You can use one or multiple buckets on Ceph to access data directly on your machine learning experiments and jobs.

Deploy Ceph

Before using Ceph, you need to deploy it and create a bucket (or several) to host your data.

Create a bucket on Ceph

You should create a bucket (e.g. plx-storage) where you will host your data.

Use the bucket in Polyaxon

In order to use the buckets with Polyaxon, you can follow the S3 DataStore integration.

Please note that you will need to set as well AWS_ENDPOINT_URL or S3_ENDPOINT_URL.