Metadata-Version: 2.1
Name: napolab
Version: 1.0.0
Summary: Natural Portuguese Language Benchmark
Author-email: Ruan Chaves Rodrigues <ruanchaves93@gmail.com>
License: MIT License
        
        Copyright (c) 2023 Ruan Chaves
        
        Permission is hereby granted, free of charge, to any person obtaining a copy
        of this software and associated documentation files (the "Software"), to deal
        in the Software without restriction, including without limitation the rights
        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
        copies of the Software, and to permit persons to whom the Software is
        furnished to do so, subject to the following conditions:
        
        The above copyright notice and this permission notice shall be included in all
        copies or substantial portions of the Software.
        
        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
        SOFTWARE.
        
Project-URL: Homepage, https://github.com/ruanchaves/napolab
Keywords: ai,llm,nlp
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python
Classifier: Programming Language :: Python :: 3
Requires-Python: >=3.9
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: datasets
Provides-Extra: dev
Requires-Dist: pytest; extra == "dev"

# Natural Portuguese Language Benchmark (Napolab)

<p align="center">
  <img width="300" height="300" src="https://raw.githubusercontent.com/ruanchaves/napolab/main/ideogram_ai_logo.png">
</p>

The Natural Portuguese Language Benchmark (Napolab) is designed to be a collection of Portuguese datasets that are:

* **Natural**: Composed, as far as possible, of natural Portuguese text or text that has undergone professional translation.
* **Reliable**: The standard metrics used to evaluate the tasks on the dataset should reliably correlate with human judgments (such as accuracy, F1 score, Pearson correlation coefficient, etc.).
* **Public**: All datasets must be available and downloadable via a publicly accessible link.
* **Human**: All datasets must be annotated by human experts without any automation.
* **General**: The datasets should not necessitate domain-specific knowledge or any preparation for an educated Portuguese speaker to solve the suggested tasks.

This repository contains links to demos, models fine-tuned on the benchmark, and instructions for using the datasets in the most convenient manner. The benchmark currently includes the following datasets:

* [assin](https://huggingface.co/datasets/assin)
* [assin2](https://huggingface.co/datasets/assin2)
* [rerelem](https://huggingface.co/datasets/ruanchaves/rerelem)
* [hatebr](https://huggingface.co/datasets/ruanchaves/hatebr)
* [reli-sa](https://huggingface.co/datasets/ruanchaves/reli-sa)
* [faquad-nli](https://huggingface.co/datasets/ruanchaves/faquad-nli)
* [porsimplessent](https://huggingface.co/datasets/ruanchaves/porsimplessent)

We are open to expanding the benchmark, and suggestions for future additions are welcome in the issues. We also welcome evaluation results from any models on this benchmark, and we are particularly curious about the outcomes of evaluating recent LLMs on these datasets.

In addition to making these datasets available in Portuguese, all datasets have also been automatically translated into **English, Spanish, Galician, and Catalan** using the `facebook/nllb-200-1.3B` model through the [Easy-Translate](https://github.com/ikergarcia1996/Easy-Translate) library.

The small `napolab.py` script in this repository includes a convenient `load_napolab_benchmark` function that will download the entire Napolab benchmark from the Hugging Face Hub, including all translated versions:

```python
from napolab import load_napolab_benchmark

napolab = load_napolab_benchmark(include_translations=True)

benchmark = napolab["datasets"]
translated_benchmark = napolab["translations"]
```

## Demos

All of our fine-tuned models have been integrated into an appropriate [Hugging Face Space](https://huggingface.co/ruanchaves).

