{"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 os\nimport re\nimport shutil\nimport tensorflow as tf\nfrom tensorflow import keras\nfrom tensorflow.keras import layers\nfrom kaggle_datasets import KaggleDatasets\nimport tensorflow_addons as tfa\n\nimport matplotlib.pyplot as plt\nimport numpy as np\n\nfrom scipy import linalg","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:07:12.284888Z","iopub.execute_input":"2021-06-13T15:07:12.285298Z","iopub.status.idle":"2021-06-13T15:07:17.553058Z","shell.execute_reply.started":"2021-06-13T15:07:12.285215Z","shell.execute_reply":"2021-06-13T15:07:17.552105Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"alpha = 0.2\nshape = [256, 256, 3]\nLambda_cycle = 15\nLambda_identity = 10\nBATCH_SIZE = 30\n#EPOCHS = 1\nAUTOTUNE = tf.data.experimental.AUTOTUNE","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:07:17.554597Z","iopub.execute_input":"2021-06-13T15:07:17.555007Z","iopub.status.idle":"2021-06-13T15:07:17.559892Z","shell.execute_reply.started":"2021-06-13T15:07:17.554968Z","shell.execute_reply":"2021-06-13T15:07:17.558635Z"},"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)","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:07:17.561890Z","iopub.execute_input":"2021-06-13T15:07:17.562694Z","iopub.status.idle":"2021-06-13T15:07:23.108184Z","shell.execute_reply.started":"2021-06-13T15:07:17.562650Z","shell.execute_reply":"2021-06-13T15:07:23.107335Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"GCS_PATH = KaggleDatasets().get_gcs_path('gan-getting-started')\n\nMONET_FILENAMES = tf.io.gfile.glob(str(GCS_PATH + '/monet_tfrec/monet*.tfrec'))\nPHOTO_FILENAMES = tf.io.gfile.glob(str(GCS_PATH + '/photo_tfrec/photo*.tfrec'))\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) #获取图片的数量\n\ndef data_augment(image): #数据增强\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 if p_crop > .5:\n image = tf.image.resize(image, [286, 286]) #resizing to 286 x 286 x 3\n image = tf.image.random_crop(image, size=[256, 256, 3]) # randomly cropping to 256 x 256 x 3\n if p_crop > .9:\n image = tf.image.resize(image, [300, 300])\n image = tf.image.random_crop(image, size=[256, 256, 3])\n \n if p_rotate > .9:\n image = tf.image.rot90(image, k=3) # rotate 270º\n elif p_rotate > .7:\n image = tf.image.rot90(image, k=2) # rotate 180º\n elif p_rotate > .5:\n image = tf.image.rot90(image, k=1) # rotate 90º\n \n ## random mirroring\n if p_spatial > .6:\n image = tf.image.random_flip_left_right(image)\n image = tf.image.random_flip_up_down(image)\n if p_spatial > .9:\n image = tf.image.transpose(image)\n \n return image\n\nIMAGE_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\n\ndef load_dataset(filenames): #数据集解码\n dataset = tf.data.TFRecordDataset(filenames)\n dataset = dataset.map(read_tfrecord, num_parallel_calls=AUTOTUNE)\n return dataset\n\ndef 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=AUTOTUNE)\n photo_ds = photo_ds.map(augment, num_parallel_calls=AUTOTUNE)\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 monet_ds = monet_ds.cache()\n photo_ds = photo_ds.cache()\n monet_ds = monet_ds.prefetch(AUTOTUNE)\n photo_ds = photo_ds.prefetch(AUTOTUNE)\n \n gan_ds = tf.data.Dataset.zip((monet_ds, photo_ds)) #经过数据增强后将数据集进行组合\n \n return gan_ds,photo_ds,monet_ds","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:07:23.109993Z","iopub.execute_input":"2021-06-13T15:07:23.110300Z","iopub.status.idle":"2021-06-13T15:07:23.587965Z","shell.execute_reply.started":"2021-06-13T15:07:23.110272Z","shell.execute_reply":"2021-06-13T15:07:23.586970Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"#获得完整的数据集\nfull_dataset,my_photo_ds,my_monet_ds = get_gan_dataset(MONET_FILENAMES, PHOTO_FILENAMES, augment=data_augment, repeat=True, shuffle=True, batch_size=BATCH_SIZE)\n\n#这段代码用于可视化数据集，调试上方的程序\nexample_monet , example_photo = next(iter(full_dataset))\nplt.subplot(121)\nplt.title('Real photo')\nplt.imshow(example_photo[2] * 0.5 + 0.5)\n\nplt.subplot(122)\nplt.title('Monet painting')\nplt.imshow(example_monet[2]* 0.5 + 0.5)","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:07:23.591519Z","iopub.execute_input":"2021-06-13T15:07:23.591836Z","iopub.status.idle":"2021-06-13T15:07:31.800250Z","shell.execute_reply.started":"2021-06-13T15:07:23.591809Z","shell.execute_reply":"2021-06-13T15:07:31.799509Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"fast_photo_ds = load_dataset(PHOTO_FILENAMES).batch(32*strategy.num_replicas_in_sync).prefetch(32)\nmonet_ds_fid = load_dataset(MONET_FILENAMES).batch(32*strategy.num_replicas_in_sync).prefetch(32)","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:07:31.801447Z","iopub.execute_input":"2021-06-13T15:07:31.801883Z","iopub.status.idle":"2021-06-13T15:07:31.849749Z","shell.execute_reply.started":"2021-06-13T15:07:31.801853Z","shell.execute_reply":"2021-06-13T15:07:31.848995Z"},"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 inception_model = tf.keras.applications.InceptionV3(input_shape=(256,256,3),pooling=\"avg\",include_top=False)\n\n mix3 = inception_model.get_layer(\"mixed3\").output\n f0 = tf.keras.layers.GlobalMaxPooling2D()(mix3)\n\n inception_model = tf.keras.Model(inputs=inception_model.input, outputs=f0)\n inception_model.trainable = False\n \ndef calculate_activation_statistics_mod(images,fid_model):\n act=fid_model.predict(images)\n mu = np.mean(act, axis=0)\n sigma = np.cov(act, rowvar=False)\n return mu, sigma\nmyFID_mu2, myFID_sigma2 = calculate_activation_statistics_mod(monet_ds_fid,inception_model)","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:07:31.850868Z","iopub.execute_input":"2021-06-13T15:07:31.851324Z","iopub.status.idle":"2021-06-13T15:07:53.957447Z","shell.execute_reply.started":"2021-06-13T15:07:31.851294Z","shell.execute_reply":"2021-06-13T15:07:53.956027Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(myFID_mu2.shape,myFID_sigma2.shape)","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:07:53.962442Z","iopub.execute_input":"2021-06-13T15:07:53.963042Z","iopub.status.idle":"2021-06-13T15:07:53.970151Z","shell.execute_reply.started":"2021-06-13T15:07:53.962990Z","shell.execute_reply":"2021-06-13T15:07:53.968955Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def calculate_frechet_distance(mu1,sigma1,mu2,sigma2):\n fid_epsilon = 1e-14\n mu1 = np.atleast_1d(mu1)\n mu2 = np.atleast_1d(mu2)\n sigma1 = np.atleast_2d(sigma1)\n sigma2 = np.atleast_2d(sigma2)\n\n assert mu1.shape == mu2.shape, 'Training and test mean vectors have different lengths'\n assert sigma1.shape == sigma2.shape, 'Training and test covariances have different dimensions'\n\n # product might be almost singular\n covmean, _ = linalg.sqrtm(sigma1.dot(sigma2), disp=False)\n if not np.isfinite(covmean).all():\n msg = f'fid calculation produces singular product; adding {fid_epsilon} to diagonal of cov estimates'\n warnings.warn(msg)\n offset = np.eye(sigma1.shape[0]) * fid_epsilon\n covmean = linalg.sqrtm((sigma1 + offset).dot(sigma2 + offset))\n \n # numerical error might give slight imaginary component\n if np.iscomplexobj(covmean):\n if not np.allclose(np.diagonal(covmean).imag, 0, atol=1e-3):\n m = np.max(np.abs(covmean.imag))\n raise ValueError(f'Imaginary component {m}')\n covmean = covmean.real\n tr_covmean = np.trace(covmean)\n return (mu1 - mu2).dot(mu1 - mu2) + np.trace(sigma1) + np.trace(sigma2) - 2 * tr_covmean\n\ndef FID(images,gen_model,inception_model=inception_model,myFID_mu2=myFID_mu2, myFID_sigma2=myFID_sigma2):\n with strategy.scope():\n inp = layers.Input(shape=[256, 256, 3], name='input_image')\n x = gen_model(inp)\n x=inception_model(x)\n fid_model = tf.keras.Model(inputs=inp, outputs=x)\n \n mu1, sigma1 = calculate_activation_statistics_mod(images,fid_model)\n\n fid_value = calculate_frechet_distance(mu1, sigma1,myFID_mu2, myFID_sigma2)\n\n return fid_value","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:07:53.972770Z","iopub.execute_input":"2021-06-13T15:07:53.976004Z","iopub.status.idle":"2021-06-13T15:07:54.119366Z","shell.execute_reply.started":"2021-06-13T15:07:53.975941Z","shell.execute_reply":"2021-06-13T15:07:54.118093Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"OUTPUT_CHANNELS = 3\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',kernel_initializer=initializer, use_bias=False))\n #这里没有做池化、Dropout，通过卷积减少图像体积，与VGG会有点不一样\n if apply_instancenorm: #是否进行标准化（InstanceNormalization优于batchNormalization）\n result.add(tfa.layers.InstanceNormalization(gamma_initializer=gamma_init))\n \n result.add(layers.LeakyReLU()) #采用LeakyReLU激活，参数并没有设置\n\n return result","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:07:54.121110Z","iopub.execute_input":"2021-06-13T15:07:54.121446Z","iopub.status.idle":"2021-06-13T15:07:54.145185Z","shell.execute_reply.started":"2021-06-13T15:07:54.121414Z","shell.execute_reply":"2021-06-13T15:07:54.143970Z"},"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,padding='same',kernel_initializer=initializer, use_bias=False))\n #基本上不用偏置\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()) #用ReLU激活\n\n return result #注意：result是个model,是可以通过model(x)来作用于数据的","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:07:54.146658Z","iopub.execute_input":"2021-06-13T15:07:54.147027Z","iopub.status.idle":"2021-06-13T15:07:54.168012Z","shell.execute_reply.started":"2021-06-13T15:07:54.146996Z","shell.execute_reply":"2021-06-13T15:07:54.167011Z"},"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 ]#首先下采样（这里层次可以添加（调整downSample的stride）！！）（参考VGG可是不止这么少的，不过down+up倒是足够）\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 ]#然后上采样\n #上面只是建立模型，下面是将模型用到数据集上\n \n initializer = tf.random_normal_initializer(0., 0.02)\n last = layers.Conv2DTranspose(OUTPUT_CHANNELS, 4,strides=2,padding='same',\n kernel_initializer=initializer,activation='tanh') # (bs, 256, 256, 3)\n\n x = inputs\n\n skips = []\n for down in down_stack:\n x = down(x)\n skips.append(x) #实现下采样\n\n skips = reversed(skips[:-1])\n\n for up, skip in zip(up_stack, skips):\n x = up(x)\n x = layers.Concatenate()([x, skip]) #这里是提高质量的核心，通过连接将当前上采样层和下采样层连接，即保留了卷积的特征，\n #又能在下采样的基础上进行上采样\n\n x = last(x) #最后一层通过上采样生成[255,255,3]的假图\n\n return keras.Model(inputs=inputs, outputs=x) #返回生成器模型","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:07:54.169450Z","iopub.execute_input":"2021-06-13T15:07:54.170046Z","iopub.status.idle":"2021-06-13T15:07:54.183707Z","shell.execute_reply.started":"2021-06-13T15:07:54.170009Z","shell.execute_reply":"2021-06-13T15:07:54.182724Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"generator_g = Generator()\ntf.keras.utils.plot_model(generator_g, show_shapes=True, dpi=64)\n\nphoto = (example_photo[0,...] * 0.5 + 0.5)\nexample_gen_output_y = generator_g(photo[tf.newaxis,...], training=False)\n\nplt.subplot(1,2,1)\nplt.imshow(photo, vmin=0, vmax=255) \n\nplt.subplot(1,2,2)\nplt.imshow(example_gen_output_y[0]) \n\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:07:54.185079Z","iopub.execute_input":"2021-06-13T15:07:54.185687Z","iopub.status.idle":"2021-06-13T15:07:56.937704Z","shell.execute_reply.started":"2021-06-13T15:07:54.185650Z","shell.execute_reply":"2021-06-13T15:07:56.936587Z"},"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 #层数上比较少，后面基本通过补零再卷积来提取特征\n norm1 = tfa.layers.InstanceNormalization(gamma_initializer=gamma_init)(conv)\n\n leaky_relu = layers.LeakyReLU()(norm1)\n\n zero_pad2 = layers.ZeroPadding2D()(leaky_relu) # (bs, 33, 33, 512)\n\n last = layers.Conv2D(1, 4, strides=1,\n kernel_initializer=initializer)(zero_pad2) # (bs, 30, 30, 1)\n\n return tf.keras.Model(inputs=inp, outputs=last)","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:07:56.939656Z","iopub.execute_input":"2021-06-13T15:07:56.940071Z","iopub.status.idle":"2021-06-13T15:07:56.949369Z","shell.execute_reply.started":"2021-06-13T15:07:56.940026Z","shell.execute_reply":"2021-06-13T15:07:56.948673Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"discriminator_y = Discriminator()\ntf.keras.utils.plot_model(discriminator_y, show_shapes=True, dpi=64) ","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:07:56.950316Z","iopub.execute_input":"2021-06-13T15:07:56.950744Z","iopub.status.idle":"2021-06-13T15:07:57.316407Z","shell.execute_reply.started":"2021-06-13T15:07:56.950702Z","shell.execute_reply":"2021-06-13T15:07:57.315116Z"},"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.status.busy":"2021-06-13T15:07:57.318692Z","iopub.execute_input":"2021-06-13T15:07:57.319106Z","iopub.status.idle":"2021-06-13T15:08:06.569094Z","shell.execute_reply.started":"2021-06-13T15:07:57.319069Z","shell.execute_reply":"2021-06-13T15:08:06.568194Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"photo = (example_photo[0,...] * 0.5 + 0.5)\nmonet = (example_monet[0,...] * 0.5 + 0.5)\n\n# From photo we generate Monet (fake) and regenerate the photo (cycle) again\nexample_gen_output_monet_fake = monet_generator(photo[tf.newaxis,...], training=False)\nexample_gen_output_photo_cycle = photo_generator(example_gen_output_monet_fake, training=False)\n\n# We run the discriminator for Monet (fake)\nexample_disc_out_monet = monet_discriminator(example_gen_output_monet_fake, training=False)\n\n\n# From Monet we generate photo (fake) and regenerate Monet (cycle) again\nexample_gen_output_photo_fake = photo_generator(monet[tf.newaxis,...], training=False)\nexample_gen_output_monet_cycle = monet_generator(example_gen_output_photo_fake, training=False)\n\n# We execute the discriminator for Photo (fake)\nexample_disc_out_photo = photo_discriminator(example_gen_output_photo_fake, training=False)\n\n\n# We present results, as the network is not trained, the outputs are not good, \n# but we modify the scala to be able to have some example images\n\nplt.figure(figsize=(10,10))\n\n# Input Photo\nplt.subplot(2,4,1)\nplt.imshow(photo, vmin=0, vmax=255) \n\n# Fake Monet\nplt.subplot(2,4,2)\n#m = example_gen_output_monet_fake[0,...].numpy()\n#print(np.min(m), np.max(m))\ncontrast = 100 \nplt.imshow(example_gen_output_monet_fake[0,...]*contrast) \n\n# Photo Cycle\nplt.subplot(2,4,3)\n#m = example_gen_output_photo_cycle[0,...].numpy()\n#print(np.min(m), np.max(m))\ncontrast = 100\nplt.imshow(example_gen_output_photo_cycle[0,...]*contrast) \n\n# Monet discriminator result\n#plt.subplot(2,4,4)\n#m = example_disc_out_monet[0,...,-1].numpy()\n#print(np.min(m), np.max(m))\n#contrast = 1000\n#im = plt.imshow(m*contrast, vmin=-20, vmax=20, cmap='RdBu_r')\n#plt.colorbar(im,fraction=0.046, pad=0.04)\n\n\n\n# Input Monet\nplt.subplot(2,4,5)\nplt.imshow(monet, vmin=0, vmax=255) \n\n# Fake Photo\nplt.subplot(2,4,6)\n#m = example_gen_output_photo_fake[0,...].numpy()\n#print(np.min(m), np.max(m))\ncontrast = 100\nplt.imshow(example_gen_output_photo_fake[0,...]