ConstToAttrReduceSumFusion

Description

Changes the ReduceSum operator that fits the graph fusion pattern to the ReduceSumD operator in static scenarios.

Restrictions

• The input data type can only be float16, float32, or bfloat16.
• The fusion pattern does not take effect when the axis is an empty tensor.
• The fusion pattern does not take effect in dynamic scenarios.
• This fusion pattern is enabled by default and cannot be disabled.

Availability

Atlas A2 training product/Atlas A2 inference product