Interact with our models in your browser by exploring our demos:

* [Portuguese Offensive Language Detection](https://ruanchaves-portuguese-offensive-language-de-d4d0507.hf.space)
* [Portuguese Question Answering](https://ruanchaves-portuguese-question-answering.hf.space)
* [Portuguese Semantic Similarity](https://ruanchaves-portuguese-semantic-similarity.hf.space)
* [Portuguese Textual Entailment](https://ruanchaves-portuguese-textual-entailment.hf.space)
* [Portuguese Text Simplification](https://ruanchaves-portuguese-text-simplification.hf.space)

## Models

### Summary of the Fine-Tuning procedure

* **Step 1**. Hyperparameter optimization is performed using quasi-random search based on Google's [Deep Learning Playbook](https://github.com/google-research/tuning_playbook) instructions. The best learning rate, weight decay, and adam beta1 parameters for each Transformer model on each task are identified.

* **Step 2**. The best hyperparameters from step 1 are used to fine-tune each model 40 times with different random seeds for up to one epoch ( [Dodge et al. (2020)](https://arxiv.org/abs/2002.06305) ). The 10 best models after the first epoch are selected for the next step.

* **Step 3**. The top 10 models from step 2 are fine-tuned for 20 epochs, generating predictions for the test set ( [Mosbach et al. (2021)](https://arxiv.org/abs/2006.04884) ). We select the model that is closest to the average of predictions (for regression tasks) or the mode of predictions (for classification tasks). This final model is then uploaded to the Hugging Face Hub and displayed in the tables below.

### Links and Results  

Our fine-tuning procedure has achieved results that are either slightly superior or at the same level as the previous state-of-the-art (if any).
Below is a summary of the results achieved on each dataset.

#### [ASSIN 2](huggingface.co/datasets/assin2) - STS ( Semantic Textual Similarity )

| Model                                                    | Pearson | MSE  |
|----------------------------------------------------------|---------|------|
| **ruanchaves/bert-large-portuguese-cased-assin2-similarity** | **0.86**    | **0.48** |
| Previous SOTA ( for Pearson ) - [Souza et al. (2020)](https://link.springer.com/chapter/10.1007/978-3-030-61377-8_28/tables/2)                           | 0.852   | 0.50 |
| SOTA ( for MSE ) - [Stilingue](https://link.springer.com/chapter/10.1007/978-3-030-41505-1_39/tables/2)                               | 0.817   | **0.47** |
| [ruanchaves/mdeberta-v3-base-assin2-similarity](https://huggingface.co/ruanchaves/mdeberta-v3-base-assin2-similarity)            | 0.847   | 0.62 |
| [ruanchaves/bert-base-portuguese-cased-assin2-similarity](https://huggingface.co/ruanchaves/bert-base-portuguese-cased-assin2-similarity)    | 0.843   | 0.54 |

#### [ASSIN 2](huggingface.co/datasets/assin2) - RTE ( Recognizing Textual Entailment )

| Model                                                    | Accuracy | F1  |
|----------------------------------------------------------|---------|------|
| **[ruanchaves/bert-large-portuguese-cased-assin2-entailment](https://huggingface.co/ruanchaves/bert-large-portuguese-cased-assin2-entailment)** | **0.90**    | **0.90** |
| [ruanchaves/mdeberta-v3-base-assin2-entailment](https://huggingface.co/ruanchaves/mdeberta-v3-base-assin2-entailment)           | **0.90**   | **0.90** |
| Previous SOTA                           | **0.90**  | **0.90** |
| [ruanchaves/bert-base-portuguese-cased-assin2-entailment](https://huggingface.co/ruanchaves/bert-base-portuguese-cased-assin2-entailment) | 0.88   | 0.88 |

#### [ASSIN](https://huggingface.co/datasets/assin) - STS ( Semantic Textual Similarity )

| Model                                                    | Pearson | MSE  |
|----------------------------------------------------------|---------|------|
| [ruanchaves/bert-large-portuguese-cased-assin-similarity](https://huggingface.co/ruanchaves/bert-large-portuguese-cased-assin-similarity) | 0.859    | 0.3 |
| [ruanchaves/mdeberta-v3-base-assin-similarity](https://huggingface.co/ruanchaves/mdeberta-v3-base-assin-similarity)            | 0.855  | 0.39 |
| [ruanchaves/bert-base-portuguese-cased-assin-similarity](https://huggingface.co/ruanchaves/bert-base-portuguese-cased-assin-similarity)  | 0.847   | 0.33 |

