-
Notifications
You must be signed in to change notification settings - Fork 2.4k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
PyTorch job prototype should contain the full job spec #1114
Labels
Comments
jlewi
added a commit
to jlewi/kubeflow
that referenced
this issue
Jul 3, 2018
* The current pattern in Kubeflow is to put the complete prototype in the jsonnet file. This way the result of ks generate is a .jsonnet file containing a full spec. This makes it easy for users to do complex modifications starting with the prototype as an example. * Configure the prototype to do mnist by default. Fix kubeflow#1114
k8s-ci-robot
pushed a commit
that referenced
this issue
Jul 5, 2018
* The current pattern in Kubeflow is to put the complete prototype in the jsonnet file. This way the result of ks generate is a .jsonnet file containing a full spec. This makes it easy for users to do complex modifications starting with the prototype as an example. * Configure the prototype to do mnist by default. Fix #1114
jlewi
added a commit
to jlewi/kubeflow
that referenced
this issue
Jul 11, 2018
* The current pattern in Kubeflow is to put the complete prototype in the jsonnet file. This way the result of ks generate is a .jsonnet file containing a full spec. This makes it easy for users to do complex modifications starting with the prototype as an example. * Configure the prototype to do mnist by default. Fix kubeflow#1114
k8s-ci-robot
pushed a commit
that referenced
this issue
Jul 12, 2018
* The current pattern in Kubeflow is to put the complete prototype in the jsonnet file. This way the result of ks generate is a .jsonnet file containing a full spec. This makes it easy for users to do complex modifications starting with the prototype as an example. * Configure the prototype to do mnist by default. Fix #1114
saffaalvi
pushed a commit
to StatCan/kubeflow
that referenced
this issue
Feb 11, 2021
* The current pattern in Kubeflow is to put the complete prototype in the jsonnet file. This way the result of ks generate is a .jsonnet file containing a full spec. This makes it easy for users to do complex modifications starting with the prototype as an example. * Configure the prototype to do mnist by default. Fix kubeflow#1114
yanniszark
pushed a commit
to arrikto/kubeflow
that referenced
this issue
Feb 15, 2021
* Close menu on click Fix close time for snackbar Add Material UI to the README Add link to developer guide * Change README
surajkota
pushed a commit
to surajkota/kubeflow
that referenced
this issue
Jun 13, 2022
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
The PyTorch job prototype
https://github.com/kubeflow/kubeflow/blob/master/kubeflow/pytorch-job/prototypes/pytorch-job.jsonnet
Currently defines the spec in a .libsonnet file which is imported.
We want to move the complete spec into the jsonnet file. This way when a user generates a component they have complete access to the spec and can easily modify it.
This is the pattern we are now following with other spec e.g.
https://github.com/kubeflow/kubeflow/blob/master/kubeflow/examples/prototypes/tf-job-simple.jsonnet
The text was updated successfully, but these errors were encountered: