PennyLaneAI / pennylane

PennyLane is a cross-platform Python library for quantum computing, quantum machine learning, and quantum chemistry. Train a quantum computer the same way as a neural network.
https://pennylane.ai
Apache License 2.0
2.38k stars 607 forks source link

Docker Image build not supported for Apple Silicon Chip and small minor bugs. #6452

Open zazabap opened 1 month ago

zazabap commented 1 month ago

Feature details

Recently during development, I tried using docker image build process written in the https://docs.pennylane.ai/en/stable/development/guide/installation.html?highlight=docker with Apple M2 chip.

However, there are issues encountered while preparing for metadata issue I consider adding one line of code should solve such issue. Also in line :

RUN python3 setup.py install

This gives error that setuptools are not included, maybe we could include it in the requirements.txt? There are also many minor issues like dependencies not found, missing libraries in building and etc.

The python3.12 in default installation should not suitable for development or docker file creation. Maybe the python3.10 stable version is the best for dockerfile creation? Or it would be better just upload the proper docker images?

Implementation

Modify these two files adding necessary packages could solve the build problem.

docker/pennylane.dockerifle
requirements.txt

If possible, I could contribute modifying the docker works with python3.10 and M2 chip.

How important would you say this feature is?

1: Not important. Would be nice to have.

Additional information

No response

albi3ro commented 1 month ago

Thanks for opening this issue @zazabap .

We do currently have docker builds actively maintained, unfortunately they are not the ones mentioned in the installation guide.

The builds can be found at:

https://hub.docker.com/r/pennylaneai/pennylane/tags

and information can be found at:

https://docs.pennylane.ai/projects/lightning/en/stable/dev/docker.html

The ones discussed in the installation guide have never been really tested or maintained, and we just haven't gotten around to getting rid of them. We will try to reprioritize fully removing them.