{"id":803617,"date":"2021-12-16T14:47:39","date_gmt":"2021-12-16T22:47:39","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/?post_type=msr-blog-post&p=803617"},"modified":"2021-12-16T14:47:39","modified_gmt":"2021-12-16T22:47:39","slug":"pymarlin-a-lightweight-library-that-improves-deep-learning-training-agility","status":"publish","type":"msr-blog-post","link":"https:\/\/www.microsoft.com\/en-us\/research\/articles\/pymarlin-a-lightweight-library-that-improves-deep-learning-training-agility\/","title":{"rendered":"PyMarlin: A lightweight library that improves deep learning training agility"},"content":{"rendered":"
By Amin Saied, Ananth Rao, Ashwin Srinivasan, Damien Jose, Eduardo Gonzalez, Han Yu, Jon Sleep, Krishan Subudhi, Shruti Gullapuram<\/strong><\/p>\n PyMarlin (opens in new tab)<\/span><\/a>\u00a0is a lightweight\u00a0PyTorch (opens in new tab)<\/span><\/a>\u00a0extension library for agile experimentation.\u00a0It was designed with the goal of simplifying the end-to-end deep learning experimentation lifecycle, agnostic of the compute environment.\u00a0In July 2021, the PyMarlin team open-sourced their internal model training library to all PyTorch users. PyMarlin abstracts out all the boilerplate code for scaling, logging, and argument parsing that are crucial for training deep learning-based models. PyMarlin can be thought of as a high-level abstraction over PyTorch. We have created a five-minute \u201cGetting Started (opens in new tab)<\/span><\/a>\u201d module for anyone interested in trying out PyMarlin. Today we\u2019ll look at how PyMarlin works, how it supports extensibility, and next steps needed to advance its functionality further.<\/p>\nHow the typical deep learning training lifecycle works<\/h3>\n