l3cube-pune/marathi-emotion-detect

MahaEmotions-BERT

MahaEmotions-BERT is a MahaBERT(l3cube-pune/marathi-bert-v2) model fine-tuned on L3Cube-MahaEmotions Corpus, a Marathi Emotion Recognition dataset.
MahaEmotions is a high-quality Marathi emotion recognition dataset designed to address the challenge of limited annotated data in low-resource languages. It features 11 fine-grained emotion labels and combines synthetically annotated training data (generated using Large Language Models like GPT-4) with manually labeled validation and test sets to establish a reliable gold-standard benchmark.
[github link] (https://github.com/l3cube-pune/MarathiNLP)

More details on the dataset, models, and baseline results can be found in our [paper] (https://arxiv.org/abs/2506.00863)
Citing:

@article{kowtal2025l3cube,
  title={L3Cube-MahaEmotions: A Marathi Emotion Recognition Dataset with Synthetic Annotations using CoTR prompting and Large Language Models},
  author={Kowtal, Nidhi and Joshi, Raviraj},
  journal={arXiv preprint arXiv:2506.00863},
  year={2025}
}