Releases: AlexanderFengler/ssm-simulators
Releases · AlexanderFengler/ssm-simulators
v0.7.0
- add support for deadlines for all simulators (provide
deadline
parameter in theta vector/dictionary and supply model name asname_deadline
) - major refactor of simulator function
- addition of new simulators for hierarchical decision models
- simulator returns more derived quantities (
choice_p
,omission_p
,nogo_p
...) - dataset generators can now produce training data relevant for
cpn
,opn
,gonogopn
networks
v0.6.1
- allow return_option argument to be set in simulator (can now choose between ram-saving option 'minimal' and 'full' output)
- add mic2 simulator which allows noise scaling of low dimensional process depending on state of high dimensional process
v0.5.4
- fix a bug where
view
was used wrongly in cython code - add a few models of the
mic2_ornstein
andmic2_leak
kind
v0.5.1
- fix random seed setting in simulator for
smooth_unif = True
setting.
v0.5.0
- bug fix, negative rts were allowed through random smoothing of rt functions, now generated rts are strictly positive
v0.4.9
Merge pull request #18 from AlexanderFengler/v049 bump version to v049
v.0.4.6
Improve documentation
Fix some naming inconsistencies between models in config and model in simulators
0.4.5
- fix bug in the simulator function that could execute a faulty if statement
0.4.3
- unit tests added
- improvements to Readme
- simulator function now also accept pd.DataFrame as inputs for parameters
0.4.1
- change setup to allow automatic publishing