This tutorial will guide you through the process of setting up and running miners and validators for the AutoML network.
If you don't have a key on the bittensor network refer to this. You can refer to taostats.io to find out where to buy TAO to cover your registration cost.
git clone https://github.com/Hivetrain/DistributedAutoML
cd DistributedAutoML
sudo apt install git-lfs
git lfs install
pip install -r requirements.txt
pip install -e .
First, you need to register your miner on the network:
btcli s register
Edit the config files in the configs/ folder to set the following configurations:
- Bittensor Config:
netuid
: Set the network UID (100 for testnet, 49 for mainnet)wallet_name
: Set your wallet namewallet_hotkey
: Set your wallet hotkeynetwork
: Set to "test" for testnet or "finney" for mainnetsubtensor_chain_endpoint
: Edit if using your own subtensor node- General Config:
hf_token
: Set your Hugging Face tokengene_repo
: Set your Hugging Face repository name for storing genes (Miner only)- Miner Config:
- Good luck ! Hyperparameter optimization here might help performance
Execute the miner script:
No autoupdate:
python neurons/miner.py
Autoupdate (Make sure pm2 is installed):
pm2 start pm2_miner.json
Execute the validator script:
No autoupdate:
python neurons/validator.py
Autoupdate (Make sure pm2 is installed):
pm2 start pm2_validator.json
- Ensure you have the required dependencies installed. You may need to run
pip install -r requirements.txt
(if a requirements file is provided). - The
metrics_file
inconfig.py
specifies where performance metrics will be saved. - For both miners and validators, make sure you have sufficient balance in your wallet to pay for transaction fees.
- Monitor the console output and log files (
miner.log
for miners,validator.log
for validators) for any errors or important information. - The
device
setting inconfig.py
determines whether to use CPU or GPU. Set it to "cuda" if you want to use a GPU.
Remember to keep your wallet information and Hugging Face token secure and never share them publicly.