Violet Subnet - Audio Processing Network

Welcome to Violet, a cutting-edge Bittensor subnet dedicated to decentralized audio processing and AI-powered speech technologies. Our network leverages the power of distributed computing to provide high-quality transcription, text-to-speech synthesis, and audio analysis services.

🌟 Overview

Violet subnet enables a decentralized ecosystem where miners provide computational power for audio processing tasks while validators ensure quality and distribute rewards. The network integrates with a hosted proxy server for seamless task management and real-time processing.

🏗️ System Architecture

Key Components

1. Miners (Computational Nodes)

   • Audio Processing: Transcription, TTS, and audio analysis
   • Task Execution: Processes assigned tasks with AI models
   • Response Submission: Sends results back to proxy server
   • Performance Monitoring: Tracks processing times and accuracy

2. Validators (Quality Assurance)

   • Response Evaluation: Assesses miner response quality
   • Reward Distribution: Calculates and distributes TAO rewards
   • Network Health: Monitors miner performance and availability

🔐 Security & Best Practices

• Never commit .env files - They contain sensitive credentials
• Use virtual environments - Isolate dependencies for each project
• Keep API keys secure - Rotate keys if compromised
• Monitor GPU usage - Ensure adequate cooling and power supply
• Regular updates - Keep dependencies and codebase up to date

💻 System Requirements

Operating System

• Linux (Ubuntu 20.04+ recommended)
• macOS (10.15+)

Hardware Requirements

For Miners

• RAM: 12 GB minimum
• Storage: 500 GB SSD
• GPU: NVIDIA GPU with 12 GB VRAM minimum (recommended for TTS tasks)
• Network: Stable internet connection (100+ Mbps)

For Validators

• RAM: 12 GB minimum
• Storage: 500 GB SSD
• GPU: NVIDIA GPU with 12 GB VRAM minimum
• Network: Stable internet connection (100+ Mbps)

Software Dependencies

• Python 3.10 (Required - Coqui TTS does not support Python 3.12)
• CUDA 11.8+ (for GPU acceleration)
• Bittensor CLI (pip install bittensor-cli)

⚠️ Important: Coqui TTS (used for text-to-speech) requires Python 3.9-3.11. Python 3.12 is NOT supported. You MUST use Python 3.10 in a virtual environment.

🚀 Quick Start

Step 1: Clone the Repository

git clone https://github.com/hivetrainai/violet.git
cd violet

Step 2: Create Python 3.10 Virtual Environment

⚠️ CRITICAL: Coqui TTS requires Python 3.10. You MUST create a Python 3.10 virtual environment before installing dependencies.

# Check if Python 3.10 is available
python3.10 --version

# If Python 3.10 is not installed, install it first:
# On macOS (using Homebrew):
# brew install python@3.10
# On Ubuntu/Debian:
# sudo apt-get install python3.10 python3.10-venv

# Create Python 3.10 virtual environment
python3.10 -m venv venv_py310

# Activate virtual environment
# On Linux/macOS:
source venv_py310/bin/activate
# On Windows:
venv_py310\Scripts\activate

# Verify Python version (should show 3.10.x)
python --version

Step 3: Install Dependencies

⚠️ IMPORTANT: Ensure you are in the Python 3.10 virtual environment before installing:

# Verify you're in Python 3.10 virtual environment
python --version  # Should show Python 3.10.x

# Upgrade pip and setuptools
pip install --upgrade pip setuptools wheel

# IMPORTANT: Handle numpy/bittensor conflict
# bittensor==9.10.0 requires numpy>=2.0.1, but Coqui TTS requires numpy<1.25
# Install in this specific order to resolve the conflict:

# 1. Install numpy first (required for Coqui TTS/numba, despite coqui-tts declaring numpy>=1.26.0)
pip install "numpy>=1.24.0,<1.25.0"

# 2. Install bittensor without dependencies to avoid numpy conflict
pip install --no-deps bittensor==9.10.0

# 3. Install bittensor's dependencies (excluding numpy which is already installed)
pip install aiohttp async-substrate-interface asyncstdlib bittensor-drand bittensor-wallet colorama fastapi msgpack-numpy-opentensor munch nest_asyncio netaddr packaging pycryptodome pydantic python-statemachine pyyaml requests retry scalecodec setuptools uvicorn wheel

# 4. Install numba and llvmlite (required by coqui-tts, but install after numpy to avoid conflicts)
pip install "numba>=0.58.0" "llvmlite>=0.40.0"

# 5. Install coqui-tts without dependencies to avoid numpy/numba conflict
#    (coqui-tts declares numpy>=1.26.0 and numba>=0.58.0, but we use numpy 1.24.x)
pip install --no-deps coqui-tts>=0.27.0

# 6. Install coqui-tts dependencies manually
pip install coqui-tts-trainer>=0.3.0 coqpit-config>=0.2.0 einops>=0.6.0 encodec>=0.1.1 monotonic-alignment-search>=0.2.0 num2words>=0.5.14 pysbd>=0.3.4 inflect>=5.6.0 anyascii>=0.3.0 matplotlib>=3.8.4 "gruut[de,es,fr]>=2.4.0" ko-speech-tools>=0.1.0 torchcodec

# 7. Install remaining requirements (bittensor and coqui-tts are commented out in requirements.txt)
pip install -r requirements.txt

# Note: Skip "pip install -e ." to avoid additional conflicts with setup.py

Step 4: Install System Dependencies

Install FFmpeg and related libraries (required for audio/video processing):

