torsional-diffusion
The premium Open Source alternative to ConfGen
🎯 Best for:Generating diverse molecular conformers for machine learning pipelines.
What is torsional-diffusion?
Replaces traditional force-field based conformer search methods like ETKDG. Utilizes score-based generative modeling on the torsional manifold to rapidly generate high-quality molecular ensembles.
Tech Stack
PythonChemistry & Materials
Why torsional-diffusion?
- • Superior sampling accuracy
- • Fast inference on GPUs
- • NeurIPS-validated architecture
Limitations
- • Requires GPU hardware
- • Complex ML environment setup
- • Limited to small/medium molecules
2/16/2026
Last Update
46
Forks
11
Issues
MIT
License
Financial Leak Detected
Stop the "SaaS Tax"
Your team could be burning cash. Switching to torsional-diffusion instantly boosts your runway.
Competitor Cost
-$1,440
/ year (est. based on ConfGen)
Self-Hosted
$0
/ year
Team Size10 Users
150+
SAVE 100%