activity
20182022
most citedLAION-5B: An open large-scale dataset for training next generation image-text models

1k citations · 1.1k across the 4 of their papers we have counts for

collaborators
Showing cs.LGShow all

9 papers · 1 filter

cs.LG20241 cited

Resolving Discrepancies in Compute-Optimal Scaling of Language Models

Tomer Porian, Mitchell Wortsman, Jenia Jitsev +2

Kaplan et al. and Hoffmann et al. developed influential scaling laws for the optimal model size as a function of the compute budget, but these laws yield substantially different pr…

cs.LG20247 cited

DataComp-LM: In search of the next generation of training sets for language models

Jeffrey Li, Alex Fang, Georgios Smyrnis +56

We introduce DataComp for Language Models (DCLM), a testbed for controlled dataset experiments with the goal of improving language models. As part of DCLM, we provide a standardize…

cs.LG2023

Stable and low-precision training for large-scale vision-language models

Mitchell Wortsman, Tim Dettmers, Luke Zettlemoyer +3

We introduce new methods for 1) accelerating and 2) stabilizing training for large language-vision models. 1) For acceleration, we introduce SwitchBack, a linear layer for int8 qua…

cs.LG20227 cited

lo-fi: distributed fine-tuning without communication

Mitchell Wortsman, Suchin Gururangan, Shen Li +4

When fine-tuning large neural networks, it is common to use multiple nodes and to communicate gradients at each optimization step. By contrast, we investigate completely local fine…

cs.LG2021

Learning Neural Network Subspaces

Mitchell Wortsman, Maxwell Horton, Carlos Guestrin +2

Recent observations have advanced our understanding of the neural network optimization landscape, revealing the existence of (1) paths of high accuracy containing diverse solutions…

cs.LG2020

Deconstructing the Structure of Sparse Neural Networks

Maxwell Van Gelder, Mitchell Wortsman, Kiana Ehsani

Although sparse neural networks have been studied extensively, the focus has been primarily on accuracy. In this work, we focus instead on network structure, and analyze three popu…