qredtea 0.3.15

Quantum TEA's python tensor library beyond numpy/cupy

qredtea

The quantum red tea library of Quantum TEA provides tensor beyond the dense tensor with numpy or cupy, which could be either tensor for systems with symmetries or tensor using another backend, e.g., pytorch.

Documentation

Here is the documentation. The documentation can also be built locally via sphinx.

Backends

We list here the supported backends with the tested version:

• pytorch==2.5.1
• jax==0.4.38
• tensorflow==2.18.0

The version allow to install all four tensor backends (numpy, torch, jax, tensorflow) within python3.11 environment (for CPUs).

License

The project qredtea is hosted at the repository https://baltig.infn.it/quantum_red_tea/py_api_quantum_red_tea.git, and is licensed under the following license:

Apache License 2.0

The license applies to the files of this project as indicated in the header of each file, but not its dependencies.

Installation

The qredtea library is never used as stand-alone package; it replaces the tensor backend for one of the quantum TEA applications in agreement with a qtealeaves version. Therefore, the minimal use-case to explore the library is together with qtealeaves. Moreover, it can be used as well with qmatchatea.

Local installation via pip

The package is available via a local pip installation as pip install ., i.e., after cloning the repository.

Dependencies

The python dependencies can be found in the pyproject.toml and are required independently of the following use-cases.

Depending on your use-case, more requirements might be necessary at runtime. As we implement the API to other packages via qredtea, we do not require the user to install all of them.

qtealeaves simulations

If you want to use the qtealeaves toolchain, a compatible version of qtealeaves has to be installed. The correct version is fetched automatically if this package is installed via pip. Otherwise, see pyproject.toml.

qmatchatea simulations

Quantum circuit simulations via qmatchatea require both qredtea and qtealeaves as a dependency. Follow the instructions contained in the qmatchatea repository.