app.py

import os
import torch
import gradio as gr
import time
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, pipeline

# Define flores_codes dictionary
flores_codes = {
    "Standard Tibetan": "bod_Tibt",
    "English": "eng_Latn"
}

def load_models():
    # build model and tokenizer
    model_name_dict = {
        'nllb-biboen-1': 'TenzinGayche/nllb_600M_bi_boen',
        'nllb-biboen-2': 'TenzinGayche/nllb_600M_bi_boen_gold',
        'nllb-biboen-3': 'TenzinGayche/nllb_600M_bi_boen_3',
    }

    model_dict = {}

    for call_name, real_name in model_name_dict.items():
        print('\tLoading model: %s' % call_name)
        model = AutoModelForSeq2SeqLM.from_pretrained(real_name)
        tokenizer = AutoTokenizer.from_pretrained("facebook/nllb-200-distilled-600M")
        model_dict[call_name + '_model'] = model
        model_dict[call_name + '_tokenizer'] = tokenizer

    return model_dict


def translation(model_name, source, target, text):
    start_time = time.time()
    source = flores_codes[source]
    target = flores_codes[target]

    model = model_dict[model_name + '_model']
    tokenizer = model_dict[model_name + '_tokenizer']

    # Check if a GPU is available and set device accordingly
    device = 0 if torch.cuda.is_available() else -1

    translator = pipeline('translation', model=model, tokenizer=tokenizer, src_lang=source, tgt_lang=target, device=device)
    output = translator(text, max_length=400)

    end_time = time.time()

    output = output[0]['translation_text']
    return output


if __name__ == '__main__':
    print('\tinit models')

    global model_dict

    model_dict = load_models()
    
    # Define gradio demo
    lang_codes = list(flores_codes.keys())
    with gr.Blocks() as demo:
        gr.Markdown("# NLLB Distilled 600M Translation Demo")
        gr.Markdown("This demo allows you to test the translation models for English to Standard Tibetan and vice versa.")

        with gr.Row():
            with gr.Column():
                model_input = gr.Radio(['nllb-biboen-1', 'nllb-biboen-2','nllb-biboen-3'], label='Select NLLB Model')
                source_lang = gr.Dropdown(lang_codes, value='English', label='Source Language')
                target_lang = gr.Dropdown(lang_codes, value='Standard Tibetan', label='Target Language')
                input_text = gr.Textbox(lines=5, label="Input Text", placeholder="Enter the text you want to translate")

            with gr.Column():
                output_text = gr.Textbox(lines=5, label="Translated Text", interactive=False, placeholder="The translated text will appear here")

        def update_output(model_name, source, target, text):
            result = translation(model_name, source, target, text)
            output_text.value = result
            return result

        translate_button = gr.Button("Translate")
        translate_button.click(update_output, inputs=[model_input, source_lang, target_lang, input_text], outputs=output_text)

    demo.launch()