Collection of small todos

This is a collection of small todo's, that can be altered and items can be removed when finished.

[x] add "yields" as an argument to get dependents to clarify whether to include yields or not
[x] add __iter__ to space wrapping iter_limits and return Spaces
[x] rename NameMinimizer to simply Name
[ ] better estimate step_size or params, in minimizer?
[x] change default of reset graph to true, add explicit reset in run?
[x] remove nll from each single point to move to 0
[x] cache sample etc tensor?
[x] allow minimiser to restart from FitResult, allow stepwise
[x] dtype for functors
[ ] split loss into term calculation and reduce operation to allow for iterating
[x] what if composite parameter is given to minimize params? get dependents? And anyway, build set of params?
[ ] loading from root with spaces?
[x] check space combine
[x] change default tolerance in minimizer
[x] refactor space to allow multiple limits being composition of several spaces
[x] add initial_value (and reset_value?)
[x] ztf.convert: make iterative

zfit / zfit-development