search(filename).group(1)) for filename in filenames]\n    return np.sum(n)\n\nn_monet_samples = count_data_items(MONET_FILENAMES)\nn_photo_samples = count_data_items(PHOTO_FILENAMES)","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:09.749576Z","iopub.execute_input":"2021-09-23T19:37:09.749831Z","iopub.status.idle":"2021-09-23T19:37:09.760013Z","shell.execute_reply.started":"2021-09-23T19:37:09.749805Z","shell.execute_reply":"2021-09-23T19:37:09.759269Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# LOAD AND PREPARE DATASET","metadata":{}},{"cell_type":"markdown","source":"Load dataset and plot images","metadata":{}},{"cell_type":"code","source":"def load_dataset(filenames, labeled=True, ordered=False):\n    dataset = tf.data.TFRecordDataset(filenames)\n    dataset = dataset.map(read_tfrecord, num_parallel_calls=AUTOTUNE)\n    return dataset","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:09.762226Z","iopub.execute_input":"2021-09-23T19:37:09.762627Z","iopub.status.idle":"2021-09-23T19:37:09.76874Z","shell.execute_reply.started":"2021-09-23T19:37:09.762589Z","shell.execute_reply":"2021-09-23T19:37:09.76792Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"monet_dataset = load_dataset(MONET_FILENAMES).batch(1)\nphoto_dataset = load_dataset(PHOTO_FILENAMES).batch(1)","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:09.770052Z","iopub.execute_input":"2021-09-23T19:37:09.770529Z","iopub.status.idle":"2021-09-23T19:37:11.692035Z","shell.execute_reply.started":"2021-09-23T19:37:09.770494Z","shell.execute_reply":"2021-09-23T19:37:11.691181Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"## Used for the metric calculation\nfid_photo_ds = load_dataset(PHOTO_FILENAMES).take(1024).batch(32*strategy.num_replicas_in_sync).prefetch(32)\nfid_monet_ds = load_dataset(MONET_FILENAMES).batch(32*strategy.num_replicas_in_sync).prefetch(32)","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:11.694302Z","iopub.execute_input":"2021-09-23T19:37:11.694644Z","iopub.status.idle":"2021-09-23T19:37:11.73112Z","shell.execute_reply.started":"2021-09-23T19:37:11.694609Z","shell.execute_reply":"2021-09-23T19:37:11.730385Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"example_monet = next(iter(monet_dataset))\nexample_photo = next(iter(photo_dataset))","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:11.732488Z","iopub.execute_input":"2021-09-23T19:37:11.732816Z","iopub.status.idle":"2021-09-23T19:37:12.56333Z","shell.execute_reply.started":"2021-09-23T19:37:11.732783Z","shell.execute_reply":"2021-09-23T19:37:12.562462Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"plt.subplot(121)\nplt.title('Photo')\nplt.imshow(example_photo[0] * 0.5 + 0.5)\n# plt.imshow(example_photo[0])\n\nplt.subplot(122)\nplt.title('Monet')\nplt.imshow(example_monet[0] * 0.5 + 0.5)\n# plt.imshow(example_monet[0])","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:12.567244Z","iopub.execute_input":"2021-09-23T19:37:12.567597Z","iopub.status.idle":"2021-09-23T19:37:13.126149Z","shell.execute_reply.started":"2021-09-23T19:37:12.567562Z","shell.execute_reply":"2021-09-23T19:37:13.12532Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Load and 'prepare' the dataset","metadata":{}},{"cell_type":"code","source":"def prepare_dataset(monet_filenames, photo_filenames):\n    \n    monet_dataset = load_dataset(monet_filenames)\n    photo_dataset = load_dataset(photo_filenames)  \n        \n#     monet_dataset = monet_dataset.shuffle(100000, seed=42)\n#     photo_dataset = photo_dataset.shuffle(100000, seed=42)\n    \n    monet_dataset = monet_dataset.repeat()\n    photo_dataset = photo_dataset.repeat()\n    \n#     monet_dataset = monet_dataset.batch(16)\n#     photo_dataset = photo_dataset.batch(16)\n    \n    monet_dataset = monet_dataset.batch(4)\n    photo_dataset = photo_dataset.batch(4)\n    \n    monet_dataset = monet_dataset.prefetch(AUTOTUNE)\n    photo_dataset = photo_dataset.prefetch(AUTOTUNE)\n    \n    final_dataset = tf.data.Dataset.zip((monet_dataset, photo_dataset))\n    \n    return final_dataset","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:13.127206Z","iopub.execute_input":"2021-09-23T19:37:13.12757Z","iopub.status.idle":"2021-09-23T19:37:13.134702Z","shell.execute_reply.started":"2021-09-23T19:37:13.127533Z","shell.execute_reply":"2021-09-23T19:37:13.133574Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"final_dataset = prepare_dataset(MONET_FILENAMES, PHOTO_FILENAMES)","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:13.136379Z","iopub.execute_input":"2021-09-23T19:37:13.136796Z","iopub.status.idle":"2021-09-23T19:37:13.180415Z","shell.execute_reply.started":"2021-09-23T19:37:13.13676Z","shell.execute_reply":"2021-09-23T19:37:13.179716Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# FRECHET INCEPTION DISTANCE","metadata":{}},{"cell_type":"code","source":"with strategy.scope():\n\n    inception_model = tf.keras.applications.InceptionV3(input_shape=(256,256,3),pooling=\"avg\",include_top=False)\n\n    mix3  = inception_model.get_layer(\"mixed9\").output\n    f0 = tf.keras.layers.GlobalAveragePooling2D()(mix3)\n\n    inception_model = tf.keras.Model(inputs=inception_model.input, outputs=f0)\n    inception_model.trainable = False\n\n    \n    \n    def calculate_activation_statistics_mod(images,fid_model):\n\n            act=tf.cast(fid_model.predict(images), tf.float32)\n\n            mu = tf.reduce_mean(act, axis=0)\n            mean_x = tf.reduce_mean(act, axis=0, keepdims=True)\n            mx = tf.matmul(tf.transpose(mean_x), mean_x)\n            vx = tf.matmul(tf.transpose(act), act)/tf.cast(tf.shape(act)[0], tf.float32)\n            sigma = vx - mx\n            return mu, sigma\n    myFID_mu2, myFID_sigma2 = calculate_activation_statistics_mod(fid_monet_ds,inception_model)\n    fids=[]","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:13.181684Z","iopub.execute_input":"2021-09-23T19:37:13.182027Z","iopub.status.idle":"2021-09-23T19:37:24.252455Z","shell.execute_reply.started":"2021-09-23T19:37:13.181994Z","shell.execute_reply":"2021-09-23T19:37:24.251424Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    def calculate_frechet_distance(mu1,sigma1,mu2,sigma2):\n        fid_epsilon = 1e-14\n       \n        covmean = tf.linalg.sqrtm(tf.cast(tf.matmul(sigma1,sigma2),tf.complex64))\n\n        covmean = tf.cast(tf.math.real(covmean),tf.float32)\n  \n        tr_covmean = tf.linalg.trace(covmean)\n\n        return tf.matmul(tf.expand_dims(mu1 - mu2, axis=0),tf.expand_dims(mu1 - mu2, axis=1)) + tf.linalg.trace(sigma1) + tf.linalg.trace(sigma2) - 2 * tr_covmean\n    \n    \n    def FID(images,gen_model,inception_model=inception_model,myFID_mu2=myFID_mu2, myFID_sigma2=myFID_sigma2):\n        inp = layers.Input(shape=[256, 256, 3], name='input_image')\n        x  = gen_model(inp)\n        x=inception_model(x)\n        fid_model = tf.keras.Model(inputs=inp, outputs=x)\n\n        mu1, sigma1= calculate_activation_statistics_mod(images,fid_model)\n\n        fid_value = calculate_frechet_distance(mu1, sigma1,myFID_mu2, myFID_sigma2)\n\n\n        return fid_value","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:24.254094Z","iopub.execute_input":"2021-09-23T19:37:24.254431Z","iopub.status.idle":"2021-09-23T19:37:24.265294Z","shell.execute_reply.started":"2021-09-23T19:37:24.254393Z","shell.execute_reply":"2021-09-23T19:37:24.264362Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# GENERATOR","metadata":{}},{"cell_type":"code","source":"OUTPUT_CHANNELS = 3\n\ndef downsample(filters, size, apply_instancenorm=True):\n    initializer = tf.random_normal_initializer(0., 0.02)\n    gamma_init = tf.keras.initializers.RandomNormal(mean=0., stddev=0.02)\n    \n    model_down = tf.keras.Sequential()\n    model_down.add(layers.Conv2D(filters, size, strides=2, padding='same', \n                                kernel_initializer=initializer, use_bias=False))\n    \n    if apply_instancenorm:\n        model_down.add(tfa.layers.InstanceNormalization(gamma_initializer=gamma_init))\n    \n    model_down.add(layers.LeakyReLU())\n    \n    return model_down\n    ","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:24.266743Z","iopub.execute_input":"2021-09-23T19:37:24.267094Z","iopub.status.idle":"2021-09-23T19:37:24.276163Z","shell.execute_reply.started":"2021-09-23T19:37:24.267061Z","shell.execute_reply":"2021-09-23T19:37:24.275446Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def upsample(filters, size, apply_dropout=False):\n    initializer = tf.random_normal_initializer(0., 0.02)\n    gamma_init = tf.keras.initializers.RandomNormal(mean=0., stddev=0.02)\n    \n    model_up = tf.keras.Sequential()\n    model_up.add(layers.Conv2DTranspose(filters, size, strides=2, padding='same', \n                                kernel_initializer=initializer, use_bias=False))\n    \n    model_up.add(tfa.layers.InstanceNormalization(gamma_initializer=gamma_init))\n    \n    if apply_dropout:\n        model_up.add(layers.Dropout(0.5))\n    \n    model_up.add(layers.ReLU())\n    \n    return model_up   ","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:24.277639Z","iopub.execute_input":"2021-09-23T19:37:24.278014Z","iopub.status.idle":"2021-09-23T19:37:24.28587Z","shell.execute_reply.started":"2021-09-23T19:37:24.27798Z","shell.execute_reply":"2021-09-23T19:37:24.284934Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def Generator():\n    inputs = layers.Input([256,256,3])\n    \n    down_stack = [\n        downsample(64, 4, apply_instancenorm=False),\n        downsample(128, 4),\n        downsample(256, 4),\n        downsample(512, 4),\n        downsample(512, 4),\n        downsample(512, 4),\n        downsample(512, 4),\n        downsample(512, 4),\n    ]\n    \n    up_stack = [\n        upsample(512, 4, apply_dropout=True),\n        upsample(512, 4, apply_dropout=True),\n        upsample(512, 4, apply_dropout=True),\n        upsample(512, 4),\n        upsample(256, 4),\n        upsample(128, 4),\n        upsample(64, 4),\n    ]\n    \n    initializer = tf.random_normal_initializer(0., 0.02)\n    last = layers.Conv2DTranspose(OUTPUT_CHANNELS, 4, strides=2, padding='same', \n                                kernel_initializer=initializer, activation='tanh')\n    \n    x = inputs\n    \n    skips = []\n    for down in down_stack:\n        x = down(x)\n        skips.append(x)\n    \n    skips = reversed(skips[:-1])\n    \n    for up, skip in zip(up_stack, skips):\n        x = up(x)\n        x = layers.Concatenate()([x, skip])\n    \n    x = last(x)\n    \n    return tf.keras.Model(inputs=inputs, outputs=x)\n    ","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:24.290308Z","iopub.execute_input":"2021-09-23T19:37:24.290669Z","iopub.status.idle":"2021-09-23T19:37:24.301266Z","shell.execute_reply.started":"2021-09-23T19:37:24.290636Z","shell.execute_reply":"2021-09-23T19:37:24.300455Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"Generator()","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:24.303522Z","iopub.execute_input":"2021-09-23T19:37:24.303898Z","iopub.status.idle":"2021-09-23T19:37:25.57447Z","shell.execute_reply.started":"2021-09-23T19:37:24.303863Z","shell.execute_reply":"2021-09-23T19:37:25.573657Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# DISCRIMINATOR","metadata":{}},{"cell_type":"code","source":"def Discriminator():\n    initializer = tf.random_normal_initializer(0., 0.02)\n    gamma_init = tf.keras.initializers.RandomNormal(mean=0., stddev=0.02)\n    \n    inp = layers.Input(shape=[256, 256, 3], name='input_image')\n    \n    x = inp\n    \n    down1 = downsample(64, 4, False)(x)\n    down2 = downsample(128, 4)(down1)\n    down3 = downsample(256, 4)(down2)\n    \n    zero_pad1 = layers.ZeroPadding2D()(down3)\n    \n    conv = layers.Conv2D(512, 4, strides=1, \n                        kernel_initializer=initializer, use_bias=False)(zero_pad1)\n    \n    norm1 = tfa.layers.InstanceNormalization(gamma_initializer=gamma_init)(conv)\n    \n    leaky_relu = layers.LeakyReLU()(norm1)\n    \n    zero_pad2 = layers.ZeroPadding2D()(leaky_relu)\n    \n    last = layers.Conv2D(1, 4, strides=1, \n                        kernel_initializer=initializer)(zero_pad2)\n    \n    return tf.keras.Model(inputs=inp, outputs=last)","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:25.575632Z","iopub.execute_input":"2021-09-23T19:37:25.575972Z","iopub.status.idle":"2021-09-23T19:37:25.585871Z","shell.execute_reply.started":"2021-09-23T19:37:25.575943Z","shell.execute_reply":"2021-09-23T19:37:25.585054Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"Discriminator()","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:25.586989Z","iopub.execute_input":"2021-09-23T19:37:25.587506Z","iopub.status.idle":"2021-09-23T19:37:25.751572Z","shell.execute_reply.started":"2021-09-23T19:37:25.587468Z","shell.execute_reply":"2021-09-23T19:37:25.750654Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    monet_generator = Generator()\n    photo_generator = Generator()\n    \n    monet_discriminator = Discriminator()\n    photo_discriminator = Discriminator()","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:25.752849Z","iopub.execute_input":"2021-09-23T19:37:25.753192Z","iopub.status.idle":"2021-09-23T19:37:27.738798Z","shell.execute_reply.started":"2021-09-23T19:37:25.753156Z","shell.execute_reply":"2021-09-23T19:37:27.738004Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"to_monet = monet_generator(example_photo)\n\nplt.subplot(1, 2, 1)\nplt.title(\"Original Photo\")\nplt.imshow(example_photo[0] * 0.5 + 0.5)\n\nplt.subplot(1, 2, 2)\nplt.title(\"Monet-esque Photo\")\nplt.imshow(to_monet[0] * 0.5 + 0.5)\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:27.740081Z","iopub.execute_input":"2021-09-23T19:37:27.740425Z","iopub.status.idle":"2021-09-23T19:37:28.201757Z","shell.execute_reply.started":"2021-09-23T19:37:27.740388Z","shell.execute_reply":"2021-09-23T19:37:28.20071Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# CYCLEGAN MODEL","metadata":{}},{"cell_type":"markdown","source":"CycleGAN model with custom model.fit","metadata":{}},{"cell_type":"code","source":"class CycleGan(tf.keras.Model):\n    def __init__(\n        self,\n        monet_generator,\n        photo_generator,\n        monet_discriminator,\n        photo_discriminator, \n        lambda_cycle=10):\n        super(CycleGan, self).