import os import torch import spaces import psycopg2 import gradio as gr from threading import Thread from collections.abc import Iterator from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer MAX_MAX_NEW_TOKENS = 4096 MAX_INPUT_TOKEN_LENGTH = 4096 DEFAULT_MAX_NEW_TOKENS = 2048 HF_TOKEN = os.environ["HF_TOKEN"] model_id = "ai4bharat/IndicTrans3-beta" model = AutoModelForCausalLM.from_pretrained( model_id, torch_dtype=torch.float16, device_map="auto", token=HF_TOKEN ) tokenizer = AutoTokenizer.from_pretrained("meta-llama/Llama-3.2-3B-Instruct") LANGUAGES = [ "Hindi", "Bengali", "Telugu", "Marathi", "Tamil", "Urdu", "Gujarati", "Kannada", "Odia", "Malayalam", "Punjabi", "Assamese", "Maithili", "Santali", "Kashmiri", "Nepali", "Sindhi", "Konkani", "Dogri", "Manipuri", "Bodo", ] def format_message_for_translation(message, target_lang): return f"Translate the following text to {target_lang}: {message}" def store_feedback(rating, feedback_text, chat_history, tgt_lang): try: if not rating: gr.Warning("Please select a rating before submitting feedback.", duration=5) return None if not feedback_text or feedback_text.strip() == "": gr.Warning("Please provide some feedback before submitting.", duration=5) return None if not chat_history: gr.Warning( "Please provide the input text before submitting feedback.", duration=5 ) return None if len(chat_history[0]) < 2: gr.Warning( "Please translate the input text before submitting feedback.", duration=5, ) return None conn = psycopg2.connect( host=os.getenv("DB_HOST"), database=os.getenv("DB_NAME"), user=os.getenv("DB_USER"), password=os.getenv("DB_PASSWORD"), port=os.getenv("DB_PORT"), ) cursor = conn.cursor() insert_query = """ INSERT INTO feedback (tgt_lang, rating, feedback_txt, chat_history) VALUES (%s, %s, %s, %s) """ cursor.execute( insert_query, (tgt_lang, int(rating), feedback_text, chat_history) ) conn.commit() cursor.close() conn.close() gr.Info("Thank you for your feedback! 🙏", duration=5) except: gr.Error( "An error occurred while storing feedback. Please try again later.", duration=5, ) def store_output(tgt_lang, input_text, output_text): conn = psycopg2.connect( host=os.getenv("DB_HOST"), database=os.getenv("DB_NAME"), user=os.getenv("DB_USER"), password=os.getenv("DB_PASSWORD"), port=os.getenv("DB_PORT"), ) cursor = conn.cursor() insert_query = """ INSERT INTO translation (input_txt, output_txt, tgt_lang) VALUES (%s, %s, %s) """ cursor.execute(insert_query, (input_text, output_text, tgt_lang)) conn.commit() cursor.close() @spaces.GPU def translate_message( message: str, chat_history: list[dict], target_language: str = "Hindi", max_new_tokens: int = 1024, temperature: float = 0.6, top_p: float = 0.9, top_k: int = 50, repetition_penalty: float = 1.2, ) -> Iterator[str]: conversation = [] translation_request = format_message_for_translation(message, target_language) conversation.append({"role": "user", "content": translation_request}) input_ids = tokenizer.apply_chat_template( conversation, return_tensors="pt", add_generation_prompt=True ) if input_ids.shape[1] > MAX_INPUT_TOKEN_LENGTH: input_ids = input_ids[:, -MAX_INPUT_TOKEN_LENGTH:] gr.Warning( f"Trimmed input as it was longer than {MAX_INPUT_TOKEN_LENGTH} tokens." ) input_ids = input_ids.to(model.device) streamer = TextIteratorStreamer( tokenizer, timeout=240.0, skip_prompt=True, skip_special_tokens=True ) generate_kwargs = dict( {"input_ids": input_ids}, streamer=streamer, max_new_tokens=max_new_tokens, do_sample=True, top_p=top_p, top_k=top_k, temperature=temperature, num_beams=1, repetition_penalty=repetition_penalty, ) t = Thread(target=model.generate, kwargs=generate_kwargs) t.start() outputs = [] for text in streamer: outputs.append(text) yield "".join(outputs) store_output(target_language, message, "".join(outputs)) css = """ # body { # background-color: #f7f7f7; # } .feedback-section { margin-top: 30px; border-top: 1px solid #ddd; padding-top: 20px; } .container { max-width: 90%; margin: 0 auto; } .language-selector { margin-bottom: 20px; padding: 10px; background-color: #ffffff; border-radius: 8px; box-shadow: 0 2px 5px rgba(0,0,0,0.1); } .advanced-options { margin-top: 20px; } """ DESCRIPTION = """\ IndicTrans3 is the latest state-of-the-art (SOTA) translation model from AI4Bharat, designed to handle translations across 22 Indic languages with high accuracy. It supports document-level machine translation (MT) and is built to match the performance of other leading SOTA models.
