Data

Tools

Indexes

Applications

Experiments

Open Source Science

Recent Releases of FitBenchmarking

FitBenchmarking - v1.3.0

What's Changed

Fitting Software

  • Added basic support for GALAHAD by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1290 ### Tables
  • Updated comparison plots and added residuals plots by @Letizia97 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1433
  • Added a dropdown radio button menu to select the runtime metric for the compare and runtime tables by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1461
  • Added new dropdown menus for filtering the software and cost function results by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1476 ### Other new features
  • Added more mantid data examples. by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1344
  • Added support for SpinW 1D cuts of 2D powder data by @jess-farmer in https://github.com/fitbenchmarking/fitbenchmarking/pull/1405
  • Added slicing of SpinW data given Q by @Letizia97 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1422
  • Added check to compare jacobian computed by fitbenchmarking with that obtained through finite differences by @Letizia97 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1411
  • Support 2D SpinW data fitting by @jess-farmer in https://github.com/fitbenchmarking/fitbenchmarking/pull/1468
  • Added functionality to benchmark minimizer performance with multiple starting conditions for mantid problems by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1486 ### Breaking changes
  • Refactored the FitbenchmarkingParser class and added more unit tests by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1419
  • Refactored the MantidParser and MantidDevParser class and added more unit tests by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1455
  • Refactored the MantidController and added unit tests by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1463 ### Bug Fixes
  • Fixed corruption in checkpoint file after an exception by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1271
  • Added check for nfev and nit attributes in scipy controller by @jess-farmer in https://github.com/fitbenchmarking/fitbenchmarking/pull/1401

Full Changelog: https://github.com/fitbenchmarking/fitbenchmarking/compare/v1.2.0...v1.3.0

Scientific Software - Peer-reviewed - Python
Published by jess-farmer 9 months ago

FitBenchmarking - v1.2.0

What's Changed

Fitting Software

  • Added new controller for Scipy leastsq by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1281 ### Tables
  • Added number of iterations/function evaluations to fitting report by @jess-farmer in https://github.com/fitbenchmarking/fitbenchmarking/pull/1307 ### Other new features
  • Added parser for inputting sparse jacobian by @Letizia97 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1257
  • Measure energy usage instead of emissions by @jess-farmer in https://github.com/fitbenchmarking/fitbenchmarking/pull/1377
  • Added synthetic Back2BackExponential data to test mantid fitting by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1382
  • Added synthetic Gaussian data to test mantid fitting by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1376
  • Updated minimum supported version of python to 3.9.x by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1334
  • Added support for python 3.12 by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1301
  • Added support for python 3.13.x by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1372
  • Added python and numpy version tags to checkpoint.json by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1365 ### Bug Fixes
  • Fixed emissions display when not measured by @jess-farmer in https://github.com/fitbenchmarking/fitbenchmarking/pull/1287
  • Fixed Bayes bugs by @jess-farmer in https://github.com/fitbenchmarking/fitbenchmarking/pull/1278
  • Fix issue where perf profiles could be out of sync in dash by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1337
  • Added error message for online plots that are unavailable by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1338
  • Fix issue where failures were incorrectly reported in performance profile by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1336

Full Changelog: https://github.com/fitbenchmarking/fitbenchmarking/compare/v1.1.0...v1.2.0

Scientific Software - Peer-reviewed - Python
Published by jess-farmer over 1 year ago

