AConv2dMulFusion

Description

Fuses the Conv2d+mul or Conv3d+mul operators into one Conv operator.

Before: After:

Restrictions

• The conv node can be Conv2D or Conv3D.
• When the data node has dynamic inputs, fusion is supported.
• When the three inputs of filter, bias, and mul are all of const type, fusion is supported.
• For Conv3D nodes, fusion is supported only when the data input is in NDHWC format, the Mul operator's const mul input is 1D, and its size matches the C dimension of the Conv3D output.
• For Conv3D nodes, fusion is supported only when the data input is in NDHWC format, the Mul operator's const mul input is also in NDHWC format, the NDHW dimensions are all 1, and the C dimension matches the C dimension of the Conv3D output.

Availability

Atlas inference product

Atlas 200I/500 A2 inference product

Atlas training product

Atlas A2 training product/Atlas A2 inference product

Atlas A3 training product/Atlas A3 inference product