*contrast) \n\n# Monet Cycle\nplt.subplot(2,4,7)\n#m = example_gen_output_monet_cycle[0,...].numpy()\n#print(np.min(m), np.max(m))\ncontrast = 100\nplt.imshow(example_gen_output_monet_cycle[0,...]*contrast) \n\n# Photo discriminator result \n#plt.subplot(2,4,8)\n#m = example_disc_out_photo[0,...,-1].numpy()\n#print(np.min(m), np.max(m))\n#contrast = 1000\n#im = plt.imshow(contrast, vmin=-20, vmax=20, cmap='RdBu_r')\n#plt.colorbar(im,fraction=0.046, pad=0.04)","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:08:06.570234Z","iopub.execute_input":"2021-06-13T15:08:06.570638Z","iopub.status.idle":"2021-06-13T15:08:08.529550Z","shell.execute_reply.started":"2021-06-13T15:08:06.570596Z","shell.execute_reply":"2021-06-13T15:08:08.528301Z"},"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=15,\n ):\n super(CycleGan, self).__init__()\n self.m_gen = monet_generator\n self.p_gen = photo_generator\n self.m_disc = monet_discriminator\n self.p_disc = photo_discriminator\n self.lambda_cycle = lambda_cycle\n \n def compile(\n self,\n m_gen_optimizer,\n p_gen_optimizer,\n m_disc_optimizer,\n p_disc_optimizer,\n gen_loss_fn,\n disc_loss_fn,\n cycle_loss_fn,\n identity_loss_fn\n ):\n super(CycleGan, self).compile()\n self.m_gen_optimizer = m_gen_optimizer\n self.p_gen_optimizer = p_gen_optimizer\n self.m_disc_optimizer = m_disc_optimizer\n self.p_disc_optimizer = p_disc_optimizer\n self.gen_loss_fn = gen_loss_fn\n self.disc_loss_fn = disc_loss_fn\n self.cycle_loss_fn = cycle_loss_fn\n self.identity_loss_fn = identity_loss_fn\n \n def train_step(self, batch_data):\n real_monet, real_photo = batch_data\n \n with tf.GradientTape(persistent=True) as tape:\n # photo to monet back to photo\n fake_monet = self.m_gen(real_photo, training=True)\n cycled_photo = self.p_gen(fake_monet, training=True)\n\n # monet to photo back to monet\n fake_photo = self.p_gen(real_monet, training=True)\n cycled_monet = self.m_gen(fake_photo, training=True)\n\n # generating itself\n same_monet = self.m_gen(real_monet, training=True)\n same_photo = self.p_gen(real_photo, training=True)\n\n # discriminator used to check, inputing real images\n disc_real_monet = self.m_disc(real_monet, training=True)\n disc_real_photo = self.p_disc(real_photo, training=True)\n\n # discriminator used to check, inputing fake images\n disc_fake_monet = self.m_disc(fake_monet, training=True)\n disc_fake_photo = self.p_disc(fake_photo, training=True)\n\n # evaluates generator loss\n monet_gen_loss = self.gen_loss_fn(disc_fake_monet)\n photo_gen_loss = self.gen_loss_fn(disc_fake_photo)\n\n # evaluates total cycle consistency loss\n total_cycle_loss = self.cycle_loss_fn(real_monet, cycled_monet, self.lambda_cycle) + self.cycle_loss_fn(real_photo, cycled_photo, self.lambda_cycle)\n\n # evaluates total generator loss\n total_monet_gen_loss = monet_gen_loss + total_cycle_loss + self.identity_loss_fn(real_monet, same_monet, self.lambda_cycle)\n total_photo_gen_loss = photo_gen_loss + total_cycle_loss + self.identity_loss_fn(real_photo, same_photo, self.lambda_cycle)\n\n # evaluates discriminator loss\n monet_disc_loss = self.disc_loss_fn(disc_real_monet, disc_fake_monet)\n photo_disc_loss = self.disc_loss_fn(disc_real_photo, disc_fake_photo)\n\n # Calculate the gradients for generator and discriminator\n monet_generator_gradients = tape.gradient(total_monet_gen_loss,\n self.m_gen.trainable_variables)\n photo_generator_gradients = tape.gradient(total_photo_gen_loss,\n self.p_gen.trainable_variables)\n\n monet_discriminator_gradients = tape.gradient(monet_disc_loss,\n self.m_disc.trainable_variables)\n photo_discriminator_gradients = tape.gradient(photo_disc_loss,\n self.p_disc.trainable_variables)\n\n # Apply the gradients to the optimizer\n self.m_gen_optimizer.apply_gradients(zip(monet_generator_gradients,\n self.m_gen.trainable_variables))\n\n self.p_gen_optimizer.apply_gradients(zip(photo_generator_gradients,\n self.p_gen.trainable_variables))\n\n self.m_disc_optimizer.apply_gradients(zip(monet_discriminator_gradients,\n self.m_disc.trainable_variables))\n\n self.p_disc_optimizer.apply_gradients(zip(photo_discriminator_gradients,\n self.p_disc.trainable_variables))\n \n return {\n \"monet_gen_loss\": total_monet_gen_loss,\n \"photo_gen_loss\": total_photo_gen_loss,\n \"monet_disc_loss\": monet_disc_loss,\n \"photo_disc_loss\": photo_disc_loss\n }","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:08:08.531251Z","iopub.execute_input":"2021-06-13T15:08:08.531570Z","iopub.status.idle":"2021-06-13T15:08:08.550360Z","shell.execute_reply.started":"2021-06-13T15:08:08.531539Z","shell.execute_reply":"2021-06-13T15:08:08.549321Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():#并行运算\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 #这里用交叉熵，但吴恩达最新教程推荐Mse(可以试着换)\n total_disc_loss = real_loss + generated_loss\n\n return total_disc_loss * 0.4","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:08:08.551434Z","iopub.execute_input":"2021-06-13T15:08:08.551809Z","iopub.status.idle":"2021-06-13T15:08:08.569485Z","shell.execute_reply.started":"2021-06-13T15:08:08.551769Z","shell.execute_reply":"2021-06-13T15:08:08.568100Z"},"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,\n reduction=tf.keras.losses.Reduction.NONE)(tf.ones_like(generated), generated)\nwith 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_cycle * loss1\n\nwith 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_identity * loss","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:08:08.571756Z","iopub.execute_input":"2021-06-13T15:08:08.572216Z","iopub.status.idle":"2021-06-13T15:08:08.583194Z","shell.execute_reply.started":"2021-06-13T15:08:08.572147Z","shell.execute_reply":"2021-06-13T15:08:08.582197Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"with strategy.scope():\n monet_generator_optimizer = tf.keras.optimizers.Adam(0.002, decay = 0, beta_1=0.5)\n photo_generator_optimizer = tf.keras.optimizers.Adam(0.002, beta_1=0.5)\n\n monet_discriminator_optimizer = tf.keras.optimizers.Adam(0.002, beta_1=0.5)\n photo_discriminator_optimizer = tf.keras.optimizers.Adam(0.002, beta_1=0.5)","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:08:08.584325Z","iopub.execute_input":"2021-06-13T15:08:08.584763Z","iopub.status.idle":"2021-06-13T15:08:08.600897Z","shell.execute_reply.started":"2021-06-13T15:08:08.584734Z","shell.execute_reply":"2021-06-13T15:08:08.599816Z"},"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 )\n \n#cycle_gan_model.fit(\n #full_dataset,\n #epochs=EPOCHS,\n #steps_per_epoch=(max(n_monet_samples, n_photo_samples)//BATCH_SIZE),\n#)#这里是训练啦","metadata":{"execution":{"iopub.status.busy":"2021-06-13T15:08:08.602236Z","iopub.execute_input":"2021-06-13T15:08:08.602514Z","iopub.status.idle":"2021-06-13T15:08:08.663174Z","shell.execute_reply.started":"2021-06-13T15:08:08.602488Z","shell.execute_reply":"2021-06-13T15:08:08.662076Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"%%time\nfids=[]\nbest_fid=999999999\nfor epoch in range(1,32):\n\n print(\"Epoch = \",epoch)\n hist=cycle_gan_model.fit(full_dataset,steps_per_epoch=1500, epochs=1).history\n cur_fid=FID(fast_photo_ds,monet_generator)\n# disc_m_loss1.append(hist[\"monet_disc_loss1\"][0][0][0])\n# disc_p_loss1.append(hist[\"photo_disc_loss1\"][0][0][0])\n fids.append(cur_fid)\n print(\"After epoch #{} FID = {}\\n\".format(epoch,cur_fid))\n if cur_fid