Dvc.org: use-cases: training on a remote machine (SSH)

[ghost]

for example, ssh to a remote machine, git clone / pull to get your code, thendvc pull to download your dataset, and dvc repro or dvc run to start training your models

[dnabanita7]

@mroutis An almost similar example is https://dvc.org/doc/use-cases/data-and-model-files-versioning and https://dvc.org/doc/use-cases/share-data-and-model-files and versioning in getting-started explains it.
I am sorry Can you explain this a little bit more?

[shcheklein]

@Naba7 the point here is in connecting your local environment with a remote one. DVC helps running an expensive training remotely, while you can keep doing some stuff locally. DVC commands in this scenario help you moving pulling/pushing data across different machines.

[dnabanita7]

Can I take this up? I am not sure if I can do it,but I'll give it a try.

[shcheklein]

@Naba7 sure! You can take our NLP tutorial or get-started as a base example, and show step by step how to run it remotely via SSH or on S3 EC2 using tools like spotty.

[dashohoxha]

Maybe these two interactive examples are related to this:

• https://katacoda.com/dvc/courses/examples/shared-server
• https://katacoda.com/dvc/courses/examples/ssh-storage

@mroutis what do you think?
Can we close this issue?

[shcheklein]

@dashohoxha I don't see how are they related, tbh. Also, I'm not sure I understood what https://katacoda.com/dvc/courses/examples/ssh-storage is about from reading the description - and it lacks any additional information inside that can help to "feel" the motivation behind it.

[shcheklein]

This is not a use case actually by the way it's written. Title should be changed to make it a use case, or it can be part of the user guide.

[shcheklein]

E.g. use case can be something like: "Training on a remote machine", "Organizing CI system for ML", etc

[shcheklein]

Not sure what @mroutis had initially in mind.

[ghost]

Sorry for the late response!

@shcheklein , changing the title