Releases: jgidi/ComplexSPSA.jl
Releases · jgidi/ComplexSPSA.jl
v0.2.1
What's Changed
- Fix calibration for complex variable type by @jgidi in #36
- Use dictionary to specify number of resamplings per iteration by @jgidi in #37
- fix update when blocking is false by @jgidi in #38
- Fix blocking by @jgidi in #39
- Add keyword for constant perturbation size by @jgidi in #40
- Fix hessian_postprocess_spall by @jgidi in #41
- correct hessian_postprocess_spall by @jgidi in #42
- calibrate blocking with 25 its per default by @jgidi in #43
- Clean code by @jgidi in #44
Full Changelog: v0.2.0...v0.2.1
v0.2.0
Notes
This is the first consistent version of the package where the optimizers have been unified to only two core functions. Also, many options -explained on the documentation- have been added.
testing is yet required to ensure that the optimizers provided can fully reproduce the results of the first release.
What's Changed
Full Changelog: v0.1.0...v0.2.0
First working release
Notes
- This version is functional but not complete.
What's Changed
- Add docs by @jgidi in #1
- add Qiskit submodule by @jgidi in #2
- Development by @jgidi in #3
- Development by @jgidi in #4
- Development by @jgidi in #5
- Use standard gains by default by @jgidi in #6
- Expand documentation and go public by @jgidi in #7
- Development by @jgidi in #8
- Fix QN docs by @jgidi in #9
- add SPSA QN scalar on complex by @jgidi in #10
- add SPSA QN scalar on complex to docpages by @jgidi in #11
- mod example 5 by @jgidi in #12
- update example 05 by @jgidi in #13
- Update example 5 by @jgidi in #15
- Add postprocess and correct gains by @jgidi in #16
- simulate_experiment: force 'refvalue' between 0 and 1 by @jgidi in #17
- correct previous pr by @jgidi in #18
- add module PaperPlot by @jgidi in #19
- Correct dependencies for previous PR by @jgidi in #20
- Correct shading for IQR and widen margins by @jgidi in #21
- correct vector concatenation by @jgidi in #22
- Scale and margins for paper plot by @jgidi in #23
- fix kwargs by @jgidi in #24
- accumulate Hessian info for iter < hessian_delay by @jgidi in #25
- Add option to use constant learning rates by @jgidi in #27
- Add experimental second order functions by @jgidi in #28
- add keyword "initial_iteration=1" by @jgidi in #29
- add scalar methods by @jgidi in #30
- Export scalar methods (amend last PR) by @jgidi in #31
- Add 'a2=1.0' keyword to control second order stepsize by @jgidi in #32
- Include files for scalar methods con src/ComplexSPSA.jl by @jgidi in #33
New Contributors
Full Changelog: https://github.com/jgidi/ComplexSPSA.jl/commits/v0.1.0