metadata

language:
  - ca
license: apache-2.0
tags:
  - catalan
  - text classification
  - WikiCAT_ca
  - CaText
  - Catalan Textual Corpus
datasets:
  - projecte-aina/WikiCAT_ca
metrics:
  - f1
model-index:
  - name: roberta-base-ca-v2-cased-wikicat-ca
    results:
      - task:
          type: text-classification
        dataset:
          type: projecte-aina/WikiCAT_ca
          name: WikiCAT_ca
        metrics:
          - name: F1
            type: f1
            value: 77.82
widget:
  - text: >-
      La ressonància magnètica és una prova diagnòstica clau per a moltes
      malalties.

Catalan BERTa-v2 (roberta-base-ca-v2) finetuned for Text Classification.

Model Description
Intended Uses and Limitations
How to Use
Training
- Training Data
- Training Procedure
Evaluation
- Variable and Metrics
- Evaluation Results
Licensing Information
Citation Information
Funding
Contributions

Model description

The roberta-base-ca-v2-cased-wikicat-ca is a Text Classification model for the Catalan language fine-tuned from the roberta-base-ca-v2 model, a RoBERTa base model pre-trained on a medium-size corpus collected from publicly available corpora and crawlers (check the roberta-base-ca-v2 model card for more details).

Intended Uses and Limitations

roberta-base-ca-v2-cased-wikicat-ca model can be used to classify texts. The model is limited by its training dataset and may not generalize well for all use cases.

How to Use

Here is how to use this model:

from transformers import pipeline
from pprint import pprint

nlp = pipeline("text-classification", model="roberta-base-ca-v2-cased-wikicat-ca")
example = "La ressonància magnètica és una prova diagnòstica clau per a moltes malalties."

tc_results = nlp(example)
pprint(tc_results)

Training

Training data

We used the TC dataset in Catalan called WikiCAT_ca for training and evaluation.

Training Procedure

The model was trained with a batch size of 4 and three learning rates (1e-5, 3e-5, 5e-5) for 10 epochs. We then selected the best learning rate (3e-5) and checkpoint (epoch 3, step 1857) using the downstream task metric in the corresponding development set.

Evaluation

Variable and Metrics

This model was finetuned maximizing F1 (weighted) score.

Evaluation results

We evaluated the roberta-base-ca-v2-cased-wikicat-ca on the WikiCAT_ca dev set:

Model	WikiCAT_ca (F1)
roberta-base-ca-v2-cased-wikicat-ca	77.82

For more details, check the fine-tuning and evaluation scripts in the official GitHub repository.

Licensing Information

Apache License, Version 2.0

Citation Information

Funding

This work was funded by the Departament de la Vicepresidència i de Polítiques Digitals i Territori de la Generalitat de Catalunya within the framework of Projecte AINA.

Contributions

[N/A]

projecte-aina
/

roberta-base-ca-v2-cased-wikicat-ca