Guy Moss
Guy Moss
The scale of the gradient shouldn't really matter for an Adam optimizer, so I'm not sure if that is the issue. I do think taking the mean instead of the...
Hi @anastasiakrouglova! Is there an update from your end on the status of this PR, or any input from our end that could help at this stage?
I am closing this as the prior wrapping issue was addressed by #1286 - @paarth-dudani, please re-open this issue if your issue is still unresolved.
Are you using the default `zuko_maf` and `maf` builders? I am asking because the problematic parameters seem to be the ones where the parameter range is narrowest - I am...
Hi @satwiksps! Glad you want to work on this issue, and happy for you to make a PR on this. Your suggested fix seems reasonable to me.
I see your point @michaeldeistler. I still think the `max_sampling_time` is the better way to go, because given a warning, some script would need to be rerun after changing the...
> Thanks @satwiksps looks very good! > > Just made suggestions for the error message. > > Additionally, I am tending towards introducing an actual time out by default, e.g.,...
Hi @satwiksps! We had a quick discussion now about what behaviour we exactly want to have after this PR, and these are the resolutions: - The default `max_sampling_time` should be...