MANTIS

Clone the MANTIS subnet repository

git clone https://github.com/opentensor/validators.git && cd validators/subnet_123_mantis

Install dependencies including timelock-wasm from source (NOT via pip)

./install_reqs.sh

Install Python dependencies (torch 2.7.1, scikit-learn, boto3, etc.)

pip install -r requirements.txt

Configure environment: HOTKEY, WALLET, R2 credentials (AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY), FLOW_COLLATERAL_ADDRESS if trading FLOW

cp .env.example .env && nano .env

Generate and encrypt embeddings for a challenge; writes V2 JSON with dual encryption (owner + timelock drand)

python generate_and_encrypt.py --challenge ETH-1H-BINARY --payload-file embeddings.json

Run miner: samples block chain, encrypts embeddings, posts commits to R2 on schedule

python neurons/miner.py --netuid 123

Daemonize miner with PM2 (optional but recommended for production)(optional)

pm2 start neurons/miner.py --name miner -- --netuid 123

Open a FLOW collateral slot and fund it (required only if submitting FLOW bracket trades)(optional)

python flow_post.py fund --hotkey myhotkey --amount 0.5 --chain mainnet

Clone the MANTIS subnet repository

git clone https://github.com/opentensor/validators.git && cd validators/subnet_123_mantis

Install dependencies including timelock-wasm from source

./install_reqs.sh

Install Python dependencies

pip install -r requirements.txt

Configure: HOTKEY, WALLET, FLOW_COLLATERAL_RPC (archive endpoint), FLOW_COLLATERAL_ADDRESS (live pool on chain 964)

cp .env.example .env && nano .env

Initialize SQLite datalog with challenge metadata and download remote archives (datalog.db, flow_datalog.db, latest_prices.json)

mkdir -p .storage && python -c 'from ledger import DataLog; DataLog.init_db()'

Run validator: samples miner payloads every SAMPLE_EVERY blocks, decrypts after maturity, computes per-challenge salience via model.multi_salience(), applies EMA, sets weights on-chain

python neurons/validator.py --netuid 123

Daemonize validator with PM2 for production(optional)

pm2 start neurons/validator.py --name validator -- --netuid 123

Set custom weight-setting interval (default WEIGHT_SET_INTERVAL=360 blocks)(optional)

pm2 start python neurons/validator.py -- --netuid 123 --update-interval 360