{"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":"code","source":"import tensorflow as tf\nfrom tensorflow import keras\nfrom tensorflow.keras import layers\nimport tensorflow_addons as tfa\n\nfrom keras.preprocessing.image import load_img\nfrom keras.preprocessing.image import img_to_array\nfrom keras.applications.vgg16 import preprocess_input\nfrom keras.applications.vgg16 import decode_predictions\nfrom keras.applications.vgg16 import VGG16\nfrom keras.models import Model\nfrom pickle import dump\n\n\nfrom kaggle_datasets import KaggleDatasets\nimport matplotlib.pyplot as plt\nimport numpy as np\n\nis_vgg = False\nvgg_frozen_layers = 20 # integer [0-13]\n\n\ntry:\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\n    \nprint(tf.__version__)","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","execution":{"iopub.status.busy":"2021-06-20T06:45:56.353211Z","iopub.execute_input":"2021-06-20T06:45:56.353877Z","iopub.status.idle":"2021-06-20T06:46:06.830274Z","shell.execute_reply.started":"2021-06-20T06:45:56.353789Z","shell.execute_reply":"2021-06-20T06:46:06.829445Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"GCS_PATH = KaggleDatasets().get_gcs_path()\n","metadata":{"execution":{"iopub.status.busy":"2021-06-20T06:46:06.831456Z","iopub.execute_input":"2021-06-20T06:46:06.831704Z","iopub.status.idle":"2021-06-20T06:46:07.183359Z","shell.execute_reply.started":"2021-06-20T06:46:06.831679Z","shell.execute_reply":"2021-06-20T06:46:07.182187Z"},"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":"2021-06-20T06:46:07.185064Z","iopub.execute_input":"2021-06-20T06:46:07.18536Z","iopub.status.idle":"2021-06-20T06:46:07.372192Z","shell.execute_reply.started":"2021-06-20T06:46:07.185316Z","shell.execute_reply":"2021-06-20T06:46:07.371327Z"},"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":"2021-06-20T06:46:07.373421Z","iopub.execute_input":"2021-06-20T06:46:07.373666Z","iopub.status.idle":"2021-06-20T06:46:07.380181Z","shell.execute_reply.started":"2021-06-20T06:46:07.373641Z","shell.execute_reply":"2021-06-20T06:46:07.379264Z"},"trusted":true},"execution_count":null,"outputs":[]},{"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":"2021-06-20T06:46:07.38177Z","iopub.execute_input":"2021-06-20T06:46:07.382159Z","iopub.status.idle":"2021-06-20T06:46:07.391428Z","shell.execute_reply.started":"2021-06-20T06:46:07.38212Z","shell.execute_reply":"2021-06-20T06:46:07.390736Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"monet_ds = load_dataset(MONET_FILENAMES, labeled=True).batch(1)\nphoto_ds = load_dataset(PHOTO_FILENAMES, labeled=True).batch(1)\nexample_monet = next(iter(monet_ds))\nexample_photo = next(iter(photo_ds))","metadata":{"execution":{"iopub.status.busy":"2021-06-20T06:46:07.392681Z","iopub.execute_input":"2021-06-20T06:46:07.392948Z","iopub.status.idle":"2021-06-20T06:46:08.616476Z","shell.execute_reply.started":"2021-06-20T06:46:07.392922Z","shell.execute_reply":"2021-06-20T06:46:08.615578Z"},"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\nplt.subplot(122)\nplt.title('Monet')\nplt.imshow(example_monet[0] * 0.5 + 0.5)","metadata":{"execution":{"iopub.status.busy":"2021-06-20T06:46:08.619706Z","iopub.execute_input":"2021-06-20T06:46:08.619994Z","iopub.status.idle":"2021-06-20T06:46:08.900243Z","shell.execute_reply.started":"2021-06-20T06:46:08.61997Z","shell.execute_reply":"2021-06-20T06:46:08.899375Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Build the generator\n# We'll be using a UNET architecture for our CycleGAN. To build our generator, let's first define our downsample and upsample methods.