Zobrazeno 1 - 10
of 787
pro vyhledávání: '"Agarwal, Saurabh"'
Autor:
Elhoushi, Mostafa, Shrivastava, Akshat, Liskovich, Diana, Hosmer, Basil, Wasti, Bram, Lai, Liangzhen, Mahmoud, Anas, Acun, Bilge, Agarwal, Saurabh, Roman, Ahmed, Aly, Ahmed A, Chen, Beidi, Wu, Carole-Jean
We present LayerSkip, an end-to-end solution to speed-up inference of large language models (LLMs). First, during training we apply layer dropout, with low dropout rates for earlier layers and higher dropout rates for later layers, and an early exit
Externí odkaz:
http://arxiv.org/abs/2404.16710
Autor:
Agarwal, Saurabh, Acun, Bilge, Hosmer, Basil, Elhoushi, Mostafa, Lee, Yejin, Venkataraman, Shivaram, Papailiopoulos, Dimitris, Wu, Carole-Jean
Large Language Models (LLMs) with hundreds of billions of parameters have transformed the field of machine learning. However, serving these models at inference time is both compute and memory intensive, where a single request can require multiple GPU
Externí odkaz:
http://arxiv.org/abs/2403.08058
Speculative Decoding is a widely used technique to speed up inference for Large Language Models (LLMs) without sacrificing quality. When performing inference, speculative decoding uses a smaller draft model to generate speculative tokens and then use
Externí odkaz:
http://arxiv.org/abs/2402.01528
Chest X-ray imaging is a critical diagnostic tool for identifying pulmonary diseases. However, manual interpretation of these images is time-consuming and error-prone. Automated systems utilizing convolutional neural networks (CNNs) have shown promis
Externí odkaz:
http://arxiv.org/abs/2401.00728
Deep Learning (DL) workloads have rapidly increased in popularity in enterprise clusters and several new cluster schedulers have been proposed in recent years to support these workloads. With rapidly evolving DL workloads, it is challenging to quickl
Externí odkaz:
http://arxiv.org/abs/2312.12621
Autor:
Wang, Hongyi, Agarwal, Saurabh, U-chupala, Pongsakorn, Tanaka, Yoshiki, Xing, Eric P., Papailiopoulos, Dimitris
Recent research has shown that training low-rank neural networks can effectively reduce the total number of trainable parameters without sacrificing predictive accuracy, resulting in end-to-end speedups. However, low-rank model training necessitates
Externí odkaz:
http://arxiv.org/abs/2305.02538
Autor:
Agarwal, Saurabh
Publikováno v:
The CASE Journal, 2023, Vol. 20, Issue 2, pp. 405-421.
Externí odkaz:
http://www.emeraldinsight.com/doi/10.1108/TCJ-04-2021-0056
Autor:
Sunny, Nibin1 nibinsunny1996@gmail.com, Agarwal, Saurabh2, Navadaya, Mahesh3, Singh, Mayank Kumar4
Publikováno v:
Asian Journal of Medical Sciences. Jul2024, Vol. 15 Issue 7, p133-136. 4p.
Autor:
Agarwal, Saurabh
Publikováno v:
Freely available online.
Thesis (M.S.)--University of Missouri--Rolla, 2007.
Vita. The entire thesis text is included in file. Title from title screen of thesis/dissertation PDF file (viewed May 13, 2008) Includes bibliographical references (p. 108-113).
Vita. The entire thesis text is included in file. Title from title screen of thesis/dissertation PDF file (viewed May 13, 2008) Includes bibliographical references (p. 108-113).