SMAC3 v0.2.2

• FIX 124: SMAC could crash if the number of instances was less than seven
• FIX 126: Memory limit was not correctly passed to the target algorithm
  evaluator
• Local search is now started from the configurations with highest EI, drawn by
  random sampling
• Reduce the number of trees to 10 to allow faster predictions (as in SMAC2)
• Do an adaptive number of stochastic local search iterations instead of a fixd
  number (a5914a1)
• FIX a bug which didn't make SMAC run at least two configurations per call to
  intensify
• ADD more efficient data structure to update the cost of a configuration
• FIX do only count a challenger as a run if it actually was run
  (and not only considered)(a993c29)

SMAC3 v0.2.1

This is a bugfix, performance and similarity (to SMAC2) release.

• CI: travis-ci continuous integration on OSX
• ADD: initial design for mulitple configurations, initial design for a random configuration
• MAINT: use sklearn PCA if more than 7 instance features are available (as in SMAC 1 and 2)
• MAINT: use same minimum step size for the stochastic local search as in SMAC2.
• MAINT: use same number of imputation iterations as in SMAC2.
• FIX 98: automatically seed the configuration space object based on the SMAC seed.

SMAC3 v0.2

Bugfix and feature release:

• ADD 55: Separate modules for the initial design and a more flexible constructor for the SMAC class
• ADD 41: Add ROAR (random online adaptive racing) class
• ADD 82: Add fmin_smac, a scipy.optimize.fmin_l_bfgs_b-like interface to the SMAC algorithm
• NEW documentation at https://automl.github.io/SMAC3/stable and https://automl.github.io/SMAC3/dev
• FIX 62: intensification previously used a random seed from np.random instead of from SMAC's own random number generator
• FIX 42: class RunHistory can now be pickled
• FIX 48: stats and runhistory objects are now injected into the target algorithm execution classes
• FIX 72: it is now mandatory to either specify a configuration space or to pass the path to a PCS file
• FIX 49: allow passing a callable directly to SMAC. SMAC will wrap the callable with the appropriate target algorithm runner

0.1.3

Bugfix release

SMAC3-v0.1.2

bug fix release

SMAC3-v0.1.0

Bug fix release