tiny-model
An 184M-parameter decision engine delivering calibrated distributions from one 22ms forward pass while matching models 2.3x its size.
Fracast-0 is an 85K-parameter pretrained Time Series Foundation Model, small enough to load directly in a web page.
A tiny byte-level multi-head content classifier (~1.5M params, ~200KB ONNX, <6ms). Classifies code, text, markup, config, images, binary, secrets, 62 code languages, 30 text languages, 90 MIME types from raw bytes — no tokenizer needed.
