README.md

---
license: apache-2.0
library_name: transformers
pipeline_tag: video-text-to-text
---

# DAR-R1

This is the official model checkpoint for **DAR-R1** (built on top of Qwen2.5-VL-3B-Instruct), presented in the paper:
[Benchmarking Dynamic Affective Reasoning: A Viewer-Centric Video Emotion Dataset](https://arxiv.org/abs/2607.10238).

## Overview

**Dynamic Affective Reasoning (DAR)** is a viewer-centric video emotion benchmark. Instead of assigning a single static label to a whole clip, DAR asks a model to identify when the viewer's emotion changes (affective segmentation), what the fine-grained emotion is (fine-grained emotion classification), and why the visual event triggers that affective reaction (affective reasoning).

The benchmark contains 15,087 videos, 36,908 event-aligned affective segments, and 27 emotion categories. Each segment includes a temporal span, an emotion label, and a visually grounded causal rationale.

**DAR-R1** is trained using a two-stage framework:
1. **Cold-Start SFT**: Adapts Qwen2.5-VL-3B-Instruct to the structured DAR output format.
2. **GRPO Training**: Uses Group Relative Policy Optimization (GRPO) to refine temporal localization, emotion prediction, and reasoning quality.

## Resources

- **GitHub Repository:** [Zhang-Zhiyan/DAR](https://github.com/Zhang-Zhiyan/DAR)
- **Hugging Face Dataset:** [aiaiaizzy/DAR-R1](https://huggingface.co/datasets/aiaiaizzy/DAR-R1)
- **Paper:** [arXiv:2607.10238](https://arxiv.org/abs/2607.10238)

## Quick Start (Evaluation)

To run evaluation using the official script from the repository:

```bash
python test.py \
  --model-path /path/to/DAR-R1 \
  --test-jsonl /path/to/DAR/test.jsonl \
  --video-root /path/to/DAR/videos \
  --output-jsonl /path/to/outputs/dar_r1_test_predictions.jsonl \
  --batch-size 8
```

## Citation

If you find this model or the DAR benchmark useful in your research, please consider citing:

```bibtex
@misc{zhang2026benchmarkingdynamicaffectivereasoning,
      title={Benchmarking Dynamic Affective Reasoning: A Viewer-Centric Video Emotion Dataset}, 
      author={Zhiyan Zhang and Peipei Song and Jinpeng Hu and Jingyang Jia and Xun Yang and Xiaojun Chang},
      year={2026},
      eprint={2607.10238},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2607.10238}, 
}
```