ESP32S3 cluster running 1.58-bit (BitNet) Language model

This project implements a distributed inference engine for a 0.5B parameter 1.58-bit BitNet model across a cluster of seven ESP32-S3 microcontrollers. A master node handles tokenization and embeddings, while six compute nodes execute transformer blocks via a high-speed SPI daisy-chain. The system utilizes assembly-optimized MAC operations to process ternary weights on resource-constrained hardware.