{"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\nimport tensorflow.keras.backend as K\nfrom tensorflow.keras import layers\nfrom tensorflow.keras.models import Model\nimport tensorflow_addons as tfa\nfrom tensorflow.keras.optimizers.legacy import Adam\nimport matplotlib.pyplot as plt\nimport numpy as np\nimport re\nimport os\nimport math\nfrom kaggle_datasets import KaggleDatasets\nimport warnings","metadata":{"execution":{"iopub.status.busy":"2023-07-25T02:23:20.852096Z","iopub.execute_input":"2023-07-25T02:23:20.852441Z","iopub.status.idle":"2023-07-25T02:24:03.749113Z","shell.execute_reply.started":"2023-07-25T02:23:20.852408Z","shell.execute_reply":"2023-07-25T02:24:03.747933Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"tpu = tf.distribute.cluster_resolver.TPUClusterResolver.connect(tpu=\"local\")\nstrategy = tf.distribute.TPUStrategy(tpu)\nAUTO = tf.data.experimental.AUTOTUNE","metadata":{"execution":{"iopub.status.busy":"2023-07-25T02:24:03.751069Z","iopub.execute_input":"2023-07-25T02:24:03.751848Z","iopub.status.idle":"2023-07-25T02:24:11.507784Z","shell.execute_reply.started":"2023-07-25T02:24:03.751801Z","shell.execute_reply":"2023-07-25T02:24:11.506869Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"BATCH_SIZE = 32\ndef seed_everything(seed=36):\n    np.random.seed(seed)\n    tf.random.set_seed(seed)\nseed_everything(36)\nwarnings.filterwarnings('ignore')","metadata":{"execution":{"iopub.status.busy":"2023-07-25T02:24:11.508900Z","iopub.execute_input":"2023-07-25T02:24:11.509181Z","iopub.status.idle":"2023-07-25T02:24:11.514398Z","shell.execute_reply.started":"2023-07-25T02:24:11.509156Z","shell.execute_reply":"2023-07-25T02:24:11.513430Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def accelerator():\n    tf.config.optimizer.set_jit(True)\n    print(tf.config.optimizer.get_jit())\n    \n    tf.config.optimizer.set_experimental_options({'scoped_allocator_optimization': True,'implementation_selector': True, 'auto_parallel':True,'constant_folding': True, 'shape_optimization':True, 'remapping':True,'arithmetic_optimization': True, 'dependency_optimization':True, 'function_optimization':True, 'disable_model_pruning':True, 'loop_optimization':True})\n    print(tf.config.optimizer.get_experimental_options())\naccelerator()","metadata":{"execution":{"iopub.status.busy":"2023-07-25T02:24:11.516631Z","iopub.execute_input":"2023-07-25T02:24:11.516936Z","iopub.status.idle":"2023-07-25T02:24:11.530880Z","shell.execute_reply.started":"2023-07-25T02:24:11.516909Z","shell.execute_reply":"2023-07-25T02:24:11.529984Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"GCS_PATH = KaggleDatasets().get_gcs_path(\"gan-getting-started\")\nMONET_FILENAMES = tf.io.gfile.glob(str(GCS_PATH + '/monet_tfrec/*.tfrec'))\nPHOTO_FILENAMES = tf.io.gfile.glob(str(GCS_PATH + '/photo_tfrec/*.tfrec'))\nIMAGE_SIZE = [256, 256]","metadata":{"execution":{"iopub.status.busy":"2023-07-25T02:24:11.531978Z","iopub.execute_input":"2023-07-25T02:24:11.532284Z","iopub.status.idle":"2023-07-25T02:24:11.553739Z","shell.execute_reply.started":"2023-07-25T02:24:11.532257Z","shell.execute_reply":"2023-07-25T02:24:11.552789Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n\n    def data_augment_flip(image):\n        image = tf.image.random_flip_left_right(image)\n        return image\n\n    def 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\n    def read_tfrecord(example):\n        tfrecord_format = {\"image\": tf.io.FixedLenFeature([], tf.string)}\n        example = tf.io.parse_single_example(example, tfrecord_format)\n        image = decode_image(example['image'])\n        return image\n\n    def load_dataset(filenames):\n        dataset = tf.data.TFRecordDataset(filenames)\n        dataset = dataset.map(read_tfrecord, num_parallel_calls=AUTO)\n        return dataset\n\n    def get_gan_dataset(monet_files, photo_files, augment=None, repeat=True, shuffle=True, batch_size=1):\n        