mathurinm

In a fresh python 3.8 env `pip install git+https://github.com/mathurinm/celer.git@main` works but `pip install git+https://github.com/mathurinm/celer.git@test_pyproject` does not

Hello @idc9, thanks for the interest. For the reweighting scheme, I plan to have a few string options corresponding to known penalites (Lq, log), but the user will also be...

Changing the seed in the above script to 1 gives me a different error: ```python ValueError Traceback (most recent call last) ValueError: unexpected array size: new_size=1, got array with arr_size=0...

@MarcAntoineSchmidtQC I may be wrong in the way I do it, but I am trying to fit a Lasso. Statistical analysis of the Lasso shows that it recovers good variables...

The output now is a bit better but still shows "not installed" ``` (base) ➜ lasso git:(main) ✗ benchopt run -d simulated -s celer Benchopt is running Simulated[n_features=10000,n_samples=100,rho=0] |--Lasso Regression[fit_intercept=True,reg=0.5]...

shouldn't this rather be done through a config file?

Good idea, we could use `temp_benchmark` for that - and maybe decorators are not even needed. Can you provide an example of single .py file containing the syntax that you'd...

@tomMoral fixed in #451 ? I can no longer reproduce on the bilevel benchmark with benchopt 1.3.2dev22

yes, it's what happens in these two lines in `callback.py`: https://github.com/benchopt/benchopt/blob/main/benchopt/callback.py#L71 https://github.com/benchopt/benchopt/blob/main/benchopt/callback.py#L84 `x` is fed to `Objective.__call__()` which itself calls `Objective.compute()`. So it should be of the type of `solver.get_result`,...

Passing options via command line may clutter the CLI, maybe we could have plot options passed through a `plot_config.yml` file ?