CapRL-Qwen3VL-4B Image Captioning

This Space demonstrates CapRL-Qwen3VL-4B, a 4B parameter multimodal vision-language model fine-tuned using reinforcement learning for dense image captioning.

Model

• Model: internlm/CapRL-Qwen3VL-4B
• Architecture: Based on Qwen3-VL
• Task: Dense Image Captioning

Usage

Upload an image or select from the examples to generate a detailed caption.

Links

• Paper
• GitHub
• CapRL-3B Model
• CapRL-2M Dataset
• CapRL Collection

Citation

@article{xing2025caprl,
  title={{CapRL}: Stimulating Dense Image Caption Capabilities via Reinforcement Learning},
  author={Xing, Long and Dong, Xiaoyi and Zang, Yuhang and Cao, Yuhang and Liang, Jianze and Huang, Qidong and Wang, Jiaqi and Wu, Feng and Lin, Dahua},
  journal={arXiv preprint arXiv:2509.22647},
  year={2025}
}

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference