This project demonstrates how to fine-tune the Flan-T5 base model using LoRA (Low-Rank Adaptation) to generate summaries from dialogue text.
The project guides you through:
Loading and exploring SAMsum dataset
Tokenizing inputs and preparing labels
Applying PEFT + LoRA to reduce memory and computation cost
Training the model using HuggingFace Trainer
Evaluating the generated summaries using ROUGE
