ddpw 2.0.0

A utility package to scaffold PyTorch's DDP

DDPW

The Distributed Data Parallel wrapper (DDPW) is created as a utility package to encapsulate the scaffolding for PyTorch's Distributed Data Parallel.

This code is written in Python 3.8. The DDPW documentation contains details on how to use this package.

Overview

Installation

conda install -c tvsujal ddpw # with conda
pip install ddpw # with pip from PyPI

Usage

  from ddpw.platform import PlatformConfig
  from ddpw.artefacts import ArtefactsConfig
  from ddpw.trainer import TrainingConfig
  from ddpw.wrapper import Wrapper

  from src import CustomTrainer

  p = PlatformConfig(...)
  a = ArtefactsConfig(...)
  t = TrainingConfig(...)

  d = Wrapper(p, a)
  j = CustomTrainer(t)

  d.start(j)