Get hands-on experience with 20+ free Google Cloud products and $300 in free credit for new customers.

Mount gcsfuse in gcloud ai custom-jobs local-run

When locally testing my custom-job through "gcloud ai custom-jobs local-run" command, I would like to have access to a bucket mounted though gcsFuse as it happens when I launch the same containerized job from GCloud console. Is there the option to have the same access locally?

Thank you for helping

Solved Solved
0 1 792
1 ACCEPTED SOLUTION

What you could do is use cloud storage as a file system within ai training, since while using fuse your training jobs on both of the platforms can access your data that is stored on Cloud Storage as files on your local file system, also the documentation I shared provides you useful information as the problems you might encounter, permissions, a brief description of cloud storage fuse, performance related information, the restrictions this method has and also how you can make use of the logs.

View solution in original post

1 REPLY 1

What you could do is use cloud storage as a file system within ai training, since while using fuse your training jobs on both of the platforms can access your data that is stored on Cloud Storage as files on your local file system, also the documentation I shared provides you useful information as the problems you might encounter, permissions, a brief description of cloud storage fuse, performance related information, the restrictions this method has and also how you can make use of the logs.