FitBenchmarking - v1.1.0

What's Changed

Fitting Software

  • Added paramonte software and Bayesian comparison framework by @joshuadkitenge in https://github.com/fitbenchmarking/fitbenchmarking/pull/1108
  • Added bumps DREAM to Bayesian comparison framework by @jess-farmer in https://github.com/fitbenchmarking/fitbenchmarking/pull/1236
  • Added emcee and FABADA as Bayesian minimizers by @jess-farmer in https://github.com/fitbenchmarking/fitbenchmarking/pull/1272 ### Tables
  • Added problem size info to results tables, fitting reports and problem summary pages by @Letizia97 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1187
  • Updated tables to add blank cells in case of error by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1275 ### Other new features
  • Replaced matplotlib with plotly as the engine for generating plots of the fits by @tyronerees in https://github.com/fitbenchmarking/fitbenchmarking/pull/1164
  • Added support for automated release notes by @tyronerees in https://github.com/fitbenchmarking/fitbenchmarking/pull/1185
  • Added subcommand to fitbenchmarking-cp to allow merging checkpoint files by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1165
  • Added multiple ways of reporting time for benchmarking results by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1186
  • Replaced matplotlib with plotly as the engine for generating performance profiles by @Letizia97 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1192
  • Added support for HOGBEN by @jess-farmer in https://github.com/fitbenchmarking/fitbenchmarking/pull/1204
  • Introduced a limit to the number of solvers in performance profile plots. by @Letizia97 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1218
  • Updated live performance profiles to dynamically show correct plot for the lines shown by @Letizia97 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1243
  • Added code carbon task-based tracking by @jess-farmer in https://github.com/fitbenchmarking/fitbenchmarking/pull/1244
  • Switch to pyproject.toml by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1239
  • Added support for Mantid Jacobians and created a Mantid development parser by @AnthonyLim23 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1231
  • Added best_available jacobian option to allow for sets with only some analytic by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1251 ### Breaking changes
  • Updated default method for computing jacobians and hessians to "analytic" by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1163
  • Added a more interactive (Dash) version of performance profiles by @Letizia97 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1200 ### Bug Fixes
  • Changed rendering of html plotly plots (to reduce size of html files) by @Letizia97 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1189
  • Fixed issue with parser overwriting any runtime metric passed through options file. by @Letizia97 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1198
  • Fixed function for writing plots to html and added test by @Letizia97 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1203
  • Fixed bug where the emissions table showed a runtime performance profile by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1216
  • Fixed issue with back page button not working after introducing Dash. by @Letizia97 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1215
  • Bumped sphinx version for docs by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1224
  • Updated bumps and lmfit controllers to be more robust against parameter names by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1270
  • Fixed bug where relative runtime was always based on mean by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1274
  • Added cost func to the list of options displayed in log file by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1249
  • Improved performance of codecarbon calls on Windows machines by @jess-farmer in https://github.com/fitbenchmarking/fitbenchmarking/pull/1285 ### Other Changes
  • Add a readthedocs configuration file for online documentation by @tyronerees in https://github.com/fitbenchmarking/fitbenchmarking/pull/1179
  • Fix bash syntax by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1184

New Contributors

  • @Letizia97 made their first contribution in https://github.com/fitbenchmarking/fitbenchmarking/pull/1187

Full Changelog: https://github.com/fitbenchmarking/fitbenchmarking/compare/v1.0.0...v1.1.0

Scientific Software - Peer-reviewed - Python
Published by AndrewLister-STFC almost 2 years ago

FitBenchmarking - FitBenchmarking v1.0.0

What's Changed

In the year since the previous release of FitBenchmarking we've made the following changes:

New Features

  • New environmental impact table
  • Tables can now be regenerated from a checkpoint file
  • Options can now be set from the command line
  • An optional progress bar has been added
  • More RALFit options are now exposed to FitBenchmarking
  • The text output is now CSV
  • Auto-displaying of results is now optional
  • Tutorial videos have been added
  • The FitBenchmarking log now prints the date and time
  • Plots can now be chosen to have log or linear scales
  • Fonts on the tables have been changed to improve contrast
  • Text in table cells no longer wrap ### Parsers
  • HORACE is now supported ### Fitting Software
  • NLOPT is supported
  • GOFit is supported
  • Google's Ceres is now supported
  • SciPy solvers that require Hessians are now supported
  • LMFit is supported
  • The DREAM minimizer from BUMPS is now supported
  • Meta's Theseus is now supported
  • HERBERT's LM is supported
  • MINUIT's SIMPLEX is supported

Scientific Software - Peer-reviewed - Python
Published by AndrewLister-STFC over 2 years ago

FitBenchmarking - FitBenchmarking v1.0.0 (release candidate 1)

