AscendQuantV2ScatterFusionPass

Description

Fuses the structure that meets the following pattern into the QuantUpdateScatter operator.

Restrictions

• There is only one AscendQuantV2 output, which is provided to the Scatter node. The output is of the int8 data type.
• For AscendQuantV2, sqrt_mode is set to false, round_mode is set to round, and reduce of the Scatter node is set to update.
• When the third input of AscendQuantV2 exists, the data type must be bfloat16.
• For AscendQuantV2, input0 must be of the float16 or bfloat16 type, and input1 must be of the bfloat16 or float32 type.
• The number of input0 dimensions of AscendQuantV2 must be the same as the number of var dimensions of Scatter. The size of axis must be less than or equal to that of axis of var. The sizes of other axes must be the same, the size of the first axis of input0 of AscendQuantV2 is equal to the size of the first axis of indices of Scatter.
• The total number of elements in input1 of AscendQuantV2 is equal to the size of the last dimension of input0. If input2 exists, the total number of elements is equal to the size of the last dimension of input0.
• The indices dimensions of Scatter can only be 1 or 2. If the dimension is 2, the size of the second dimension must be 2.
• The axis attribute of Scatter is -2.

Availability

Atlas A2 training product/Atlas A2 inference product