simplegan.autoencoder

The autoencoder sub-package provdies various autoencoder models that can be trained off the shelf with minimal fuss.

Vanilla Autoencoder

class simplegan.autoencoder.VanillaAutoencoder(interm_dim=64, enc_units=[256, 128], dec_units=[128, 256], activation='relu', kernel_initializer='glorot_uniform', kernel_regularizer=None)

Vanilla Autoencoder model

Parameters:

• interm_dim (int, optional) – represents the dimension of the bottleneck layer. Defaults to 64
• enc_units (int, list, optional) – represents the number of units/neurons in the encoder part of the network. Defaults to [256, 128]
• dec_units (int, list, optional) – represents the number of units/neurons in the decoder part of the network. Defaults to [128, 256]`
• activation (str, optional) – type of non-linearity to be applied. Defaults to relu
• kernel_initializer (str, optional) – initialization of kernel weights. Defaults to glorot_uniform
• kernel_regularizer (str, optional) – type of regularization to be applied to the weights. Defaults to None

fit(train_ds=None, epochs=100, optimizer='Adam', verbose=1, learning_rate=0.001, tensorboard=False, save_model=None)

Function to train the model

Parameters:

• train_ds (tf.data object) – training data
• epochs (int, optional) – number of epochs to train the model. Defaults to 100
• optimizer (str, optional) – optimizer used to train the model. Defaults to Adam
• verbose (int, optional) – 1 - prints training outputs, 0 - no outputs. Defaults to 1
• learning_rate (float, optional) – learning rate of the optimizer. Defaults to 0.001
• tensorboard (bool, optional) – if true, writes loss values to logs/gradient_tape directory which aids visualization. Defaults to False
• save_model (str, optional) – Directory to save the trained model. Defaults to None

generate_samples(test_ds=None, save_dir=None)

Generate samples using the trained model

Parameters:

• test_ds (tf.data object) – test data object used to generate samples
• save_dir (str, optional) – directory to save the generated images. Defaults to None

Returns:

returns None if save_dir is not None, otherwise returns a numpy array with generated samples

get_sample(data=None, n_samples=1, save_dir=None)

View sample of the data

Parameters:

• data (tf.data object) – dataset to load samples from
• n_samples (int, optional) – number of samples to load. Defaults to 1
• save_dir (str, optional) – directory to save the sample images. Defaults to None

Returns:

None if save_dir is not None, otherwise returns numpy array of samples with shape (n_samples, img_shape)

load_data(data_dir=None, use_mnist=False, use_cifar10=False, batch_size=32, img_shape=(64, 64))

Load data to train the model

Parameters:

• data_dir (str, optional) – string representing the directory to load data from. Defaults to None
• use_mnist (bool, optional) – use the MNIST dataset to train the model. Defaults to False
• use_cifar10 (bool, optional) – use the CIFAR10 dataset to train the model. Defaults to False
• batch_size (int, optional) – mini batch size for training the model. Defaults to 32
• img_shape (int, tuple, optional) – shape of the image when loading data from custom directory. Defaults to (64, 64)

Returns:

two tensorflow dataset objects representing the train and test datset

Convolutional Autoencoder

class simplegan.autoencoder.ConvolutionalAutoencoder(interm_dim=128, enc_channels=[32, 64], dec_channels=[64, 32], kernel_size=(5, 5), activation='relu', kernel_initializer='glorot_uniform', kernel_regularizer=None)

Convolutional Autoencoder model

Parameters:

• interm_dim (int, optional) – represents the dimension of the bottleneck layer. Defaults to 128
• enc_channels (int, list, optional) – represents the number of filters in the encoder part of the network. Defaults to [32, 64]
• dec_channels (int, list, optional) – represents the number of filters in the decoder part of the network. Defaults to [64, 32]`
• kernel_size (int, tuple, optional) – repersents the size of the kernel to perform the convolution. Defaults to (5, 5)
• activation (str, optional) – type of non-linearity to be applied. Defaults to relu
• kernel_initializer (str, optional) – initialization of kernel weights. Defaults to glorot_uniform
• kernel_regularizer (str, optional) – type of regularization to be applied to the weights. Defaults to None

fit(train_ds=None, epochs=100, optimizer='Adam', verbose=1, learning_rate=0.001, tensorboard=False, save_model=None)

