Model Card for lucio94c/ita_to_eng

This model is a fine-tuned neural machine translation (NMT) model for translating text from Italian to English using the Hugging Face 🤗 transformers library.

Model Details

Model Description

This model is based on the pre-trained Helsinki-NLP/opus-mt-it-en MarianMT architecture and has been fine-tuned on additional parallel Italian–English data for improved translation quality on general and informal texts.

Developed by: Lucio Ciracì
Model type: Sequence-to-sequence neural translation (MarianMT)
Language(s) (NLP): Italian → English
License: MIT (or match the original model’s license if not MIT)
Finetuned from model: Helsinki-NLP/opus-mt-it-en

Model Sources

Repository: https://huggingface.co/lucio94c/ita_to_eng
Paper [optional]: MarianMT: Fast Neural Machine Translation in C++

Uses

Direct Use

Translate Italian text into English for documents, emails, or general understanding.
Useful for language learners, travelers, or anyone needing fast Italian-to-English translation.

Downstream Use

Can be fine-tuned further for specific domains (e.g., legal, medical, technical).

Out-of-Scope Use

Not suitable for translating highly domain-specific or sensitive content without further fine-tuning.
Not designed for translating code or non-natural language text.

Bias, Risks, and Limitations

The model may reflect biases present in the training data.
Quality may decrease on slang, dialects, or specialized terminology.
Automatic translations can contain errors; always review important text.

Recommendations

Review translations, especially for professional or sensitive contexts.
Do not use for critical translations (medical, legal, etc.) without human verification.

How to Get Started with the Model

from transformers import AutoModelForSeq2SeqLM, AutoTokenizer

model = AutoModelForSeq2SeqLM.from_pretrained("lucio94c/ita_to_eng")
tokenizer = AutoTokenizer.from_pretrained("lucio94c/ita_to_eng")

text = "Ciao, come stai?"
inputs = tokenizer(text, return_tensors="pt")
output = model.generate(**inputs)
print(tokenizer.decode(output[0], skip_special_tokens=True))