bion howard

Here is the result for all of the ProcGen envs: > Observation Space: > Box(64, 64, 3) > dtype = "uint8" > high[1] = 255 > low[1] = 0 >...

sorry, i'm not working in julia right now, but you're welcome to reuse the code

Similar problem on my first attempts to use this library. I get pumped and write a big migration, then gotta debug it. But the error message reads: error at or...

I'm not using this library, just GET requests, and sorry to necro the thread here, but doesn't this screw up a bunch of the technical indicator calculations if sometimes we...

I did get compile time bounds checking of named tensors working in rust but without type aliases the errors are unreadable HLists of TypeNum, const generics are an interesting approach...

Storage is literally sql so adding multiple objectives is just adding columns to a table, right? Here's why this is hard: we're using dict instead of dataframe. A dataframe is...

This approach would require breaking changes but would unify single and multiple objective code paths. Let me know if you want me to make a PR and I'll do it!

e.g. ```diff def __init__( self, number: int, state: TrialState, - value: Optional[float], datetime_start: Optional[datetime.datetime], datetime_complete: Optional[datetime.datetime], params: Dict[str, Any], distributions: Dict[str, BaseDistribution], user_attrs: Dict[str, Any], system_attrs: Dict[str, Any], - intermediate_values:...

example multiobjective pruner: ```py def decide_if_pruning( population_trial_n: int, db: SQLDatabase = default_db, rendering: bool = False, patience_level: int = 1, objective_names: Tuple[str] = ( "mean_alpha", "mean_surprisal", "mean_frequency", ), directions: Tuple[str]...

Agreed. In that case, one idea would be to lift single values to multi-objective fully internally within the trial and then users could keep using the system the exact same...