hitty28/mbart-en-te

mBART Fine-tuned: English ↔ Telugu

This model is a fine-tuned version of facebook/mbart-large-50-many-to-many-mmt for English ↔ Telugu machine translation.

---

🚀 Usage

from transformers import MBartForConditionalGeneration, MBart50TokenizerFast

model_name = "your-username/mbart-en-te"

tokenizer = MBart50TokenizerFast.from_pretrained(model_name)
model = MBartForConditionalGeneration.from_pretrained(model_name)

# English → Telugu
text = "How are you?"
inputs = tokenizer(text, return_tensors="pt")
tokenizer.src_lang = "en_XX"
generated = model.generate(**inputs, forced_bos_token_id=tokenizer.lang_code_to_id["te_IN"])
print(tokenizer.decode(generated[0], skip_special_tokens=True))
# "మీరు ఎలా ఉన్నారు?"

# Telugu → English
text = "మీరు ఎలా ఉన్నారు?"
inputs = tokenizer(text, return_tensors="pt")
tokenizer.src_lang = "te_IN"
generated = model.generate(**inputs, forced_bos_token_id=tokenizer.lang_code_to_id["en_XX"])
print(tokenizer.decode(generated[0], skip_special_tokens=True))
# "How are you?"
Training

Base model: facebook/mbart-large-50-many-to-many-mmt

Task: English ↔ Telugu translation

Framework: Hugging Face Transformers (PyTorch)

Training setup: [Add epochs, batch size, learning rate here]

Dataset: [Mention dataset name or source]

⚠️ Limitations

Works best on short to medium sentences.

May struggle with idiomatic or domain-specific text.