def forward(self, x_input: Tensor) -> Tensor:
"""Compute the forward pass."""
# pylint: disable=arguments-differ,arguments-renamed
if self.analog_tile_count() == 1:
out = AnalogFunction.apply(
self.analog_tile_array[0][0].get_analog_ctx(), x_input,
self.analog_tile_array[0][0].shared_weights, not self.training)
if self.weight_scaling_omega:
out = out * self.analog_tile_array[0][0].out_scaling_alpha
if self.digital_bias:
return out + self.bias
return out
# mapped version
last_dim = x_input.ndim - 1
splits = split(x_input, self.in_sizes, dim=last_dim)
result = None # type: Tensor
for idx, (x, in_tiles) in enumerate(zip(splits, self.analog_tile_array)):
out_result = []
for analog_tile in in_tiles:
output = AnalogFunction.apply(
analog_tile.get_analog_ctx(), x,
analog_tile.shared_weights, not self.training)
if self.weight_scaling_omega:
output = output * analog_tile.out_scaling_alpha
out_result.append(output)
if idx == 0:
result = cat(out_result, last_dim)
else:
result.add_(cat(out_result, last_dim))
# add bias to final result
if self.digital_bias:
return result.add_(self.bias)
return result
def forward(self, x_input: Tensor) -> Tensor:
"""Compute the forward pass."""
# pylint: disable=arguments-differ, arguments-renamed
out = AnalogFunction.apply(
self.analog_tile.get_analog_ctx(), x_input,
self.analog_tile.shared_weights, not self.training)
out = self.analog_tile.apply_out_scaling(out, (-1, ))
if self.digital_bias:
return out + self.bias
return out
def forward(self, x_input: Tensor) -> Tensor:
"""Computes the forward pass."""
# pylint: disable=arguments-differ
def get_size(size: int, i: int) -> int:
"""Calculate the output image sizes"""
nom = (size + 2 * self.padding[i] - self.dilation[i] *
(self.kernel_size[i] - 1) - 1)
return nom // self.stride[i] + 1
input_size = x_input.numel() / x_input.size(0)
if not self.fold_indices.numel() or self.input_size != input_size:
# pytorch just always uses NCHW order?
fold_indices = arange(2, input_size + 2, dtype=float64).detach()
shape = [1] + list(x_input.shape[1:])
fold_indices = fold_indices.reshape(*shape)
unfold = Unfold(kernel_size=self.kernel_size,
stride=self.stride,
padding=self.padding,
dilation=self.dilation)
fold_indices = unfold(fold_indices).flatten().round().to(
dtype=int32)
if self.use_bias:
out_image_size = fold_indices.numel() // self.out_channels
fold_indices = cat(
(fold_indices, ones(out_image_size, dtype=int32)), 0)
self.fold_indices = fold_indices.to(x_input.device)
x_height = x_input.size(2)
x_width = x_input.size(3)
d_height = get_size(x_height, 0)
d_width = get_size(x_width, 1)
image_sizes = (self.in_channels, x_height, x_width, d_height,
d_width)
self.input_size = input_size
self.analog_tile.set_indexed(self.fold_indices,
image_sizes) # type: ignore
return AnalogFunction.apply(self.analog_tile, x_input, self.weight,
self.bias)
def forward(self, x_input: Tensor) -> Tensor:
"""Compute the forward pass."""
# pylint: disable=arguments-differ,arguments-renamed
return AnalogFunction.apply(self.analog_tile.get_analog_ctx(), x_input,
self.analog_tile.shared_weights,
not self.training)
def forward(self, x_input: Tensor) -> Tensor:
"""Computes the forward pass."""
# pylint: disable=arguments-differ
return AnalogFunction.apply(self.analog_tile, x_input, self.weight,
self.bias, not self.training)