anirudh2290 commented on a change in pull request #10538: [MXNET-318] Allow dot
for fp16 on GPU
URL: https://github.com/apache/incubator-mxnet/pull/10538#discussion_r181274419
##########
File path: src/operator/tensor/dot-inl.h
##########
@@ -69,9 +69,10 @@ void DotForward_(const nnvm::NodeAttrs& attrs,
<< "Binary function only support input/output with the same type";
CHECK_EQ(outputs[0].type_flag_, inputs[1].type_flag_)
<< "Binary function only support input/output with the same type";
- CHECK(outputs[0].type_flag_ == kFloat32 || outputs[0].type_flag_ == kFloat64)
- << "dot only supports float32 and float64";
- MSHADOW_SGL_DBL_TYPE_SWITCH(outputs[0].type_flag_, DType, {
+ CHECK(outputs[0].type_flag_ == kFloat32 || outputs[0].type_flag_ == kFloat64
||
Review comment:
don't we need to do the same for BatchDotForward?
----------------------------------------------------------------
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:
us...@infra.apache.org
With regards,
Apache Git Services
