[GitHub] reminisce commented on a change in pull request #12230: Update Operator Implementation Tutorial

[GitBox <gi...@apache.org>]

reminisce commented on a change in pull request #12230: Update Operator Implementation Tutorial
URL: https://github.com/apache/incubator-mxnet/pull/12230#discussion_r212173680
 
 

 ##########
 File path: docs/faq/add_op_in_backend.md
 ##########
 @@ -543,22 +545,21 @@ If you use python, when you type `import mxnet as mx`, two python
 functions for invoking your backend implementation are
 generated on the fly: one is for imperative programming
 registered as `mxnet.ndarray.quadratic` or `mx.nd.quadratic` for short;
-the other one is for symbolic
-programming registered under module `mxnet.symbol.quadratic`
-or `mx.sym.quadratic` for short.
+the other one is for symbolic programming registered under
+module `mxnet.symbol.quadratic` or `mx.sym.quadratic` for short.
 
 In order to unit test it in frontend, we need to add the following code
-to the python file `test_operator.py`. Note that while testing the
-forward pass is straightforward using `mx.nd.quadratic`, testing
-the backward involves a bit of more efforts. We create a
-`quadratic` symbol and feed it into the utility function `check_numeric_gradient`.
-The utility function will perform a perturbation on the input
-and calculate the response rate of the output using the
-[finite difference method](https://en.wikipedia.org/wiki/Finite_difference_method).
-Then it will compare the gradient from the backward pass with
-the values from the finite difference method. The test
-will be successful once the comparison satisfies user specified
-relative and absolute thresholds.
+to the python file `test_operator.py`. A typical operator implementation
+tests for both the `symbol` API and the `ndarray` API. In this tutorial 
+we will go over the symbol API test. The `symbol` API test, tests for the
+complete functionality of the operator - the forward pass, the backward
+pass and the gradient computation in the operator implementation. To
 
 Review comment:
   Backward pass and gradient computation mean the same thing.

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services