Memory Math
TinyRustLM explains quantization as product architecture: q8_0 and q4_0 determine which tiny models can fit into browser memory without hiding compute costs.
01
The runtime data paths are intentionally limited. f32 is useful for tests and baselines, q8_0 is the balanced path for quality and compactness, and q4_0 exists for maximum model capacity under tight browser memory limits.
02
q4_0 saves bytes but costs instructions. Standard WASM does not provide native 4-bit tensor arithmetic, so the runtime must unpack nibbles with masks and shifts before accumulation. That makes q4_0 a memory win but not a free speed win.
03
The source analysis identifies candidates such as llama2.c-stories15M, Aurelius 10M, Pythia-14M, Pythia-31M, tiny-gpt2, and larger q4-only stress candidates. The site does not claim all are production-ready; it explains why each is useful for validation or browser experiments.
No WordPress bootstrap, theme system, database requirement, composer package, npm build, CDN script, Bootstrap class dependency, or jQuery dependency.
Routes are handled by index.php and server rewrites. Core content lives in data/pages.php. HTML and Markdown share the same page data.
Upload the package to the subdomain root, point Apache/Nginx to the folder, and keep .htaccess enabled for clean routes on Apache.