matmul

fun matmul(input: Tensor<FP32, Float>, ternaryWeights: TernaryTensorData, ctx: ExecutionContext): Tensor<FP32, Float>(source)

Perform matrix multiplication with ternary weights.

Return

FP32 output tensor of shape batch, outputDim or outputDim

Parameters

input

FP32 input tensor of shape batch, inputDim or inputDim

ternaryWeights

Ternary weight tensor of shape inputDim, outputDim

ctx

ExecutionContext for creating the output tensor