Fine-tuned spanish language models with EmoEvent dataset. A project for VIHrtual-App chatbot to detect emotions in STI/HIV related conversations.
- RoBERTa-base-BNE-FineTunedEmovent
- RoBERTa-large-BNE-FineTunedEmovent
- RoBERTa-base-biomedical-es-FineTunedEmovent
- GPT2-base-BNE-FineTunedEmovent
- GPT2-large-BNE-FineTunedEmovent
Model | F1 | Accuracy |
---|---|---|
RoBERTa-base-BNE-FineTunedEmovent | 0.7089 | 0.7071 |
RoBERTa-large-BNE-FineTunedEmovent | 0.7638 🔥 | 0.7714 🔥 |
RoBERTa-base-biomedical-es-FineTunedEmovent | 0.6909 | 0.7000 |
GPT2-base-BNE-FineTunedEmovent | 0.6256 | 0.6214 |
GPT2-large-BNE-FineTunedEmovent | 0.4598 | 0.4857 |
Evaluation performed with STI dataset. See more details at metrics docs.
from transformers import pipeline
from transformers import AutoModelForSequenceClassification
from transformers import AutoTokenizer
# Load the model
load_model = AutoModelForSequenceClassification.from_pretrained("joancipria/roberta-large-bne-FineTunedEmoEvent")
load_tokenizer = AutoTokenizer.from_pretrained("joancipria/roberta-large-bne-FineTunedEmoEvent")
# Setup pipeline
my_pipeline = pipeline("sentiment-analysis", model=load_model, tokenizer=load_tokenizer)
# Predict sentiment for the following text
text = ["me encuentro genial con la nueva medicación"]
print(my_pipeline(text))
Tested with Python 3.10.7
Clone repository
git clone https://github.com/joancipria/sentiment-analysis && cd sentiment-analysis
Create virtual environment
python -m venv ./venv && source ./venv/bin/activate
Install requirements
pip install -r requirements.txt
Run python train.py
to fine-tune the models.
Edit and run python predict.py
to test recognition.
Edit and run python eval.py
to evaluate a model.
To cite this resource in a publication, please use the following:
@inproceedings{moreno2022conversational,
title={A Conversational Agent for Medical Disclosure of Sexually Transmitted Infections},
author={Moreno, Joan C and S{\'a}nchez-Anguix, Victor and Alberola, Juan M and Juli{\'a}n, Vicente and Botti, Vicent},
booktitle={International Conference on Hybrid Artificial Intelligence Systems},
pages={431--442},
year={2022},
organization={Springer}
}
Licensed under GNU General Public License v3.