FusedGemm

 Graphs in : cluster, ONNX model architecture.

 A (heterogeneous) – T : object, input tensor A. The shape of A should be (M, K) if transA is 0, or (K, M) if transA is non-zero.
 B (heterogeneous) – T : object, input tensor B. The shape of B should be (K, N) if transB is 0, or (N, K) if transB is non-zero.
 C (optional, heterogeneous) – T : object, input tensor C. The shape of C should be unidirectional broadcastable to (M, N).

 Parameters : cluster,

activation : enum, activation function.
Default value “Relu”.
activation_alpha : float, scalar forwarded to the fused activation; meaning depends on the selected activation (e.g., LeakyRelu: negative slope α; Elu: α; HardSigmoid: α). Ignored if not applicable.
Default value “0”.
activation_beta : float, scalar forwarded to the fused activation for activations that define a second parameter (e.g., HardSigmoid: β). Ignored if not applicable.
Default value “0”.
activation_gamma : float, extra scalar for activations that require a third parameter (provider-specific/scaled variants). Ignored if not applicable.
Default value “0”.
alpha : float, scalar multiplier for the product of input tensors A * B.
Default value “1”.
beta : float, scalar multiplier for input tensor C.
Default value “1”.
 transA : boolean, whether A should be transposed.
Default value “False”.
transB : boolean, whether B should be transposed.
Default value “False”.
 training? : boolean, whether the layer is in training mode (can store data for backward).
Default value “True”.
 lda coeff : float, defines the coefficient by which the loss derivative will be multiplied before being sent to the previous layer (since during the backward run we go backwards).
Default value “1”.

 name (optional) : string, name of the node.