AttributeError in GP modeling with jaxns integration

[fcotizelati]

While trying to perform GP modeling using the GPResult.sample function integrated with jaxns for nested sampling, I encountered the following AttributeError:

AttributeError: 'ArrayImpl' object has no attribute 'next_sample_idx'

It seems to me this is due to the fact that the state object returned from the nested sampling process executed by exact_ns apparently lacks the expected attribute next_sample_idx?

Here are the versions of Python and the involved libraries I am using:

• Python version: 3.11.8
• stingray version: 2.0.0
• jax version: 0.4.23
• jaxns version: 2.4.12
• tensorflow_probability version: 0.22.1

Thanks for the help!

[matteobachetti]

Hi @fcotizelati, thanks for reporting the issue
@Gaurav17Joshi, @dhuppenkothen : any ideas?