Hey all, and thank you for the development of this benchmark :)
I tried to run arlbench as a benchmark for my current project and encountered some problems.
I believe they are mainly caused by the version in the pyproject.toml. It appears that all version numbers are pinned to certain versions. This makes the current arlbench version hard to combine with existing environments.
For example, numpy and pytest are pinned to certain version numbers.
I tried to resolve these dependency issues myself, but ran into further problems because of the optional hypersweeper dependency, which again contains pinned version numbers https://github.com/automl/hypersweeper/blob/main/pyproject.toml.
Is there a reason for pinning these version numbers that I am missing?
Hey all, and thank you for the development of this benchmark :)
I tried to run arlbench as a benchmark for my current project and encountered some problems.
I believe they are mainly caused by the version in the pyproject.toml. It appears that all version numbers are pinned to certain versions. This makes the current arlbench version hard to combine with existing environments.
For example, numpy and pytest are pinned to certain version numbers.
I tried to resolve these dependency issues myself, but ran into further problems because of the optional hypersweeper dependency, which again contains pinned version numbers https://github.com/automl/hypersweeper/blob/main/pyproject.toml.
Is there a reason for pinning these version numbers that I am missing?