A readable PyTorch decoder-only Transformer for learning the full training and inference pipeline. Defaults: 2 layers, 4 attention heads, width 64, context 64 bytes, about 140k parameters. This is an ...
A complete AI and Web3 curriculum built from scratch. 158 AI phases. Built in Python and NumPy. Every neuron, every gradient, every matrix multiplication visible. 71 Web3 phases. Built in Rust and ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results