🌐 LinguaNet

Welcome to LinguaNet, the cutting-edge translation subnet developed under Commune AI! 🌟 LinguaNet specializes in multimodal cross-transform translation between text and speech. This means LinguaNet facilitates seamless communication by allowing text-to-speech, speech-to-text, and language translation in a variety of forms. 💬🎙️

📂 Repository

Explore the source code and contribute at: LinguaNet GitHub Repository

🌟 Why LinguaNet Matters

In an increasingly globalized world, effective communication across languages and mediums is essential. LinguaNet offers solutions for:

• 🌍 Global Accessibility: Break language barriers by providing accurate translations for both text and speech.
• 💡 Multimodal Communication: Enable applications to handle dynamic modes of interaction, such as voice assistants understanding and responding in multiple languages.
• ✨ Enhanced User Experience: Power multilingual customer support systems, enabling businesses to reach diverse audiences.
• 🤝 Universal Inclusion: Provide tools for the hearing or visually impaired by converting speech to text or text to speech.
• 🌟 Support for 103 Languages: LinguaNet supports an impressive 103 diverse languages, making it one of the most versatile translation subnets available. Whether you’re working with widely spoken languages like English and Spanish or lesser-known ones, LinguaNet ensures high-quality, real-time translations. 🌏🗣️

LinguaNet is your ultimate partner in building smarter apps, expanding your reach to global audiences, and enabling seamless interactions across cultures and regions. 🚀

🚀 How to Run

🛠️ Running the Validator

To run the validator, execute:

python3 -m src.validator.cli <name-of-your-com-key> [--netuid <number>] [--call_timeout <number>] [--use-testnet]

⚙️ Running the Miner

To run the miner, execute:

python3 -m src.miner.cli <name-of-your-com-key> [--netuid <number>] [--ip <text>] [--port <number>] [--use-testnet]

🔄 Running with PM2

To ensure reliable and continuous execution of your validator or miner using PM2:

1. Install PM2 if not already installed:

   npm install -g pm2

2. Start the validator or miner using PM2:

   pm2 start "python3 -m src.validator.cli <name-of-your-com-key> --netuid <number>" --name linguanet-validator

   pm2 start "python3 -m src.miner.cli <name-of-your-com-key> --netuid <number>" --name linguanet-miner

3. Monitor your processes:

   pm2 list

🖥️ Device Requirements

Miner

• Minimum GPU: GeForce RTX 3090 🎮

Validator

• Minimum GPU: GeForce RTX 3090 🎮
• Recommended GPU: NVIDIA RTX A6000 💪

🔑 Key Features

⚡ Advanced Translation Models

• Miner Models: The default model for miners is facebook/seamless-M4T-V2-large, known for its versatility in multilingual and multimodal translation tasks.
• Query Generation: Validators generate queries using LLM models, specifically Llama (with plans to integrate more models in the future). 🦙
• Scoring Mechanism: Miners are scored based on responses evaluated by Llama models, ensuring high-quality and accurate translations.

🛤️ Roadmap

We are committed to expanding LinguaNet’s capabilities by integrating additional LLM models and refining the scoring mechanisms to better serve diverse use cases. 🚀

🤝 How to Contribute

Interested in joining the LinguaNet revolution? Here’s how you can contribute:

1. Fork the repository from GitHub. 🍴
2. Create feature branches to implement new models, optimize translation processes, or enhance validator and miner tools. 🌟
3. Submit pull requests with detailed descriptions of your contributions. 📜
4. Collaborate with the team to ensure seamless integration of your features. 🤝

🎉 Cheers to You!

Thank you for exploring LinguaNet! 🌟 Together, we can create a truly global and inclusive communication system. Let’s break down barriers and connect the world, one translation at a time! 🌍💬