ConvNets

class pythae.models.nn.benchmarks.celeba.Encoder_Conv_AE_CELEBA(args)

A Convolutional encoder Neural net suited for CELEBA-64 and Autoencoder-based models.

It can be built as follows:

>>> from pythae.models.nn.benchmarks.celeba import Encoder_Conv_AE_CELEBA
>>> from pythae.models import AEConfig
>>> model_config = AEConfig(input_dim=(3, 64, 64), latent_dim=64)
>>> encoder = Encoder_Conv_AE_CELEBA(model_config)
>>> encoder
... Encoder_Conv_AE_CELEBA(
...   (layers): ModuleList(
...     (0): Sequential(
...       (0): Conv2d(3, 128, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
...       (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
...       (2): ReLU()
...     )
...     (1): Sequential(
...       (0): Conv2d(128, 256, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
...       (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
...       (2): ReLU()
...     )
...     (2): Sequential(
...       (0): Conv2d(256, 512, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
...       (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
...       (2): ReLU()
...     )
...     (3): Sequential(
...       (0): Conv2d(512, 1024, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
...       (1): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
...       (2): ReLU()
...     )
...   )
...   (embedding): Linear(in_features=16384, out_features=64, bias=True)
... )

and then passed to a pythae.models instance

>>> from pythae.models import AE
>>> model = AE(model_config=model_config, encoder=encoder)
>>> model.encoder == encoder
... True

Note

Please note that this encoder is only suitable for Autoencoder based models since it only outputs the embeddings of the input data under the key embedding.

>>> import torch
>>> input = torch.rand(2, 3, 64, 64)
>>> out = encoder(input)
>>> out.embedding.shape
... torch.Size([2, 64])

forward(x, output_layer_levels=None)

Forward method

Parameters

output_layer_levels (List[int]) – The levels of the layers where the outputs are extracted. If None, the last layer’s output is returned. Default: None.

Returns

An instance of ModelOutput containing the embeddings of the input data under the key embedding. Optional: The outputs of the layers specified in output_layer_levels arguments are available under the keys embedding_layer_i where i is the layer’s level.

Return type

ModelOutput

class pythae.models.nn.benchmarks.celeba.Encoder_Conv_VAE_CELEBA(args)

A Convolutional encoder Neural net suited for CELEBA-64 and Variational Autoencoder-based models.

It can be built as follows:

>>> from pythae.models.nn.benchmarks.celeba import Encoder_Conv_VAE_CELEBA
>>> from pythae.models import VAEConfig
>>> model_config = VAEConfig(input_dim=(3, 64, 64), latent_dim=64)
>>> encoder = Encoder_Conv_VAE_CELEBA(model_config)
>>> encoder
... Encoder_Conv_VAE_CELEBA(
...   (layers): ModuleList(
...     (0): Sequential(
...       (0): Conv2d(3, 128, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
...       (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
...       (2): ReLU()
...     )
...     (1): Sequential(
...       (0): Conv2d(128, 256, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
...       (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
...       (2): ReLU()
...     )
...     (2): Sequential(
...       (0): Conv2d(256, 512, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
...       (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
...       (2): ReLU()
...     )
...     (3): Sequential(
...       (0): Conv2d(512, 1024, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
...       (1): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
...       (2): ReLU()
...     )
...   )
...   (embedding): Linear(in_features=16384, out_features=64, bias=True)
...   (log_var): Linear(in_features=16384, out_features=64, bias=True)
... )

and then passed to a pythae.models instance

>>> from pythae.models import VAE
>>> model = VAE(model_config=model_config, encoder=encoder)
>>> model.encoder == encoder
... True

Note

Please note that this encoder is only suitable for Variational Autoencoder based models since it outputs the embeddings and the log of the covariance diagonal coefficients of the input data under the key embedding and log_covariance.

>>> import torch
>>> input = torch.rand(2, 3, 64, 64)
>>> out = encoder(input)
>>> out.embedding.shape
... torch.Size([2, 64])
>>> out.log_covariance.shape
... torch.Size([2, 64])

forward(x, output_layer_levels=None)

Forward method

Parameters

output_layer_levels (List[int]) – The levels of the layers where the outputs are extracted. If None, the last layer’s output is returned. Default: None.

Returns

An instance of ModelOutput containing the embeddings of the input data under the key embedding and the log of the diagonal coefficient of the covariance matrices under the key log_covariance. Optional: The outputs of the layers specified in output_layer_levels arguments are available under the keys embedding_layer_i where i is the layer’s level.

Return type

ModelOutput

class pythae.models.nn.benchmarks.celeba.Encoder_Conv_SVAE_CELEBA(args)

A Convolutional encoder Neural net suited for CELEBA-64 and Hyperspherical autoencoder Variational Autoencoder.

It can be built as follows:

>>> from pythae.models.nn.benchmarks.celeba import Encoder_Conv_SVAE_CELEBA
>>> from pythae.models import SVAEConfig
>>> model_config = SVAEConfig(input_dim=(3, 64, 64), latent_dim=64)
>>> encoder = Encoder_Conv_SVAE_CELEBA(model_config)
>>> encoder
... Encoder_Conv_SVAE_CELEBA(
...   (layers): ModuleList(
...     (0): Sequential(
...       (0): Conv2d(3, 128, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
...       (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
...       (2): ReLU()
...     )
...     (1): Sequential(
...       (0): Conv2d(128, 256, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
...       (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
...       (2): ReLU()
...     )
...     (2): Sequential(
...       (0): Conv2d(256, 512, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
...       (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
...       (2): ReLU()
...     )
...     (3): Sequential(
...       (0): Conv2d(512, 1024, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
...       (1): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
...       (2): ReLU()
...     )
...   )
...   (embedding): Linear(in_features=16384, out_features=64, bias=True)
...   (log_concentration): Linear(in_features=16384, out_features=1, bias=True)
... )

and then passed to a pythae.models instance

>>> from pythae.models import SVAE
>>> model = SVAE(model_config=model_config, encoder=encoder)
>>> model.encoder == encoder
... True

Note

Please note that this encoder is only suitable for Hyperspherical Variational Autoencoder models since it outputs the embeddings and the log of the concentration in the Von Mises Fisher distributions under the key embedding and log_concentration.

>>> import torch
>>> input = torch.rand(2, 3, 64, 64)
>>> out = encoder(input)
>>> out.embedding.shape
... torch.Size([2, 64])
>>> out.log_concentration.shape
... torch.Size([2, 1])

forward(x, output_layer_levels=None)

Forward method

Parameters

output_layer_levels (List[int]) – The levels of the layers where the outputs are extracted. If None, the last layer’s output is returned. Default: None.

Returns

An instance of ModelOutput containing the embeddings of the input data under the key embedding and the log of the diagonal coefficient of the covariance matrices under the key log_covariance. Optional: The outputs of the layers specified in output_layer_levels arguments are available under the keys embedding_layer_i where i is the layer’s level.

Return type

ModelOutput

class pythae.models.nn.benchmarks.celeba.Decoder_Conv_AE_CELEBA(args)

A Convolutional decoder Neural net suited for CELEBA-64 and Autoencoder-based models.

It can be built as follows:

>>> from pythae.models.nn.benchmarks.celeba import Decoder_Conv_AE_CELEBA
>>> from pythae.models import VAEConfig
>>> model_config = VAEConfig(input_dim=(3, 64, 64), latent_dim=64)
>>> decoder = Decoder_Conv_AE_CELEBA(model_config)
>>> decoder
... Decoder_Conv_AE_CELEBA(
...   (layers): ModuleList(
...     (0): Sequential(
...       (0): Linear(in_features=64, out_features=65536, bias=True)
...     )
...     (1): Sequential(
...       (0): ConvTranspose2d(1024, 512, kernel_size=(5, 5), stride=(2, 2), padding=(2, 2))
...       (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
...       (2): ReLU()
...     )
...     (2): Sequential(
...       (0): ConvTranspose2d(512, 256, kernel_size=(5, 5), stride=(2, 2), padding=(1, 1))
...       (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
...       (2): ReLU()
...     )
...     (3): Sequential(
...       (0): ConvTranspose2d(256, 128, kernel_size=(5, 5), stride=(2, 2), padding=(2, 2), output_padding=(1, 1))
...       (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
...       (2): ReLU()
...     )
...     (4): Sequential(
...       (0): ConvTranspose2d(128, 3, kernel_size=(5, 5), stride=(1, 1), padding=(1, 1))
...       (1): Sigmoid()
...     )
...   )
... )

and then passed to a pythae.models instance

>>> from pythae.models import VAE
>>> model = VAE(model_config=model_config, decoder=decoder)
>>> model.decoder == decoder
... True

Note

Please note that this decoder is suitable for all models.

>>> import torch
>>> input = torch.randn(2, 64)
>>> out = decoder(input)
>>> out.reconstruction.shape
... torch.Size([2, 3, 64, 64])

forward(z, output_layer_levels=None)

Forward method

Parameters

output_layer_levels (List[int]) – The levels of the layers where the outputs are extracted. If None, the last layer’s output is returned. Default: None.

Returns

An instance of ModelOutput containing the reconstruction of the latent code under the key reconstruction. Optional: The outputs of the layers specified in output_layer_levels arguments are available under the keys reconstruction_layer_i where i is the layer’s level.

Return type

ModelOutput

class pythae.models.nn.benchmarks.celeba.Discriminator_Conv_CELEBA(args)

A Convolutional discriminator Neural net suited for CELEBA.

It can be built as follows:

>>> from pythae.models.nn.benchmarks.celeba import Discriminator_Conv_CELEBA
>>> from pythae.models import VAEGANConfig
>>> model_config = VAEGANConfig(input_dim=(3, 64, 64), latent_dim=64)
>>> discriminator = Discriminator_Conv_CELEBA(model_config)
>>> discriminator
... Discriminator_Conv_CELEBA(
...   (layers): ModuleList(
...     (0): Sequential(
...       (0): Conv2d(3, 128, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
...       (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
...       (2): ReLU()
...     )
...     (1): Sequential(
...       (0): Conv2d(128, 256, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
...       (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
...       (2): Tanh()
...     )
...     (2): Sequential(
...       (0): Conv2d(256, 512, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
...       (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
...       (2): ReLU()
...     )
...     (3): Sequential(
...       (0): Conv2d(512, 1024, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
...       (1): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
...       (2): ReLU()
...     )
...     (4): Sequential(
...       (0): Linear(in_features=16384, out_features=1, bias=True)
...       (1): Sigmoid()
...     )
...   )
... )

and then passed to a pythae.models instance

>>> from pythae.models import VAEGAN
>>> model = VAEGAN(model_config=model_config, discriminator=discriminator)
>>> model.discriminator == discriminator
... True

forward(x, output_layer_levels=None)

Forward method

Parameters

output_layer_levels (List[int]) – The levels of the layers where the outputs are extracted. If None, the last layer’s output is returned. Default: None.

Returns

An instance of ModelOutput containing the adversarial score of the input under the key embedding. Optional: The outputs of the layers specified in output_layer_levels arguments are available under the keys embedding_layer_i where i is the layer’s level.

Return type

ModelOutput