Saved in:
Bibliographic Details
Main Authors: Xue, Tengfei, Li, Xuefeng, Smirnov, Roman, Azim, Tahir, Sadrieh, Arash, Pahlavan, Babak
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2407.12057
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916327776583680
author Xue, Tengfei
Li, Xuefeng
Smirnov, Roman
Azim, Tahir
Sadrieh, Arash
Pahlavan, Babak
author_facet Xue, Tengfei
Li, Xuefeng
Smirnov, Roman
Azim, Tahir
Sadrieh, Arash
Pahlavan, Babak
contents Retrieval-augmented generation (RAG) techniques are widely used today to retrieve and present information in a conversational format. This paper presents a set of enhancements to traditional RAG techniques, focusing on large language models (LLMs) fine-tuned and hosted on AWS Trainium and Inferentia2 AI chips via SageMaker. These chips are characterized by their elasticity, affordability, and efficient performance for AI compute tasks. Besides enabling deployment on these chips, this work aims to improve tool usage, add citation capabilities, and mitigate the risks of hallucinations and unsafe responses due to context bias. We benchmark our RAG system's performance on the Natural Questions and HotPotQA datasets, achieving an accuracy of 62% and 59% respectively, exceeding other models such as DBRX and Mixtral Instruct.
format Preprint
id arxiv_https___arxiv_org_abs_2407_12057
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle NinjaLLM: Fast, Scalable and Cost-effective RAG using Amazon SageMaker and AWS Trainium and Inferentia2
Xue, Tengfei
Li, Xuefeng
Smirnov, Roman
Azim, Tahir
Sadrieh, Arash
Pahlavan, Babak
Computation and Language
Artificial Intelligence
I.2.7
Retrieval-augmented generation (RAG) techniques are widely used today to retrieve and present information in a conversational format. This paper presents a set of enhancements to traditional RAG techniques, focusing on large language models (LLMs) fine-tuned and hosted on AWS Trainium and Inferentia2 AI chips via SageMaker. These chips are characterized by their elasticity, affordability, and efficient performance for AI compute tasks. Besides enabling deployment on these chips, this work aims to improve tool usage, add citation capabilities, and mitigate the risks of hallucinations and unsafe responses due to context bias. We benchmark our RAG system's performance on the Natural Questions and HotPotQA datasets, achieving an accuracy of 62% and 59% respectively, exceeding other models such as DBRX and Mixtral Instruct.
title NinjaLLM: Fast, Scalable and Cost-effective RAG using Amazon SageMaker and AWS Trainium and Inferentia2
topic Computation and Language
Artificial Intelligence
I.2.7
url https://arxiv.org/abs/2407.12057