{"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\nimport matplotlib.pyplot as plt\nimport numpy as np\nimport tensorflow_addons as tfa\nimport time\nimport os\nimport random\n\nfrom tensorflow import keras\nfrom tensorflow.keras import layers\n\nfrom kaggle_datasets import KaggleDatasets\nfrom numpy import cov\nfrom numpy import trace\nfrom numpy import iscomplexobj\nfrom numpy import asarray\nfrom numpy.random import randint\nfrom tqdm import tqdm\nfrom scipy.linalg import sqrtm\nfrom tensorflow.keras.applications.inception_v3 import InceptionV3\nfrom tensorflow.keras.applications.inception_v3 import preprocess_input\nfrom tensorflow.keras.datasets.mnist import load_data\nfrom skimage.transform import resize\nfrom scipy import linalg","metadata":{"id":"WZKbyU2-AiY-","execution":{"iopub.status.busy":"2022-06-12T09:59:06.546426Z","iopub.execute_input":"2022-06-12T09:59:06.546986Z","iopub.status.idle":"2022-06-12T09:59:06.554271Z","shell.execute_reply.started":"2022-06-12T09:59:06.546950Z","shell.execute_reply":"2022-06-12T09:59:06.553552Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"GCS_PATH = KaggleDatasets().get_gcs_path()","metadata":{"id":"cDaqQgyhpbUl","outputId":"1ba6f392-863f-44ea-bef2-15e88bcbe9ea","execution":{"iopub.status.busy":"2022-06-12T09:59:06.648508Z","iopub.execute_input":"2022-06-12T09:59:06.649253Z","iopub.status.idle":"2022-06-12T09:59:07.463233Z","shell.execute_reply.started":"2022-06-12T09:59:06.649215Z","shell.execute_reply":"2022-06-12T09:59:07.462355Z"},"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)\n# except:\n#     strategy = tf.distribute.get_strategy()\n# print('Number of replicas:', strategy.num_replicas_in_sync)","metadata":{"id":"4_cwVUVmCPMW","outputId":"e030f6ca-a836-4d0f-93d0-ea3ff894c981","execution":{"iopub.status.busy":"2022-06-12T09:59:07.465115Z","iopub.execute_input":"2022-06-12T09:59:07.465478Z","iopub.status.idle":"2022-06-12T09:59:07.471812Z","shell.execute_reply.started":"2022-06-12T09:59:07.465433Z","shell.execute_reply":"2022-06-12T09:59:07.468903Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"HEIGHT = 256\nWIDTH = 256\nIMAGE_SIZE = [HEIGHT, WIDTH]\nCHANNELS = 3\n\nBATCH_SIZE = 6","metadata":{"id":"AcEn9slMCPMX","execution":{"iopub.status.busy":"2022-06-12T09:59:07.473642Z","iopub.execute_input":"2022-06-12T09:59:07.474449Z","iopub.status.idle":"2022-06-12T09:59:07.479540Z","shell.execute_reply.started":"2022-06-12T09:59:07.474407Z","shell.execute_reply":"2022-06-12T09:59:07.478729Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Load and prepare datasets","metadata":{"id":"iYn4MdZnKCey"}},{"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","metadata":{"id":"a4fYMGxGhrna","outputId":"0f2d6a9c-de7e-4cb8-ce8d-5e6d442a1597","execution":{"iopub.status.busy":"2022-06-12T09:59:07.482693Z","iopub.execute_input":"2022-06-12T09:59:07.483612Z","iopub.status.idle":"2022-06-12T09:59:09.385235Z","shell.execute_reply.started":"2022-06-12T09:59:07.483554Z","shell.execute_reply":"2022-06-12T09:59:09.384450Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"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, [HEIGHT, WIDTH, CHANNELS])\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\n\ndef load_dataset(filenames, labeled=True, ordered=False):\n    dataset = tf.data.TFRecordDataset(filenames)\n    dataset = dataset.map(read_tfrecord, num_parallel_calls=tf.data.experimental.AUTOTUNE)\n    return dataset","metadata":{"id":"NFC2ghIdiZYE","execution":{"iopub.status.busy":"2022-06-12T09:59:09.389318Z","iopub.execute_input":"2022-06-12T09:59:09.391316Z","iopub.status.idle":"2022-06-12T09:59:09.402756Z","shell.execute_reply.started":"2022-06-12T09:59:09.391277Z","shell.execute_reply":"2022-06-12T09:59:09.401934Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"monet_ds = load_dataset(MONET_FILENAMES, labeled=True).batch(BATCH_SIZE)\nphoto_ds = load_dataset(PHOTO_FILENAMES, labeled=True).batch(BATCH_SIZE)","metadata":{"id":"S4PIDhoDLbsZ","outputId":"f8f36487-bc27-4b68-a41c-9e4729687ff2","execution":{"iopub.status.busy":"2022-06-12T09:59:09.404332Z","iopub.execute_input":"2022-06-12T09:59:09.404984Z","iopub.status.idle":"2022-06-12T09:59:11.870419Z","shell.execute_reply.started":"2022-06-12T09:59:09.404939Z","shell.execute_reply":"2022-06-12T09:59:11.869650Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Model","metadata":{"id":"2dA29Rcs0k70"}},{"cell_type":"markdown","source":"## Generator & discriminator","metadata":{"id":"THY-sZMiQ4UV"}},{"cell_type":"code","source":"def 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', 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\n\n\ndef 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, padding='same', kernel_initializer=initializer, 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":{"id":"DA1lM8I4CPMe","execution":{"iopub.status.busy":"2022-06-12T09:59:11.871739Z","iopub.execute_input":"2022-06-12T09:59:11.872258Z","iopub.status.idle":"2022-06-12T09:59:11.882475Z","shell.execute_reply.started":"2022-06-12T09:59:11.872218Z","shell.execute_reply":"2022-06-12T09:59:11.881277Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def generator_build():\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(CHANNELS, 4, strides=2, padding='same', kernel_initializer=initializer, 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":{"id":"MMPBZqpBCPMf","execution":{"iopub.status.busy":"2022-06-12T09:59:11.884919Z","iopub.execute_input":"2022-06-12T09:59:11.885445Z","iopub.status.idle":"2022-06-12T09:59:11.898146Z","shell.execute_reply.started":"2022-06-12T09:59:11.885406Z","shell.execute_reply":"2022-06-12T09:59:11.897295Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"\ndef discriminator_build():\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":{"id":"nExPVDbtCPMh","execution":{"iopub.status.busy":"2022-06-12T09:59:11.899514Z","iopub.execute_input":"2022-06-12T09:59:11.900287Z","iopub.status.idle":"2022-06-12T09:59:11.911380Z","shell.execute_reply.started":"2022-06-12T09:59:11.900246Z","shell.execute_reply":"2022-06-12T09:59:11.909854Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"monet_generator = generator_build() \nphoto_generator = generator_build()\n\nmonet_discriminator = discriminator_build()\nphoto_discriminator = discriminator_build()","metadata":{"id":"ce0KhoO3CPMi","execution":{"iopub.status.busy":"2022-06-12T09:59:11.915220Z","iopub.execute_input":"2022-06-12T09:59:11.915791Z","iopub.status.idle":"2022-06-12T09:59:14.256435Z","shell.execute_reply.started":"2022-06-12T09:59:11.915754Z","shell.execute_reply":"2022-06-12T09:59:14.255629Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Loss and optimizers","metadata":{"id":"D1Ke2_tOCPMj"}},{"cell_type":"code","source":"cross_entropy = tf.keras.losses.BinaryCrossentropy(from_logits=True, reduction=tf.keras.losses.Reduction.NONE)\n\ndef discriminator_loss(real_output, fake_output):\n    real_loss = cross_entropy(tf.ones_like(real_output), real_output)\n    fake_loss = cross_entropy(tf.zeros_like(fake_output), fake_output)\n    total_loss = real_loss + fake_loss\n    return total_loss\n\ndef generator_loss(fake_output):\n    return cross_entropy(tf.ones_like(fake_output), fake_output)\n\ndef calc_cycle_loss(real_image, cycled_image, LAMBDA):\n    loss1 = tf.reduce_mean(tf.abs(real_image - cycled_image))\n    return LAMBDA * loss1\n\ndef 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":{"id":"os3I73C-CPMj","execution":{"iopub.status.busy":"2022-06-12T09:59:14.257832Z","iopub.execute_input":"2022-06-12T09:59:14.258182Z","iopub.status.idle":"2022-06-12T09:59:14.266839Z","shell.execute_reply.started":"2022-06-12T09:59:14.258146Z","shell.execute_reply":"2022-06-12T09:59:14.266097Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"monet_generator_optimizer = tf.keras.optimizers.Adam(2e-4, beta_1=0.5)\nphoto_generator_optimizer = tf.keras.optimizers.Adam(2e-4, beta_1=0.5)\n\nmonet_discriminator_optimizer = tf.keras.optimizers.Adam(2e-4, beta_1=0.5)\nphoto_discriminator_optimizer = tf.keras.optimizers.Adam(2e-4, beta_1=0.5)","metadata":{"id":"kvy3FZq0pm46","execution":{"iopub.status.busy":"2022-06-12T09:59:14.268696Z","iopub.execute_input":"2022-06-12T09:59:14.269381Z","iopub.status.idle":"2022-06-12T09:59:14.279666Z","shell.execute_reply.started":"2022-06-12T09:59:14.269341Z","shell.execute_reply":"2022-06-12T09:59:14.278901Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## CycleGAN model","metadata":{"id":"1ppN6Chy4ZEN"}},{"cell_type":"code","source":"class CycleGan(keras.Model):\n    def __init__(self, monet_generator, photo_generator, monet_discriminator, photo_discriminator, lambda_cycle=10):\n        super(CycleGan, self).