Skip to content
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

Is it possible to write custom dataloader and embed it in the architecture of the QuartzNet model for training? #4313

Closed
psydok opened this issue Jun 2, 2022 · 1 comment
Assignees
Labels
feature request/PR for a new feature

Comments

@psydok
Copy link

psydok commented Jun 2, 2022

Good evening! I have 2 questions...

  1. Is it possible to write your own dataloader and embed it in the architecture of the QuartzNet model for training?
    It seems to me that I saw somewhere how the default dataloader is replaced with a custom dataloader. But I can't find an example of how to do it.
  2. Is it possible to somehow see which dataloader is used from the loaded model in the code?

Many thanks in advance for any insights!

@psydok psydok added the feature request/PR for a new feature label Jun 2, 2022
@titu1994
Copy link
Collaborator

titu1994 commented Jun 2, 2022

You can potentially create your own dataloader and pass it to trainer.fit(). See the lightning documention for such use case, but note that we have never tried it.

You can access model._train_dl, _validation_dl and _test_dl I think for most Nemo models.

@titu1994 titu1994 closed this as completed Jun 4, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
feature request/PR for a new feature
Projects
None yet
Development

No branches or pull requests

3 participants