Skip to content

Latest commit

 

History

History
171 lines (151 loc) · 6.33 KB

VALIDATOR.md

File metadata and controls

171 lines (151 loc) · 6.33 KB

Validator Setup

Validator Diagram

Minimum Requirements

Step-by-step setup

  1. Install repository:
git clone https://github.com/corcel-api/cortex.t
cd cortex.t
pip install uv
uv venv
. .venv/bin/activate
uv sync
  1. Configure env variables:
export SUBTENSOR_NETWORK=finney
export SUBTENSOR_NETUID=18
export WALLET_NAME=default
export WALLET_HOTKEY=default
export AXON_PORT=8000
export OPENAI_API_KEY=your_openai_api_key
export OPENROUTER_API_KEY=your_openrouter_api_key
export CLAUDE_API_KEY=your_claude_api_key

To verify the exported variables:

python -m cortext
  1. Spin up necessary services:
Service Default Port Environment Variable Process Command
Redis 6379 REDIS_PORT & REDIS__PORT . scripts/install_redis.sh
ExifTool N/A N/A . scripts/install_exiftool.sh
Subtensor Sync 8104 W_SUBTENSOR__PORT pm2 start python --name "cortex_w_subtensor" -- -m services.subtensor_syncing.server
Scoring 8101 SCORE__PORT pm2 start python --name "cortex_scoring" -- -m services.scoring.server
Synthesizing 8102 SYNTHESIZE__PORT pm2 start python --name "cortex_synthesizing" -- -m services.synthesizing.server
Managing 8103 MINER_MANAGER__PORT pm2 start python --name "cortex_managing" -- -m services.managing.server
Synthesizing Worker N/A N/A pm2 start python --name "cortex_synthesizing_worker" -- -m services.synthesizing.refill_worker

You can modify the default ports by setting the corresponding environment variables before starting the services.

  1. Run main validating proccess:
pm2 start python --name "cortex_validating" -- -m neurons.validator
  1. Run the auto update script, it will check for updates every 30 minutes
pm2 start auto-update.sh --name "auto_updater"

Organic Serving [Optional]

  1. Set admin api key
export ADMIN_API_KEY=your_admin_api_key
  1. Run organic proxy server

By default, the organic proxy server is set to listen on port localhost:8105. You can modify the port by setting the ORGANIC__PORT environment variable.

pm2 start python --name "cortex_organic" -- -m services.organic.server
  1. Optional: Run organic frontend for managing API keys and Test generation
pm2 start python --name "cortex_organic_frontend" -- -m services.organic.frontend
Endpoint Method Description Request Parameters Authentication Response
/api/v1/chat/completions POST Process chat completion requests request: MinerPayload object API Key required Streaming response of chat completion chunks
/api/v1/keys POST Create new API key user_id: string
initial_credits: float (default: 100.0)
monthly_reset: boolean (default: true)		 Admin API Key required New APIKey object
/api/v1/keys GET Retrieve all API keys None Admin API Key required List of APIKey objects
/api/v1/keys/{key}/add-credits POST Add credits to existing API key key: string (path)
amount: float		 Admin API Key required Updated APIKey object
/api/v1/keys/{key}/status PATCH Update API key status key: string (path)
status_update: dict with is_active boolean		 Admin API Key required Updated APIKey object
/api/v1/keys/{key} DELETE Delete an API key key: string (path) Admin API Key required Success message

APIKey Object Structure:

{
    "key": string,
    "user_id": string,
    "created_at": datetime,
    "is_active": boolean,
    "permissions": list[string],
    "total_credits": decimal,
    "used_credits": decimal,
    "credit_reset_date": datetime (optional)
}

Notes:

  • All endpoints require authentication via the X-API-Key header
  • Admin operations require the admin API key (set via ADMIN_API_KEY environment variable)
  • The chat completions endpoint consumes credits based on the model configuration
  • Credits can be set to reset monthly if specified during key creation

Default Config

{
    'redis': {
        'host': 'localhost',
        'port': 6379,
        'db': 0,
        'organic_queue_key': 'organic_queue',
        'synthetic_queue_key': 'synthetic_queue',
        'miner_manager_key': 'node_manager'
    },
    'bandwidth': {
        'interval': 60,
        'min_stake': 10000,
        'model_configs': {
            'gpt-4o': {
                'credit': 1,
                'model': 'gpt-4o',
                'max_tokens': 8096,
                'synapse_type': 'streaming-chat',
                'timeout': 12,
                'allowed_params': ['messages', 'temperature', 'max_tokens', 'stream', 'model', 'seed']
            },
            'dall-e-3': {
                'credit': 1,
                'model': 'dall-e-3',
                'max_tokens': 1024,
                'synapse_type': 'streaming-chat',
                'timeout': 32,
                'allowed_params': ['prompt', 'n', 'size', 'response_format', 'user']
            },
            'claude-3-5-sonnet-20241022': {
                'credit': 1,
                'model': 'claude-3-5-sonnet-20241022',
                'max_tokens': 8096,
                'synapse_type': 'streaming-chat',
                'timeout': 12,
                'allowed_params': ['messages', 'temperature', 'max_tokens', 'stream', 'model']
            }
        },
        'min_credit': 128,
        'max_credit': 1024
    },
    'score': {'decay_factor': 0.9, 'host': 'localhost', 'port': 8101},
    'sql': {'url': 'sqlite:///miner_metadata.db'},
    'network': 'mainnet',
    'synthesize': {'host': 'localhost', 'port': 8102, 'synthetic_pool_size': 1024, 'organic_pool_size': 1024},
    'miner_manager': {'port': 8103, 'host': 'localhost'},
    'validating': {'synthetic_threshold': 0.2, 'synthetic_batch_size': 4, 'synthetic_concurrent_batches': 1},
    'w_subtensor': {'host': 'localhost', 'port': 8104},
    'organic': {'host': 'localhost', 'port': 8105},
    'subtensor_network': 'test',
    'subtensor_netuid': 245,
    'wallet_name': 'tnv',
    'wallet_hotkey': '0',
    'subtensor_tempo': 360,
    'axon_port': 8000
}