{"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\nimport matplotlib.pyplot as plt\nimport numpy as np\nimport re\ntry:\n    from kaggle_datasets import KaggleDatasets\nexcept:\n    pass","metadata":{"execution":{"iopub.execute_input":"2020-12-15T08:46:45.936344Z","iopub.status.busy":"2020-12-15T08:46:45.935143Z","iopub.status.idle":"2020-12-15T08:46:52.62082Z","shell.execute_reply":"2020-12-15T08:46:52.620001Z"},"papermill":{"duration":6.718946,"end_time":"2020-12-15T08:46:52.620974","exception":false,"start_time":"2020-12-15T08:46:45.902028","status":"completed"},"tags":[],"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\n    \nprint(tf.__version__)","metadata":{"execution":{"iopub.execute_input":"2020-12-15T08:46:52.776258Z","iopub.status.busy":"2020-12-15T08:46:52.719948Z","iopub.status.idle":"2020-12-15T08:46:56.719295Z","shell.execute_reply":"2020-12-15T08:46:56.719905Z"},"papermill":{"duration":4.075298,"end_time":"2020-12-15T08:46:56.720058","exception":false,"start_time":"2020-12-15T08:46:52.64476","status":"completed"},"tags":[],"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.execute_input":"2020-12-15T08:46:56.780729Z","iopub.status.busy":"2020-12-15T08:46:56.779946Z","iopub.status.idle":"2020-12-15T08:46:56.783428Z","shell.execute_reply":"2020-12-15T08:46:56.782784Z"},"papermill":{"duration":0.038868,"end_time":"2020-12-15T08:46:56.78355","exception":false,"start_time":"2020-12-15T08:46:56.744682","status":"completed"},"tags":[],"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.execute_input":"2020-12-15T08:46:56.840272Z","iopub.status.busy":"2020-12-15T08:46:56.839471Z","iopub.status.idle":"2020-12-15T08:46:56.842758Z","shell.execute_reply":"2020-12-15T08:46:56.842012Z"},"papermill":{"duration":0.033535,"end_time":"2020-12-15T08:46:56.842903","exception":false,"start_time":"2020-12-15T08:46:56.809368","status":"completed"},"tags":[],"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"try:\n    GCS_PATH = KaggleDatasets().get_gcs_path()\nexcept:\n    GCS_PATH = \"data\"","metadata":{"execution":{"iopub.execute_input":"2020-12-15T08:46:56.910978Z","iopub.status.busy":"2020-12-15T08:46:56.898106Z","iopub.status.idle":"2020-12-15T08:46:57.403955Z","shell.execute_reply":"2020-12-15T08:46:57.403308Z"},"papermill":{"duration":0.536752,"end_time":"2020-12-15T08:46:57.404107","exception":false,"start_time":"2020-12-15T08:46:56.867355","status":"completed"},"tags":[],"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))\n\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.execute_input":"2020-12-15T08:46:57.62092Z","iopub.status.busy":"2020-12-15T08:46:57.555389Z","iopub.status.idle":"2020-12-15T08:46:57.628057Z","shell.execute_reply":"2020-12-15T08:46:57.6273Z"},"papermill":{"duration":0.199054,"end_time":"2020-12-15T08:46:57.628257","exception":false,"start_time":"2020-12-15T08:46:57.429203","status":"completed"},"tags":[],"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)","metadata":{"execution":{"iopub.execute_input":"2020-12-15T08:46:57.691233Z","iopub.status.busy":"2020-12-15T08:46:57.688744Z","iopub.status.idle":"2020-12-15T08:46:58.008768Z","shell.execute_reply":"2020-12-15T08:46:58.008065Z"},"papermill":{"duration":0.354609,"end_time":"2020-12-15T08:46:58.008932","exception":false,"start_time":"2020-12-15T08:46:57.654323","status":"completed"},"tags":[],"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"example_monet = next(iter(monet_ds))\nexample_photo = next(iter(photo_ds))","metadata":{"execution":{"iopub.execute_input":"2020-12-15T08:46:58.066788Z","iopub.status.busy":"2020-12-15T08:46:58.065978Z","iopub.status.idle":"2020-12-15T08:46:59.124177Z","shell.execute_reply":"2020-12-15T08:46:59.123542Z"},"papermill":{"duration":1.08965,"end_time":"2020-12-15T08:46:59.124316","exception":false,"start_time":"2020-12-15T08:46:58.034666","status":"completed"},"tags":[],"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.execute_input":"2020-12-15T08:46:59.183696Z","iopub.status.busy":"2020-12-15T08:46:59.182866Z","iopub.status.idle":"2020-12-15T08:46:59.598003Z","shell.execute_reply":"2020-12-15T08:46:59.59862Z"},"papermill":{"duration":0.448273,"end_time":"2020-12-15T08:46:59.598815","exception":false,"start_time":"2020-12-15T08:46:59.150542","status":"completed"},"tags":[],"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"\n# reference - data augmentation from kaggle user\n\ndef data_augment(image):\n    \n    # p_rotate = tf.random.uniform([], 0, 1.0, dtype=tf.float32)\n    p_spatial = tf.random.uniform([], 0, 1.0, dtype=tf.float32)\n    p_crop = tf.random.uniform([], 0, 1.0, dtype=tf.float32)\n    \n    # Apply jitter\n    if p_crop > .5: \n        image = tf.image.resize(image, [286, 286])\n        image = tf.image.random_crop(image, size=[256, 256, 3])\n\n    # Random mirroring\n    if p_spatial > .6: \n        image = tf.image.random_flip_left_right(image)\n\n    return image","metadata":{"execution":{"iopub.execute_input":"2020-12-15T08:46:59.676865Z","iopub.status.busy":"2020-12-15T08:46:59.675961Z","iopub.status.idle":"2020-12-15T08:46:59.678716Z","shell.execute_reply":"2020-12-15T08:46:59.679378Z"},"papermill":{"duration":0.050301,"end_time":"2020-12-15T08:46:59.679541","exception":false,"start_time":"2020-12-15T08:46:59.62924","status":"completed"},"tags":[],"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def get_gan_dataset(monet_files, photo_files, augment=None, repeat=True, shuffle=True, batch_size=1):\n\n    monet_ds = load_dataset(monet_files)\n    photo_ds = load_dataset(photo_files)\n    \n    if augment:\n        monet_ds = monet_ds.map(augment, num_parallel_calls=tf.data.experimental.AUTOTUNE)\n        photo_ds = photo_ds.map(augment, num_parallel_calls=tf.data.experimental.AUTOTUNE)\n\n    if repeat:\n        monet_ds = monet_ds.repeat()\n        photo_ds = photo_ds.repeat()\n        \n    if shuffle:\n        monet_ds = monet_ds.shuffle(2048)\n        photo_ds = photo_ds.shuffle(2048)\n        \n    monet_ds = monet_ds.batch(batch_size, drop_remainder=True)\n    photo_ds = photo_ds.batch(batch_size, drop_remainder=True)\n    monet_ds = monet_ds.cache()\n    photo_ds = photo_ds.cache()\n    monet_ds = monet_ds.prefetch(tf.data.experimental.AUTOTUNE)\n    photo_ds = photo_ds.prefetch(tf.data.experimental.AUTOTUNE)\n    \n    gan_ds = tf.data.Dataset.zip((monet_ds, photo_ds))\n    \n    return gan_ds\n\n","metadata":{"execution":{"iopub.execute_input":"2020-12-15T08:46:59.755348Z","iopub.status.busy":"2020-12-15T08:46:59.754467Z","iopub.status.idle":"2020-12-15T08:46:59.75847Z","shell.execute_reply":"2020-12-15T08:46:59.757565Z"},"papermill":{"duration":0.047957,"end_time":"2020-12-15T08:46:59.758636","exception":false,"start_time":"2020-12-15T08:46:59.710679","status":"completed"},"tags":[],"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"final_dataset = get_gan_dataset(MONET_FILENAMES, PHOTO_FILENAMES, augment=data_augment, repeat=True, shuffle=True, batch_size=1)","metadata":{"execution":{"iopub.execute_input":"2020-12-15T08:46:59.829024Z","iopub.status.busy":"2020-12-15T08:46:59.827825Z","iopub.status.idle":"2020-12-15T08:47:01.22301Z","shell.execute_reply":"2020-12-15T08:47:01.222322Z"},"papermill":{"duration":1.432885,"end_time":"2020-12-15T08:47:01.223143","exception":false,"start_time":"2020-12-15T08:46:59.790258","status":"completed"},"tags":[],"trusted":true},"execution_count":null,"outputs":[]},{"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 = 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.execute_input":"2020-12-15T08:47:01.295321Z","iopub.status.busy":"2020-12-15T08:47:01.294123Z","iopub.status.idle":"2020-12-15T08:47:01.298039Z","shell.execute_reply":"2020-12-15T08:47:01.29862Z"},"papermill":{"duration":0.044329,"end_time":"2020-12-15T08:47:01.298792","exception":false,"start_time":"2020-12-15T08:47:01.254463","status":"completed"},"tags":[],"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.execute_input":"2020-12-15T08:47:01.37104Z","iopub.status.busy":"2020-12-15T08:47:01.370144Z","iopub.status.idle":"2020-12-15T08:47:01.373612Z","shell.execute_reply":"2020-12-15T08:47:01.37302Z"},"papermill":{"duration":0.043951,"end_time":"2020-12-15T08:47:01.373745","exception":false,"start_time":"2020-12-15T08:47:01.329794","status":"completed"},"tags":[],"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=False), # (bs, 2, 2, 1024)\n        upsample(512, 4, apply_dropout=False), # (bs, 4, 4, 1024)\n        upsample(512, 4, apply_dropout=False), # (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.execute_input":"2020-12-15T08:47:01.448449Z","iopub.status.busy":"2020-12-15T08:47:01.443016Z","iopub.status.idle":"2020-12-15T08:47:01.455983Z","shell.execute_reply":"2020-12-15T08:47:01.4552Z"},"papermill":{"duration":0.051864,"end_time":"2020-12-15T08:47:01.456115","exception":false,"start_time":"2020-12-15T08:47:01.404251","status":"completed"},"tags":[],"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.execute_input":"2020-12-15T08:47:01.53131Z","iopub.status.busy":"2020-12-15T08:47:01.530545Z","iopub.status.idle":"2020-12-15T08:47:01.534022Z","shell.execute_reply":"2020-12-15T08:47:01.534585Z"},"papermill":{"duration":0.048001,"end_time":"2020-12-15T08:47:01.534753","exception":false,"start_time":"2020-12-15T08:47:01.486752","status":"completed"},"tags":[],"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    monet_discriminator = Discriminator() # differentiates real Monet paintings and generated Monet paintings\n    photo_discriminator = Discriminator() # differentiates real photos and generated photos","metadata":{"execution":{"iopub.execute_input":"2020-12-15T08:47:01.602925Z","iopub.status.busy":"2020-12-15T08:47:01.60216Z","iopub.status.idle":"2020-12-15T08:47:12.417371Z","shell.execute_reply":"2020-12-15T08:47:12.416705Z"},"papermill":{"duration":10.852014,"end_time":"2020-12-15T08:47:12.417516","exception":false,"start_time":"2020-12-15T08:47:01.565502","status":"completed"},"tags":[],"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.execute_input":"2020-12-15T08:47:12.490359Z","iopub.status.busy":"2020-12-15T08:47:12.489406Z","iopub.status.idle":"2020-12-15T08:47:13.174097Z","shell.execute_reply":"2020-12-15T08:47:13.174675Z"},"papermill":{"duration":0.726212,"end_time":"2020-12-15T08:47:13.174822","exception":false,"start_time":"2020-12-15T08:47:12.44861","status":"completed"},"tags":[],"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"class 