Liu Hui created SINGA-435:
-----------------------------
Summary: Rafiki--Can't create a train job with 'ENABLE_GPU'
Key: SINGA-435
URL: https://issues.apache.org/jira/browse/SINGA-435
Project: Singa
Issue Type: Bug
Reporter: Liu Hui
Attachments: rafiki_admin001.png
>>https://nginyc.github.io/rafiki/docs/latest/docs/src/user/quickstart.html
I followed the quickstart and tried to create a train job with using GPU。
So I changed parameters to "budget=\{'ENABLE_GPU':1, 'MODEL_TRIAL_COUNT': 2 }"
.when I create a train job.
But the container of rafiki_worker didn't start.
I entered the container of rafiki_admin, and found an error in log file.
Finally I found that, in rafiki/rafiki/container/docker_swarm.py, the function
of _if_any_node_has_gpu always return False.
I doubt that what should I do to do a training with GPU in container. Which
steps have I missed, setting up docker's environment or others?
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