monet_ds = load_dataset(monet_files)\n        photo_ds = load_dataset(photo_files)\n\n        if repeat:\n            monet_ds = monet_ds.repeat()\n            photo_ds = photo_ds.repeat()\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\n        if augment:\n            monet_ds = monet_ds.map(augment, num_parallel_calls=AUTO)\n            photo_ds = photo_ds.map(augment, num_parallel_calls=AUTO)\n\n        monet_ds = monet_ds.prefetch(AUTO)\n        photo_ds = photo_ds.prefetch(AUTO)\n        gan_ds = tf.data.Dataset.zip((monet_ds, photo_ds))\n        return gan_ds","metadata":{"execution":{"iopub.status.busy":"2023-07-25T02:24:11.554834Z","iopub.execute_input":"2023-07-25T02:24:11.555219Z","iopub.status.idle":"2023-07-25T02:24:11.559560Z","shell.execute_reply.started":"2023-07-25T02:24:11.555192Z","shell.execute_reply":"2023-07-25T02:24:11.558616Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"monet_ds = load_dataset(MONET_FILENAMES).batch(1)\nphoto_ds = load_dataset(PHOTO_FILENAMES).batch(1)\nfast_photo_ds = load_dataset(PHOTO_FILENAMES).batch(32*strategy.num_replicas_in_sync).prefetch(32)\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)\nfinal_dataset = get_gan_dataset(MONET_FILENAMES, PHOTO_FILENAMES, augment=data_augment_flip, repeat=True, shuffle=True, batch_size=BATCH_SIZE)","metadata":{"execution":{"iopub.status.busy":"2023-07-25T02:24:11.560677Z","iopub.execute_input":"2023-07-25T02:24:11.561028Z","iopub.status.idle":"2023-07-25T02:24:11.821423Z","shell.execute_reply.started":"2023-07-25T02:24:11.561001Z","shell.execute_reply":"2023-07-25T02:24:11.820344Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    inception_model = tf.keras.applications.InceptionV3(input_shape=(256,256,3),pooling=\"avg\",include_top=False)\n    mix3  = inception_model.get_layer(\"mixed9\").output\n    f0 = tf.keras.layers.GlobalAveragePooling2D()(mix3)\n    inception_model = tf.keras.Model(inputs=inception_model.input, outputs=f0)\n    inception_model.trainable = False\n\n    def calculate_activation_statistics_mod(images,fid_model):\n        act=tf.cast(fid_model.predict(images), tf.float32)\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=[]\n\n    def calculate_frechet_distance(mu1,sigma1,mu2,sigma2):\n        fid_epsilon = 1e-14\n        covmean = tf.linalg.sqrtm(tf.cast(tf.matmul(sigma1,sigma2),tf.complex64))\n        covmean = tf.cast(tf.math.real(covmean),tf.float32)\n        tr_covmean = tf.linalg.trace(covmean)\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    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        mu1, sigma1= calculate_activation_statistics_mod(images,fid_model)\n        fid_value = calculate_frechet_distance(mu1, sigma1,myFID_mu2, myFID_sigma2)\n        return fid_value","metadata":{"execution":{"iopub.status.busy":"2023-07-25T02:24:11.822936Z","iopub.execute_input":"2023-07-25T02:24:11.823224Z","iopub.status.idle":"2023-07-25T02:24:11.831395Z","shell.execute_reply.started":"2023-07-25T02:24:11.823198Z","shell.execute_reply":"2023-07-25T02:24:11.830456Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"final_dataset = get_gan_dataset(MONET_FILENAMES, PHOTO_FILENAMES, augment = data_augment_flip, repeat=True, shuffle=True, batch_size=BATCH_SIZE)","metadata":{"execution":{"iopub.status.busy":"2023-07-25T02:24:11.832545Z","iopub.execute_input":"2023-07-25T02:24:11.832905Z","iopub.status.idle":"2023-07-25T02:24:11.972750Z","shell.execute_reply.started":"2023-07-25T02:24:11.832880Z","shell.execute_reply":"2023-07-25T02:24:11.971679Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"OUTPUT_CHANNELS = 3\nwith strategy.scope():\n    \n    def down_sample(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        layer = keras.Sequential()\n        layer.add(layers.Conv2D(filters, size, strides=2, padding='same', kernel_initializer=initializer, use_bias=False))\n        if