conv

Contents

conv#

class braintrace.conv(x, kernel, bias=None, *, strides=(1,), padding='SAME', lhs_dilation=None, rhs_dilation=None, feature_group_count=1, batch_group_count=1, dimension_numbers=None)[source]#

ETP-aware convolution.

Computes \(y = \mathrm{conv}(x, kernel) \; (+ b)\). Always expects a batch dimension on x.

Parameters:

  • x – Input tensor with batch dimension.

  • kernel – Convolution kernel.

  • bias – Optional bias.

  • strides (Sequence[int]) – Window strides.

  • padding (str) – Padding mode.

  • lhs_dilation (Sequence[int] | None) – Left-hand-side dilation.

  • rhs_dilation (Sequence[int] | None) – Right-hand-side dilation.

  • feature_group_count (int) – Feature group count.

  • batch_group_count (int) – Batch group count.

  • dimension_numbers (Any) – Convolution dimension numbers.

Returns:

Convolution output.

Contents