meta-pytorch / meta-pytorch/data

Add documentation about custom random DataPipe

Open
#1,105 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
1.3k
Forks
179
Avg merge
6d 1h
Merged PRs (30d)
2

Description

📚 The doc issue

Similar to #1104, DataLoader2 has built-in functionality to set random seed for DataPipes when they have set_seed method implemented.

https://github.com/pytorch/data/blob/main/torchdata/dataloader2/graph/settings.py#L22

We should provide official guidance about what we expect users to implement in cases where a custom random DataPipe is implemented/necessary.

Suggest a potential alternative/fix

No response

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reading torchdata/dataloader2/graph/settings.py at the linked lines and compare the related discussion in issue #1104. Document the guidance for implementing a custom random DataPipe, including the expected use of set_seed; done means the official documentation clearly explains this behavior and user expectations.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
documentation
Issue type
Documentation
Difficulty
3/5
Estimated time
Half a day
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.