StoryLLaVA: Enhancing Visual Storytelling with Multi-Modal Large Language Models

This work proposes a Topic-Driven Narrative Optimizer (TDNO) that improves both the training data and MLLM models by integrating image descriptions, topic generation, and GPT-4-based refinements, and employs a preference-based ranked story sampling method that aligns model outputs with human storytelling preferences through positive-negative pairing.