# On Ubuntu/Debian:
sudo apt install -y ffmpeg libavutil-dev libavcodec-dev libavformat-dev libavdevice-dev

# On macOS (using Homebrew):
# brew install ffmpeg

Step 5: Set Up Environment Variables

Create a .env file in the project root directory with the following required variables:

# Required: Hugging Face Token for model downloads
HF_TOKEN=your_huggingface_token_here

# Required: API Key for miner (get from subnet admin)
MINER_API_KEY=your_miner_api_key_here

# Required: API Key for validator (get from subnet admin)
VALIDATOR_API_KEY=your_validator_api_key_here

# Optional: Proxy server URL (defaults to production)
PROXY_SERVER_URL=https://violet-proxy-bl4w.onrender.com

Getting Your Tokens and API Keys

1. Hugging Face Token:

   • Go to https://huggingface.co/settings/tokens
   • Create a new token with "read" permissions
   • Copy the token and add it to your .env file as HF_TOKEN

2. Miner/Validator API Keys:

   • Contact the subnet administrator to obtain your API keys
   • Add them to your .env file as MINER_API_KEY and/or VALIDATOR_API_KEY

Important: Never commit your .env file to version control. It contains sensitive credentials.

Step 6: Create Wallets

Create a coldkey and hotkey for your subnet wallet:

# Install bittensor CLI (if not already installed)
pip install bittensor-cli

# Create a coldkey for the validator/miner
btcli wallet new_coldkey --wallet.name <your_wallet_name>

# Create a hotkey for the validator/miner
btcli wallet new_hotkey --wallet.name <your_wallet_name> --wallet.hotkey default

Step 7: Register on the Subnet

Register as a miner or validator on the subnet:

Mainnet (NetUID TBC):

btcli subnet register --netuid 49 --subtensor.network finney --wallet.name <your_wallet_name> --wallet.hotkey default

Testnet (NetUID 292):

btcli subnet register --netuid 292 --subtensor.network test --wallet.name <your_wallet_name> --wallet.hotkey default

Note: Registration requires 0.005 TAO for mainnet or test tokens for testnet.

Step 8: Run the Miner

⚠️ IMPORTANT: Before running the miner, ensure you are in the Python 3.10 virtual environment:

# Activate Python 3.10 virtual environment
source venv_py310/bin/activate  # Linux/macOS
# or
venv_py310\Scripts\activate  # Windows

# Verify you're using Python 3.10
python --version  # Should show Python 3.10.x

Mainnet:

python neurons/miner.py \
  --netuid TBC \
  --subtensor.network finney \
  --wallet.name <your_wallet_name> \
  --wallet.hotkey <your_hotkey> \
  --logging.debug \
  --axon.ip 0.0.0.0 \
  --axon.port <PORT> \
  --axon.external_ip <YOUR_PUBLIC_IP> \
  --axon.external_port <PORT>

Testnet:

python neurons/miner.py \
  --netuid 292 \
  --subtensor.network test \
  --wallet.name <your_wallet_name> \
  --wallet.hotkey <your_hotkey> \
  --logging.debug \
  --axon.ip 0.0.0.0 \
  --axon.port <PORT> \
  --axon.external_ip <YOUR_PUBLIC_IP> \
  --axon.external_port <PORT>

Step 9: Run the Validator

⚠️ IMPORTANT: Before running the validator, ensure you are in the Python 3.10 virtual environment:

# Activate Python 3.10 virtual environment
source venv_py310/bin/activate  # Linux/macOS
# or
venv_py310\Scripts\activate  # Windows

# Verify you're using Python 3.10
python --version  # Should show Python 3.10.x

Mainnet:

python neurons/validator.py \
  --netuid TBC \
  --subtensor.network finney \
  --wallet.name <your_wallet_name> \
  --wallet.hotkey <your_hotkey> \
  --logging.debug \
  --axon.ip 0.0.0.0 \
  --axon.port <PORT> \
  --axon.external_ip <YOUR_PUBLIC_IP> \
  --axon.external_port <PORT>

Testnet:

python neurons/validator.py \
  --netuid 292 \
  --subtensor.network test \
  --wallet.name <your_wallet_name> \
  --wallet.hotkey <your_hotkey> \
  --logging.debug \
  --axon.ip 0.0.0.0 \
  --axon.port <PORT> \
  --axon.external_ip <YOUR_PUBLIC_IP> \
  --axon.external_port <PORT>

📋 Network Configuration

Network	NetUID	Network Name	Registration Cost
Mainnet	TBC	finney	0.005 TAO
Testnet	292	test	0.005 test tokens

🔧 Configuration

GPU Configuration

Ensure your GPU is properly configured for CUDA:

# Check CUDA installation
nvidia-smi

# Verify PyTorch CUDA support
python -c "import torch; print(f'CUDA available: {torch.cuda.is_available()}')"
python -c "import torch; print(f'CUDA version: {torch.version.cuda}')"

Environment Variables Summary

Variable	Required	Description
HF_TOKEN	✅ Yes	Hugging Face token for downloading AI models
MINER_API_KEY	✅ Yes (for miners)	API key for miner authentication
VALIDATOR_API_KEY	✅ Yes (for validators)	API key for validator authentication
PROXY_SERVER_URL	❌ No	Proxy server URL (defaults to production)

📄 License

This project is licensed under the MIT License.

🆘 Support

For support and questions:

• Create an issue on GitHub
• Check the troubleshooting section
• Review the API documentation

---

Built with ❤️ for the Bittensor community