MT-Informal-Languages/Helsinki_lg_inf_en

Luganda to English Informal Translation Model

This model translates informal Luganda sentences to English. It was trained on a dataset of Luganda proverbs with their English translations. The dataset consists of 3135 examples.

Data

Train: The training data consists of 3135 Luganda proverbs and their corresponding English translations. Eval: The evaluation data is part of the training data and consists of informal sentences.

Model

Architecture: Seq2Seq Pretrained Model: Helsinki-NLP/opus-mt-ug-en Fine-tuning: The model was fine-tuned for 50 epochs with a learning rate of 2e-5.

Translation

Source Language: Luganda Target Language: English Domain: Informal sentences and proverbs

Usage

Here is an example of how to load and use the model for translation:

from transformers import AutoModelForSeq2SeqLM, AutoTokenizer

model_name = 'your_model_name_on_hf_hub'
model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name)

# Example input sentence in Luganda
input_sentence = 'Olutalo lwa nsi yonna lwazibwa omwaka oguwedde.'

inputs = tokenizer(input_sentence, return_tensors='pt')
outputs = model.generate(**inputs)
translation = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(translation)