Function to train the model

Parameters:

• train_ds (tf.data object) – training data
• epochs (int, optional) – number of epochs to train the model. Defaults to 100
• optimizer (str, optional) – optimizer used to train the model. Defaults to Adam
• verbose (int, optional) – 1 - prints training outputs, 0 - no outputs. Defaults to 1
• learning_rate (float, optional) – learning rate of the optimizer. Defaults to 0.001
• tensorboard (bool, optional) – if true, writes loss values to logs/gradient_tape directory which aids visualization. Defaults to False
• save_model (str, optional) – Directory to save the trained model. Defaults to None

generate_samples(test_ds=None, save_dir=None)

Generate samples using the trained model

Parameters:

• test_ds (tf.data object) – test data object used to generate samples
• save_dir (str, optional) – directory to save the generated images. Defaults to None

Returns:

returns None if save_dir is not None, otherwise returns a numpy array with generated samples

get_sample(data=None, n_samples=1, save_dir=None)

View sample of the data

Parameters:

• data (tf.data object) – dataset to load samples from
• n_samples (int, optional) – number of samples to load. Defaults to 1
• save_dir (str, optional) – directory to save the sample images. Defaults to None

Returns:

None if save_dir is not None, otherwise returns numpy array of samples with shape (n_samples, img_shape)

load_data(data_dir=None, use_mnist=False, use_cifar10=False, batch_size=32, img_shape=(64, 64))

Load data to train the model

Parameters:

• data_dir (str, optional) – string representing the directory to load data from. Defaults to None
• use_mnist (bool, optional) – use the MNIST dataset to train the model. Defaults to False
• use_cifar10 (bool, optional) – use the CIFAR10 dataset to train the model. Defaults to False
• batch_size (int, optional) – mini batch size for training the model. Defaults to 32
• img_shape (int, tuple, optional) – shape of the image when loading data from custom directory. Defaults to (64, 64)

Returns:

two tensorflow dataset objects representing the train and test datset

Variational Autoencoder

class simplegan.autoencoder.VAE(interm_dim=256, latent_dim=32, enc_units=[256, 128], dec_units=[128, 256], activation='relu', kernel_initializer='glorot_uniform', kernel_regularizer=None)

Variational Autoencoder model

Parameters:

• interm_dim (int, optional) – represents the dimension of the bottleneck layer. Defaults to 256
• latent_dim (int, optional) – represents the dimension of the distribution to sample from. Defaults to 32
• enc_units (int, list, optional) – represents the number of units/neurons in the encoder part of the network. Defaults to [256, 128]
• dec_units (int, list, optional) – represents the number of units/neurons in the decoder part of the network. Defaults to [128, 256]`
• activation (str, optional) – type of non-linearity to be applied. Defaults to relu
• kernel_initializer (str, optional) – initialization of kernel weights. Defaults to glorot_uniform
• kernel_regularizer (str, optional) – type of regularization to be applied to the weights. Defaults to None

fit(train_ds=None, epochs=100, optimizer='Adam', verbose=1, learning_rate=0.001, tensorboard=False, save_model=None)

Function to train the model

Parameters:

• train_ds (tf.data object) – training data
• epochs (int, optional) – number of epochs to train the model. Defaults to 100
• optimizer (str, optional) – optimizer used to train the model. Defaults to Adam
• verbose (int, optional) – 1 - prints training outputs, 0 - no outputs. Defaults to 1
• learning_rate (float, optional) – learning rate of the optimizer. Defaults to 0.001
• tensorboard (bool, optional) – if true, writes loss values to logs/gradient_tape directory which aids visualization. Defaults to False
• save_model (str, optional) – Directory to save the trained model. Defaults to None

generate_samples(test_ds=None, save_dir=None)

Generate samples using the trained model

Parameters:

• test_ds (tf.data object) – test data object used to generate samples
• save_dir (str, optional) – directory to save the generated images. Defaults to None

