Code

History

Commits on master20
  1. see changelogecholine committed 7 years ago077f34c
  2. oops, and the source...echoline committed 8 years agoce6d1bb
  3. adamaxecholine committed 8 years ago27815e7
  4. test both training algorithmsecholine committed 8 years ago3c9ce44
  5. truerand() initialized weights converge fasterecholine committed 8 years ago55edf9f
  6. added "adam" training algorithmecholine committed 8 years ago768a52b
  7. !?!?!echoline committed 8 years agodc1f413
  8. namespace scriptecholine committed 8 years ago1a6871e
  9. other activation and gradient functions, and a small formatting error fixecholine committed 8 years ago66828d0
  10. clean up xor testecholine committed 8 years ago8bf2091
  11. return total network MSE (before training) from anntrainecholine committed 8 years ago9cc0923
  12. weightsinit.cecholine committed 8 years ago96fcd39
  13. initialize larger weightsecholine committed 8 years agof2f2484
  14. clean up xor testecholine committed 8 years ago4638c0a
  15. correctly initialize deltas matrix in anntrain()echoline committed 8 years agoc52bc40
  16. pass Neuron* to activation and gradient for a bit more efficiencyecholine committed 8 years agod98b3b5
  17. bias neuronsecholine committed 8 years agoc1e7d1b
  18. broke up code a bitecholine committed 8 years ago42edbcc
  19. initial commitecholine committed 8 years ago71c7f11
  20. Initial commitecholine committed 8 years agobe40133