Model Card for [CVPR2026] NanoEmoX

our model is first compact (2.2B) emotion intelligence Multimodel Langugage Model. It integrates a pretrained LLM with modality-specific encoders and experts-based fusion encoder to handle a broad spectrum of affective tasks in one model.

Nano-EmoX supports six core tasks within one model: 1), Multimodal Sentiment Analysis. 2), Multimodal Emotion Recognition. 3), Open-Vocabulary Multimodal Emotion Recognition. 4), Multimodal Intention Recognition. 5), Emotion Reason Inference. 6), Empathic Response Generation

Model Details

Model Sources

• Repository: GitHub
• Paper [optional]: Arxiv

Citation

If this work has been helpful or inspiring to your research, please consider cite our article:

BibTeX:

@InProceedings{Huang_2026_CVPR,
    author    = {Huang, Jiahao and Lin, Fengyan and Yang, Xuechao and Feng, Chen and Zhu, Kexin and Yang, Xu and Chen, Zhide},
    title     = {Nano-EmoX: Unifying Multimodal Emotional Intelligence from Perception to Empathy},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
    month     = {June},
    year      = {2026},
    pages     = {22986-22997}
}