Returns:

returns None if save_dir is not None, otherwise returns a numpy array with generated samples

get_sample(data=None, n_samples=1, save_dir=None)

View sample of the data

Parameters:

• data (tf.data object) – dataset to load samples from
• n_samples (int, optional) – number of samples to load. Defaults to 1
• save_dir (str, optional) – directory to save the sample images. Defaults to None

Returns:

None if save_dir is not None, otherwise returns numpy array of samples with shape (n_samples, img_shape)

load_data(data_dir=None, use_mnist=False, use_cifar10=False, batch_size=32, img_shape=(64, 64))

Load data to train the model

Parameters:

• data_dir (str, optional) – string representing the directory to load data from. Defaults to None
• use_mnist (bool, optional) – use the MNIST dataset to train the model. Defaults to False
• use_cifar10 (bool, optional) – use the CIFAR10 dataset to train the model. Defaults to False
• batch_size (int, optional) – mini batch size for training the model. Defaults to 32
• img_shape (int, tuple, optional) – shape of the image when loading data from custom directory. Defaults to (64, 64)

Returns:

two tensorflow dataset objects representing the train and test datset

Vector-Quantized Variational Autoencoder

class simplegan.autoencoder.VQ_VAE(num_hiddens=128, num_residual_hiddens=32, num_residual_layers=2, num_embeddings=512, embedding_dim=64, commiment_cost=0.25)

Vector-Quantized Variational Autoencoder model

Parameters:

• num_hiddens (int, optional) – number of filters in the convolution operation in residual block. Defaults to 128
• num_residual_hiddens (int, optional) – number of filters in the convolution operation in residual block. Defaults to 32
• num_residual_layers (int, optional) – number of residual blocks. Defaults to 2
• num_embeddings (int, optional) – capacity of information bottleneck. Defaults to 512
• embedding_dim (int, optional) – size of the embedding. does not affect peformance of the model much. Defaults to 64
• commiment_cost (float, optional) – scale the embedding latent loss. Defaults to 0.25

fit(train_ds=None, epochs=100, optimizer='Adam', verbose=1, learning_rate=0.0003, tensorboard=False, save_model=None)

Function to train the model

Parameters:

• train_ds (tf.data object) – training data
• epochs (int, optional) – number of epochs to train the model. Defaults to 100
• optimizer (str, optional) – optimizer used to train the model. Defaults to Adam
• verbose (int, optional) – 1 - prints training outputs, 0 - no outputs. Defaults to 1
• learning_rate (float, optional) – learning rate of the optimizer. Defaults to 0.001
• tensorboard (bool, optional) – if true, writes loss values to logs/gradient_tape directory which aids visualization. Defaults to False
• save_model (str, optional) – Directory to save the trained model. Defaults to None

generate_samples(test_ds=None, save_dir=None)

Generate samples using the trained model

Parameters:

• test_ds (tf.data object) – test data object used to generate samples
• save_dir (str, optional) – directory to save the generated images. Defaults to None

Returns:

returns None if save_dir is not None, otherwise returns a numpy array with generated samples

get_sample(data=None, n_samples=1, save_dir=None)

View sample of the data

Parameters:

• data (tf.data object) – dataset to load samples from
• n_samples (int, optional) – number of samples to load. Defaults to 1
• save_dir (str, optional) – directory to save the sample images. Defaults to None

Returns:

None if save_dir is not None, otherwise returns numpy array of samples with shape (n_samples, img_shape)

load_data(data_dir=None, use_mnist=False, use_cifar10=False, batch_size=32, img_shape=(64, 64))

Load data to train the model

Parameters:

• data_dir (str, optional) – string representing the directory to load data from. Defaults to None
• use_mnist (bool, optional) – use the MNIST dataset to train the model. Defaults to False
• use_cifar10 (bool, optional) – use the CIFAR10 dataset to train the model. Defaults to False
• batch_size (int, optional) – mini batch size for training the model. Defaults to 32
• img_shape (int, tuple, optional) – shape of the image when loading data from custom directory. Defaults to (64, 64)

Returns:

two tensorflow dataset objects representing the train and test datset