apply_instancenorm:\n            layer.add(tfa.layers.InstanceNormalization(gamma_initializer=gamma_init))\n        layer.add(layers.LeakyReLU())\n        return layer\n\n    def up_sample(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        layer = keras.Sequential()\n        layer.add(layers.Conv2DTranspose(filters, size, strides=2, padding='same', kernel_initializer=initializer,use_bias=False))\n        layer.add(tfa.layers.InstanceNormalization(gamma_initializer=gamma_init))\n        if apply_dropout:\n            layer.add(layers.Dropout(0.5))\n        layer.add(layers.ReLU())\n        return layer\n\n    def Generator():\n        inputs = layers.Input(shape=[256,256,3])\n        down_stack = [\n            down_sample(64, 4, apply_instancenorm=False),# (size, 128, 128, 64)\n            down_sample(128, 4),                         # (size, 64, 64, 128)\n            down_sample(256, 4),                         # (size, 32, 32, 256)\n            down_sample(512, 4),                         # (size, 16, 16, 512)\n            down_sample(512, 4),                         # (size, 8, 8, 512)\n            down_sample(512, 4),                         # (size, 4, 4, 512)\n            down_sample(512, 4),                         # (size, 2, 2, 512)\n            down_sample(512, 4),                         # (size, 1, 1, 512)\n        ]\n\n        up_stack = [\n            up_sample(512, 4, apply_dropout=True),       # (size, 2, 2, 1024)\n            up_sample(512, 4, apply_dropout=True),       # (size, 4, 4, 1024)\n            up_sample(512, 4, apply_dropout=True),       # (size, 8, 8, 1024)\n            up_sample(512, 4),                           # (size, 16, 16, 1024)\n            up_sample(256, 4),                           # (size, 32, 32, 512)\n            up_sample(128, 4),                           # (size, 64, 64, 256)\n            up_sample(64, 4),                            # (size, 128, 128, 128)\n        ]\n\n        initializer = tf.random_normal_initializer(0., 0.02)\n        last = layers.Conv2DTranspose(3, 4, strides=2, padding='same', kernel_initializer=initializer, activation='tanh') \n        x = inputs\n        skips = []\n        for down in down_stack:\n            x = down(x)\n            skips.append(x)\n        skips = reversed(skips[:-1])\n        for up, skip in zip(up_stack, skips):\n            x = up(x)\n            x = layers.Concatenate()([x, skip])\n        x = last(x)\n        return keras.Model(inputs=inputs, outputs=x)\n\n    def Discriminator():\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        down1 = down_sample(64, 4, False)(x)       # (size, 128, 128, 64)\n        down2 = down_sample(128, 4)(down1)         # (size, 64, 64, 128)\n        down3 = down_sample(256, 4)(down2)         # (size, 32, 32, 256)\n        zero_pad1 = layers.ZeroPadding2D()(down3) # (size, 34, 34, 256)\n        conv = layers.Conv2D(512, 4, strides=1, kernel_initializer=initializer, use_bias=False)(zero_pad1) # (size, 31, 31, 512)\n        norm1 = tfa.layers.InstanceNormalization(gamma_initializer=gamma_init)(conv)\n        leaky_relu = layers.LeakyReLU()(norm1)\n        zero_pad2 = layers.ZeroPadding2D()(leaky_relu) # (size, 33, 33, 512)\n        last = layers.Conv2D(1, 4, strides=1, kernel_initializer=initializer)(zero_pad2)\n        return tf.keras.Model(inputs=inp, outputs=last)\n    \n    def vgg():\n        vgg = tf.keras.applications.VGG19(include_top=False, weights='imagenet', pooling=None)\n        loss_model = Model(inputs=vgg.input, outputs=vgg.get_layer('block5_conv2').output)\n        loss_model.trainable = False\n        return loss_model","metadata":{"execution":{"iopub.status.busy":"2023-07-25T02:24:11.976315Z","iopub.execute_input":"2023-07-25T02:24:11.976650Z","iopub.status.idle":"2023-07-25T02:24:11.983751Z","shell.execute_reply.started":"2023-07-25T02:24:11.976621Z","shell.execute_reply":"2023-07-25T02:24:11.982920Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\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    