keisuke umezawa

@Yard1 Thank you for the comment. > the PR is backwards compatible If the meaning of the backwards compatibility here is that we can run pycaret with both Optuna v2.x...

@Yard1 > Yes, we would like to be able to keep compatibility with Optuna v2. Could we perhaps have 2 cases - one for v2, and one for v3 (if...

@knshnb Although I am not a native English speaker, I doubt that the `Code fix` is a common wording to refer to these kinds of issues. If we use it...

@knshnb In that case, how about changing the description of this issue template as `Suggest a code fix that does not change the behaviors of Optuna, such as code refactoring.?...

@knshnb Could you rebase against master branch to pass all tests? If all tests pass, you can merge it by yourself.

The committers concluded that we need to do fundamental changes to the storages API before making it public. e.g. - Currently, the storages have two responsibility, one is for CRUD...

@jrbourbeau I also checked that [this script](https://gist.github.com/keisuke-umezawa/e58940ca71c9f2e061837c8b8017d7a0) runs correctly with both inmemory and sqllite. By the way, how can I load study from the existing sql? When I ran the...

The major difference is this part. And this fix was added by this PR https://github.com/optuna/optuna/pull/720 . ``` "x":1.0, "xpad":40 ``` In optuna dashboard, we only show a graph for one...

@c-bata I updated it. Could you check it later?

@c-bata Thank you. Updated.