Skip to main content

Conv

Conv(in_channels, out_channels, kernel, *, stride=None, padding=None, dilation=None, groups=1, bias=True, padding_mode="zeros", value=None, activation=None, dtype=None, device=None)

A grouped, padded, dilated convolution of any rank over channels-last
input; ``kernel`` fixes the rank.

Args:
    in_channels: channels of the input.
    out_channels: channels of the output.
    kernel: the kernel extents, one per spatial dimension.
    stride: per-dimension strides; ``None`` means 1.
    padding: an int or ``n`` ints (per-dimension symmetric), or ``2n``
        ints as interleaved ``(low, high)`` pairs, the runtime's own
        per-side form; ``None`` means 0.
    dilation: per-dimension dilations; ``None`` means 1.
    groups: input-to-output channel groups.
    bias: declare the bias slot.
    padding_mode: how the input is padded (``"zeros"``, ``"reflect"``,
        ``"replicate"``, ``"circular"``, or a
        :class:`~clika_runtime.PadMode` value).
    value: the constant a ``"zeros"`` / ``"constant"`` pad fills with
        (``None`` means 0).
    activation: a fused epilogue applied to every output (a name such
        as ``"relu"`` or an :class:`~clika_runtime.Activation` value);
        ``None`` for none.
    dtype: the weight dtype the layer declares; ``None`` declares the
        default dtype (:func:`~clika_runtime.get_default_dtype`). A
        ``load_state_dict`` casts the checkpoint to the declared dtype;
        ``assign=True`` adopts the checkpoint's own dtype instead.
    device: where the layer's tensors live; ``None`` declares on the
        cpu.

__init__​

__init__(self, in_channels: 'int', out_channels: 'int', kernel: 'Sequence[int]', *, stride: 'Sequence[int] | None' = None, padding: 'Sequence[int] | None' = None, dilation: 'Sequence[int] | None' = None, groups: 'int' = 1, bias: 'bool' = True, padding_mode: 'PaddingModeLike' = 'zeros', value: 'float | None' = None, activation: 'ActivationLike' = None, dtype: 'DtypeLike | None' = None, device: 'Device | str | None' = None) -> 'None'

Initialize self. See help(type(self)) for accurate signature.

extra_repr​

extra_repr(self) -> 'str'

extra_repr() -> str

One line of per-class detail for :meth:__repr__; a layer prints its geometry here (in_features=64, out_features=256).

forward​

forward(self, input: 'Tensor') -> 'Tensor'

forward(input) -> Tensor

Channels-last [batch, *spatial, in_channels] to [batch, *spatial', out_channels].

set_weights​

set_weights(self, weight: 'Tensor', bias: 'Tensor | None' = None) -> 'None'

set_weights(weight, bias=None) -> None

Bind the declared slots directly (weight in the [out_channels, *kernel, in_channels // groups] layout); raises RuntimeError on a shape mismatch.