Quyet-1.2-Tiny
Quyet-1.2-Tiny is a decision model: given a state (any text, JSON or conversation) and one or more typed questions, it returns temperature-scaled probabilities over the supplied options. Question types: choice (pick one option), score (an ordered scale) and noul (true / false). It is part of the Quyet 1.2 family (Large, Small, Tiny), released by Chinh Nguyen under Apache-2.0.
| Base model | jhu-clsp/mmBERT-small |
| Architecture | mmBERT-small encoder (first 16 of 22 layers) + wide option-query head |
| Parameters | 183M (182,924,673) |
| Languages | Multilingual |
| Input | 8,192 complete tokens |
| Weights | 732 MB (F32) |
| License | Apache-2.0 (see LICENSE and NOTICE) |
| Homepage | quyet.ai |
New in 1.2
- Up to 256 options per choice question; 1.0 allowed 10. Choices with more than 10 options use the default temperature and return a warning.
How to use
pip install quyet==1.2.0
import quyet
m = quyet.load("chinhnc/Quyet-1.2-Tiny") # downloads from Hugging Face
r = m.predict(
{"message": "Please close my card, I lost it yesterday."},
{"intent": {"type": "choice", "instructions": "What does the customer want?",
"criteria": {"cancel": "close the card", "limit": "change the limit", "other": None}},
"urgent": {"type": "noul", "instructions": "The request is urgent."},
"mood": {"type": "score", "instructions": "How upset is the customer?", "criteria": ["calm", "annoyed", "angry"]}},
)
print(r["answers"]) # {"intent": {"choice": ..., "confidence": ..., "probabilities": {...}}, "urgent": {"noul": P(true)}, ...}
Runs on CPU in float32 or on GPU with BF16 autocast over F32 weights. It reads each question, its options and the state in one encoder pass and scores each option. The package applies the configured temperatures; choices above 10 options use the default temperature and return an uncalibrated_option_count warning.
Answers contain choice (with confidence and probabilities), score (expected level, confidence, probabilities, legend) or noul (P(true), with confidence). Each choice question accepts 2 to 256 options, score 2 to 10 levels, and noul exactly 2 options. Only the state is ever truncated: conversation lists keep their most recent tokens, other states keep their beginning. strict=True rejects truncation; oversized questions or options are rejected.
Not evaluated for safety, bias or adversarial robustness; do not use as the only control for high-stakes decisions. In half precision, results can differ slightly between hardware and batch sizes on near-tie questions.
Credits
- mmBERT-small by the mmBERT authors, JHU CLSP (MIT); first 16 of its 22 layers: https://huggingface.co/jhu-clsp/mmBERT-small.
Citation
@misc{quyet2026,
title = {Quyet 1.2: calibrated decision models},
author = {Chinh Nguyen},
year = {2026},
url = {https://huggingface.co/chinhnc/Quyet-1.2-Tiny}
}
License
Apache-2.0. Keep the NOTICE file (it starts with "Quyet by Chinh Nguyen") when you redistribute this model or anything derived from it. Questions and issues: email@quyet.ai. Package source: ncchinh/quyet.