__init__()\n        self.m_gen = monet_generator\n        self.p_gen = photo_generator\n        self.m_disc = monet_discriminator\n        self.p_disc = photo_discriminator\n        self.lambda_cycle = lambda_cycle\n    \n    def compile(\n        self,\n        m_gen_optimizer,\n        p_gen_optimizer,\n        m_disc_optimizer,\n        p_disc_optimizer,\n        gen_loss_fn,\n        disc_loss_fn,\n        cycle_loss_fn,\n        identity_loss_fn,\n        aug_fn\n        ):\n        super(CycleGan, self).compile()\n        self.m_gen_optimizer = m_gen_optimizer\n        self.p_gen_optimizer = p_gen_optimizer\n        self.m_disc_optimizer = m_disc_optimizer\n        self.p_disc_optimizer = p_disc_optimizer\n        self.gen_loss_fn = gen_loss_fn\n        self.disc_loss_fn = disc_loss_fn\n        self.cycle_loss_fn = cycle_loss_fn\n        self.identity_loss_fn = identity_loss_fn\n        self.aug_fn = aug_fn\n        \n    def train_step(self, batch_data):\n        real_monet, real_photo = batch_data\n        batch_size = tf.shape(real_monet)[0]\n        \n        with tf.GradientTape(persistent=True) as tape:\n            fake_monet = self.m_gen(real_photo, training=True)\n            cycled_photo = self.p_gen(fake_monet, training=True)\n            \n            fake_photo = self.p_gen(real_monet, training=True)\n            cycled_monet = self.m_gen(fake_photo, training=True)\n            \n            same_monet = self.m_gen(real_monet, training=True)\n            same_photo = self.p_gen(real_photo, training=True)\n            \n            # With DiffAugment\n            both_monet = tf.concat([real_monet, fake_monet], axis=0)\n            aug_monet = self.aug_fn(both_monet)\n\n            aug_real_monet = aug_monet[:batch_size]\n            aug_fake_monet = aug_monet[batch_size:]\n\n            disc_real_monet = self.m_disc(aug_real_monet, training=True)\n            disc_real_photo = self.p_disc(real_photo, training=True)\n            \n            disc_fake_monet = self.m_disc(aug_fake_monet, training=True)\n            disc_fake_photo = self.p_disc(fake_photo, training=True)\n\n            monet_gen_loss = self.gen_loss_fn(disc_fake_monet)\n            photo_gen_loss = self.gen_loss_fn(disc_fake_photo)\n            \n            total_cycle_loss = self.cycle_loss_fn(real_monet, cycled_monet, self.lambda_cycle) + self.cycle_loss_fn(real_photo, cycled_photo, self.lambda_cycle)\n            \n            total_monet_gen_loss = total_cycle_loss + monet_gen_loss + self.identity_loss_fn(real_monet,same_monet, self.lambda_cycle)\n            total_photo_gen_loss = total_cycle_loss + photo_gen_loss + self.identity_loss_fn(real_photo,same_photo, self.lambda_cycle)\n            \n            monet_disc_loss = self.disc_loss_fn(disc_real_monet, disc_fake_monet)\n            photo_disc_loss = self.disc_loss_fn(disc_real_photo, disc_fake_photo)\n\n        monet_generator_gradients = tape.gradient(total_monet_gen_loss, self.m_gen.trainable_variables) \n        photo_generator_gradients = tape.gradient(total_photo_gen_loss, self.p_gen.trainable_variables) \n\n        monet_discriminator_gradients = tape.gradient(monet_disc_loss, self.m_disc.trainable_variables) \n        photo_discriminator_gradients = tape.gradient(photo_disc_loss, self.p_disc.trainable_variables)\n\n        self.m_gen_optimizer.apply_gradients(zip(monet_generator_gradients, self.m_gen.trainable_variables))\n        self.p_gen_optimizer.apply_gradients(zip(photo_generator_gradients, self.p_gen.trainable_variables))\n        self.m_disc_optimizer.apply_gradients(zip(monet_discriminator_gradients, self.m_disc.trainable_variables))\n        self.p_disc_optimizer.apply_gradients(zip(photo_discriminator_gradients, self.p_disc.trainable_variables))\n\n        return {\n            \"monet_gen_loss\":total_monet_gen_loss,\n            \"photo_gen_loss\":total_photo_gen_loss,\n            \"monet_disc_loss\":monet_disc_loss,\n            \"photo_disc_loss\":photo_disc_loss\n        }","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:28.203267Z","iopub.execute_input":"2021-09-23T19:37:28.203631Z","iopub.status.idle":"2021-09-23T19:37:28.221259Z","shell.execute_reply.started":"2021-09-23T19:37:28.203596Z","shell.execute_reply":"2021-09-23T19:37:28.220245Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# LOSS