NLLB-200 Fine-Tuned for Kimbundu → Portuguese

Model Description

• Base model: facebook/nllb-200-distilled-600M
• Task: Neural Machine Translation
• Languages: Kimbundu → Portuguese
• Fine-tuned by: Lirio Ramalheira

Intended Uses

Primary Intended Uses

• Automatic translation from Kimbundu to Portuguese
• Linguistic and NLP research on Bantu and Angolan languages
• Educational and cultural preservation tools

Out-of-Scope Uses

• Legal, medical, or governmental translation without human review
• High-stakes or safety-critical applications
• Dialects or orthographies not covered in the training data

Training Data

• Dataset type: Parallel corpus
• Languages: Kimbundu (kmb), Portuguese (pt)
• Source: Public-domain, educational, and community-translated texts
• Synthetic data: Not used

Data Splits

Preprocessing

• Sentence-level alignment
• Unicode normalization
• Duplicate and noisy pair removal
• Enforcement of NLLB language codes (kmb_Latn, por_Latn)

Training Procedure

Training Setup

• Framework: Hugging Face Transformers
• Platform: Google Colab (GPU)
• Mixed precision: FP16

Hyperparameters

• Optimizer: AdamW
• Learning rate: 1.5 × 10−4
• Batch size: 16
• Epochs: 20
• Max sequence length: 256
• Gradient accumulation: Enabled

Evaluation

Metrics

• BLEU
• chrF++
• COMET
• AfriCOMET
• BERTScore
• Len. Ratio

Results (Kimbundu → Portuguese)

Comparison with Base Model

• Lexical choice and fluency
• Morphological agreement
• Preservation of named entities
• Translation of culturally specific expressions

Limitations

• Limited corpus size due to data scarcity
• Partial coverage of Kimbundu dialectal variation
• Reduced performance on informal or spoken language

Ethical Considerations

• The dataset does not contain personal or sensitive data
• Cultural and topical biases present in source texts may be reflected in outputs
• Users should validate translations before real-world deployment
• Due to the scarcity of digital data in the Kimbundu language, this work used diverse sources for the purposes of cultural preservation and non-profit linguistic research. The dataset is made available under a scientific use license, respecting the integrity of the original sources through data fragmentation.
• Purpose: This dataset was compiled exclusively for the purposes of linguistic preservation, scientific research and development of Natural Language Processing (NLP) technologies for the Kimbundu language.
• Copyright: This dataset contains fragments of texts from different sources (literary, musical works and information platforms). Original copyright belongs to their respective holders.
• Responsibility: The author of this dataset does not claim ownership of third-party content. By downloading or using this data, you assume full responsibility for any copyright infringements that may arise from commercial use or improper redistribution.
• Removal: If you own the rights to any content here and wish to have it removed, please contact lirio.ramalheira1@gmail.com.

Recommendations

• Apply domain-specific fine-tuning for specialized applications
• Use human post-editing for high-quality translation
• Extend the dataset with dialectal and oral data

Reproducibility

• Training scripts: Google Colab notebooks
• Dataset: Hugging Face Datasets
• Base model: facebook/nllb-200-distilled-600M

Citation