paraphernalia.torch.modules module¶

A collection of utility PyTorch modules.

class AdaptiveMultiLoss(num_losses)[source]¶

Automatic loss balancing.

See also

Parameters

forward(losses)[source]¶

class Parallel(components)[source]¶

A module that runs a number of submodule in parallel and collects their outputs into a list.

Parameters: components (List[nn.Module]) – a list of submdodules to run in parallel

forward(*inputs)[source]¶

Run each submodule on input, and accumulate the outputs into a list.

class Constant(value)[source]¶

A module that returns a constant value, ignoring any inputs.

Initializes internal Module state, shared by both nn.Module and ScriptModule.

forward(*ignored)[source]¶

Return the constant value, ignoring any inputs.

class WeightedSum(**components)[source]¶

More or less a weighted sum of named module outputs, but with special handling for negative weights.

In order for the weighting to make sense, the components needs outputs with the same shape and meaning.

For loss functions outputs should be in the range [0,1]

set_weight(name, value)[source]¶

Set the weight associated with a module.

Parameters

Return type

None

Compute the weighted loss.

class SimilarTo(targets)[source]¶

Cosine similarity test with mean pooling.

Parameters: targets (torch.Tensor) – A tensor of dimension (targets, channels)

class SimilarToAny(targets)[source]¶

Cosine similarity test with max pooling.

Parameters: targets (torch.Tensor) – A tensor of dimension (targets, channels)

paraphernalia.torch.lightning module

paraphernalia.torch.noise module