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    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    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=0.1):\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    def rand_cutout(x, ratio=0.1):\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\n    def aug_fn(image):\n        return DiffAugment(image,'color,translation,cutout')","metadata":{"execution":{"iopub.status.busy":"2023-07-25T02:24:11.984815Z","iopub.execute_input":"2023-07-25T02:24:11.985098Z","iopub.status.idle":"2023-07-25T02:24:12.002754Z","shell.execute_reply.started":"2023-07-25T02:24:11.985066Z","shell.execute_reply":"2023-07-25T02:24:12.001800Z"},"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    monet_discriminator = Discriminator() # differentiates real Monet paintings and generated Monet paintings\n    photo_discriminator = Discriminator() # differentiates real photos and generated photos","metadata":{"execution":{"iopub.status.busy":"2023-07-25T02:24:12.003875Z","iopub.execute_input":"2023-07-25T02:24:12.004192Z","iopub.status.idle":"2023-07-25T02:24:12.017311Z","shell.execute_reply.started":"2023-07-25T02:24:12.004158Z","shell.execute_reply":"2023-07-25T02:24:12.016258Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"class CycleGan(keras.Model):\n\n    def __init__(\n        self,\n        monet_generator,\n        photo_generator,\n        monet_discriminator,\n        photo_discriminator,   \n        lambda_cycle=3,\n        lambda_id=3,\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        self.lambda_id = lambda_id\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_fn1,\n        gen_loss_fn2,\n        gen_loss_fn3,\n        disc_loss_fn1,\n        disc_loss_fn2,\n        disc_loss_fn3,\n        cycle_loss_fn,\n        identity_loss_fn,\n        aug_fn,\n\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_fn1 = gen_loss_fn1\n        self.gen_loss_fn2 = gen_loss_fn2\n        self.gen_loss_fn3 = gen_loss_fn3\n        self.disc_loss_fn1 = disc_loss_fn1\n        self.disc_loss_fn2 = disc_loss_fn2\n        self.disc_loss_fn3 = disc_loss_fn3\n        self.cycle_loss_fn = cycle_loss_fn\n        self.identity_loss_fn = identity_loss_fn\n        self.aug_fn = aug_fn\n        self.step_num = 0\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\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            both_monet = tf.concat([real_monet, fake_monet], axis=0)            \n            aug_monet = self.aug_fn(both_monet)\n            aug_real_monet = aug_monet[:batch_size]\n            aug_fake_monet = aug_monet[batch_size:]\n\n            # two-objective discriminator\n            disc_fake_monet1 = self.m_disc(aug_fake_monet, training=True)\n            disc_real_monet1 = self.m_disc(aug_real_monet, training=True)\n            disc_fake_monet2 = self.m_disc(aug_fake_monet, training=True)\n            disc_real_monet2 = self.m_disc(aug_real_monet, training=True)\n            #disc_fake_monet3 = self.m_disc(aug_fake_monet, training=True)\n            #disc_real_monet3 = self.m_disc(aug_real_monet, training=True)\n            #disc_fake_monet4 = self.vgg(aug_fake_monet)\n            #disc_real_monet4 = self.vgg(aug_real_monet)\n            \n            monet_gen_loss1 = self.gen_loss_fn1(disc_fake_monet1) \n            monet_gen_loss2 = self.gen_loss_fn2(disc_fake_monet2)\n            #monet_gen_loss3 = self.gen_loss_fn3(disc_fake_monet4, disc_real_monet4)\n            monet_head_loss1 = self.disc_loss_fn1(disc_real_monet1, disc_fake_monet1)\n            monet_head_loss2 = self.disc_loss_fn2(disc_real_monet2, disc_fake_monet2)\n            #monet_head_loss3 = self.disc_loss_fn3(disc_real_monet3, disc_fake_monet3, aug_real_monet, aug_fake_monet)\n           \n            