BoTorch is a library for Bayesian Optimization built on PyTorch.

BoTorch is currently in beta and under active development!

BoTorch

• Provides a modular and easily extensible interface for composing Bayesian optimization primitives, including probabilistic models, acquisition functions, and optimizers.
• Harnesses the power of PyTorch, including auto-differentiation, native support for highly parallelized modern hardware (e.g. GPUs) using device-agnostic code, and a dynamic computation graph.
• Supports Monte Carlo-based acquisition functions via the reparameterization trick, which makes it straightforward to implement new ideas without having to impose restrictive assumptions about the underlying model.
• Enables seamless integration with deep and/or convolutional architectures in PyTorch.
• Has first-class support for state-of-the art probabilistic models in GPyTorch, including support for multi-task Gaussian Processes (GPs) deep kernel learning, deep GPs, and approximate inference.

The primary audience for hands-on use of BoTorch are researchers and sophisticated practitioners in Bayesian Optimization and AI. We recommend using BoTorch as a low-level API for implementing new algorithms for Ax. Ax has been designed to be an easy-to-use platform for end-users, which at the same time is flexible enough for Bayesian Optimization researchers to plug into for handling of feature transformations, (meta-)data management, storage, etc. We recommend that end-users who are not actively doing research on Bayesian Optimization simply use Ax.

Installation Requirements

• Python >= 3.10
• PyTorch >= 2.0.1
• gpytorch == 1.13
• linear_operator == 0.5.3
• pyro-ppl >= 1.8.4
• scipy
• multiple-dispatch

The latest release of BoTorch is easily installed via pip:

pip install botorch

Note: Make sure the pip being used is actually the one from the newly created Conda environment. If you're using a Unix-based OS, you can use which pip to check.

BoTorch stopped publishing an official Anaconda package to the pytorch channel after the 0.12 release. However, users can still use the package published to the conda-forge channel and install botorch via

conda install botorch -c gpytorch -c conda-forge

If you would like to try our bleeding edge features (and don't mind potentially running into the occasional bug here or there), you can install the latest development version directly from GitHub. If you want to also install the current gpytorch and linear_operator development versions, you will need to ensure that the ALLOW_LATEST_GPYTORCH_LINOP environment variable is set:

pip install --upgrade git+https://github.com/cornellius-gp/linear_operator.git
pip install --upgrade git+https://github.com/cornellius-gp/gpytorch.git
export ALLOW_LATEST_GPYTORCH_LINOP=true
pip install --upgrade git+https://github.com/pytorch/botorch.git

If you want to contribute to BoTorch, you will want to install editably so that you can change files and have the changes reflected in your local install.

If you want to install the current gpytorch and linear_operator development versions, as in Option 2, do that before proceeding.

git clone https://github.com/pytorch/botorch.git
cd botorch
pip install -e .

git clone https://github.com/pytorch/botorch.git
cd botorch
export ALLOW_BOTORCH_LATEST=true
pip install -e ".[dev, tutorials]"

• dev: Specifies tools necessary for development (testing, linting, docs building; see Contributing below).
• tutorials: Also installs all packages necessary for running the tutorial notebooks.
• You can also install either the dev or tutorials dependencies without installing both, e.g. by changing the last command to pip install -e ".[dev]".

Here's a quick run down of the main components of a Bayesian optimization loop. For more details see our Documentation and the Tutorials.

1. Fit a Gaussian Process model to data

import torch
from botorch.models import SingleTaskGP
from botorch.models.transforms import Normalize, Standardize
from botorch.fit import fit_gpytorch_mll
from gpytorch.mlls import ExactMarginalLogLikelihood

# Double precision is highly recommended for GPs.
# See https://github.com/pytorch/botorch/discussions/1444
train_X = torch.rand(10, 2, dtype=torch.double) * 2
Y = 1 - (train_X - 0.5).norm(dim=-1, keepdim=True)  # explicit output dimension
Y += 0.1 * torch.rand_like(Y)

gp = SingleTaskGP(
    train_X=train_X,
    train_Y=Y,
    input_transform=Normalize(d=2),
    outcome_transform=Standardize(m=1),
)
mll = ExactMarginalLogLikelihood(gp.likelihood, gp)
fit_gpytorch_mll(mll)

2. Construct an acquisition function

from botorch.acquisition import LogExpectedImprovement

logEI = LogExpectedImprovement(model=gp, best_f=Y.max())

3. Optimize the acquisition function

from botorch.optim import optimize_acqf

bounds = torch.stack([torch.zeros(2), torch.ones(2)]).to(torch.double)
candidate, acq_value = optimize_acqf(
    logEI, bounds=bounds, q=1, num_restarts=5, raw_samples=20,
)

If you use BoTorch, please cite the following paper:

M. Balandat, B. Karrer, D. R. Jiang, S. Daulton, B. Letham, A. G. Wilson, and E. Bakshy. BoTorch: A Framework for Efficient Monte-Carlo Bayesian Optimization. Advances in Neural Information Processing Systems 33, 2020.

@inproceedings{balandat2020botorch,
  title={{BoTorch: A Framework for Efficient Monte-Carlo Bayesian Optimization}},
  author={Balandat, Maximilian and Karrer, Brian and Jiang, Daniel R. and Daulton, Samuel and Letham, Benjamin and Wilson, Andrew Gordon and Bakshy, Eytan},
  booktitle = {Advances in Neural Information Processing Systems 33},
  year={2020},
  url = {http://arxiv.org/abs/1910.06403}
}

See here for an incomplete selection of peer-reviewed papers that build off of BoTorch.

See the CONTRIBUTING file for how to help out.

BoTorch is MIT licensed, as found in the LICENSE file.