google-deepmind/tracr

A compiler that transforms computation graphs into the weights of a transformer

Compile programs directly into transformer weights. Includes a 2D convex-hull KV cache with O(log n) inference.

LLM-based, self-correcting transpiler. Supports JAX, PyTorch, Rust, PyMC, Stan.

A complete transformer engine in C — inference, training, chat, vision.

A place to store reusable transformer components of my own creation or found on the interwebs

RoNet 2026: Build Transformer Models and Byte-Level BPE Tokenizers in Pure Tensors

An transformer based LLM. Written completely in Rust

A neuro-symbolic micro-Transformer with speculative decoding, constraint pruning, recurrent attention, and adaptive test-time scaling — built in Rust.