Jan
Jan
Yes, it would be great to have these methods in SBI 👍
@conormdurkan we need your approval as well because it's all based on your MDN implementation :)
@harry-fuyu we kindly ask for your approval because you contributed during your internship with us.
Hi Harry, thanks for your reply! While you are at it, could you please also approve the change in SBI? https://github.com/sbi-dev/sbi/pull/997 Thanks, Jan On Apr 2, 2024 at 14:52 +0200,...
Given that @conormdurkan agreed to the re-licensing of the code in `sbi`, and the code used here is actually the same code, I would opt for merging this pull request...
Hi @alexgarciaduran as mentioned in my email, there is a corresponding open issue in the SBI toolbox> https://github.com/mackelab/sbi/issues/717 I will work on fixing this soon!
Thanks for the updates @jnsbck ! Good catch @dgedon ! I looked into this a big and noticed that actually the first two dimensions in the output of the forward...
another option would be switching to `zuko` instead, e.g., use their https://github.com/probabilists/zuko/blob/master/zuko/flows/autoregressive.py instead of `nflows` MADE. They use a `MaskedMLP` as a conditioning net. I could not figure out the...
@jnsbck this is now fixed by a patch within `sbi`, see https://github.com/sbi-dev/sbi/pull/1398 so you can proceed finishing this PR 🚀
closing this due to inactivity. Please feel free to re-open if this is still relevant.