ThomasDelteil commented on a change in pull request #13099: [MXNET-1194] Reenable nightly tutorials tests for Python2 and Python3 URL: https://github.com/apache/incubator-mxnet/pull/13099#discussion_r230839624
########## File path: docs/tutorials/onnx/fine_tuning_gluon.md ########## @@ -122,7 +122,7 @@ We need to transform the images to a format accepted by the network EDGE = 224 SIZE = (EDGE, EDGE) BATCH_SIZE = 32 -NUM_WORKERS = multiprocessing.cpu_count() +NUM_WORKERS = max(multiprocessing.cpu_count()-6, 0) Review comment: @anirudhacharya 0 workers for data loading means synchronous data loading. I have noticed unstability when using too many workers in the tutorials in those tests. I am suspecting something to do with starting processes (workers) inside a process (kernel) inside a process (jupyter notebook) inside a process (python) inside a process (nosetests) inside a container (docker). Can't reproduce the instability outside of docker. @vishaalkapoor I will update to -3, 2 for multi-threaded engine and 1 for gpu copy thread. ---------------------------------------------------------------- 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: [email protected] With regards, Apache Git Services
