Let’s look at an example of how you can use Polyaxon for your deep learning experiments. This example assumes that both Polyaxon is installed and running. And you are logged in to your Polyaxon account through the polyaxon-cli
Check that your cli is configured and you are logged in
The first step is to check that you are logged in and that you are running a version of CLI compatible with the platform.
polyaxon version --cli
polyaxon version --platform
polyaxon whoami
Check the cluster discovery
Check that polyaxon is reporting your cluster correctly
$ polyaxon cluster
Cluster info:
-------------- ----------------------------------------
build_date 2017-11-20T05:17:43Z
major 1
go_version go1.8.3
git_version v1.8.4
platform linux/amd64
git_commit 9befc2b8928a9426501d3bf62f72849d5cbcd5a3
git_tree_state clean
minor 8
compiler gc
-------------- ----------------------------------------
Cluster Nodes:
id name hostname role memory n_cpus n_gpus
---- ------------------------- ------------------------- ------ -------- -------- --------
1 k8s-agentpool1-13475325-0 k8s-agentpool1-13475325-0 agent 6.7 Gb 2 0
2 k8s-agentpool2-13475325-0 k8s-agentpool2-13475325-0 agent 54.93 Gb 6 1
3 k8s-master-13475325-0 k8s-master-13475325-0 master 6.7 Gb 2 0
Create a project
Now we can create a project, you can do that with Polyaxon Dashboard
or with polyaxon CLI
$ polyaxon project create --name=quick-start --description='Polyaxon quick start.'
Prepare your code
Polyaxon offers 2 options for tracking code, an in-cluster git server or tracking code on external platforms (Github, GitLab, Bitbucket, ...).
A project can only be linked to one repo (either in-cluster or on Github for example), in both cases, your workflow will be more or less similar.
We have created 2 quick-start guides to show how you can use each one of this options: