Removing numpy requirement from all files in examples/pytorch/domain_templates

[Bhavay-2001]

What does this PR do?

Part of #16649

Before submitting

• Was this discussed/agreed via a GitHub issue? (not for typos and docs)
• Did you read the contributor guideline, Pull Request section?
• Did you make sure your PR does only one thing, instead of bundling different changes together?
• Did you make sure to update the documentation with your changes? (if necessary)
• Did you write any new necessary tests? (not for typos and docs)
• Did you verify new and existing tests pass locally with your changes?
• Did you list all the breaking changes introduced by this pull request?
• Did you update the CHANGELOG? (not for typos, docs, test updates, or minor internal changes/refactors)

PR review

Anyone in the community is welcome to review the PR.
@Borda @carmocca @awaelchli
Before you start reviewing, make sure you have read the review guidelines. In short, see the following bullet-list:

Reviewer checklist

• Is this pull request ready for review? (if not, please submit in draft mode)
• Check that all items from Before submitting are resolved
• Make sure the title is self-explanatory and the description concisely explains the PR
• Add labels and milestones (and optionally projects) to the PR so it can be classified

---

📚 Documentation preview 📚: https://pytorch-lightning--19947.org.readthedocs.build/en/19947/

[Bhavay-2001]

Hi, I will be contributing and changing all of the files in the domain_templates folder.

[Bhavay-2001]

Hi @awaelchli, I was looking at this line as it has numpy dependency. I tried to look for a torch alternative for this but there aren't any direct functions but some workarounds possible. Should I try to change it with that or are there any other libraries supporting this function?

[awaelchli]

The equivalent function from the Python random standard library would be

random.sample(range(len(self.buffer)), batch_size)

Generally speaking, for examples it's fine to use additional dependencies. If you can't replace the numpy requirement somewhere, that's fine. The original issue #16649 is more concerned with removing the dependency from the package.

[Bhavay-2001]

The equivalent function from the Python random standard library would be

random.sample(range(len(self.buffer)), batch_size)

Generally speaking, for examples it's fine to use additional dependencies. If you can't replace the numpy requirement somewhere, that's fine. The original issue #16649 is more concerned with removing the dependency from the package.

Once this gets completed, I will try to change numpy to torch requirements in src/lightning module as well. Thanks

[awaelchli]

Sounds good, thanks for the effort.
@Peiffap was also working on that issue #16649, you may want to coordinate in case both of you are planning to submit PRs :) Thanks

[Bhavay-2001]

Alright. I will connect with him in the main issue. Thanks