MNIST training example not exiting on CPU instances after training completion

[sampathchanda]

Hi,

Am running MNIST training on 2 CPU instances using MXNET with 10 epochs (Default script, provided). However, even after the training is all done ( I could see that the execution reaches the end of train_mnist.py script), execution isn't exiting.

Can anyone help me with this issue.

Environment info

Operating System: Amazon Linux
Package used (Python/R/Scala/Julia): Python
MXNet Installed from sources
MXNet commit hash (git rev-parse HEAD): 266e439
Python version and distribution: Python 2.7.12

Steps to reproduce

1. cd $HOME/mxnet/examples/image_classification/
2. ../../tools/launch.py -n 1 -H hosts python train_mnist.py

hosts file consists of the worker hostnames (2 workers in my case, as follows):
deeplearning-worker1
deeplearning-worker2

[Soonhwan-Kwon]

If you launched it by ../../tools/launch.py, there is ../../tools/kill-mxnet.py to kill the process.

[sampathchanda]

@Soonhwan-Kwon Thanks for the comment. But it seems that there was some issue with the way i setup my environment for running the example. I fixed the issue, with help from @qiyuangong in issue #5094.

For those who are also facing this issue, please follow the steps mentioned by @qiyuangong in the issue #5094 to get the scripts working as expected. Am now able to successfully run the example given on multi CPU cluster, and hence closing this issue.