#### [ASSIN](https://huggingface.co/datasets/assin) - RTE ( Recognizing Textual Entailment )

| Model                                                    | Accuracy | F1  |
|----------------------------------------------------------|---------|------|
| [ruanchaves/mdeberta-v3-base-assin-entailment](https://huggingface.co/ruanchaves/mdeberta-v3-base-assin-entailment)           | 0.927   | 0.862 |
| [ruanchaves/bert-large-portuguese-cased-assin-entailment](https://huggingface.co/ruanchaves/bert-large-portuguese-cased-assin-entailment) | 0.92    | 0.828 |
| [ruanchaves/bert-base-portuguese-cased-assin-entailment](https://huggingface.co/ruanchaves/bert-base-portuguese-cased-assin-entailment) | 0.92   | 0.827 |

#### [HateBR](https://huggingface.co/datasets/ruanchaves/hatebr) ( Offensive Language Detection )


| Model                                                    | Accuracy | F1  |
|----------------------------------------------------------|---------|------|
| [ruanchaves/bert-large-portuguese-cased-hatebr](https://huggingface.co/ruanchaves/bert-large-portuguese-cased-hatebr) | 0.928    | 0.928 |
| [ruanchaves/mdeberta-v3-base-hatebr](https://huggingface.co/ruanchaves/mdeberta-v3-base-hatebr)          | 0.916 | 0.916 |
| [ruanchaves/bert-base-portuguese-cased-hatebr](https://huggingface.co/ruanchaves/bert-base-portuguese-cased-hatebr) | 0.914 | 0.914 |

#### [FaQUaD-NLI](https://huggingface.co/datasets/ruanchaves/faquad-nli) ( Question Answering )

| Model                                                    | Accuracy | F1  |
|----------------------------------------------------------|---------|------|
| [ruanchaves/bert-large-portuguese-cased-faquad-nli](https://huggingface.co/ruanchaves/bert-large-portuguese-cased-faquad-nli) | 0.929  | 0.93 |
| [ruanchaves/mdeberta-v3-base-faquad-nli](https://huggingface.co/ruanchaves/mdeberta-v3-base-faquad-nli)          | 0.926 | 0.926 |
| [ruanchaves/bert-base-portuguese-cased-faquad-nli](https://huggingface.co/ruanchaves/bert-base-portuguese-cased-faquad-nli) | 0.92 | 0.883 |

#### [PorSimplesSent](https://huggingface.co/datasets/ruanchaves/porsimplessent) ( Text Simplification )

| Model                                                    | Accuracy | F1  |
|----------------------------------------------------------|---------|------|
| [ruanchaves/mdeberta-v3-base-porsimplessent](https://huggingface.co/ruanchaves/mdeberta-v3-base-porsimplessent)          | 0.96 | 0.956 |
| [ruanchaves/bert-base-portuguese-cased-porsimplessent](https://huggingface.co/ruanchaves/bert-base-portuguese-cased-porsimplessent) | 0.942 | 0.937 |
| [ruanchaves/bert-large-portuguese-cased-porsimplessent](https://huggingface.co/ruanchaves/bert-large-portuguese-cased-porsimplessent)  | 0.921  | 0.913 |

## Citation

Our research is ongoing, and we are currently working on describing our experiments in a paper, which will be published soon. In the meanwhile, if you would like to cite our work or models before the publication of the paper, please use the following BibTeX citation for this repository: 

```
@software{Chaves_Rodrigues_eplm_2023,
author = {Chaves Rodrigues, Ruan and Tanti, Marc and Agerri, Rodrigo},
doi = {10.5281/zenodo.7781848},
month = {3},
title = {{Evaluation of Portuguese Language Models}},
url = {https://github.com/ruanchaves/napolab},
version = {1.0.0},
year = {2023}
}
```