monet_gen_loss = (monet_gen_loss1 + monet_gen_loss2) * 0.4\n            monet_disc_loss = monet_head_loss1 + monet_head_loss2\n\n            disc_real_photo = self.p_disc(real_photo, training=True)\n            disc_fake_photo = self.p_disc(fake_photo, training=True)\n            photo_gen_loss = self.gen_loss_fn1(disc_fake_photo)\n            photo_disc_loss = self.disc_loss_fn1(disc_real_photo, disc_fake_photo)\n\n            monet_cycle_loss = self.cycle_loss_fn(real_monet, cycled_monet, self.lambda_cycle/ tf.cast(batch_size,tf.float32))\n            photo_cycle_loss = self.cycle_loss_fn(real_photo, cycled_photo, self.lambda_cycle/ tf.cast(batch_size,tf.float32))\n            total_cycle_loss = monet_cycle_loss + photo_cycle_loss\n\n            monet_identity_loss = self.identity_loss_fn(real_monet, same_monet, self.lambda_id / tf.cast(batch_size,tf.float32))\n            photo_identity_loss = self.identity_loss_fn(real_photo, same_photo, self.lambda_id/ tf.cast(batch_size,tf.float32))\n            total_identity_loss = monet_identity_loss + photo_identity_loss\n\n            # evaluates total loss\n            total_monet_gen_loss = monet_gen_loss + total_cycle_loss + monet_identity_loss\n            total_photo_gen_loss = photo_gen_loss + total_cycle_loss + photo_identity_loss\n\n            total_loss = 2/(1/total_monet_gen_loss+1/total_photo_gen_loss)\n\n        # Calculate the gradients for generator and discriminator\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        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        # Apply the gradients to the optimizer\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            \"gen_loss\": monet_gen_loss, \n            \"identity_loss\": monet_identity_loss,\n            \"cycle_loss\": monet_cycle_loss,\n            \"disc_loss\": monet_disc_loss,\n            \"total_loss\": total_loss,\n            }","metadata":{"execution":{"iopub.status.busy":"2023-07-25T02:24:12.018477Z","iopub.execute_input":"2023-07-25T02:24:12.018822Z","iopub.status.idle":"2023-07-25T02:24:12.033395Z","shell.execute_reply.started":"2023-07-25T02:24:12.018794Z","shell.execute_reply":"2023-07-25T02:24:12.032456Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope(): # for TPU\n    def gradient_penalty2(x_real_score, x_fake_score, x_real, x_fake):\n        d_loss = tf.reduce_mean(x_real_score - x_fake_score)\n        real_grad = tf.gradients(x_real_score, [x_real])[0]\n        fake_grad = tf.gradients(x_fake_score, [x_fake])[0]\n        real_grad_norm = tf.pow(tf.reduce_sum(tf.pow(real_grad,2), axis=[1, 2, 3]),3)\n        fake_grad_norm = tf.pow(tf.reduce_sum(tf.pow(fake_grad,2), axis=[1, 2, 3]),3)\n        return real_grad_norm + fake_grad_norm + d_loss\n    \n    def discriminator_loss1(real, generated):\n        real_loss = tf.math.minimum(tf.zeros_like(real), real-tf.ones_like(real))\n        generated_loss = tf.math.minimum(tf.zeros_like(generated), -generated-tf.ones_like(generated))\n        total_disc_loss = real_loss + generated_loss\n        return tf.reduce_mean(-total_disc_loss * 0.5)\n    \n    def discriminator_loss2(real, generated):\n        generated_loss = tf.keras.losses.BinaryCrossentropy(from_logits=True, reduction=tf.keras.losses.Reduction.NONE)(tf.ones_like(generated), generated)\n        real_loss = tf.keras.losses.BinaryCrossentropy(from_logits=True, reduction=tf.keras.losses.Reduction.NONE)(tf.zeros_like(real), real)\n        total_disc_loss = real_loss + generated_loss\n        return tf.reduce_mean(total_disc_loss * 0.5)\n    \n    def discriminator_loss3(real, generated, x, y):\n        wgan_div_loss = gradient_penalty2(real, generated, x, y)\n        return tf.reduce_mean(wgan_div_loss * 0.5)\n    \n    def generator_loss1(generated):\n        return  tf.reduce_mean(-generated)\n\n    def generator_loss2(generated):\n        return