Data and code for: Resource-Efficient Transformers for Edge Intelligence: A Systematic Literature Review of Vision Transformers and Large Language Models

Data and code supporting a PRISMA 2020 systematic literature review of resource-efficient vision transformers (ViTs) and large language models (LLMs) for edge intelligence (156 included studies, 136 primary). The record contains: (i) the coded full-text records of all 238 reports assessed for eligibility in the two identification arms (database search and citation searching), with decisions, exclusion codes, model strand, full-text theme, the platform on which efficiency results were obtained, technique flags, reported metrics and quality ratings; (ii) the PRISMA flow counts; (iii) Python scripts that recompute every count reported in the article and redraw its six figures; and (iv) the article's five Online Resources (PRISMA checklist, extraction and quality-assessment workbook, reference database, search protocol, supplementary tables). See README.md for the contents and the coding scheme.