\n\n# The downsample, as the name suggests, reduces the 2D dimensions, the width and height, of the image by the stride. The stride is the length of the step the filter takes. Since the stride is 2, the filter is applied to every other pixel, hence reducing the weight and height by 2.\n\n# We'll be using an instance normalization instead of batch normalization. As the instance normalization is not standard in the TensorFlow API, we'll use the layer from TensorFlow Add-ons.\n\nOUTPUT_CHANNELS = 3\n\ndef downsample(filters, size, apply_instancenorm=True):\n    initializer = tf.random_normal_initializer(0., 0.02)\n    gamma_init = keras.initializers.RandomNormal(mean=0.0, stddev=0.02)\n\n    result = keras.Sequential()\n    result.add(layers.Conv2D(filters, size, strides=2, padding='same',\n                             kernel_initializer=initializer, use_bias=False))\n\n    if apply_instancenorm:\n        result.add(tfa.layers.InstanceNormalization(gamma_initializer=gamma_init))\n\n    result.add(layers.LeakyReLU())\n\n    return result","metadata":{"execution":{"iopub.status.busy":"2021-06-20T06:46:08.903122Z","iopub.execute_input":"2021-06-20T06:46:08.90357Z","iopub.status.idle":"2021-06-20T06:46:08.911082Z","shell.execute_reply.started":"2021-06-20T06:46:08.903528Z","shell.execute_reply":"2021-06-20T06:46:08.910444Z"},"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 = keras.initializers.RandomNormal(mean=0.0, stddev=0.02)\n\n    result = keras.Sequential()\n    result.add(layers.Conv2DTranspose(filters, size, strides=2,\n                                      padding='same',\n                                      kernel_initializer=initializer,\n                                      use_bias=False))\n\n    result.add(tfa.layers.InstanceNormalization(gamma_initializer=gamma_init))\n\n    if apply_dropout:\n        result.add(layers.Dropout(0.5))\n\n    result.add(layers.ReLU())\n\n    return result","metadata":{"execution":{"iopub.status.busy":"2021-06-20T06:46:08.912752Z","iopub.execute_input":"2021-06-20T06:46:08.913223Z","iopub.status.idle":"2021-06-20T06:46:08.921964Z","shell.execute_reply.started":"2021-06-20T06:46:08.913178Z","shell.execute_reply":"2021-06-20T06:46:08.921294Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def Generator():\n    inputs = layers.Input(shape=[256,256,3])\n\n    # bs = batch size\n    down_stack = [\n        downsample(64, 4, apply_instancenorm=False), # (bs, 128, 128, 64)\n        downsample(128, 4), # (bs, 64, 64, 128)\n        downsample(256, 4), # (bs, 32, 32, 256)\n        downsample(512, 4), # (bs, 16, 16, 512)\n        downsample(512, 4), # (bs, 8, 8, 512)\n        downsample(512, 4), # (bs, 4, 4, 512)\n        downsample(512, 4), # (bs, 2, 2, 512)\n        downsample(512, 4), # (bs, 1, 1, 512)\n    ]\n\n    up_stack = [\n        upsample(512, 4, apply_dropout=True), # (bs, 2, 2, 1024)\n        upsample(512, 4, apply_dropout=True), # (bs, 4, 4, 1024)\n        upsample(512, 4, apply_dropout=True), # (bs, 8, 8, 1024)\n        upsample(512, 4), # (bs, 16, 16, 1024)\n        upsample(256, 4), # (bs, 32, 32, 512)\n        upsample(128, 4), # (bs, 64, 64, 256)\n        upsample(64, 4), # (bs, 128, 128, 128)\n    ]\n\n    initializer = tf.random_normal_initializer(0., 0.02)\n    last = layers.Conv2DTranspose(OUTPUT_CHANNELS, 4,\n                                  strides=2,\n                                  padding='same',\n                                  kernel_initializer=initializer,\n                                  