Narrow Transformer: StarCoder-Based Java-LM For Desktop

TL;DR

This paper introduces NT-Java-1.1B, a specialized Java code language model based on StarCoder, optimized for desktop deployment, achieving state-of-the-art performance among small models on Java benchmarks.

Contribution

Development of NT-Java-1.1B, a small, efficient Java code model that surpasses similar-sized models and is suitable for desktop use, filling a research gap in small language models for specific programming languages.

Findings

NT-Java-1.1B outperforms comparable models on MultiPL-E Java benchmark.

Quantized versions maintain high performance, enabling desktop deployment.

The model establishes a foundation for language-specific, size-optimized code models.

Abstract

This paper presents NT-Java-1.1B, an open-source specialized code language model built on StarCoderBase-1.1B, designed for coding tasks in Java programming. NT-Java-1.1B achieves state-of-the-art performance, surpassing its base model and majority of other models of similar size on MultiPL-E Java code benchmark. While there have been studies on extending large, generic pre-trained models to improve proficiency in specific programming languages like Python, similar investigations on small code models for other programming languages are lacking. Large code models require specialized hardware like GPUs for inference, highlighting the need for research into building small code models that can be deployed on developer desktops. This paper addresses this research gap by focusing on the development of a small Java code model, NT-Java-1.1B, and its quantized versions, which performs comparably to open models around 1.1B on MultiPL-E Java code benchmarks, making them ideal for desktop deployment. This paper establishes the foundation for specialized models across languages and sizes for a family of NT Models.