NOTICE

dohnuts.cpp
Copyright 2026 DreamBlooms and dohnuts.cpp contributors.

This project is a native C++ inference port of Dohnuts. The source is licensed
under Apache-2.0. Model weights are separate artifacts with their own terms.

Dohnuts is licensed under Apache-2.0. This port reuses its prompt template,
calibration, and decision semantics.
https://github.com/PsiACE/dohnuts

The HTTP transport (src/http.cpp, include/dohnuts/http.hpp) is adapted from
laya.cpp, licensed under MIT. See third_party/laya.LICENSE for the full text.
https://github.com/lkarlslund/laya.cpp

cpp-httplib is vendored as a single header (third_party/httplib.h) and is
licensed under MIT.
https://github.com/yhirose/cpp-httplib

nlohmann/json is vendored as a single header (third_party/nlohmann/json.hpp)
and is licensed under MIT.
https://github.com/nlohmann/json

llama.cpp is used as a pinned submodule and is licensed under MIT.
https://github.com/ggml-org/llama.cpp

Qwen3.5-0.8B is developed by the Qwen team at Alibaba Cloud and is licensed
under Apache-2.0. Dohnuts trains language LoRA adapters and a decision head over
this base. Base weights are downloaded from the original publisher.
https://huggingface.co/Qwen/Qwen3.5-0.8B

Model weights, source datasets, and third-party model files are not distributed
with this repository. Reference names do not imply endorsement by their authors.