Improve L-BFGS

[HumbertHumbert7]

Hi, I need to achieve better precision using L-BFGS, how can I modify the source code and impose a smaller epsilon?

[lululxvi]

https://deepxde.readthedocs.io/en/latest/modules/deepxde.optimizers.html#deepxde.optimizers.config.set_LBFGS_options

[HumbertHumbert7]

Thank you Lu, but it doesn't cange nothing and if use the command
dde.config.set_default_float("float64")
as indicated, I have an error beacuse there are also some float32 that I am not able to fix.

[lululxvi]

It is possible that there are some data stored by default in float32, but this should be easy to fix.