aihwkit.simulator.tiles_indexed module

High level analog tiles (indexed).

class aihwkit.simulator.tiles_indexed.CudaIndexedAnalogTile(out_size, in_size, resistive_device=None, bias=False, in_trans=False, out_trans=False)

Bases: aihwkit.simulator.tiles_indexed.IndexedAnalogTile

Analog tile (CUDA, indexed).

cuda(device=None)

Return a copy of this tile in CUDA memory.

Parameters

device (Optional[Union[torch.device, str, int]]) – CUDA device

Return type

aihwkit.simulator.tiles_indexed.CudaIndexedAnalogTile

is_cuda = True

class aihwkit.simulator.tiles_indexed.CudaIndexedFloatingPointTile(out_size, in_size, resistive_device=None, bias=False, in_trans=False, out_trans=False)

Bases: aihwkit.simulator.tiles_indexed.IndexedFloatingPointTile

Floating point tile (CUDA, indexed).

cuda(device=None)

Return a copy of this tile in CUDA memory.

Parameters

device (Optional[Union[torch.device, str, int]]) – CUDA device

Return type

aihwkit.simulator.tiles_indexed.CudaIndexedFloatingPointTile

is_cuda = True

class aihwkit.simulator.tiles_indexed.IndexedAnalogTile(out_size, in_size, resistive_device=None, bias=False, in_trans=False, out_trans=False)

Bases: aihwkit.simulator.tiles.AnalogTile

Analog tile (indexed).

backward(d_input)

Perform the backward pass.

Parameters

d_input (torch.Tensor) – [N, out_size] tensor. If out_trans is set, transposed.

Returns

[N, in_size] tensor. If in_trans is set, transposed.

Return type

torch.Tensor

cuda(device=None)

Return a copy of this tile in CUDA memory.

Parameters

device (Optional[Union[torch.device, str, int]]) – CUDA device

Return type

aihwkit.simulator.tiles_indexed.CudaIndexedAnalogTile

forward(x_input, is_test=False)

Perform the forward pass.

Parameters

• x_input (torch.Tensor) – [N, in_size] tensor. If in_trans is set, transposed.

• is_test (bool) – whether to assume testing mode.

Returns

[N, out_size] tensor. If out_trans is set, transposed.

Return type

torch.Tensor

set_indexed(indices, image_sizes)

Sets the index matrix for convolutions ans switches to indexed forward/backward/update versions.

Parameters

• indices (torch.Tensor) – torch.tensor with int indices

• image_sizes (list) – [C_in, H_in, W_in, H_out, W_out] sizes

Return type

None

update(x_input, d_input)

Perform the update pass.

Parameters

• x_input (torch.Tensor) – [N, in_size] tensor. If in_trans is set, transposed.

• d_input (torch.Tensor) – [N, out_size] tensor. If out_trans is set, transposed.

Return type

None

class aihwkit.simulator.tiles_indexed.IndexedFloatingPointTile(out_size, in_size, resistive_device=None, bias=False, in_trans=False, out_trans=False)

Bases: aihwkit.simulator.tiles.FloatingPointTile

Floating point tile (indexed).

backward(d_input)

Perform the backward pass.

Parameters

d_input (torch.Tensor) – [N, out_size] tensor. If out_trans is set, transposed.

Returns

[N, in_size] tensor. If in_trans is set, transposed.

Return type

torch.Tensor

cuda(device=None)

Return a copy of this tile in CUDA memory.

Parameters

device (Optional[Union[torch.device, str, int]]) – CUDA device

Return type

aihwkit.simulator.tiles_indexed.CudaIndexedFloatingPointTile

forward(x_input, is_test=False)

Perform the forward pass.

Parameters

• x_input (torch.Tensor) – [N, in_size] tensor. If in_trans is set, transposed.

• is_test (bool) – whether to assume testing mode.

Returns

[N, out_size] tensor. If out_trans is set, transposed.

Return type

torch.Tensor

set_indexed(indices, image_sizes)

Sets the index matrix for convolutions ans switches to indexed forward/backward/update versions.

Parameters

• indices (torch.Tensor) – torch.tensor with int indices

• image_sizes (list) – [C_in, H_in, W_in, H_out, W_out] sizes

Return type

None

update(x_input, d_input)

Perform the update pass.

Parameters

• x_input (torch.Tensor) – [N, in_size] tensor. If in_trans is set, transposed.

• d_input (torch.Tensor) – [N, out_size] tensor. If out_trans is set, transposed.

Return type

None