anko-intel commented on a change in pull request #20719:
URL: https://github.com/apache/incubator-mxnet/pull/20719#discussion_r752321675
##########
File path: src/operator/nn/dnnl/dnnl_base-inl.h
##########
@@ -258,7 +258,7 @@ static inline int get_mxnet_type(dnnl_data_type_t dtype) {
case dnnl::memory::data_type::u8:
return mshadow::kUint8;
default:
- LOG(FATAL) << "unknown DNNL type";
+ LOG(FATAL) << "unknown oneDNN type";
Review comment:
```suggestion
LOG(FATAL) << "unknown oneDNN data type";
```
##########
File path: tests/nightly/test_np_large_array.py
##########
@@ -2066,7 +2066,7 @@ def test_rnn_dim_check():
@use_np
[email protected](reason='runs without DNNL, wtih is not default behavior')
[email protected](reason='runs without oneDNN, with is not default behavior')
Review comment:
```suggestion
@pytest.mark.skip(reason='runs without oneDNN, which is not default
behavior')
```
I am not sure what doesn't mean. This test pass when run in mxnet-native
version?
##########
File path: src/operator/nn/dnnl/dnnl_fully_connected.cc
##########
@@ -65,7 +65,7 @@ dnnl::inner_product_forward::primitive_desc
GetFCFwdImpl(const DNNLFCFullParam&
return dnnl::inner_product_forward::primitive_desc(desc, attr, engine);
} catch (dnnl::error& e) {
if (e.status == dnnl_unimplemented && full_param.dnnl_param.quantized) {
- LOG(ERROR) << "AVX512-BW support or DNNL v0.18 is required for INT8
fully_connected.";
+ LOG(ERROR) << "AVX512-BW support or oneDNN v0.18 is required for INT8
fully_connected.";
Review comment:
```suggestion
LOG(ERROR) << "AVX512-BW support or oneDNN v0.18 or later is
required for INT8 fully_connected.";
```
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