tf.reduce_mean(tf.keras.losses.BinaryCrossentropy(from_logits=True, reduction=tf.keras.losses.Reduction.NONE)(tf.zeros_like(generated), generated))\n    \n    def generator_loss3(generated, good):\n        return tf.reduce_mean(tf.square(good - generated))\n    \n    def calc_cycle_loss(real_image, cycled_image, LAMBDA):\n        loss1 = tf.reduce_sum(tf.abs(real_image - cycled_image))\n        return LAMBDA * loss1 * 0.0000152587890625\n\n    def identity_loss(real_image, same_image, LAMBDA):\n        loss = tf.reduce_sum(tf.abs(real_image - same_image))\n        return LAMBDA * 0.5 * loss * 0.0000152587890625","metadata":{"execution":{"iopub.status.busy":"2023-07-25T02:24:12.034502Z","iopub.execute_input":"2023-07-25T02:24:12.034834Z","iopub.status.idle":"2023-07-25T02:24:12.048873Z","shell.execute_reply.started":"2023-07-25T02:24:12.034806Z","shell.execute_reply":"2023-07-25T02:24:12.048052Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"os.makedirs('checkpoints', exist_ok=True)\ncallbacks = [\n        #tf.keras.callbacks.EarlyStopping(monitor = 'total_loss', verbose=2, patience=10,restore_best_weights=True, mode='min'),\n        #tf.keras.callbacks.LearningRateScheduler(lrfn, verbose = True),\n        #tf.keras.callbacks.ModelCheckpoint(filepath = f'checkpoints/best.h5',monitor = 'total_loss',save_best_only = True, save_weights_only = True),\n        tf.keras.callbacks.TerminateOnNaN(),\n        #tf.keras.callbacks.ReduceLROnPlateau(monitor = 'val_loss', factor = 0.1, patience = 10, verbose = 3)\n    ]","metadata":{"execution":{"iopub.status.busy":"2023-07-25T02:24:12.049973Z","iopub.execute_input":"2023-07-25T02:24:12.050261Z","iopub.status.idle":"2023-07-25T02:24:12.061242Z","shell.execute_reply.started":"2023-07-25T02:24:12.050236Z","shell.execute_reply":"2023-07-25T02:24:12.060334Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    monet_generator_optimizer = tf.keras.optimizers.legacy.Adam(2e-4,beta_1=0.5)\n    photo_generator_optimizer = tf.keras.optimizers.legacy.Adam(2e-4,beta_1=0.5)\n    monet_discriminator_optimizer = tf.keras.optimizers.legacy.Adam(2e-4,beta_1=0.5)\n    photo_discriminator_optimizer = tf.keras.optimizers.legacy.Adam(2e-4,beta_1=0.5)","metadata":{"execution":{"iopub.status.busy":"2023-07-25T02:24:12.062349Z","iopub.execute_input":"2023-07-25T02:24:12.062761Z","iopub.status.idle":"2023-07-25T02:24:12.073300Z","shell.execute_reply.started":"2023-07-25T02:24:12.062735Z","shell.execute_reply":"2023-07-25T02:24:12.072339Z"},"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)","metadata":{"execution":{"iopub.status.busy":"2023-07-25T02:24:12.074378Z","iopub.execute_input":"2023-07-25T02:24:12.074754Z","iopub.status.idle":"2023-07-25T02:24:35.626235Z","shell.execute_reply.started":"2023-07-25T02:24:12.074727Z","shell.execute_reply":"2023-07-25T02:24:35.625062Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\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_fn1 = generator_loss1,\n        gen_loss_fn2 = generator_loss2,\n        gen_loss_fn3 = generator_loss3,\n        disc_loss_fn1 = discriminator_loss1,\n        disc_loss_fn2 = discriminator_loss2,\n        disc_loss_fn3 = discriminator_loss3,\n        cycle_loss_fn = calc_cycle_loss,\n        identity_loss_fn = identity_loss,\n        aug_fn = aug_fn,\n    )\n    \ncycle_gan_model.fit(final_dataset, steps_per_epoch=1407, epochs=12, callbacks=[callbacks])","metadata":{"execution":{"iopub.status.busy":"2023-07-25T02:24:35.627529Z","iopub.execute_input":"2023-07-25T02:24:35.627817Z","iopub.status.idle":"2023-07-25T02:24:35.662484Z","shell.execute_reply.started":"2023-07-25T02:24:35.627792Z","shell.execute_reply":"2023-07-25T02:24:35.661596Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    p = FID(fid_photo_ds,monet_generator) \n    