FUNCTIONS","metadata":{}},{"cell_type":"code","source":"with strategy.scope():\n    def discriminator_loss(real, generated):\n        real_loss = tf.keras.losses.BinaryCrossentropy(from_logits=True, reduction=tf.keras.losses.Reduction.NONE)(tf.ones_like(real), real)\n        \n        generator_loss = tf.keras.losses.BinaryCrossentropy(from_logits=True, reduction=tf.keras.losses.Reduction.NONE)(tf.zeros_like(generated), generated)\n        \n        total_disc_loss = real_loss + generator_loss\n        \n        return 0.5*total_disc_loss","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:28.222733Z","iopub.execute_input":"2021-09-23T19:37:28.22309Z","iopub.status.idle":"2021-09-23T19:37:28.234121Z","shell.execute_reply.started":"2021-09-23T19:37:28.223053Z","shell.execute_reply":"2021-09-23T19:37:28.233282Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    def generator_loss(generated):\n        return tf.keras.losses.BinaryCrossentropy(from_logits=True, reduction=tf.keras.losses.Reduction.NONE)(tf.ones_like(generated), generated)","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:28.23526Z","iopub.execute_input":"2021-09-23T19:37:28.235694Z","iopub.status.idle":"2021-09-23T19:37:28.246321Z","shell.execute_reply.started":"2021-09-23T19:37:28.235657Z","shell.execute_reply":"2021-09-23T19:37:28.245521Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    def cycle_loss(real_image, cycled_image, LAMBDA):\n        return LAMBDA*(tf.reduce_mean(tf.abs(real_image-cycled_image)))","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:28.247676Z","iopub.execute_input":"2021-09-23T19:37:28.248189Z","iopub.status.idle":"2021-09-23T19:37:28.255843Z","shell.execute_reply.started":"2021-09-23T19:37:28.248153Z","shell.execute_reply":"2021-09-23T19:37:28.255083Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    def identity_loss(real_image, gen_image, LAMBDA):\n        return 0.5*LAMBDA*(tf.reduce_mean(tf.abs(real_image-gen_image)))","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:28.257026Z","iopub.execute_input":"2021-09-23T19:37:28.257415Z","iopub.status.idle":"2021-09-23T19:37:28.26495Z","shell.execute_reply.started":"2021-09-23T19:37:28.25738Z","shell.execute_reply":"2021-09-23T19:37:28.264239Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# DiffAugment","metadata":{}},{"cell_type":"code","source":"with strategy.scope():\n# Differentiable Augmentation for Data-Efficient GAN Training\n# Shengyu Zhao, Zhijian Liu, Ji Lin, Jun-Yan Zhu, and Song Han\n# https://arxiv.org/pdf/2006.10738\n# from https://github.com/mit-han-lab/data-efficient-gans/blob/master/DiffAugment_tf.py\n\n    def DiffAugment(x, policy='', channels_first=False):\n        if policy:\n            if channels_first:\n                x = tf.transpose(x, [0, 2, 3, 1])\n            for p in policy.split(','):\n                for f in AUGMENT_FNS[p]:\n                    x = f(x)\n            if channels_first:\n                x = tf.transpose(x, [0, 3, 1, 2])\n        return x\n\n\n    def rand_brightness(x):\n        magnitude = tf.random.uniform([tf.shape(x)[0], 1, 1, 1]) - 0.5\n        x = x + magnitude\n        return x\n\n\n    def rand_saturation(x):\n        magnitude = tf.random.uniform([tf.shape(x)[0], 1, 1, 1]) * 2\n        x_mean = tf.reduce_sum(x, axis=3, keepdims=True) * 0.3333333333333333333\n        x = (x - x_mean) * magnitude + x_mean\n        return x\n\n\n    def rand_contrast(x):\n        magnitude = tf.random.uniform([tf.shape(x)[0], 1, 1, 1]) + 0.5\n        x_mean = tf.reduce_sum(x, axis=[1, 2, 3], keepdims=True) * 5.086e-6\n        x = (x - x_mean) * magnitude + x_mean\n        return x\n\n    