📢 Training data will be released soon!

🔹 Features

✅ Supports 22 Indic languages
✅ Enables document-level translation
✅ Achieves SOTA performance in Indic MT
✅ Optimized for real-world applications

🚀 Try It Out!

1️⃣ Enter text in any supported language
2️⃣ Select the target language
3️⃣ Click Translate and get high-quality results!
Built for linguistic diversity and accessibility, IndicTrans3 is a major step forward in Indic language AI. 💡 Source: AI4Bharat | Powered by Hugging Face """ with gr.Blocks(css=css) as demo: with gr.Column(elem_classes="container"): gr.Markdown( "# 🌏 IndicTrans3-beta 🚀: Multilingual Translation for 22 Indic Languages " ) gr.Markdown(DESCRIPTION) target_language = gr.Dropdown( LANGUAGES, value="Hindi", label="Which language would you like to translate to?", elem_id="language-dropdown", ) chatbot = gr.Chatbot( height=400, elem_id="chatbot", show_copy_button=True, avatar_images=["avatars/user_logo.png", "avatars/ai4bharat_logo.png"] ) with gr.Row(): msg = gr.Textbox( placeholder="Enter a long paragraph to translate...", show_label=False, container=False, scale=9, ) submit_btn = gr.Button("Translate", scale=1) gr.Examples( examples=[ "The Taj Mahal, an architectural marvel of white marble, stands majestically along the banks of the Yamuna River in Agra, India. Built by Mughal Emperor Shah Jahan in memory of his beloved wife, Mumtaz Mahal, it symbolizes eternal love and devotion. The monument, a UNESCO World Heritage site, attracts millions of visitors each year, who admire its intricate carvings, calligraphy, and symmetrical gardens. At sunrise and sunset, the marble dome glows in hues of pink and golden, creating a breathtaking spectacle. The Taj Mahal is not only a masterpiece of Mughal architecture but also a timeless representation of romance and artistry.", "Kumbh Mela, the world’s largest spiritual gathering, is a significant Hindu festival held at four sacred riverbanks—Prayagraj, Haridwar, Nashik, and Ujjain—at intervals of 12 years. Millions of devotees, including sadhus, ascetics, and pilgrims, gather to take a holy dip in the river, believing it washes away sins and grants salvation. The festival is marked by grand processions, religious discourses, and vibrant cultural events. With its rich traditions, ancient rituals, and immense scale, Kumbh Mela is not just a religious event but also a profound representation of India’s spiritual and cultural heritage, fostering faith and unity among millions worldwide.", "India's classical dance forms, such as Bharatanatyam, Kathak, Odissi, Kuchipudi, and Kathakali, are deeply rooted in tradition and storytelling. These dance styles blend intricate footwork, graceful hand gestures, and expressive facial expressions to narrate mythological tales and historical legends. Bharatanatyam, originating from Tamil Nadu, is known for its rhythmic precision, while Kathak, from North India, features rapid spins and foot-tapping movements. Odissi, from Odisha, showcases fluid postures inspired by temple sculptures. Each form carries a distinct cultural essence, preserving centuries-old traditions while continuing to evolve in contemporary performances, keeping India’s rich artistic heritage alive and thriving.", "Ayurveda, India’s ancient medical system, emphasizes a holistic approach to health by balancing the mind, body, and spirit. Rooted in nature, it promotes well-being through herbal medicines, dietary guidelines, yoga, and meditation. Ayurveda classifies individuals based on three doshas—Vata, Pitta, and Kapha—determining their physical and mental constitution. Remedies include plant-based treatments, detox therapies, and rejuvenation practices to prevent