# Cheemly NMR model notices

NMRNet: https://github.com/Colin-Jay/NMRNet (MIT code).
Inference implementation: ai2-kit, Cheng group, https://github.com/chenggroup/ai2-kit (MIT).
Uni-Core: DP Technology and contributors, https://github.com/dptech-corp/Uni-Core (MIT).
The image retains the upstream repositories and their original license files.

Pretrained NMRNet assets: https://zenodo.org/records/19142375, released 2026-03-21.
Creators: Fanjie Xu, Wentao Guo, Feng Wang, Lin Yao, Hongshuai Wang, Fujie Tang,
Zhifeng Gao, Linfeng Zhang, Weinan E, Zhong-Qun Tian, and Jun Cheng.
Title: Towards a Unified Benchmark and Framework for Deep Learning-Based Prediction
of Nuclear Magnetic Resonance Chemical Shifts. DOI: 10.5281/zenodo.19142375.
Full bibliographic attribution is retained in the upstream record,
also preserved with the mounted assets. The deposit is licensed CC BY 4.0:
https://creativecommons.org/licenses/by/4.0/ .
Cheemly selects the liquid-state hydrogen/carbon fold-0 weights, uses CPU inference,
and adds input validation, atom mapping and approximate symmetry grouping.
No endorsement by the original creators is implied.

These are predicted chemical shifts, not experimentally measured spectra.

Mol* viewer: molstar 5.9.0, MIT. Copyright Mol* contributors.
Full license: /notices/molstar-LICENSE.txt
https://github.com/molstar/molstar

Presenton source exporter: Apache-2.0.
https://github.com/presenton/presenton/tree/cfc7233447a0c0b114be58a71f6bb0484dc5a9e2
Cheemly embeds the Python PPTX exporter with adapted slide scenes, local checked
figures and Chromium PDF/PNG rendering. Network downloads are disabled and
upstream print calls use logging. The isolated image retains LICENSE and NOTICE.
This does not include the upstream drag/drop editor or opaque export packages.

GPT Researcher HTML source reader: Apache-2.0, Assaf Elovic and contributors.
https://github.com/assafelovic/gpt-researcher/tree/5cdad9cb434754188b78bd998df18dd8d502cf7e
Cheemly reuses the BeautifulSoup reader and utilities with a supplied bounded
network session, adjusted relative import and redacted logging. Original LICENSE,
NOTICE and hashes remain in the chemistry image. Full public license:
/notices/gpt-researcher-LICENSE.txt
This does not include GPT Researcher's autonomous LLM/report or PDF dependency stack.

Chart UI components adapted from shadcn/ui, Copyright (c) 2023 shadcn, MIT License. https://github.com/shadcn-ui/ui — full license: https://github.com/shadcn-ui/ui/blob/main/LICENSE.md

Voice mode voice activity detection: Silero VAD v5 model, MIT License, Silero Team.
https://github.com/snakers4/silero-vad — run in the browser through @ricky0123/vad-web
(ISC License, Ricky Samore) and ONNX Runtime Web (MIT License, Microsoft).
Speech-to-text and text-to-speech in voice mode are provided by Deepgram.