papers

Publications (25)

cs.LG2019

Sparse Networks from Scratch: Faster Training without Losing Performance

Tim Dettmers, Luke Zettlemoyer

We demonstrate the possibility of what we call sparse learning: accelerated training of deep neural networks that maintain sparse weights throughout training while achieving dense…

cs.LG2024

MatFormer: Nested Transformer for Elastic Inference

Devvrit, Sneha Kudugunta, Aditya Kusupati +8

Foundation models are applied in a broad spectrum of settings with different inference constraints, from massive multi-accelerator clusters to resource-constrained standalone mobil…

cs.CL2024

Scaling Retrieval-Based Language Models with a Trillion-Token Datastore

Rulin Shao, Jacqueline He, Akari Asai +5

Scaling laws with respect to the amount of training data and the number of parameters allow us to predict the cost-benefit trade-offs of pretraining language models (LMs) in differ…

cs.CL2026

AdaExplore: Failure-Driven Adaptation and Diversity-Preserving Search for Efficient Kernel Generation

Weihua Du, Jingming Zhuo, Yixin Dong +9

Recent large language model (LLM) agents have shown promise in using execution feedback for test-time adaptation. However, robust self-improvement remains far from solved: most app…

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.LG2022

8-bit Optimizers via Block-wise Quantization

Tim Dettmers, Mike Lewis, Sam Shleifer +1

Stateful optimizers maintain gradient statistics over time, e.g., the exponentially smoothed sum (SGD with momentum) or squared sum (Adam) of past gradient values. This state can b…