Overview
BoTorch supplies modular components for Bayesian optimization on PyTorch. Its official examples fit a probabilistic surrogate model, construct an acquisition function and optimize that function to propose new candidates. The library is aimed at researchers and advanced practitioners; Ax is the higher-level experimentation platform built with BoTorch.
BoTorch is general optimization software. A chemistry workflow must supply its own experimental objective, feasible parameter ranges, constraints and measurement loop. The repository documents Python installation under the MIT license.
Key Features
- Probabilistic surrogate models and acquisition functions
- Candidate generation through acquisition optimization
Use Cases
- Prototype a Bayesian optimization loop for a measured chemistry objective after defining the search space and constraints.
How to Use
Follow the official installation and getting-started guide. Define a bounded objective and representative observations, fit a surrogate model, and inspect proposed candidates before any laboratory run.
Sources and verification
This entry describes capabilities documented by the linked project sources. Catalog edits are not evidence of a successful run, compatibility test or scientific performance evaluation.
Read the editorial and verification policy