What's Changed

  • Add abort test for max runtime by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1026
  • Update minimizer names in dropdown by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1027
  • Ensure max runtime is checked in matlab controllers by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1028
  • Add a minimum python version by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1030
  • Remove unneeded noqa by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1035
  • 882 lm herbert by @tyronerees in https://github.com/fitbenchmarking/fitbenchmarking/pull/1033
  • Support for bumps 0.9.0 (remove mp, add scipy.leastsq) by @tyronerees in https://github.com/fitbenchmarking/fitbenchmarking/pull/1041
  • Move start_matlab call so it is only called once by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1034
  • Fix install after testing on iPython notebook by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1042
  • Enable max runtime in curve fitting by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1036
  • Fix the poisson cost func test by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1050
  • Add GOFit controller by @jfowkes in https://github.com/fitbenchmarking/fitbenchmarking/pull/1054
  • Enable mistune bump from 0.8.4 to 2.x in /docs by @dependabot in https://github.com/fitbenchmarking/fitbenchmarking/pull/1057
  • Set certain options input via argparse by @Shanrahan16 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1061
  • Add a progress bar for the benchmark problems (#1055) by @joshuadkitenge in https://github.com/fitbenchmarking/fitbenchmarking/pull/1063
  • Fix "external software" marked as "default" #1056 by @Shanrahan16 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1073
  • Fix docstring formatting by @tyronerees in https://github.com/fitbenchmarking/fitbenchmarking/pull/1075
  • Change the ralfit minimizer used for regression tests to gn by @tyronerees in https://github.com/fitbenchmarking/fitbenchmarking/pull/1080
  • Headings in options #678 by @Shanrahan16 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1076
  • Add ceres software by @joshuadkitenge in https://github.com/fitbenchmarking/fitbenchmarking/pull/1072
  • Update the versions of our github actions by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1088
  • reorder dockerfile and update docs by @tyronerees in https://github.com/fitbenchmarking/fitbenchmarking/pull/1091
  • CSV output by @Shanrahan16 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1094
  • Make auto display results optional by @Shanrahan16 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1093
  • Option to turn off the progress bar by @Shanrahan16 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1099
  • Tutorial videos upload by @Shanrahan16 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1103
  • Rename test data folder from mockdata to testfiles by @Shanrahan16 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1092
  • Remove failing pygsl from linting ci by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1100
  • Add date and time to fitbenchmarking log #1090 by @joshuadkitenge in https://github.com/fitbenchmarking/fitbenchmarking/pull/1095
  • Tutorial videos improvements for new users by @Shanrahan16 in https://github.com/fitbenchmarking/fitbenchmarking/pull/1107
  • Add scipy Hessian solvers #1016 by @joshuadkitenge in https://github.com/fitbenchmarking/fitbenchmarking/pull/1110
  • Add lmfit software #1096 by @joshuadkitenge in https://github.com/fitbenchmarking/fitbenchmarking/pull/1109
  • add bumps dream minimizer #1017 by @joshuadkitenge in https://github.com/fitbenchmarking/fitbenchmarking/pull/1105
  • Support use of BUMPS with Hellinger cost functions by @tyronerees in https://github.com/fitbenchmarking/fitbenchmarking/pull/1116
  • Add support for bounds (Ceres solver) #1111 by @joshuadkitenge in https://github.com/fitbenchmarking/fitbenchmarking/pull/1112
  • Remove hessian required minimizers for scipy default #1114 by @joshuadkitenge in https://github.com/fitbenchmarking/fitbenchmarking/pull/1119
  • Avoid parsing Mantid function strings by @jfowkes in https://github.com/fitbenchmarking/fitbenchmarking/pull/1077
  • allow plot scale options #1064 by @joshuadkitenge in https://github.com/fitbenchmarking/fitbenchmarking/pull/1121
  • Fix badges by @jfowkes in https://github.com/fitbenchmarking/fitbenchmarking/pull/1127
  • theseus controller #1052 by @joshuadkitenge in https://github.com/fitbenchmarking/fitbenchmarking/pull/1089
  • update ralfit options by @tyronerees in https://github.com/fitbenchmarking/fitbenchmarking/pull/1085
  • Enhancement/883 Add horace parser by @joshuadkitenge in https://github.com/fitbenchmarking/fitbenchmarking/pull/1120
  • remove redundant herbert_off by @joshuadkitenge in https://github.com/fitbenchmarking/fitbenchmarking/pull/1133
  • Improve the matlab installation docs #983 by @joshuadkitenge in https://github.com/fitbenchmarking/fitbenchmarking/pull/1143
  • Remove python2 code by @joshuadkitenge in https://github.com/fitbenchmarking/fitbenchmarking/pull/1139
  • Add more minuit minimizers #1135 by @joshuadkitenge in https://github.com/fitbenchmarking/fitbenchmarking/pull/1136
  • Add missing help spaces #1125 by @joshuadkitenge in https://github.com/fitbenchmarking/fitbenchmarking/pull/1134
  • Add more tests for fb minimizers #1137 by @joshuadkitenge in https://github.com/fitbenchmarking/fitbenchmarking/pull/1138
  • Omit matlab files fom coverage reports by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1147
  • Add checkpointing by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1058
  • Add prefix to title of webpage #1123 by @joshuadkitenge in https://github.com/fitbenchmarking/fitbenchmarking/pull/1142
  • Updates the output tables font colours by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1141
  • Updates the table text span by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1151
  • Improve support for missing data in profiler and local_min by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1152
  • Pass fitting_link as absolute path internally by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1150
  • Fix empty cell scrolling behaviour in tables by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1154
  • add nlopt software #1037 by @joshuadkitenge in https://github.com/fitbenchmarking/fitbenchmarking/pull/1104
  • Validate cost_func against initial problem by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1160
  • Add joss badge and CFF file by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1162
  • Add environmental impact table by @jess-farmer in https://github.com/fitbenchmarking/fitbenchmarking/pull/1161
  • Pin codecarbon by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1174
  • Add instructions for levmar on newer python versions by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1167
  • Store all benchmarking runtimes into the FittingResults class by @RabiyaF in https://github.com/fitbenchmarking/fitbenchmarking/pull/1172
  • Handle CITATION.cff in release code and docs by @AndrewLister-STFC in https://github.com/fitbenchmarking/fitbenchmarking/pull/1176

