Data-Prep-Kit: getting your data ready for LLM application development

TL;DR

Data Prep Kit (DPK) is an open-source, scalable toolkit designed to facilitate efficient data preparation for Large Language Model development, supporting local and cluster-based processing with extensible modules.

Contribution

Introduces DPK, a flexible, scalable, and extensible data preparation toolkit for LLMs, enabling easy customization and large-scale processing.

Findings

DPK scales from local to cluster environments.

Modules effectively transform natural language and code data.

Used in preparing data for Granite Models.

Abstract

Data preparation is the first and a very important step towards any Large Language Model (LLM) development. This paper introduces an easy-to-use, extensible, and scale-flexible open-source data preparation toolkit called Data Prep Kit (DPK). DPK is architected and designed to enable users to scale their data preparation to their needs. With DPK they can prepare data on a local machine or effortlessly scale to run on a cluster with thousands of CPU Cores. DPK comes with a highly scalable, yet extensible set of modules that transform natural language and code data. If the user needs additional transforms, they can be easily developed using extensive DPK support for transform creation. These modules can be used independently or pipelined to perform a series of operations. In this paper, we describe DPK architecture and show its performance from a small scale to a very large number of CPUs. The modules from DPK have been used for the preparation of Granite Models [1] [2]. We believe DPK is a valuable contribution to the AI community to easily prepare data to enhance the performance of their LLM models or to fine-tune models with Retrieval-Augmented Generation (RAG).