---
language:
- pt
language_creators:
- found
annotations_creators:
- no-annotation
multilinguality:
- monolingual
pretty_name: Camoniana
license: other
task_categories:
- text-generation
size_categories:
- 10K<n<100K
tags:
- text
- portuguese
- portuguese-literature
- literature
- poetry
- camoes
- luis-de-camoes
- character-level
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train.txt
  - split: validation
    path: data/validation.txt
- config_name: full
  data_files:
  - split: train
    path: data/full.txt
---

# Camoniana

Camoniana is a small Portuguese literary corpus of works attributed to Luís de
Camões, prepared for character-level language modeling.

## Dataset Details

- **Language:** Portuguese (`pt`)
- **Authorial focus:** works attributed to Luís de Camões.
- **Modality:** text
- **Primary task:** text generation / character-level language modeling
- **Annotations:** none
- **Tokenizer assumption:** one Unicode character is one token.

## Files

```text
README.md
metadata.json
vocab.json
data/full.txt
data/train.txt
data/validation.txt
```

- `data/full.txt`: full corpus text.
- `data/train.txt`: first 90% of the corpus by character offset.
- `data/validation.txt`: final 10% of the corpus by character offset.
- `metadata.json`: corpus statistics and split metadata.
- `vocab.json`: character vocabulary and integer mappings for character-level modeling.

Binary training artifacts are intentionally not included here. The canonical dataset is the text plus metadata in this repository.

## Dataset Structure

The default configuration exposes the training and validation splits:

```python
from datasets import load_dataset

ds = load_dataset("vreabernardo/camoniana")
```

To load the full corpus as one split:

```python
from datasets import load_dataset

full = load_dataset("vreabernardo/camoniana", "full")
```

For raw-file workflows, download `data/full.txt` directly.

## Splits and Statistics

```text
characters: 1,046,084
lines: 34,241
vocab size: 119
train characters: 941,475
validation characters: 104,609
split: contiguous 90/10 by character offset
```

The split is contiguous rather than shuffled so character-level training runs can reproduce the same setup.

## Intended Uses

- Training small character-level language models.
- Testing Portuguese text generation experiments.
- Reproducing character-level dataset preparation.

## Out-of-Scope Uses

- Treating the validation split as a benchmark for unrelated Portuguese language modeling systems.
- Assuming the corpus represents modern Portuguese usage.
- Using generated outputs as factual or scholarly editions of the source text.

## Data Creation

The corpus was prepared as a single UTF-8 text file, then split by character offset into train and validation text files. No labels or annotations were added.

## Limitations and Biases

This is a small literary corpus. It reflects the language, spelling, style, and editorial choices of its source material and should not be treated as representative of contemporary Portuguese.

## License and Provenance

The corpus contains works attributed to Luís de Camões (c. 1524-1580), including
Portuguese epic and lyric poetry. The authorial works are public-domain by age in
many jurisdictions.

This dataset is a UTF-8 plain-text compilation prepared for language-modeling
experiments. It includes normalized text files, train/validation splits,
character-vocabulary metadata, and corpus statistics.

The exact transcription and edition lineage for this release has not been
fully reconstructed. For that reason, the dataset is marked with `license:
other` rather than a permissive SPDX license. Users who need a legally precise
source chain should verify the transcription provenance and applicable local
public-domain rules before redistribution or commercial use.

## Citation

Use the exact Hugging Face revision you used. Replace `<revision-sha>` with the
commit hash shown on the dataset page.

### APA

```text
Vrea, B. (2026). Camoniana: A Portuguese literary corpus for character-level
language modeling [Data set]. Hugging Face.
https://huggingface.co/datasets/vreabernardo/camoniana
Revision: <revision-sha>
```

### BibTeX

```bibtex
@misc{vrea2026camoniana,
  title     = {Camoniana: A Portuguese Literary Corpus for Character-Level Language Modeling},
  author    = {Vrea, Bernardo},
  year      = {2026},
  publisher = {Hugging Face},
  url       = {https://huggingface.co/datasets/vreabernardo/camoniana},
  note      = {Revision: <revision-sha>}
}
```