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        self.step_num = 0\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        \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 \\\n                            + self.cycle_loss_fn(real_monet, cycled_monet, self.lambda_cycle) * self.cycle_loss_factor \\\n                            + self.identity_loss_fn(real_monet, same_monet, self.lambda_cycle)\n                    \n            total_photo_gen_loss = photo_gen_loss \\\n                            + self.cycle_loss_fn(real_photo, cycled_photo, self.lambda_cycle) * self.cycle_loss_factor \\\n                            + 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.execute_input":"2020-12-15T08:47:13.26764Z","iopub.status.busy":"2020-12-15T08:47:13.262291Z","iopub.status.idle":"2020-12-15T08:47:13.276781Z","shell.execute_reply":"2020-12-15T08:47:13.276027Z"},"papermill":{"duration":0.067471,"end_time":"2020-12-15T08:47:13.276925","exception":false,"start_time":"2020-12-15T08:47:13.209454","status":"completed"},"tags":[],"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.execute_input":"2020-12-15T08:47:13.354574Z","iopub.status.busy":"2020-12-15T08:47:13.353485Z","iopub.status.idle":"2020-12-15T08:47:13.356313Z","shell.execute_reply":"2020-12-15T08:47:13.356845Z"},"papermill":{"duration":0.045495,"end_time":"2020-12-15T08:47:13.357043","exception":false,"start_time":"2020-12-15T08:47:13.311548","status":"completed"},"tags":[],"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.execute_input":"2020-12-15T08:47:13.433344Z","iopub.status.busy":"2020-12-15T08:47:13.432296Z","iopub.status.idle":"2020-12-15T08:47:13.435611Z","shell.execute_reply":"2020-12-15T08:47:13.434864Z"},"papermill":{"duration":0.04517,"end_time":"2020-12-15T08:47:13.435734","exception":false,"start_time":"2020-12-15T08:47:13.390564","status":"completed"},"tags":[],"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.execute_input":"2020-12-15T08:47:13.511803Z","iopub.status.busy":"2020-12-15T08:47:13.51083Z","iopub.status.idle":"2020-12-15T08:47:13.513838Z","shell.execute_reply":"2020-12-15T08:47:13.513119Z"},"papermill":{"duration":0.044017,"end_time":"2020-12-15T08:47:13.513993","exception":false,"start_time":"2020-12-15T08:47:13.469976","status":"completed"},"tags":[],"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.execute_input":"2020-12-15T08:47:13.588683Z","iopub.status.busy":"2020-12-15T08:47:13.587894Z","iopub.status.idle":"2020-12-15T08:47:13.591127Z","shell.execute_reply":"2020-12-15T08:47:13.590521Z"},"papermill":{"duration":0.043457,"end_time":"2020-12-15T08:47:13.591268","exception":false,"start_time":"2020-12-15T08:47:13.547811","status":"completed"},"tags":[],"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.execute_input":"2020-12-15T08:47:13.668516Z","iopub.status.busy":"2020-12-15T08:47:13.66772Z","iopub.status.idle":"2020-12-15T08:47:13.670645Z","shell.execute_reply":"2020-12-15T08:47:13.669853Z"},"papermill":{"duration":0.045115,"end_time":"2020-12-15T08:47:13.670774","exception":false,"start_time":"2020-12-15T08:47:13.625659","status":"completed"},"tags":[],"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.execute_input":"2020-12-15T08:47:13.757335Z","iopub.status.busy":"2020-12-15T08:47:13.750001Z","iopub.status.idle":"2020-12-15T08:47:13.79159Z","shell.execute_reply":"2020-12-15T08:47:13.790562Z"},"papermill":{"duration":0.08604,"end_time":"2020-12-15T08:47:13.791739","exception":false,"start_time":"2020-12-15T08:47:13.705699","status":"completed"},"tags":[],"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"EPOCHS_NUM = 54","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"class LossCallback(tf.keras.callbacks.Callback):\n    def __init__(self, model):\n        super(LossCallback, self).