CamemBERT-EmoTextToKids

Classification model for

  • input: a sentence + the previous and following sentences
  • output: 20 labels
    • is emotional
    • mode of expression
    • type of emotion(s) (basic, complex)
    • category of emotion(s)

Input format

The prompt template is: before:{previous_sentence}</s>current: {target_sentence}</s>after:{next_sentence}</s>

Output format

Labels are returned in the following order: 0. sentence is emotional

  1. mode is behavioral
  2. mode is labeled
  3. mode is displayed
  4. mode is suggested
  5. type is basic
  6. type is complex
  7. category is admiration
  8. category is other
  9. category is anger
  10. category is guilt
  11. category is disgust
  12. category is embarassement
  13. category is pride
  14. category is jealousy
  15. category is fear
  16. category is joy
  17. categoy is fear
  18. category is surprise
  19. category is sadness

See the original paper for details about training.

Dataset

See EmoTextToKids-sentences

Citation information

@inproceedings{etienne2024emotion,
  title={Emotion Identification for French in Written Texts: Considering Modes of Emotion Expression as a Step Towards Text Complexity Analysis},
  author={{\'E}tienne, Aline and Battistelli, Delphine and Lecorv{\'e}, Gw{\'e}nol{\'e}},
  booktitle={Proceedings of the 14th ACL Workshop on Computational Approaches to Subjectivity, Sentiment & Social Media Analysis (WASSA)},
  year={2024}
}
Downloads last month
12
Safetensors
Model size
111M params
Tensor type
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.