lihaofd commented on a change in pull request #14713: MKLDNN RNN Inference
Integration(fp32 LSTM and vRNN with tanh and relu)
URL: https://github.com/apache/incubator-mxnet/pull/14713#discussion_r277275799
##########
File path: src/operator/rnn.cc
##########
@@ -138,6 +138,21 @@ static bool RNNType(const nnvm::NodeAttrs& attrs,
return true;
}
+inline static bool RNNStorageType(const nnvm::NodeAttrs& attrs,
+ const int dev_mask,
+ DispatchMode* dispatch_mode,
+ std::vector<int> *in_attrs,
+ std::vector<int> *out_attrs) {
+ DispatchMode wanted_mode = DispatchMode::kFCompute;
+
+ #if MXNET_USE_MKLDNN_RNN
Review comment:
If using USE_MKLDNN here, several GPU case in ci will fail
----------------------------------------------------------------
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