__init__()\n        model.cycle_loss_factor = tf.Variable(1.0, trainable=False, name='weight1', dtype=tf.float32)\n\n    def on_epoch_begin(self, epoch, logs=None):\n        tf.keras.backend.set_value(self.model.cycle_loss_factor, (tf.constant(1.0) - tf.constant(epoch / EPOCHS_NUM)))\n\n    def on_epoch_end(self, epoch, logs=None):\n        logs = logs or {}\n        logs['cycle_loss_factor'] = tf.keras.backend.get_value(self.model.cycle_loss_factor)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"callback = LossCallback(cycle_gan_model)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"BATCH_SIZE = 4\ncycle_gan_model.fit(\n    #tf.data.Dataset.zip((monet_ds, photo_ds)),\n    final_dataset,\n    epochs=EPOCHS_NUM,\n    steps_per_epoch=(max(n_monet_samples, n_photo_samples)//BATCH_SIZE),\n    callbacks=[callback]\n)","metadata":{"execution":{"iopub.execute_input":"2020-12-15T08:47:13.880247Z","iopub.status.busy":"2020-12-15T08:47:13.877576Z","iopub.status.idle":"2020-12-15T11:58:13.216734Z","shell.execute_reply":"2020-12-15T11:58:13.217298Z"},"papermill":{"duration":11459.386194,"end_time":"2020-12-15T11:58:13.217509","exception":false,"start_time":"2020-12-15T08:47:13.831315","status":"completed"},"tags":[],"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import PIL\n! mkdir ../images","metadata":{"execution":{"iopub.execute_input":"2020-12-15T11:59:27.685215Z","iopub.status.busy":"2020-12-15T11:59:27.684138Z","iopub.status.idle":"2020-12-15T11:59:28.499141Z","shell.execute_reply":"2020-12-15T11:59:28.497455Z"},"papermill":{"duration":38.090999,"end_time":"2020-12-15T11:59:28.499356","exception":false,"start_time":"2020-12-15T11:58:50.408357","status":"completed"},"tags":[],"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"i = 1\nfor img in photo_ds:\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.execute_input":"2020-12-15T12:00:42.761825Z","iopub.status.busy":"2020-12-15T12:00:42.760985Z","iopub.status.idle":"2020-12-15T12:39:03.463029Z","shell.execute_reply":"2020-12-15T12:39:03.4639Z"},"papermill":{"duration":2338.08136,"end_time":"2020-12-15T12:39:03.464089","exception":false,"start_time":"2020-12-15T12:00:05.382729","status":"completed"},"tags":[],"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import shutil\nshutil.make_archive(\"/kaggle/working/images\", 'zip', \"/kaggle/images\")","metadata":{"execution":{"iopub.execute_input":"2020-12-15T12:40:17.354836Z","iopub.status.busy":"2020-12-15T12:40:17.354068Z","iopub.status.idle":"2020-12-15T12:40:22.663172Z","shell.execute_reply":"2020-12-15T12:40:22.662231Z"},"papermill":{"duration":41.847712,"end_time":"2020-12-15T12:40:22.66332","exception":false,"start_time":"2020-12-15T12:39:40.815608","status":"completed"},"tags":[],"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"cycle_gan_model.p_gen.save_weights(\"p_gen_weights.h5\")\ncycle_gan_model.m_gen.save_weights(\"m_gen_weights.h5\")\n\ncycle_gan_model.p_disc.save_weights(\"p_disc_weights.h5\")\ncycle_gan_model.m_disc.save_weights(\"m_disc_weights.h5\")","metadata":{"trusted":true},"execution_count":null,"outputs":[]}]}