def rand_translation(x, ratio=0.125):\n        batch_size = tf.shape(x)[0]\n        image_size = tf.shape(x)[1:3]\n        shift = tf.cast(tf.cast(image_size, tf.float32) * ratio + 0.5, tf.int32)\n        translation_x = tf.random.uniform([batch_size, 1], -shift[0], shift[0] + 1, dtype=tf.int32)\n        translation_y = tf.random.uniform([batch_size, 1], -shift[1], shift[1] + 1, dtype=tf.int32)\n        grid_x = tf.clip_by_value(tf.expand_dims(tf.range(image_size[0], dtype=tf.int32), 0) + translation_x + 1, 0, image_size[0] + 1)\n        grid_y = tf.clip_by_value(tf.expand_dims(tf.range(image_size[1], dtype=tf.int32), 0) + translation_y + 1, 0, image_size[1] + 1)\n        x = tf.gather_nd(tf.pad(x, [[0, 0], [1, 1], [0, 0], [0, 0]]), tf.expand_dims(grid_x, -1), batch_dims=1)\n        x = tf.transpose(tf.gather_nd(tf.pad(tf.transpose(x, [0, 2, 1, 3]), [[0, 0], [1, 1], [0, 0], [0, 0]]), tf.expand_dims(grid_y, -1), batch_dims=1), [0, 2, 1, 3])\n        return x\n\n\n    def rand_cutout(x, ratio=0.5):\n        batch_size = tf.shape(x)[0]\n        image_size = tf.shape(x)[1:3]\n        cutout_size = tf.cast(tf.cast(image_size, tf.float32) * ratio + 0.5, tf.int32)\n        offset_x = tf.random.uniform([tf.shape(x)[0], 1, 1], maxval=image_size[0] + (1 - cutout_size[0] % 2), dtype=tf.int32)\n        offset_y = tf.random.uniform([tf.shape(x)[0], 1, 1], maxval=image_size[1] + (1 - cutout_size[1] % 2), dtype=tf.int32)\n        grid_batch, grid_x, grid_y = tf.meshgrid(tf.range(batch_size, dtype=tf.int32), tf.range(cutout_size[0], dtype=tf.int32), tf.range(cutout_size[1], dtype=tf.int32), indexing='ij')\n        cutout_grid = tf.stack([grid_batch, grid_x + offset_x - cutout_size[0] // 2, grid_y + offset_y - cutout_size[1] // 2], axis=-1)\n        mask_shape = tf.stack([batch_size, image_size[0], image_size[1]])\n        cutout_grid = tf.maximum(cutout_grid, 0)\n        cutout_grid = tf.minimum(cutout_grid, tf.reshape(mask_shape - 1, [1, 1, 1, 3]))\n        mask = tf.maximum(1 - tf.scatter_nd(cutout_grid, tf.ones([batch_size, cutout_size[0], cutout_size[1]], dtype=tf.float32), mask_shape), 0)\n        x = x * tf.expand_dims(mask, axis=3)\n        return x\n\n\n    AUGMENT_FNS = {\n        'color': [rand_brightness, rand_saturation, rand_contrast],\n        'translation': [rand_translation],\n        'cutout': [rand_cutout],\n}\n    def aug_fn(image):\n        return DiffAugment(image,\"color,translation,cutout\")\n","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:28.26754Z","iopub.execute_input":"2021-09-23T19:37:28.267795Z","iopub.status.idle":"2021-09-23T19:37:28.291986Z","shell.execute_reply.started":"2021-09-23T19:37:28.267771Z","shell.execute_reply":"2021-09-23T19:37:28.290987Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# TRAIN THE NETWORK","metadata":{}},{"cell_type":"code","source":"with strategy.scope():\n    monet_generator_optimizer = tf.keras.optimizers.Adam(2e-4, beta_1=0.5)\n    photo_generator_optimizer = tf.keras.optimizers.Adam(2e-4, beta_1=0.5)\n    \n    monet_discriminator_optimizer = tf.keras.optimizers.Adam(2e-4, beta_1=0.5)\n    photo_discriminator_optimizer = tf.keras.optimizers.Adam(2e-4, beta_1=0.5)","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:28.293188Z","iopub.execute_input":"2021-09-23T19:37:28.293584Z","iopub.status.idle":"2021-09-23T19:37:28.302422Z","shell.execute_reply.started":"2021-09-23T19:37:28.293552Z","shell.execute_reply":"2021-09-23T19:37:28.30164Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    cycle_gan_model = CycleGan(monet_generator, photo_generator, monet_discriminator, photo_discriminator)\n    \n    cycle_gan_model.compile(\n        m_gen_optimizer = monet_generator_optimizer,\n        p_gen_optimizer = photo_generator_optimizer,\n        m_disc_optimizer = monet_discriminator_optimizer,\n        p_disc_optimizer = photo_discriminator_optimizer,\n        gen_loss_fn = generator_loss,\n        disc_loss_fn = discriminator_loss,\n        cycle_loss_fn = cycle_loss,\n        identity_loss_fn = identity_loss,\n        aug_fn = aug_fn\n    )","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:28.303822Z","iopub.execute_input":"2021-09-23T19:37:28.304181Z","iopub.status.idle":"2021-09-23T19:37:28.331569Z","shell.execute_reply.started":"2021-09-23T19:37:28.304146Z","shell.execute_reply":"2021-09-23T19:37:28.330877Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# cycle_gan_model.fit(\n#     final_dataset, epochs=45, \n#     steps_per_epoch=(max(n_monet_samples, n_photo_samples)//4)\n# )\n# cycle_gan_model.fit(\n#     final_dataset, \n#     epochs=45\n# )\ncycle_gan_model.fit(\n    final_dataset, epochs=20, \n    steps_per_epoch=(max(n_monet_samples, n_photo_samples)//4)\n)\nFID(fid_photo_ds,monet_generator)","metadata":{"execution":{"iopub.status.busy":"2021-09-23T19:37:28.332743Z","iopub.execute_input":"2021-09-23T19:37:28.333085Z","iopub.status.idle":"2021-09-24T00:23:16.486292Z","shell.execute_reply.started":"2021-09-23T19:37:28.333052Z","shell.execute_reply":"2021-09-24T00:23:16.485519Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# TEST MODEL AND VISUALIZE PHOTOS","metadata":{}},{"cell_type":"code","source":"ds_iter = iter(photo_dataset)\nfor n_sample in range(8):\n    example_sample = next(ds_iter)\n    generated_sample = monet_generator(example_sample)\n     \n    f = plt.figure(figsize=(32, 32))\n  \n    plt.subplot(121)\n    plt.title('Input image')\n    plt.imshow(example_sample[0] * 0.5 + 0.5)\n    plt.axis('off')\n\n    plt.subplot(122)\n    plt.title('Generated image')\n    plt.imshow(generated_sample[0] * 0.5 + 0.5)\n    plt.axis('off')\n    plt.show()","metadata":{"execution":{"iopub.status.busy":"2021-09-24T00:23:16.487524Z","iopub.execute_input":"2021-09-24T00:23:16.487833Z","iopub.status.idle":"2021-09-24T00:23:19.292599Z","shell.execute_reply.started":"2021-09-24T00:23:16.487798Z","shell.execute_reply":"2021-09-24T00:23:19.291545Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# SAVE PHOTS CONVERTED INTO MONET-STYLE IMAGES","metadata":{}},{"cell_type":"code","source":"import PIL\n! mkdir ../images","metadata":{"execution":{"iopub.status.busy":"2021-09-24T00:23:19.30014Z","iopub.execute_input":"2021-09-24T00:23:19.300659Z","iopub.status.idle":"2021-09-24T00:23:20.003373Z","shell.execute_reply.started":"2021-09-24T00:23:19.300622Z","shell.execute_reply":"2021-09-24T00:23:20.002038Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"i = 1\nfor img in photo_dataset:\n    prediction = monet_generator(img, training=False)[0].numpy()\n    prediction = (prediction * 127.5 + 127.5).astype(np.uint8)\n    im = PIL.Image.fromarray(prediction)\n    im.save(\"../images/\" + str(i) + \".jpg\")\n    i += 1","metadata":{"execution":{"iopub.status.busy":"2021-09-24T00:23:20.00523Z","iopub.execute_input":"2021-09-24T00:23:20.005629Z","iopub.status.idle":"2021-09-24T00:27:42.359336Z","shell.execute_reply.started":"2021-09-24T00:23:20.00559Z","shell.execute_reply":"2021-09-24T00:27:42.3585Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import shutil\nshutil.make_archive(\"/kaggle/working/images\", 'zip', \"/kaggle/images\")","metadata":{"execution":{"iopub.status.busy":"2021-09-24T00:27:42.360947Z","iopub.execute_input":"2021-09-24T00:27:42.361265Z","iopub.status.idle":"2021-09-24T00:27:45.957411Z","shell.execute_reply.started":"2021-09-24T00:27:42.361229Z","shell.execute_reply":"2021-09-24T00:27:45.956613Z"},"trusted":true},"execution_count":null,"outputs":[]}]}