__init__()\n        self.m_gen = monet_generator\n        self.p_gen = photo_generator\n        self.m_disc = monet_discriminator\n        self.p_disc = photo_discriminator\n        self.lambda_cycle = lambda_cycle\n\n    def compile(self, m_gen_optimizer, p_gen_optimizer, m_disc_optimizer, p_disc_optimizer, gen_loss_fn, disc_loss_fn, cycle_loss_fn, identity_loss_fn):\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\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            # 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 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            # 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, 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\n        return {\"monet_gen_loss\": total_monet_gen_loss, \"photo_gen_loss\": total_photo_gen_loss, \"monet_disc_loss\": monet_disc_loss, \"photo_disc_loss\": photo_disc_loss}","metadata":{"id":"ixTTlol64dob","execution":{"iopub.status.busy":"2022-06-12T09:59:14.280869Z","iopub.execute_input":"2022-06-12T09:59:14.281328Z","iopub.status.idle":"2022-06-12T09:59:14.300734Z","shell.execute_reply.started":"2022-06-12T09:59:14.281292Z","shell.execute_reply":"2022-06-12T09:59:14.299840Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Biuild & compile","metadata":{"id":"t35kvuyr5DHb"}},{"cell_type":"code","source":"tf.config.run_functions_eagerly(True)\ncycle_gan_model = CycleGan(monet_generator, photo_generator, monet_discriminator, photo_discriminator)\n\ncycle_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":{"id":"IjYpfpcW5G6Q","execution":{"iopub.status.busy":"2022-06-12T09:59:14.301843Z","iopub.execute_input":"2022-06-12T09:59:14.303115Z","iopub.status.idle":"2022-06-12T09:59:14.326464Z","shell.execute_reply.started":"2022-06-12T09:59:14.303076Z","shell.execute_reply":"2022-06-12T09:59:14.325870Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Evaluation by FID","metadata":{"id":"oCB4kWx-zrKn"}},{"cell_type":"code","source":"# https://wandb.ai/ayush-thakur/gan-evaluation/reports/How-to-Evaluate-GANs-using-Frechet-Inception-Distance-FID---Vmlldzo0MTAxOTI\n\ndef calculate_fid(real_embeddings, generated_embeddings):\n    mu1, sigma1 = real_embeddings.mean(axis=0), np.cov(real_embeddings, rowvar=False)\n    mu2, sigma2 = generated_embeddings.mean(axis=0), np.cov(generated_embeddings,  rowvar=False)\n    ssdiff = np.sum((mu1 - mu2)**2.0)\n    covmean = linalg.sqrtm(sigma1.dot(sigma2))\n    if np.iscomplexobj(covmean):\n       covmean = covmean.real\n    fid = ssdiff + np.trace(sigma1 + sigma2 - 2.0 * covmean)\n    return fid","metadata":{"id":"K_SKUdfyRVZT","execution":{"iopub.status.busy":"2022-06-12T09:59:14.327751Z","iopub.execute_input":"2022-06-12T09:59:14.328356Z","iopub.status.idle":"2022-06-12T09:59:14.335175Z","shell.execute_reply.started":"2022-06-12T09:59:14.328316Z","shell.execute_reply":"2022-06-12T09:59:14.334356Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"inception_model = tf.keras.applications.InceptionV3(include_top=False, weights=\"imagenet\", pooling='avg')\n\ndef compute_embeddings(dataloader, count):\n    image_embeddings = []\n\n    for _ in range(count):\n        images = next(iter(dataloader))\n        embeddings = inception_model.predict(images)\n\n        image_embeddings.extend(embeddings)\n\n    return np.array(image_embeddings)\n","metadata":{"id":"tdX5IaRqxNsW","execution":{"iopub.status.busy":"2022-06-12T09:59:14.336573Z","iopub.execute_input":"2022-06-12T09:59:14.336949Z","iopub.status.idle":"2022-06-12T09:59:17.042907Z","shell.execute_reply.started":"2022-06-12T09:59:14.336914Z","shell.execute_reply":"2022-06-12T09:59:17.042066Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def generate_and_calculate_fid(samples):\n    generated = []\n    for photo in photo_ds.take(samples):\n        generated.append(cycle_gan_model.m_gen(photo))\n\n    generated_image_embeddings = compute_embeddings(generated, samples)\n    real_image_embeddings = compute_embeddings(monet_ds, samples)\n\n    return calculate_fid(generated_image_embeddings, real_image_embeddings)","metadata":{"id":"ntdYIf4z2lnQ","execution":{"iopub.status.busy":"2022-06-12T09:59:17.044219Z","iopub.execute_input":"2022-06-12T09:59:17.044633Z","iopub.status.idle":"2022-06-12T09:59:17.050701Z","shell.execute_reply.started":"2022-06-12T09:59:17.044547Z","shell.execute_reply":"2022-06-12T09:59:17.049768Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"class FidCallback(tf.keras.callbacks.Callback):\n    def __init__(self):\n        super().__init__()\n\n    def on_epoch_end(self, epoch, logs=None):\n        #if epoch%5 == 1:\n        logs['fid'] = generate_and_calculate_fid(50)\n","metadata":{"id":"VjlY_5rbzqZR","execution":{"iopub.status.busy":"2022-06-12T09:59:17.051834Z","iopub.execute_input":"2022-06-12T09:59:17.052708Z","iopub.status.idle":"2022-06-12T09:59:17.059804Z","shell.execute_reply.started":"2022-06-12T09:59:17.052609Z","shell.execute_reply":"2022-06-12T09:59:17.058996Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Train","metadata":{"id":"RUwPzi-S00WH"}},{"cell_type":"code","source":"cycle_gan_model.fit(tf.data.Dataset.zip((monet_ds, photo_ds)), epochs=40)","metadata":{"id":"dKTxn6rM5UFz","outputId":"5c52a52a-4b35-454d-9965-f864b8fef383","execution":{"iopub.status.busy":"2022-06-12T09:59:17.061024Z","iopub.execute_input":"2022-06-12T09:59:17.061489Z","iopub.status.idle":"2022-06-12T10:48:31.499226Z","shell.execute_reply.started":"2022-06-12T09:59:17.061455Z","shell.execute_reply":"2022-06-12T10:48:31.498467Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Example","metadata":{"id":"oC3CPo9ssu15"}},{"cell_type":"code","source":"iterator = iter(photo_ds)\n\nN = 4\nplt.figure(figsize=(10,50), dpi=100)\n\nfor i in range(1, N+1):\n    example_photo = next(iterator)\n    generated_image = cycle_gan_model.m_gen(example_photo, training=False)\n\n    plt.subplot(N*2, 2, (i*2)-1)\n    plt.title(\"Original\")\n    plt.imshow(example_photo[0] * 0.5 + 0.5, interpolation='nearest')\n\n    plt.subplot(N*2, 2, (i*2))\n    plt.title(\"Monet\")\n    plt.imshow(generated_image[0] * 0.5 + 0.5, interpolation='nearest')\n\nplt.show()","metadata":{"id":"vUktGkCJCPMm","execution":{"iopub.status.busy":"2022-06-12T10:48:31.500419Z","iopub.execute_input":"2022-06-12T10:48:31.500871Z","iopub.status.idle":"2022-06-12T10:48:33.563777Z","shell.execute_reply.started":"2022-06-12T10:48:31.500832Z","shell.execute_reply":"2022-06-12T10:48:33.552174Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Create submition file","metadata":{"id":"whutA7woZbCo"}},{"cell_type":"code","source":"import PIL\n! mkdir ../images/\n\nphoto_to_convert = load_dataset(PHOTO_FILENAMES, labeled=True).batch(1)","metadata":{"id":"33dwODk8ZdrB","execution":{"iopub.status.busy":"2022-06-12T10:50:08.027225Z","iopub.execute_input":"2022-06-12T10:50:08.027622Z","iopub.status.idle":"2022-06-12T10:50:08.761728Z","shell.execute_reply.started":"2022-06-12T10:50:08.027569Z","shell.execute_reply":"2022-06-12T10:50:08.760648Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"i = 1\nfor img in photo_to_convert:\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\n    \nprint(i)","metadata":{"id":"1xvSWnaKZihk","execution":{"iopub.status.busy":"2022-06-12T14:34:09.612814Z","iopub.execute_input":"2022-06-12T14:34:09.613270Z","iopub.status.idle":"2022-06-12T14:34:09.701041Z","shell.execute_reply.started":"2022-06-12T14:34:09.613183Z","shell.execute_reply":"2022-06-12T14:34:09.699695Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import shutil\nshutil.make_archive(\"/kaggle/working/images\", 'zip', \"/kaggle/images\")","metadata":{"id":"U5kiIRH6Zk4P","execution":{"iopub.status.busy":"2022-06-12T10:52:12.196672Z","iopub.execute_input":"2022-06-12T10:52:12.197069Z","iopub.status.idle":"2022-06-12T10:52:12.814801Z","shell.execute_reply.started":"2022-06-12T10:52:12.197035Z","shell.execute_reply":"2022-06-12T10:52:12.814039Z"},"trusted":true},"execution_count":null,"outputs":[]}]}