Sirius Quantum

Welcome to the official Hugging Face org for Sirius Quantum, the quantum data layer for Physical AI.

The accuracy of AI in chemistry, biology and materials is limited by the cost of quantum-mechanical data, not by model size. We build representations that carry quantum mechanics directly into machine learning, so models learn more from far less labelled data.

Canis M

Canis M is the first molecular model trained on quantum tokens: compact representations of a molecule's electronic structure, read by the model in place of atomic coordinates. Trained on 50 molecules, it is more accurate than the standard approach trained on 16,000.

• 🧪 Model: SiriusQuantum/canis-m
• 📦 Token pack: SiriusQuantum/canis-m-token-pack
• 📄 Paper: Tokenising quantum data for label-efficient training of molecular models
• 🌐 Announcement: siriusquantum.com/canis-m

Open datasets

Machine-learning datasets with labels from exact quantum computation, across quantum chemistry, drug discovery, quantum many-body physics and quantitative finance. Browse them all in the datasets list below.

Open tools

• ⚡ Zilver: a distributed quantum simulator for integrated GPUs, open source under Apache 2.0.

Learn more

• 🌐 Website: siriusquantum.com
• 💻 GitHub: Sirius-Quantum
• 𝕏 : @SiriusQuantum
• 🤖 For AI agents: llms.txt
• 📧 Contact: info@siriusquantum.com

Each model and dataset in this organization carries its own license. See the card of each repository for terms.