FilesLast commitLast update
4 months ago
4 months ago
4 months ago
19 days ago
4 days ago
2 months ago
16 days ago
4 days ago
1 month ago
18 days ago
README

ReluGrad

English | 简体中文

Applicable Products

Product Support
Ascend 950PR&950DT products √
Atlas A3 products √
Atlas A2 products √
Atlas 200I/500 A2 inference products ×
Atlas inference products √
Atlas training products √

Function

  • Operator function: the backpropagation gradient corresponding to the Relu operation.

  • Formula:

    gradInput=gradOutput∗{1,x>00,x≤0 gradInput = gradOutput * \begin{cases} 1, \quad x > 0\\ 0, \quad x \leq 0 \end{cases}

Parameters

Parameter name Input/Output/Attribute Description Data type Data format
gradients Input The backpropagation gradient passed to the corresponding Relu operation INT8, INT32, UINT8, INT64, FLOAT, FLOAT16, BFLOAT16 ND
features Input The features passed as input to the corresponding Relu operation INT8, INT32, UINT8, INT64, FLOAT, FLOAT16, BFLOAT16 ND
backprops Output The output tensor in the formula INT8, INT32, UINT8, INT64, FLOAT, FLOAT16, BFLOAT16 ND

Constraints

None

Invocation Description

Invocation method Sample code Description
Graph mode test_geir_relu_backward.cpp Calls the ReluGrad operator by building the graph through the operator IR.