An incentivized intelligence template for running P2P large language models within the Hypertensor network with a POI (Proof of Inference) consensus mechanism, and POS (Proof of Stake) consensus mechanism.
Read the whitepaper
Read the full documentation to start an incentivized subnet validator node.
- Clone repository
- Create an
.env
file in the root directory by copying.env.example
and fill in the variables. - Add virtual environment (optional):
python -m venv .venv
source .venv/bin/activate
- Install:
python -m pip install .
- Generate Ed25519 private key (Required for validator node):
python -m subnet.cli.crypto.keygen
Start your own subnet:
python -m subnet.cli.run_server_validator bigscience/bloom-560m --host_maddrs /ip4/0.0.0.0/tcp/{PORT} ip4/0.0.0.0/udp/{PORT}/quic --announce_maddrs ip4/{IP}/tcp/{PORT}/ip4/{IP}/udp/{PORT}/quic --identity_path {PRIVATE_KEY_PATH} --new_swarm
Start your own subnet validator node:
python -m subnet.cli.run_server_validator bigscience/bloom-560m --public_ip {IP} --port {PORT} --initial_peers {INITIAL_PEERS} --identity_path {PRIVATE_KEY_PATH}
Instead of using --initial_peers
, the constants.py
file can be updated to include them. Read the full documentation or the cli
directory for all available arguments.