YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
StudyMateAI FLAN-T5 LoRA Adapter
This repository contains the LoRA adapter fine-tuned for the StudyMateAI multi-PDF Retrieval-Augmented Generation (RAG) system.
Base Model
google/flan-t5-small
Fine-tuning
The adapter was trained using PEFT LoRA for sequence-to-sequence question answering.
The training data consists of contextual QA examples derived from the StudyMateAI academic PDF collection.
Configuration
- LoRA rank: 8
- LoRA alpha: 16
- LoRA dropout: 0.05
- Target modules: q, v
- Task type: SEQ_2_SEQ_LM
Intended Use
This adapter is intended to be loaded on top of google/flan-t5-small for source-grounded academic question answering within the StudyMateAI RAG pipeline.
Architecture
PDFs → preprocessing → chunking → SBERT embeddings → FAISS retrieval → FLAN-T5 + LoRA → source-grounded answer
Evaluation
The adapter is evaluated using the same 30-question evaluation set used for the original StudyMateAI baseline.
Final evaluation metrics are reported separately in the StudyMateAI project.