activation='tanh') # (bs, 256, 256, 3)\n\n    x = inputs\n\n    # Downsampling through the model\n    skips = []\n    for down in down_stack:\n        x = down(x)\n        skips.append(x)\n\n    skips = reversed(skips[:-1])\n\n    # Upsampling and establishing the skip connections\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 keras.Model(inputs=inputs, outputs=x)","metadata":{"execution":{"iopub.status.busy":"2021-06-20T06:46:08.923054Z","iopub.execute_input":"2021-06-20T06:46:08.923471Z","iopub.status.idle":"2021-06-20T06:46:08.937402Z","shell.execute_reply.started":"2021-06-20T06:46:08.923443Z","shell.execute_reply":"2021-06-20T06:46:08.936705Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def Discriminator():\n    initializer = tf.random_normal_initializer(0., 0.02)\n    gamma_init = keras.initializers.RandomNormal(mean=0.0, 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) # (bs, 128, 128, 64)\n    down2 = downsample(128, 4)(down1) # (bs, 64, 64, 128)\n    down3 = downsample(256, 4)(down2) # (bs, 32, 32, 256)\n\n    zero_pad1 = layers.ZeroPadding2D()(down3) # (bs, 34, 34, 256)\n    conv = layers.Conv2D(512, 4, strides=1,\n                         kernel_initializer=initializer,\n                         use_bias=False)(zero_pad1) # (bs, 31, 31, 512)\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) # (bs, 33, 33, 512)\n\n    last = layers.Conv2D(1, 4, strides=1,\n                         kernel_initializer=initializer)(zero_pad2) # (bs, 30, 30, 1)\n\n    return tf.keras.Model(inputs=inp, outputs=last)","metadata":{"execution":{"iopub.status.busy":"2021-06-20T06:46:08.938261Z","iopub.execute_input":"2021-06-20T06:46:08.93864Z","iopub.status.idle":"2021-06-20T06:46:08.950287Z","shell.execute_reply.started":"2021-06-20T06:46:08.938602Z","shell.execute_reply":"2021-06-20T06:46:08.94964Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def vgg_discrimnator():  \n    initializer = tf.random_normal_initializer(0., 0.02)\n    gamma_init = keras.initializers.RandomNormal(mean=0.0, stddev=0.02)\n\n    # load model without classifier layers\n    model = VGG16(include_top=False, input_shape=(256, 256, 3))  #weights=None\n    \n    # mark loaded layers as not trainable\n    for count, layer in enumerate(model.layers):\n        if count > vgg_frozen_layers:\n            break\n        layer.trainable = False\n        \n    conv = layers.Conv2D(512, 3, strides=1,\n                         kernel_initializer=initializer,\n                         use_bias=False)(model.layers[-6].output) # (bs, 31, 31, 512)\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) # (bs, 33, 33, 512)\n\n    output = layers.Conv2D(1, 3, strides=1,\n                         kernel_initializer=initializer)(zero_pad2) # (bs, 30, 30, 1)\n\n    model = Model(inputs=model.inputs, outputs=output)\n    return model","metadata":{"execution":{"iopub.status.busy":"2021-06-20T06:46:08.951408Z","iopub.execute_input":"2021-06-20T06:46:08.951772Z","iopub.status.idle":"2021-06-20T06:46:08.964654Z","shell.execute_reply.started":"2021-06-20T06:46:08.951746Z","shell.execute_reply":"2021-06-20T06:46:08.963908Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def normal_discriminator():\n    initializer = tf.random_normal_initializer(0., 0.02)\n    gamma_init = keras.initializers.RandomNormal(mean=0.0, 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) # (bs, 128, 128, 64)\n    down2 = downsample(128, 4)(down1) # (bs, 64, 64, 128)\n    down3 = downsample(256, 4)(down2) # (bs, 32, 32, 256)\n\n    zero_pad1 = layers.ZeroPadding2D()(down3) # (bs, 34, 34, 256)\n    conv = layers.Conv2D(512, 4, strides=1,\n                         kernel_initializer=initializer,\n                         use_bias=False)(zero_pad1) # (bs, 31, 31, 512)\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) # (bs, 33, 33, 512)\n\n    output = layers.Conv2D(1, 4, strides=1,\n                         kernel_initializer=initializer)(zero_pad2) # (bs, 30, 30, 1)\n\n    return tf.keras.Model(inputs=inp, outputs=output)\n#     return output","metadata":{"execution":{"iopub.status.busy":"2021-06-20T06:46:08.966436Z","iopub.execute_input":"2021-06-20T06:46:08.967152Z","iopub.status.idle":"2021-06-20T06:46:08.978047Z","shell.execute_reply.started":"2021-06-20T06:46:08.96711Z","shell.execute_reply":"2021-06-20T06:46:08.977391Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def Discriminator_vgg():\n    initializer = tf.random_normal_initializer(0., 0.02)\n    gamma_init = keras.initializers.RandomNormal(mean=0.0, stddev=0.02)\n    inp = layers.Input(shape=[256, 256, 3], name='input_image')\n    x = inp\n    normal_model = normal_discriminator()\n    vgg_model = vgg_discrimnator()\n#     mergedOut = Concatenate()([normal_model.output,vgg_model.output])\n    out1 = normal_model(x)    \n    out2 = vgg_model(x)    \n    mergedOut = layers.Concatenate()([out1,out2])    \n    mergedOut = layers.Conv2D(1, 1, strides=1,\n                         kernel_initializer=initializer,\n                         use_bias=False)(mergedOut)\n    \n    \n    merged_model = Model(inp,mergedOut)\n    return merged_model ","metadata":{"execution":{"iopub.status.busy":"2021-06-20T06:46:08.979539Z","iopub.execute_input":"2021-06-20T06:46:08.980241Z","iopub.status.idle":"2021-06-20T06:46:08.994616Z","shell.execute_reply.started":"2021-06-20T06:46:08.980195Z","shell.execute_reply":"2021-06-20T06:46:08.993795Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    monet_generator = Generator() # transforms photos to Monet-esque paintings\n    photo_generator = Generator() # transforms Monet paintings to be more like photos\n    \n    if is_vgg:\n        monet_discriminator = Discriminator() # differentiates real Monet paintings and generated Monet paintings\n        photo_discriminator = Discriminator() # differentiates real photos and generated photos\n    else:\n        monet_discriminator = Discriminator_vgg() # differentiates real Monet paintings and generated Monet paintings\n        photo_discriminator = Discriminator_vgg() # differentiates real photos and generated photos\n        ","metadata":{"execution":{"iopub.status.busy":"2021-06-20T06:46:08.996282Z","iopub.execute_input":"2021-06-20T06:46:08.996981Z","iopub.status.idle":"2021-06-20T06:46:22.418071Z","shell.execute_reply.started":"2021-06-20T06:46:08.99694Z","shell.execute_reply":"2021-06-20T06:46:22.417178Z"},"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":"2021-06-20T06:46:22.419061Z","iopub.execute_input":"2021-06-20T06:46:22.419286Z","iopub.status.idle":"2021-06-20T06:46:22.993246Z","shell.execute_reply.started":"2021-06-20T06:46:22.419262Z","shell.execute_reply":"2021-06-20T06:46:22.992456Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Build the CycleGAN model\n# We will subclass a tf.keras.Model so that we can run fit() later to train our model. \n# During the training step, the model transforms a photo to a Monet painting and then back to a photo. \n# The losses are defined in the next section.