pytorch/ignite

Support for TorchSnapshot for efficient checkpoint saving and loading

Open

#2,752 opened on Oct 24, 2022

View on GitHub
 (3 comments) (0 reactions) (0 assignees)Python (602 forks)batch import
enhancementhelp wanted

Repository metrics

Stars
 (4,313 stars)
PR merge metrics
 (Avg merge 15d 11h) (17 merged PRs in 30d)

Description

🚀 Feature

TorchSnapshot is a performant, memory-efficient checkpointing library for PyTorch applications, designed with large, complex distributed workloads in mind. It includes many optimizations to control for memory usage and optimize checkpoint writing for DDP-style workloads over torch.save/torch.load. For more information, please check out the readme: https://github.com/pytorch/torchsnapshot#why-torchsnapshot

This could be a nice addition to Ignite, similar to the existing Checkpoint handler

cc @yifuwang

Contributor guide

Research direction
Start by reviewing the existing Checkpoint handler in pytorch/ignite (likely in ignite/handlers/checkpoint.py) to understand the current implementation. Then explore the TorchSnapshot documentation and examples to understand its API and how it could replace or augment the current checkpointing. Consider contributing by implementing a new handler or modifying the existing one to support TorchSnapshot, similar to how Ignite handles other saving methods. Discuss design in the issue comments with maintainers before proceeding.
Tech stack
python
Domain
machine learning
Issue type
Feature
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
3-5 days
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Stale
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Clear
Prerequisites
PyTorch Ignite basicsUnderstanding of checkpointing concepts
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
30