benchmarks/experimental/experimental_async_approaches.py [98:110]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
        if self.src_mask is None or self.src_mask.size(0) != len(src):
            device = src.device
            mask = self._generate_square_subsequent_mask(len(src)).to(device)
            self.src_mask = mask

        return super().forward(src, self.src_mask)


class LinearLayer(nn.Linear):
    def __init__(self, ninp, ntoken, initrange):
        super().__init__(ninp, ntoken)
        self.bias.data.zero_()
        self.weight.data.uniform_(-initrange, initrange)
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



benchmarks/models/transformer_lm.py [178:192]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
        if self.src_mask is None or self.src_mask.size(0) != len(src):
            device = src.device
            mask = self._generate_square_subsequent_mask(len(src)).to(device)
            self.src_mask = mask

        return super().forward(src, self.src_mask)


class LinearLayer(nn.Linear):
    """Wrapped nn.Linear layer to allow for weight initialization."""

    def __init__(self, ninp, ntoken, initrange):
        super().__init__(ninp, ntoken)
        self.bias.data.zero_()
        self.weight.data.uniform_(-initrange, initrange)
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -