Revision d048736df973ccaf2bc8efd03f04c33893941a40 authored by st-- on 27 October 2019, 16:28:49 UTC, committed by Artem Artemev on 27 October 2019, 16:28:49 UTC
Explicitly sets the default optimization method for Scipy() to L-BFGS-B (as in gpflow1), instead of letting scipy.minimize() figure it out. This resulted in BFGS, which may be too aggressive when encountering complex optimization surfaces, so L-BFGS-B may be a more sensible default.
1 parent ca40b13
File | Mode | Size |
---|---|---|
notebooks | ||
apiAndArchitecture.rst | -rw-r--r-- | 208 bytes |
conf.py | -rw-r--r-- | 10.3 KB |
derivations.rst | -rw-r--r-- | 128 bytes |
index.rst | -rw-r--r-- | 618 bytes |
intro.rst | -rw-r--r-- | 8.0 KB |
kernel_options.rst | -rw-r--r-- | 555 bytes |
likelihood_options.rst | -rw-r--r-- | 119 bytes |
mean_function_options.rst | -rw-r--r-- | 125 bytes |
model_options.rst | -rw-r--r-- | 822 bytes |
prior_options.rst | -rw-r--r-- | 265 bytes |
Computing file changes ...