This repository contains all the relevant data and code files for our DLT Project.
Abstract
Financial reports offer critical insights into a company’s operations, yet their extensive length—typically spanning 30-40 pages—poses challenges for swift decision-making in dynamic markets. To address this, we leveraged fine-tuned Language Models (LLMs) to distill key indicators and operational metrics from these reports. We devised a method to locate critical data, and leverage the FinQA dataset to fine-tune both Llama 2 7B and T5 models for customized question answering. We achieved 65% accuracy on final numerical answer, a competitive accuracy in numerical reasoning and calculation.
Methodology
Figure 1: LLMs to do Numerical Reasoning on annual reports (Pipeline)
Important Links: