Downloads · 30 days
2.9K
1% of all-time downloads
metaphor_detection_roberta_seq
CreativeLang/metaphor_detection_roberta_seq
metaphor_detection_roberta_seq is a token classification model from CreativeLang. Use it when you need labels on individual words, such as names. It is set up for transformers. The card lists the license as cc-by-2.0.
- Paper: FrameBERT: Conceptual Metaphor Detection with Frame Embedding Learning
All-time downloads
197K
Public
Repo size
3 GB
Likes
4
Public
Repo makeup
Click a slice to open those files.
.bin496 MB · 99%
At a glance
- Task
- Token Classification
- Library
- transformers
- License
- cc-by-2.0
- Model type
- roberta
- Access
- Public
- Created
- Jul 8, 2023
- Updated
- Sep 25, 2023
- SHA
- 368f0536
Try a prompt
Datasets
Model card
From the Hugging Face model README
Metaphor_Detection_Roberta_Seq
Description
Model Summary
Creative Language Toolkit (CLTK) Metadata
- CL Type: Metaphor
- Task Type: detection
- Size: roberta-base (500MB)
- Created time: 2022
This model is a easy to use metaphor detection baseline realised with roberta-base fine-tuned on CreativeLang/vua20_metaphor dataset.
To use this model, please use the inference.py in the FrameBERT repo.
Just run:
python inference.py CreativeLang/metaphor_detection_roberta_seq
Check out inference.py to learn how to apply the model on your own data.
For the details of this model and the dataset used, we refer you to the release paper.
Metrics
| Metric | Value |
|---|---|
| eval_loss | 0.2656 |
| eval_accuracy_score | 0.9142 |
| eval_precision | 0.9142 |
| eval_recall | 0.9142 |
| eval_f1 | 0.9142 |
| eval_f1_macro | 0.7315 |
| eval_runtime | 8.9802 |
| eval_samples_per_second | 411.7960 |
| eval_steps_per_second | 51.5580 |
| epoch | 3.0000 |
Citation Information
If you find this dataset helpful, please cite:
@article{Li2023FrameBERTCM,
title={FrameBERT: Conceptual Metaphor Detection with Frame Embedding Learning},
author={Yucheng Li and Shunyu Wang and Chenghua Lin and Frank Guerin and Lo{\"i}c Barrault},
journal={ArXiv},
year={2023},
volume={abs/2302.04834}
}
Contributions
If you have any queries, please open an issue or direct your queries to mail.
- Task
- Token Classification
- Library
- transformers
- Type
- roberta
- License
- cc-by-2.0
- Languages
- en
- Created
- Jul 8, 2023
- Updated
- Sep 25, 2023