'NaturalGradient' object has no attribute '_name'

[kbryan-zjc]

Bug / performance issue / build issue

To reproduce

Minimal, reproducible example

I was trying to replicate the [Heteroskedastic Likelihood and Multi-Latent GP](https://gpflow.github.io/GPflow/2.7.1/notebooks/advanced/heteroskedastic.html) example from the GPFlow library site. Everything ran fine except for the last snippet of code under "Run Optimization Loop" and is throwing an error: AttributeError: 'NaturalGradient' object has no attribute '_name'.

# This is the place for your code that reproduces the bug.
# Please make sure it does not depend on external libraries (beyond GPflow's own requirements) or specific datasets, and the smaller, the better :)
# For help on how to write a good bug report, see https://stackoverflow.com/help/minimal-reproducible-example

Stack trace, or error message

// Paste the full stack trace/error message here

Expected behavior

System information

• GPflow version: 2.8.0
• GPflow installed from: pip install
• TensorFlow version: 2.11.0
• Python version: 3.7
• Operating system

Additional context

[Rafael-DB]

I believe this issue stems from the fact that the NaturalGradient class passes the name up through its super until the class _BaseOptimizer where it is stored as self.name. The _natgrad_steps function however tries to access self._name instead. A workaround that seems to do the job is just to set the name manually natgrad_opt._name = 'natgrad_opt'

[kbryan-zjc]

这是来自QQ邮箱的自动回复邮件。您好，您的邮件我已收到，我会尽快处理，谢谢！