NLLB English-to-Persian Psychology Translation Model

This model is a fine-tuned version of facebook/nllb-200-distilled-600M for English-to-Persian translation in the psychology and mental-health domain.

The goal of this model is to produce fluent, natural, and domain-aware Persian translations for psychology-related English text.

Model Details

Intended Use

This model is intended for translating English psychology-related content into Persian, including:

• Psychology articles
• Mental-health educational content
• Counseling-related text
• Therapy-related explanations
• Self-help and personal development content
• Academic or semi-academic psychology material

Limitations

This model should not be used as a replacement for a professional translator, psychologist, psychiatrist, therapist, or medical expert.

Possible limitations include:

• Incorrect translation of specialized clinical terms
• Loss of nuance in emotionally sensitive language
• Overly literal translations
• Reduced performance outside the psychology domain
• Bias inherited from the base model or fine-tuning data

from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

model_name = "aradmanamnaoon/nllb-en-fa-psych-translation"

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

text = "Anxiety can affect both the body and the mind."

inputs = tokenizer(text, return_tensors="pt")

translated_tokens = model.generate( **inputs, forced_bos_token_id=tokenizer.convert_tokens_to_ids("pes_Arab"), max_length=128 )

translation = tokenizer.batch_decode( translated_tokens, skip_special_tokens=True )[0]

print(translation)

Example

Input

text Anxiety can affect both the body and the mind.

Output

اضطراب می‌تواند هم بر بدن و هم بر ذهن تأثیر بگذارد.

Training Data

This model was fine-tuned on English-Persian parallel text focused on psychology and mental-health content.

More details should be added when available:

Training Procedure

Evaluation

Evaluation results should be added after testing the model on a held-out English-Persian psychology translation set.

Ethical Considerations

This model works with psychology-related language, which may include sensitive mental-health topics.

Outputs should be reviewed by humans before use in professional, clinical, academic, legal, or high-stakes settings.

This model should not be used for diagnosis, treatment recommendations, crisis intervention, or medical advice.

Author

Fine-tuned and released by aradmanamnaoon.