Installation

Installation

PyPI package Conda package Supported Python versions

The fastest way of installing this package is through PyPI or Conda:

TensorWaves can work with different computational backends. They are provided through optional dependencies. JAX is usually the fastest backend, so it’s recommended to install that as in the install examples here.

python3 -m pip install tensorwaves[jax]

And with support for amplitude analysis:

python3 -m pip install tensorwaves[jax,pwa]

conda install -c conda-forge tensorwaves jax jaxlib

And with support for amplitude analysis:

conda install -c conda-forge ampform phasespace

This installs the latest, stable release that you can find on the stable branch.

Optional dependencies can be installed as follows:

pip install tensorwaves[pwa]  # installs tensorwaves with ampform
pip install tensorwaves[jax,scipy,tf]
pip install tensorwaves[all]  # all runtime dependencies
GPU support

Computations with are fastest on a GPU. To get JAX and TensorFlow work with your graphics card, you need to install at least CUDA Toolkit 11.2 and cuDNN 8.1.

Below is an installation guide for installing TF and JAX with GPU support. These instructions may become outdated, so refer to this TF page an these JAX instructions if you run into trouble.

  1. Download and install CUDA Toolkit 11.x by following these platform-dependent instructions. There may be dependency conflicts with existing NVIDIA packages. In that case, have a look here.

  2. Tell the system where to find CUDA. In Ubuntu (or WSL-Ubuntu), this can be done by adding the following line to your .bashrc file:

    export PATH="$PATH:/usr/local/cuda/bin"

    But first, check if /usr/local/cuda/bin indeed exists!

  3. Restart your machine. Then, launch a terminal to check whether CUDA is installed:

    nvcc --version

  4. Download and install cuDNN following these instructions. Make sure that you download cuDNN for CUDA 11.x!

    In Ubuntu (Debian), there are two convenient options: (1) installing through apt or (2) using a local installer. You may need to create an NVIDIA account.

  5. [Recommended] Install JAX with GPU binaries in your virtual environment following these instructions.

    pip install --upgrade jax[cuda] jaxlib -f https://storage.googleapis.com/jax-releases/jax_releases.html

    It doesn’t matter whether you do this before or after installing TensorWaves.

  6. If TensorFlow can correctly find your GPU, the following should return a non-empty list:

    python3 -c 'import tensorflow as tf; print(tf.config.list_physical_devices("GPU"))'

    If JAX can correctly find your GPU, the following should return gpu:

    python3 -c 'from jax.lib import xla_bridge; print(xla_bridge.get_backend().platform)'

The latest version on the main branch can be installed as follows:

python3 -m pip install git+https://github.com/ComPWA/tensorwaves@main

In that case, however, we highly recommend using the more dynamic ‘editable installation’ instead. This goes as follows:

  1. Get the source code:

    git clone https://github.com/ComPWA/tensorwaves.git
cd tensorwaves

  2. [Recommended] Create a virtual environment (see here).

  3. Install the project as an ‘editable installation’ and install additional packages for the developer:

    python3 -m pip install -e .[dev]
    Pinning dependency versions

    In order to install the exact same versions of the dependencies with which the framework has been tested, use the provided constraints files for the specific Python version 3.x you are using:

    python3 -m pip install -c .constraints/py3.x.txt -e .[dev]

    See also

    Pinning dependency versions

That’s all! Have a look at the General examples page to try out the package. You can also have a look at the Help developing page for tips on how to work with this ‘editable’ developer setup!

previous

Welcome to TensorWaves!

next

General examples