and heal ailments. Unlike modern medicine, Ayurveda focuses on personalized healing and long-term wellness. With growing global interest in alternative medicine, Ayurveda continues to gain recognition for its effectiveness in promoting natural healing and overall health optimization.", "Diwali, the festival of lights, is one of India’s most celebrated festivals, symbolizing the victory of light over darkness and good over evil. Families clean and decorate their homes with colorful rangoli, oil lamps, and twinkling fairy lights. The festival marks the return of Lord Rama to Ayodhya after defeating Ravana, and it also honors Goddess Lakshmi, the deity of wealth and prosperity. Fireworks illuminate the night sky, while families exchange sweets and gifts, spreading joy and togetherness. Beyond its religious significance, Diwali fosters unity, strengthens relationships, and brings communities together in a spirit of happiness and renewal.", ], example_labels=[ "The Taj Mahal, an architectural marvel of white marble, stands majestically along the banks of the Yamuna River in Agra...", "Kumbh Mela, the world’s largest spiritual gathering, is a significant Hindu festival held at four sacred riverbanks...", "India's classical dance forms, such as Bharatanatyam, Kathak, Odissi, Kuchipudi, and Kathakali, are deeply rooted in tradition...", "Ayurveda, India’s ancient medical system, emphasizes a holistic approach to health by balancing the mind, body, and spirit...", "Diwali, the festival of lights, is one of India’s most celebrated festivals, symbolizing the victory of light over darkness...", ], inputs=msg, ) with gr.Accordion("Provide Feedback", open=True): gr.Markdown("## Rate Translation & Provide Feedback 📝") gr.Markdown( "Help us improve the translation quality by providing your feedback." ) with gr.Row(): rating = gr.Radio( ["1", "2", "3", "4", "5"], label="Translation Rating (1-5)" ) feedback_text = gr.Textbox( placeholder="Share your feedback about the translation...", label="Feedback", lines=3, ) feedback_submit = gr.Button("Submit Feedback") feedback_result = gr.Textbox(label="", visible=False) with gr.Accordion( "Advanced Options", open=False, elem_classes="advanced-options" ): max_new_tokens = gr.Slider( label="Max new tokens", minimum=1, maximum=MAX_MAX_NEW_TOKENS, step=1, value=DEFAULT_MAX_NEW_TOKENS, ) temperature = gr.Slider( label="Temperature", minimum=0.1, maximum=1.0, step=0.1, value=0.1, ) top_p = gr.Slider( label="Top-p (nucleus sampling)", minimum=0.05, maximum=1.0, step=0.05, value=0.9, ) top_k = gr.Slider( label="Top-k", minimum=1, maximum=100, step=1, value=50, ) repetition_penalty = gr.Slider( label="Repetition penalty", minimum=1.0, maximum=2.0, step=0.05, value=1.0, ) chat_state = gr.State([]) def user(user_message, history, target_lang): return "", history + [[user_message, None]] def bot( history, target_lang, max_tokens, temp, top_p_val, top_k_val, rep_penalty ): user_message = history[-1][0] history[-1][1] = "" for chunk in translate_message( user_message, history[:-1], target_lang, max_tokens, temp, top_p_val, top_k_val, rep_penalty, ): history[-1][1] = chunk yield history msg.submit( user, [msg, chatbot, target_language], [msg, chatbot], queue=False ).then( bot, [ chatbot, target_language, max_new_tokens, temperature, top_p, top_k, repetition_penalty, ], chatbot, ) submit_btn.click( user, [msg, chatbot, target_language], [msg, chatbot], queue=False ).then( bot, [ chatbot, target_language, max_new_tokens, temperature, top_p, top_k, repetition_penalty, ], chatbot, ) feedback_submit.click( fn=store_feedback, inputs=[rating, feedback_text, chatbot, target_language], ) if __name__ == "__main__": demo.launch()