Lightning-AI / Lightning-AI/pytorch-lightning

Documentation for averaging logged values when using accumulate_grad_batches

Open
#14,301 2 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

docs
Dominant language
Python
Stars
31.4k
Forks
3.8k
Avg merge
6d 7h
Merged PRs (30d)
6

Description

## 📚 Documentation
It would be very nice if in the [documentation](https://pytorch-lightning.readthedocs.io/en/stable/advanced/training_tricks.html#accumulate-gradients) there was a suggestion of how to average the logged values in one accumulation group. I'm not sure what the best way is to do this. Also, to me, it was not obvious that the logged values where not already averaged automatically.

cc @borda @rohitgr7

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 with the linked accumulate-gradients section in the documentation and review how logged values behave with accumulate_grad_batches. Document how to average values within one accumulation group and clarify whether logged values are averaged automatically, with wording that matches the current behavior.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, pytorch
Domain
documentation
Issue type
Documentation
Difficulty
2/5
Estimated time
1-3 hours
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
38/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.