src.encoders package

Subpackages

src.encoders.transforms package

Submodules

src.encoders.base module

class src.encoders.base.BaseEncoder

Bases: ABC

abstract decode(image)

abstract encode(image)

abstract update(batch_of_images)

src.encoders.vae module

class src.encoders.vae.VAE(image_channels: int = 3, image_height: int = 42, image_width: int = 144, z_dim: int = 32, load_checkpoint_from: str = '')

Bases: BaseEncoder, Module

Input should be (bsz, C, H, W) where C=3, H=42, W=144

bottleneck(h)

decode(z)

distribution(x, device='cpu')

encode(x: ndarray, device='cpu') → Tensor

forward(x)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

classmethod instantiate_from_config(config_file_location)

Initialize class from config file

Parameters: config_file_location (path) – Path to config file
Raises: ValueError – Error loading file
Returns: object from class.
Return type: cls

classmethod instantiate_from_config_dict(config)

Initialize class from config dictionary

Parameters: config (dictionary) – Create instance of class from dictionary.
Returns: Object from class and config.
Return type: cls

loss(actual, recon, mu, logvar, kld_weight=1.0)

reparameterize(mu, logvar)

representation(x)

schema = Map({Optional("image_channels"): Int(), Optional("image_height"): Int(), Optional("image_width"): Int(), Optional("z_dim"): Int(), Optional("load_checkpoint_from"): Str()})

training: bool

update(batch_of_images)

Module contents

Encoder definitions.