ZhennanQin commented on a change in pull request #16025: Numpy add numpy op
left_shift and right_shift
URL: https://github.com/apache/incubator-mxnet/pull/16025#discussion_r326488169
##########
File path: src/operator/numpy/np_elemwise_broadcast_op.cc
##########
@@ -144,5 +165,21 @@
MXNET_OPERATOR_REGISTER_NP_BINARY_SCALAR(_backward_npi_rcopysign_scalar)
.set_attr<FCompute>("FCompute<cpu>",
BinaryScalarOp::Backward<cpu, mshadow_op::rcopysign_grad>);
+MXNET_OPERATOR_REGISTER_NP_BINARY_SCALAR(_npi_left_shift_scalar)
+.set_attr<FCompute>("FCompute<cpu>", BitScalarCompute<cpu,
mshadow_op::left_shift>)
+.set_attr<nnvm::FGradient>("FGradient", MakeZeroGradNodes);
Review comment:
Why the gradient is zero? as operator described, this operation is
equivalent to multiplying x1 by 2**x2. So it should have same gradient as
multiplying by 2.
----------------------------------------------------------------
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:
[email protected]
With regards,
Apache Git Services