print(p)","metadata":{"execution":{"iopub.status.busy":"2023-07-25T02:24:35.627529Z","iopub.execute_input":"2023-07-25T02:24:35.627817Z","iopub.status.idle":"2023-07-25T02:24:35.662484Z","shell.execute_reply.started":"2023-07-25T02:24:35.627792Z","shell.execute_reply":"2023-07-25T02:24:35.661596Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    monet_generator_optimizer = tf.keras.optimizers.legacy.Adam(1e-4,beta_1=0.5)\n    photo_generator_optimizer = tf.keras.optimizers.legacy.Adam(1e-4,beta_1=0.5)\n    monet_discriminator_optimizer = tf.keras.optimizers.legacy.Adam(1e-4,beta_1=0.5)\n    photo_discriminator_optimizer = tf.keras.optimizers.legacy.Adam(1e-4,beta_1=0.5)","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    cycle_gan_model = CycleGan(monet_generator, photo_generator, monet_discriminator, photo_discriminator)","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\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_fn1 = generator_loss1,\n        gen_loss_fn2 = generator_loss2,\n        gen_loss_fn3 = generator_loss3,\n        disc_loss_fn1 = discriminator_loss1,\n        disc_loss_fn2 = discriminator_loss2,\n        disc_loss_fn3 = discriminator_loss3,\n        cycle_loss_fn = calc_cycle_loss,\n        identity_loss_fn = identity_loss,\n        aug_fn = aug_fn,\n    )\n    \ncycle_gan_model.fit(final_dataset, steps_per_epoch=1407, epochs=3, callbacks=[callbacks])","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    p = FID(fid_photo_ds,monet_generator) \n    print(p)","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    monet_generator_optimizer = tf.keras.optimizers.legacy.Adam(1e-5,beta_1=0.5)\n    photo_generator_optimizer = tf.keras.optimizers.legacy.Adam(1e-5,beta_1=0.5)\n    monet_discriminator_optimizer = tf.keras.optimizers.legacy.Adam(1e-5,beta_1=0.5)\n    photo_discriminator_optimizer = tf.keras.optimizers.legacy.Adam(1e-5,beta_1=0.5)","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    cycle_gan_model = CycleGan(monet_generator, photo_generator, monet_discriminator, photo_discriminator)","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\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_fn1 = generator_loss1,\n        gen_loss_fn2 = generator_loss2,\n        gen_loss_fn3 = generator_loss3,\n        disc_loss_fn1 = discriminator_loss1,\n        disc_loss_fn2 = discriminator_loss2,\n        disc_loss_fn3 = discriminator_loss3,\n        cycle_loss_fn = calc_cycle_loss,\n        identity_loss_fn = identity_loss,\n        aug_fn = aug_fn,\n    )\n    \ncycle_gan_model.fit(final_dataset, steps_per_epoch=1407, epochs=3, callbacks=[callbacks])","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n    p = FID(fid_photo_ds,monet_generator) \n    print(p)","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"ds_iter = iter(photo_ds)\nfor n_sample in range(16):\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":"2023-07-25T02:30:44.538317Z","iopub.execute_input":"2023-07-25T02:30:44.539213Z","iopub.status.idle":"2023-07-25T02:30:44.543703Z","shell.execute_reply.started":"2023-07-25T02:30:44.539178Z","shell.execute_reply":"2023-07-25T02:30:44.542655Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import PIL\n! mkdir ../images\ni = 1\nfor img in fast_photo_ds:\n    prediction = monet_generator(img, training=False).numpy()\n    prediction = (prediction * 127.5 + 127.5).astype(np.uint8)\n    for pred in prediction:\n        im = PIL.Image.fromarray(pred)\n        im.save(\"../images/\" + str(i) + \".jpg\")\n        i += 1","metadata":{"execution":{"iopub.status.busy":"2023-07-25T02:31:37.847784Z","iopub.status.idle":"2023-07-25T02:31:37.848171Z","shell.execute_reply.started":"2023-07-25T02:31:37.847977Z","shell.execute_reply":"2023-07-25T02:31:37.847995Z"},"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":"2023-07-25T02:31:37.849669Z","iopub.status.idle":"2023-07-25T02:31:37.850065Z","shell.execute_reply.started":"2023-07-25T02:31:37.849864Z","shell.execute_reply":"2023-07-25T02:31:37.849883Z"},"trusted":true},"execution_count":null,"outputs":[]}]}