Pedestrian Intention Estimation (PIE) with fine-tuned LLaVA-Next-Video The implementation of this project relies on resources from PIE and this link. We thank the original authors for their open-sourcing.