Skip to content
Snippets Groups Projects

Repository graph

You can move around the graph by using the arrow keys.
Select Git revision
  • 120--learnernd-curvature
  • 133-use-a-itemsorteddict-for-the-loss-in-the-learnernd
  • 134-learner1d-load-throws-exception-when-file-is-empty
  • 74--add-anisotropicity-to-learnerND
  • AverageLearner2D
  • bugfix/suppress
  • ci_benchmarks
  • cython
  • function_in_runner
  • make_notebook_with_content
  • master default
  • no_overlay_plotting
  • private_methods_learnernd
  • refactor/triangulating-learner
  • renorm_2d
  • rst_readme
  • rtol_integrator
  • stable-0.7
  • v0.7.2
  • v0.8.0-dev
  • v0.7.0
  • v0.7.0-dev
  • v0.6.0
  • v0.5.0
  • v0.4.1
  • v0.5.0-dev
  • v0.4.0
  • v0.4.0-dev
  • v0.3.0
  • v0.2.1
  • v0.3.0-dev
  • v0.2.0
  • v0.2.0-dev
  • v0.1.0
34 results
Created with Raphaël 2.2.03Jan230Dec29282221201918109873130Nov29282724222120191615149873131Oct302725241312429Sep27181564131Aug302926232119181716126Jul242116Feb431Jan30make max_ivals a learner attributemake the error comparison test a bit more lenientadd assert in complete_process that checks if it can be calledremove incorrect commentremove unnecessary check, error's cannot be infinite anymoremake self.parent.depth_complete check more explicitset c00=0 in the first intervalupdate algorithm_4.py to work with the testsupdate cwg's code to 493f8810544827fa43f1977a9ca0cee993d580b6start with an interval of depth=2add doc-string to update_heuristic_errstyle fixes in the testsensure deterministic chosing of intervals to be splitcheck whether learners have identical chosen pointsuse a very big number for the error of the first intervaladd raising assert, done_leaves should not have an infinite errorif an interval is removed, propagate it down to it's childrenuse a normal set instead of error sorted set for the intervalsonly add excess igral when there are no childrentest if the error SortedSet is really sortedtest if the approximating_intervals are the sameno special treatment for the first interval in complete_processsimplify the force_split logicunveil another problem in the tests, choose_points is broken after random addingadd an extra test that is a special condition of the failing testPEP8 style fixadd a test condition that the errors are finiteset c00 in the correct placefix error updating algodo not rely on adding the first 33 points before anything elseincrementally update ndivmake cquad testing of random order more correctrename calc_ndiv_recursively -> update_ndiv_recursivelysimplify a testdon't go up if the parent is Noneadd more tests for random order point choosingupdate the ival's err and ndiv recursivelyimprove readabilityRevert "simplify _calc_coeffs and remove _zero_nans"simplify _calc_coeffs and remove _zero_nans