Expected behavior

I am getting a runtime error. eg:

Running python3 -c 'import optuna; print(f"- Optuna version:{optuna.__version__}")' is resulting in runtime error.

Running the first basic example is resulting in runtime error.

Environment

• Optuna version: 4.2.0
• Python version: 3.12.8
• OS: Ubuntu 24.04.1 LTS (and also Windows 11)

Error messages, stack traces, or logs

RuntimeError: The grpc package installed is at version 1.67.1, but the generated code in api_pb2_grpc.py depends on grpcio>=1.68.1. Please upgrade your grpc module to grpcio>=1.68.1 or downgrade your generated code using grpcio-tools<=1.67.1.

Steps to reproduce

1. Using a conda-forge install of optuna and tensorboard mamba install optuna tensorboard -c conda-forge
2. python -c 'import optuna; optuna.__version__' (for example).

Additional context (optional)

After install, grpcio is version 1.67.1. But optuna wants >=1.68.1 according to here. I see this is auto generated by grpcio but when or where is beyond me.

Seems to be a problem specifically with tensorboard keeping grpcio version below at 1.67.1. Installing optuna on its own is fine and as per error message, a pip install also fixes, ie pip install grpcio==1.70.0.

I note that the conda-forge version of grpcio is at version 1.67.1 only.

So more of a conda annoyance than a bug. Also time to stop using tensorboard. but I hope still useful.