Show HN: Trained a 12M transformer on an ML framework we built from scratch

We’re two second-year CS students, and over the last 4 months we built a ML framework in Rust+CUDA with a TypeScript API on top (and a WebGPU fallback)Our goal is to fully understand the ML stack end-to-end. To force ourselves to use it for something real, we also trained a 12M-parameter...