MG-Verilog: Multi-grained Dataset Towards Enhanced LLM-assisted Verilog Generation

TL;DR

This paper introduces MG-Verilog, a multi-grained dataset designed to improve LLM-assisted hardware design by providing diverse, detailed descriptions and code samples, along with a fine-tuning scheme that enhances model performance.

Contribution

The paper presents a novel multi-grained Verilog dataset and an open-source infrastructure, along with a balanced fine-tuning scheme to better leverage diverse hardware data for LLMs.

Findings

Enhanced LLM performance in hardware design tasks.

The dataset supports various levels of detail for flexible training.

Fine-tuning with MG-Verilog improves model accuracy.

Abstract

Large Language Models (LLMs) have recently shown promise in streamlining hardware design processes by encapsulating vast amounts of domain-specific data. In addition, they allow users to interact with the design processes through natural language instructions, thus making hardware design more accessible to developers. However, effectively leveraging LLMs in hardware design necessitates providing domain-specific data during inference (e.g., through in-context learning), fine-tuning, or pre-training. Unfortunately, existing publicly available hardware datasets are often limited in size, complexity, or detail, which hinders the effectiveness of LLMs in hardware design tasks. To address this issue, we first propose a set of criteria for creating high-quality hardware datasets that can effectively enhance LLM-assisted hardware design. Based on these criteria, we propose a Multi-Grained-Verilog (MG-Verilog) dataset, which encompasses descriptions at various levels of detail and corresponding code samples. To benefit the broader hardware design community, we have developed an open-source infrastructure that facilitates easy access, integration, and extension of the dataset to meet specific project needs. Furthermore, to fully exploit the potential of the MG-Verilog dataset, which varies in complexity and detail, we introduce a balanced fine-tuning scheme. This scheme serves as a unique use case to leverage the diverse levels of detail provided by the dataset. Extensive experiments demonstrate that the proposed dataset and fine-tuning scheme consistently improve the performance of LLMs in hardware design tasks.