pytorch/examples

Distributed doesn't work with env:// init_method

Open

#458 opened on Dec 1, 2018

View on GitHub
 (0 comments) (0 reactions) (0 assignees)Python (9,429 forks)batch import
distributedgood first issue

Repository metrics

Stars
 (21,634 stars)
PR merge metrics
 (No merged PRs in 30d)

Description

(1) if RANK or WORLD_SIZE is set in env variable, they will not be picked up (2) multiprocessing-distributed won't work with env variable: WORLD_SIZE and RANK being set since the world size needs to be recalculated.

Should be an easy fix

Contributor guide

Research direction
Examine the distributed module's handling of the 'env://' init method. Trace how environment variables RANK and WORLD SIZE are read and ensure they are used correctly. Also check the world size recalculation logic for multiprocessing distributed.
Tech stack
pythonpytorch
Domain
backendmachine learning
Issue type
Bug
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
1
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
Under 1 hour
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Active
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Clear
Prerequisites
PythonPyTorch distributed
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
75