- conv2d<T>(__namedParameters): T
Parameters
__namedParameters: {
activation?: Activation;
bias?: Tensor<Rank> | TensorLike;
dataFormat?: "NHWC" | "NCHW";
dilations?: number | [number, number];
dimRoundingMode?: "floor" | "round" | "ceil";
filter: TensorLike | Tensor4D;
leakyreluAlpha?: number;
pad: number | "same" | "valid" | ExplicitPadding;
preluActivationWeights?: Tensor<Rank>;
strides: number | [number, number];
x: TensorLike | T;
}
Optional
activation?: Activation
Optional
dataFormat?: "NHWC" | "NCHW"
Optional
dilations?: number | [number, number]
Optional
dimRoundingMode?: "floor" | "round" | "ceil"
Optional
leakyreluAlpha?: number
pad: number | "same" | "valid" | ExplicitPadding
Optional
preluActivationWeights?: Tensor<Rank>
strides: number | [number, number]
Returns T
Computes a 2D convolution over the input x, optionally fused with adding a bias and applying an activation.