\n\nclass CycleGan(keras.Model):\n    def __init__(\n        self,\n        monet_generator,\n        photo_generator,\n        monet_discriminator,\n        photo_discriminator,\n        lambda_cycle=10,\n    ):\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    ):\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        \n    def train_step(self, batch_data):\n        real_monet, real_photo = batch_data\n        \n        with tf.GradientTape(persistent=True) as tape:\n            # photo to monet back to photo\n            fake_monet = self.m_gen(real_photo, training=True)\n            cycled_photo = self.p_gen(fake_monet, training=True)\n\n            # monet to photo back to monet\n            fake_photo = self.p_gen(real_monet, training=True)\n            cycled_monet = self.m_gen(fake_photo, training=True)\n\n            # generating itself\n            same_monet = self.m_gen(real_monet, training=True)\n            same_photo = self.p_gen(real_photo, training=True)\n\n            # discriminator used to check, inputing real images\n            disc_real_monet = self.m_disc(real_monet, training=True)\n            disc_real_photo = self.p_disc(real_photo, training=True)\n\n            # discriminator used to check, inputing fake images\n            disc_fake_monet = self.m_disc(fake_monet, training=True)\n            disc_fake_photo = self.p_disc(fake_photo, training=True)\n\n            # evaluates generator loss\n            monet_gen_loss = self.gen_loss_fn(disc_fake_monet)\n            photo_gen_loss = self.gen_loss_fn(disc_fake_photo)\n\n            # evaluates total cycle consistency loss\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            # evaluates total generator loss\n            total_monet_gen_loss = monet_gen_loss + total_cycle_loss + self.identity_loss_fn(real_monet, same_monet, self.lambda_cycle)\n            total_photo_gen_loss = photo_gen_loss + total_cycle_loss + self.identity_loss_fn(real_photo, same_photo, self.lambda_cycle)\n\n            # evaluates discriminator loss\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        # Calculate the gradients for generator and discriminator\n        monet_generator_gradients = tape.gradient(total_monet_gen_loss,\n                                                  self.m_gen.trainable_variables)\n        photo_generator_gradients = tape.gradient(total_photo_gen_loss,\n                                                  self.p_gen.trainable_variables)\n\n        monet_discriminator_gradients = tape.gradient(monet_disc_loss,\n                                                      self.m_disc.trainable_variables)\n        photo_discriminator_gradients = tape.gradient(photo_disc_loss,\n                                                      self.p_disc.trainable_variables)\n\n        # Apply the gradients to the optimizer\n        self.m_gen_optimizer.apply_gradients(zip(monet_generator_gradients,\n                                                 self.m_gen.trainable_variables))\n\n        self.p_gen_optimizer.apply_gradients(zip(photo_generator_gradients,\n                                                 self.p_gen.trainable_variables))\n\n        self.m_disc_optimizer.apply_gradients(zip(monet_discriminator_gradients,\n                                                  self.m_disc.trainable_variables))\n\n        self.p_disc_optimizer.apply_gradients(zip(photo_discriminator_gradients,\n                                                  