jinhuang415 commented on a change in pull request #10433: [MXNET-290] MKLDNN
support for model quantization
URL: https://github.com/apache/incubator-mxnet/pull/10433#discussion_r191751428
##########
File path: ci/docker/runtime_functions.sh
##########
@@ -382,7 +381,6 @@ unittest_ubuntu_python3_cpu() {
#export MXNET_MKLDNN_DEBUG=1 # Ignored if not present
export MXNET_STORAGE_FALLBACK_LOG_VERBOSE=0
nosetests-3.4 --verbose tests/python/unittest
- nosetests-3.4 --verbose tests/python/quantization
Review comment:
@zhang-da @reminisce The difficulty is we need a way to separate the naive
CPU quantization and MKLDNN quantization test if we want to share the same
test_quantization.py (it's our aim to share code as @reminisce mentioned) since
they are both use CPU context, I figured out a way to set a environment
variable USE_MKLDNN in test_quantization_mkldnn.py and check this in
test_quantization.py, for naive CPU path USE_MKLDNN will not be set so will go
to naive path. We have added quantization test for CPU path back in new diff.
Please help to review and check if there is any better way to do this. Thanks.
----------------------------------------------------------------
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
