Subnet 102 is a Bittensor subnet for collaborative, decentralized training of large language models using a Mixture-of-Experts (MoE) architecture. Rather than training a single monolithic model on one machine, the network splits the model into expert groups distributed across many independent miners. Validators coordinate the process, aggregate contributions, and score miners to drive TAO rewards.

For the most up-to-date and comprehensive documentation, including architecture overviews, setup guides for validators and miners, and reward optimization, please visit our official website:

Contributions are welcome. Please open an issue first to discuss significant changes.

See LICENSE for details.