README.md

---
library_name: transformers
tags:
- roman eng2nep
- translation
- transliteration
license: mit
datasets:
- syubraj/roman2nepali-transliteration
language:
- en
- ne
base_model:
- google-t5/t5-base
pipeline_tag: translation
new_version: syubraj/RomanEng2Nep-v2
---


This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.

- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [Roman Eng, Nep]
- **License:** [MIT]
- **Finetuned from model [google-t5/t5-small]:** 

<!-- Provide the basic links for the model. -->


## How to Get Started with the Model

Use the code below to get started with the model.
```Python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

# Load your fine-tuned model and tokenizer
model_name = 'syubraj/romaneng2nep'
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSeq2SeqLM.from_pretrained(model_name)

# Set max sequence length
max_seq_len = 30

def translate(text):
    # Tokenize the input text with a max length of 30
    inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=max_seq_len)

    # Generate translation
    translated = model.generate(**inputs)

    # Decode the translated tokens back to text
    translated_text = tokenizer.decode(translated[0], skip_special_tokens=True)
    return translated_text

# Example usage
source_text = "timilai kasto cha?"  # Example Romanized Nepali text
translated_text = translate(source_text)
print(f"Translated Text: {translated_text}")

```


## Training Details
```Python
training_args = Seq2SeqTrainingArguments(
    output_dir="/kaggle/working/romaneng2nep/",
    eval_strategy="epoch",
    learning_rate=2e-5,
    per_device_train_batch_size=16,
    per_device_eval_batch_size=16,
    weight_decay=0.01,
    save_total_limit=3,
    num_train_epochs=3,
    predict_with_generate=True,
    fp16=True,
)
```

### Training Data

<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[syubraj/roman2nepali-transliteration](https://huggingface.co/datasets/syubraj/roman2nepali-transliteration)