Comprehensive Study of Parameter Efficient Fine tuning techniques on Small Language Models for Gherkin Scenario Generation

Chandra Babu, Shruthi (2025) Comprehensive Study of Parameter Efficient Fine tuning techniques on Small Language Models for Gherkin Scenario Generation. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration Manual) Download (3MB) | Preview

Abstract

Validating software against requirements is one of the critical phases in software development. In recent times, Test cases which aid in this validation, are written as gherkin scenarios, as they represent the test cases in a more structured way. This paper will be evaluating the potential of small language models, such as StarCoder, DeepSeekCoder, Phi and Tinyllama, to accelerate the gherkin scenario generation. In addition, this paper will also be evaluating the resource efficient methods of fine tuning such as DoRA and Hybrid LoRA, which are extensions of LoRA fine tuning. The primary motivation behind choosing small language models and fine tuning such as LoRA, is to study resource efficient models and fine tuning methods, which would make them a practical choice for real world deployments. Further, The quality of the generated gherkin scenarios will be evaluated on the basis of metrics such as BLEU, ROUGE score to analyse similarity and BERT-score to evaluate the semantic similarity.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Shahid, Abdul UNSPECIFIED
Uncontrolled Keywords:	Small Language Models; LoRA Fine tuning; Dora Fine tuning; IA3 fine tuning; Gherkin Scenario Generation; Chain of Thought prompting technique
Subjects:	Q Science > QA Mathematics > Computer software T Technology > T Technology (General) > Information Technology > Computer software Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence P Language and Literature > P Philology. Linguistics > Computational linguistics. Natural language processing
Divisions:	School of Computing > Master of Science in Artificial Intelligence
Depositing User:	Ciara O'Brien
Date Deposited:	28 May 2026 13:20
Last Modified:	28 May 2026 13:20
URI:	https://norma.ncirl.ie/id/eprint/9315

Actions (login required)

View Item