Regression Transformer
Generate or predict molecular structures with properties
Generative models deep learning scientific discovery language models molecular machine learning
The is an open-source platform to accelerate hypothesis generation in the scientific discovery process. The gt4sd-core provides a library for making state-of-the-art generative AI models easier to use. Our mission is to accelerate scientific discovery by creating an open community around generative models for science 🚀
This page hosts ~20 Gradio apps of generative models in GT4SD. The apps cover topics such as (targeted) generation of small molecules, polymers, peptides, and catalysts, scientific language models (e.g., patent writing or keyword extraction), text-to-image diffusion models and property prediction models for proteins, molecules and crystals 👩💻
All apps can be used without having to write/run any piece of code 🌈
Generate or predict molecular structures with properties
Generate molecular compounds based on property goals
Generate molecular structures using scaffolds and seed molecules