nmr-prediction

Predict per-atom 1H and 13C NMR chemical shifts from SMILES strings.

52|11|Updated Mar 10, 2026

Search Tags:#inference #spectrum #smiles #nmr #ppm #nmrnet

Ready to run

npx skills add https://github.com/InternScience/ChemClaw --skill nmr-prediction

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: nmr-prediction
Source: https://github.com/InternScience/ChemClaw/tree/main/skills/nmr-prediction
Command: npx skills add https://github.com/InternScience/ChemClaw --skill nmr-prediction

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires rdkit, numpy, torch, matplotlib, remotezip, and includes assets (resource) components.

What problem does it solve?

Predicts per-atom ¹H and ¹³C NMR chemical shifts and Lorentzian-broadened spectra from a SMILES string, enabling quick, interpretation-friendly insights without demanding quantum-chemical calculations.

Core Features & Use Cases

Per-atom ¹H and ¹³C chemical shifts (ppm) for input SMILES.
Generates Lorentzian-broadened spectrum PNGs for rapid visualization.
Lightweight deep-learning based inference using NMRNet to avoid DFT for quick screening.

Quick Start

Run python nmr_prediction.py 'CCO' to predict 1H/13C shifts and generate the spectrum image.

Frequently Asked Questions about nmr-prediction

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I predict NMR chemical shifts from a SMILES string?▼

To predict NMR chemical shifts from a SMILES string, provide the SMILES input to run NMRNet inference, which outputs per-atom 1H and 13C ppm values and renders spectrum PNGs without requiring quantum-chemical calculations.

Can I predict NMR spectra for large organic molecules?▼

You can predict NMR spectra for small- to medium-sized organic molecules. The Skill applies NMRNet inference to these molecule sizes to generate per-atom chemical shifts and Lorentzian-broadened spectrum visualizations.

What is the best way to get NMR-like insights without running DFT calculations?▼

Using deep-learning based inference with NMRNet is the best way to get NMR-like insights without DFT. It rapidly predicts 1H and 13C chemical shifts and generates visual spectra from SMILES strings.

How do I visualize predicted NMR shifts as a spectrum?▼

NMR shifts are visualized by generating Lorentzian-broadened spectrum PNGs. The Skill outputs these images alongside per-atom 1H and 13C ppm values after processing the input SMILES string.

Does NMR prediction with NMRNet require pre-trained weights?▼

NMR prediction requires pre-trained NMRNet weights and the Uni-Core runtime. These dependencies are necessary to perform the deep-learning inference and render the resulting NMR spectra.

What are the limitations of using deep learning for NMR shift prediction?▼

The limitation of using deep learning for NMR shift prediction is that it applies only to small- to medium-sized organic molecules, providing rapid screening insights rather than the exact accuracy of heavy quantum-chemical calculations.