https://github.com/matt-graham/continuously-tempered-hmc

Continuously tempered Hamiltonian Monte Carlo
https://github.com/matt-graham/continuously-tempered-hmc

Last synced: 27 days ago
JSON representation

Continuously tempered Hamiltonian Monte Carlo

• Host: GitHub
• URL: https://github.com/matt-graham/continuously-tempered-hmc
• Owner: matt-graham
• License: mit
• Created: 2017-04-07T21:25:14.000Z (about 7 years ago)
• Default Branch: master
• Last Pushed: 2017-04-12T00:29:51.000Z (about 7 years ago)
• Last Synced: 2024-02-18T00:31:33.412Z (4 months ago)
• Language: Jupyter Notebook
• Size: 17.2 MB
• Stars: 9
• Watchers: 3
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Lists

• awesome-stars - matt-graham/continuously-tempered-hmc - Continuously tempered Hamiltonian Monte Carlo (Jupyter Notebook)

README

        
# Continuously tempered Hamiltonian Monte Carlo

Python code accompanying the paper [Continuously tempered Hamiltonian Monte Carlo](http://arxiv.org/abs/1704.03338).

## Installation

Two Python dependencies for running the experiment are included as submodules so you will need to do a recursive clone e.g.

```

git clone --recursive https://github.com/matt-graham/continuously-tempered-hmc.git

```

The code has only been tested with Python 2.7. The [environment.yml](environment.yml) file specifies the requirements for a [Conda](https://conda.io/docs/) environment that should be able to run all the experiments. After you have cloned the repository, from within the repository directory run

```

conda env create -f environment.yml

```

to create a new Conda environment called `cthmc` in which all the external dependencies will be installed to.

Once the `cthmc` environment is set up, activate it using

```

source activate cthmc

```

and then install the two internal Python dependencies by running

```

python thermodynamic-monte-carlo/setup.py install

python boltzmann-machine-tools/setup.py install -use-openmp

```