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":"2021-06-20T06:46:22.994491Z","iopub.execute_input":"2021-06-20T06:46:22.994756Z","iopub.status.idle":"2021-06-20T06:46:23.012726Z","shell.execute_reply.started":"2021-06-20T06:46:22.99473Z","shell.execute_reply":"2021-06-20T06:46:23.011735Z"},"trusted":true},"execution_count":null,"outputs":[]},{"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        generated_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 + generated_loss\n\n        return total_disc_loss * 0.5","metadata":{"execution":{"iopub.status.busy":"2021-06-20T06:46:23.013956Z","iopub.execute_input":"2021-06-20T06:46:23.014221Z","iopub.status.idle":"2021-06-20T06:46:23.026397Z","shell.execute_reply.started":"2021-06-20T06:46:23.014195Z","shell.execute_reply":"2021-06-20T06:46:23.025479Z"},"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":"2021-06-20T06:46:23.027682Z","iopub.execute_input":"2021-06-20T06:46:23.028063Z","iopub.status.idle":"2021-06-20T06:46:23.039213Z","shell.execute_reply.started":"2021-06-20T06:46:23.028018Z","shell.execute_reply":"2021-06-20T06:46:23.038517Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    def calc_cycle_loss(real_image, cycled_image, LAMBDA):\n        loss1 = tf.reduce_mean(tf.abs(real_image - cycled_image))\n\n        return LAMBDA * loss1","metadata":{"execution":{"iopub.status.busy":"2021-06-20T06:46:23.040407Z","iopub.execute_input":"2021-06-20T06:46:23.040672Z","iopub.status.idle":"2021-06-20T06:46:23.049511Z","shell.execute_reply.started":"2021-06-20T06:46:23.040647Z","shell.execute_reply":"2021-06-20T06:46:23.048665Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    def identity_loss(real_image, same_image, LAMBDA):\n        loss = tf.reduce_mean(tf.abs(real_image - same_image))\n        return LAMBDA * 0.5 * loss","metadata":{"execution":{"iopub.status.busy":"2021-06-20T06:46:23.050559Z","iopub.execute_input":"2021-06-20T06:46:23.050794Z","iopub.status.idle":"2021-06-20T06:46:23.058474Z","shell.execute_reply.started":"2021-06-20T06:46:23.050771Z","shell.execute_reply":"2021-06-20T06:46:23.057637Z"},"trusted":true},"execution_count":null,"outputs":[]},{"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":"2021-06-20T06:46:23.059579Z","iopub.execute_input":"2021-06-20T06:46:23.05986Z","iopub.status.idle":"2021-06-20T06:46:23.067265Z","shell.execute_reply.started":"2021-06-20T06:46:23.059835Z","shell.execute_reply":"2021-06-20T06:46:23.066449Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    cycle_gan_model = CycleGan(\n        monet_generator, photo_generator, monet_discriminator, photo_discriminator\n    )\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 = calc_cycle_loss,\n        identity_loss_fn = identity_loss\n    )","metadata":{"execution":{"iopub.status.busy":"2021-06-20T06:46:23.070048Z","iopub.execute_input":"2021-06-20T06:46:23.070341Z","iopub.status.idle":"2021-06-20T06:46:23.122444Z","shell.execute_reply.started":"2021-06-20T06:46:23.070313Z","shell.execute_reply":"2021-06-20T06:46:23.121551Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"history = cycle_gan_model.fit(\n    tf.data.Dataset.zip((monet_ds, photo_ds)),\n    epochs=40\n)","metadata":{"execution":{"iopub.status.busy":"2021-06-20T06:46:23.123709Z","iopub.execute_input":"2021-06-20T06:46:23.123973Z","iopub.status.idle":"2021-06-20T07:19:39.488661Z","shell.execute_reply.started":"2021-06-20T06:46:23.123948Z","shell.execute_reply":"2021-06-20T07:19:39.487757Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"###print loss############\nmonet_gen_loss = np.asarray(history.history['monet_gen_loss'])\nmonet_gen_loss_mean = monet_gen_loss.mean(axis=(1, 2, 3))\nphoto_gen_loss = np.asarray(history.history['photo_gen_loss'])\nphoto_gen_loss_mean = photo_gen_loss.mean(axis=(1, 2, 3))\nmonet_disc_loss = np.asarray(history.history['monet_disc_loss'])\nmonet_disc_loss_mean = monet_disc_loss.mean(axis=(1, 2, 3))\nphoto_disc_loss = np.asarray(history.history['photo_disc_loss'])\nphoto_disc_loss_mean = photo_disc_loss.mean(axis=(1, 2, 3))\n\nplt.plot(monet_gen_loss_mean.flatten())\nplt.plot(photo_gen_loss_mean.flatten())\nplt.plot(monet_disc_loss_mean.flatten())\nplt.plot(photo_disc_loss_mean.flatten())\nplt.title('model loss')\nplt.ylabel('loss')\nplt.xlabel('epoch')\nplt.legend(['monet_gen_loss', 'photo_gen_loss', 'monet_disc_loss', 'photo_disc_loss'], loc='upper left')\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2021-06-20T07:19:39.489885Z","iopub.execute_input":"2021-06-20T07:19:39.490166Z","iopub.status.idle":"2021-06-20T07:19:39.650982Z","shell.execute_reply.started":"2021-06-20T07:19:39.490139Z","shell.execute_reply":"2021-06-20T07:19:39.650085Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# _, ax = plt.subplots(5, 2, figsize=(120, 120))\n# for i, img in enumerate(photo_ds.take(100)):\n#     prediction = monet_generator(img, training=False)[0].numpy()\n#     prediction = (prediction * 127.5 + 127.5).astype(np.uint8)\n#     img = (img[0] * 127.5 + 127.5).numpy().astype(np.uint8)\n\n#     ax[i, 0].imshow(img)\n#     ax[i, 1].imshow(prediction)\n#     ax[i, 0].set_title(\"Input Photo\")\n#     ax[i, 1].set_title(\"Monet-esque\")\n#     ax[i, 0].axis(\"off\")\n#     ax[i, 1].axis(\"off\")\n# plt.show()","metadata":{"execution":{"iopub.status.busy":"2021-06-20T07:19:39.652172Z","iopub.execute_input":"2021-06-20T07:19:39.652453Z","iopub.status.idle":"2021-06-20T07:19:39.655911Z","shell.execute_reply.started":"2021-06-20T07:19:39.652426Z","shell.execute_reply":"2021-06-20T07:19:39.655125Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"_, ax = plt.subplots(100, 2)\nfor i, img in enumerate(photo_ds.take(100)):\n    prediction = monet_generator(img, training=False)[0].numpy()\n    prediction = (prediction * 127.5 + 127.5).astype(np.uint8)\n    img = (img[0] * 127.5 + 127.5).numpy().astype(np.uint8)\n\n    ax[i, 0].imshow(img)\n    ax[i, 1].imshow(prediction)\n    ax[i, 0].set_title(\"Input Photo\")\n    ax[i, 1].set_title(\"Monet-esque\")\n    ax[i, 0].axis(\"off\")\n    ax[i, 1].axis(\"off\")\nplt.show()\n\n\n\n\n","metadata":{"execution":{"iopub.status.busy":"2021-06-20T07:19:39.656958Z","iopub.execute_input":"2021-06-20T07:19:39.657207Z","iopub.status.idle":"2021-06-20T07:20:17.567537Z","shell.execute_reply.started":"2021-06-20T07:19:39.657184Z","shell.execute_reply":"2021-06-20T07:20:17.566825Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"_, ax = plt.subplots(10, 2, figsize=(120, 120))\nfor i, img in enumerate(photo_ds.take(10)):\n\n    prediction = monet_generator(img, training=False)[0].numpy()\n    prediction = (prediction * 127.5 + 127.5).astype(np.uint8)\n    img = (img[0] * 127.5 + 127.5).numpy().astype(np.uint8)\n\n    ax[i, 0].imshow(img)\n    ax[i, 1].imshow(prediction)\n    ax[i, 0].set_title(\"Input Photo\")\n    ax[i, 1].set_title(\"Monet-esque\")\n    ax[i, 0].axis(\"off\")\n    ax[i, 1].axis(\"off\")\nplt.show()\n","metadata":{"execution":{"iopub.status.busy":"2021-06-20T07:25:48.682866Z","iopub.execute_input":"2021-06-20T07:25:48.683322Z","iopub.status.idle":"2021-06-20T07:25:57.075046Z","shell.execute_reply.started":"2021-06-20T07:25:48.683294Z","shell.execute_reply":"2021-06-20T07:25:57.074068Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]}]}