Lightning-AI / Lightning-AI/pytorch-lightning

Support for AdaptDL

Open
#15,119 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

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

Description

Reporting from the idea-pool channel on slack, as discussed with @carmocca.


Hi there,

On the way to solve a OOM problem with dynamic batch sizes based on sequence length, I have just discovered AdaptDL. Might be an interesting library to support.

Some core features offered by AdaptDL are:

  • Elastically schedule distributed DL training jobs in shared clusters.
  • Cost-aware resource auto-scaling in cloud computing environments (e.g. AWS).
  • Automatic batch size and learning rate scaling for distributed training.

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 the AdaptDL PyTorch documentation linked in the issue and reviewing how this project handles distributed deep-learning training. Determine the intended integration boundary for elastic scheduling, resource scaling, and automatic batch-size and learning-rate scaling. Done should mean the project has an agreed, tested support path for the requested AdaptDL capabilities.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, pytorch
Domain
cloud, distributed-systems, machine-learning
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.