DecisionTune 1.0: a 395M encoder that picks from your options offline, about 10 ms per short decision on MLX (Apache-2.0)

Disclosure: I made this. Sharing it here because it is fully local and small, and I want feedback from people who run models on their own machines. What it is: a 395M decision model (ModernBERT-large plus a 4 KB scoring head). You give it a state, a question and a list of options. It does one encode

中文全文翻译正在补齐，可先查看下方英文原文。

Disclosure: I made this. Sharing it here because it is fully local and small, and I want feedback from people who run models on their own machines. What it is: a 395M decision model (ModernBERT-large plus a 4 KB scoring head). You give it a state, a question and a list of options. It does one encode