[GitHub] [incubator-mxnet] sxjscience commented on a change in pull request #19044: Fix the error of gradient of np.pad

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

sxjscience commented on a change in pull request #19044:
URL: https://github.com/apache/incubator-mxnet/pull/19044#discussion_r479817156



##########
File path: src/operator/numpy/np_pad_op-inl.h
##########
@@ -792,15 +825,50 @@ void NumpyPadOpBackward(const nnvm::NodeAttrs& attrs,
                         const std::vector<TBlob>& inputs,
                         const std::vector<OpReqType>& req,
                         const std::vector<TBlob>& outputs) {
-  using namespace mxnet_op;
-  using namespace mshadow;
-  CHECK_EQ(inputs.size(), 1U);
-  CHECK_EQ(outputs.size(), 1U);
-  Stream<xpu> *s = ctx.get_stream<xpu>();
-  const TBlob& in_data = inputs[0];
-  const TBlob& out_data = outputs[0];
-  NumpyPadOpBackImpl<xpu>(in_data, out_data,
-                          out_data.Size(), req, s);
+  MXNET_NDIM_SWITCH(inputs[0].ndim(), NDim, {
+    using namespace mxnet_op;
+    using namespace mshadow;
+    CHECK_EQ(inputs.size(), 1U);
+    CHECK_EQ(outputs.size(), 1U);
+    CHECK_EQ(req.size(), 1U);
+    CHECK_EQ(req[0], kWriteTo);

Review comment:
       You will need to also implement kAddTo.




----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to 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