.. _l-onnx-doccom.microsoft-SoftmaxDropoutGrad:

==================================
com.microsoft - SoftmaxDropoutGrad
==================================

.. contents::
    :local:


.. _l-onnx-opcom-microsoft-softmaxdropoutgrad-1:

SoftmaxDropoutGrad - 1 (com.microsoft)
======================================

**Version**

* **name**: `SoftmaxDropoutGrad (GitHub) <https://github.com/onnx/onnx/blob/main/docs/Operators.md#com.microsoft.SoftmaxDropoutGrad>`_
* **domain**: **com.microsoft**
* **since_version**: **1**
* **function**:
* **support_level**:
* **shape inference**:

This version of the operator has been available
**since version 1 of domain com.microsoft**.

**Summary**

Gradient of BiasSoftmaxDropout Op.

**Attributes**

* **axis** (required):
  apply softmax to elements for dimensions axis or higher Default value is ``?``.

**Inputs**

Between 3 and 4 inputs.

* **dy** (heterogeneous) - **T**:
  The gradient tensor from output.
* **mask** (heterogeneous) - **tensor(bool)**:
  The mask output of the dropout.
* **softmax_y** (heterogeneous) - **T**:
  The output of Softmax.
* **ratio** (optional, heterogeneous) - **T1**:
  Same value as the ratio input supplied to the dropout op with value
  in [0, 1). If this input is not specified, a default value of 0.5 is
  used.

**Outputs**

* **dx** (heterogeneous) - **T**:
  Gradient of the input.

**Examples**