Source code for pytorchrl.agent.actors.feature_extractors.ensemble_layer

import torch
import torch.nn as nn


class EnsembleFC(nn.Module):
    def __init__(self, in_features: int, out_features: int, ensemble_size: int, weight_decay: float = 0., bias: bool = True) -> None:
        super(EnsembleFC, self).__init__()
        self.in_features = in_features
        self.out_features = out_features
        self.ensemble_size = ensemble_size
        self.weight = nn.Parameter(torch.Tensor(ensemble_size, in_features, out_features))
        torch.nn.init.xavier_uniform_(self.weight)
        #torch.nn.init.trunc_normal_(self.weight)
        self.weight_decay = weight_decay
        if bias:
            self.bias = nn.Parameter(torch.zeros(ensemble_size, out_features))
        else:
            self.register_parameter('bias', None)
        self.reset_parameters()

    def reset_parameters(self) -> None:
        pass

    def forward(self, input: torch.Tensor) -> torch.Tensor:
        w_times_x = torch.bmm(input, self.weight)
        return torch.add(w_times_x, self.bias[:, None, :])  # w times x + b

    def extra_repr(self) -> str:
        return 'in_features={}, out_features={}, bias={}'.format(
            self.in_features, self.out_features, self.bias is not None
        )