TREC 2025 Proceedings

videollama-7B-vqa

Submission Details

Organization: WHU-NERCMS
Track: Video Question Answering
Task: Multiple Choice Task
Date: 2025-09-29

Run Description

What base model(s) did you use?: VideoLLaMA3
Did you fine tune existing models or used off the shelf open source model(s) ?: Off the shelf
If fine tuned or built your own model, what training data did you used?: TRECVID2025 VQA training dataset
How long development time did it take you to finish your system?: 3 weeks

Evaluation Files

videollama-7B-vqa.vqa_mc_eval (vqa_mc_eval)