New Contributors

  • @Shanrahan16 made their first contribution in https://github.com/fitbenchmarking/fitbenchmarking/pull/1061
  • @joshuadkitenge made their first contribution in https://github.com/fitbenchmarking/fitbenchmarking/pull/1063
  • @RabiyaF made their first contribution in https://github.com/fitbenchmarking/fitbenchmarking/pull/1141

Full Changelog: https://github.com/fitbenchmarking/fitbenchmarking/compare/v0.2.0...v1.0.0-rc1

Scientific Software - Peer-reviewed - Python
Published by AndrewLister-STFC over 2 years ago

FitBenchmarking - FitBenchmarking v0.2.0

What's Changed

FitBenchmarking v0.2.0 is the result of 10 months of work, and we've introduced a lot of changes to the FitBenchmarking package.

This release now requires Python 3.7.1 or greater.

New Features

  • Support for different cost functions was added; see below
  • Improved support for different jacobians; see below
  • Better handling of conflicting options
  • More descriptions of the methods in docs
  • Box constraints supported
  • Licences of the software used documented
  • Minimizers are collected into types
  • Second derivatives (Hessians) supported, and can be compared
  • Added an option to select algorithms by type
  • Results directory can be set from the command line
  • A maximum runtime can be selected for each solver

Parsers

  • Mantid's Crystal Field objective is supported
  • Native parser has been refactored
  • IVP parser has been added

Fitting Software

  • iminuit>2.0 is supported
  • levmar supported
  • Matlab supported (basic, and curve_fitting, statistics and optimization toolboxes)
  • horace supported (Matlab package)
  • scipy_go supported
  • gradient_free_optimizers supported
  • bumps support has been updated to version 0.9.0

Cost functions

  • NLLS and WeightedNLLSs cost functions replace the defaults
  • HellingerNLLS cost function
  • Poisson cost function added
  • Multiple cost functions can be compared at once

Jacobians

  • numdifftools Jacobians have been added
  • Analytic Jacobians added for NIST and SIF file formats
  • Solver default Jacobians have been added
  • Multiple Jacobians can be compared at once

Tables

  • the generation of these has been refactored
  • Failed fits are highlighted in the tables
  • Customizable colourmaps in the tables
  • Header row and leading column of results table is frozen
  • Problem descriptions displayed in fitting report
  • Accuracy and runtime displayed in fitting report
  • Problem summary page has been added

Data sets

  • more SASView examples added
  • Data Assimilation (IVP) examples added
  • CrystalField examples added
  • Global optimization SIF file examples added

Breaking Changes

  • Support is dropped for iminuit<2.0
  • lm-scipy-no-jac option under scipy_ls in [MINIMIZERS] has been removed (this is now available under the solver default jacobian)
  • results_dir has moved from [PLOTTING] into [OUTPUT] in the options file
  • use_errors has been removed as an option, and this functionality is controlled by the choice of cost function
  • The Jacobian method SciPyFD is no longer supported; this has been renamed to scipy
  • colour_scale option in [PLOTTING] is no longer supported; this is replaced by the colour_map
  • bumps minimizers name changes: lm-bumps -> scipy-leastsq and mp -> lm-bumps

Full Changelog: https://github.com/fitbenchmarking/fitbenchmarking/compare/v0.1.5...v0.2.0

Scientific Software - Peer-reviewed - Python
Published by AndrewLister-STFC over 3 years ago

FitBenchmarking - FitBenchmarking v0.2.0 (release candidate 3)

Scientific Software - Peer-reviewed - Python
Published by AndrewLister-STFC over 3 years ago

FitBenchmarking - FitBenchmarking v0.2.0 (release candidate 2)

Scientific Software - Peer-reviewed - Python
Published by AndrewLister-STFC almost 4 years ago

FitBenchmarking - FitBenchmarking v0.2.0 (release candidate 1)

What's Changed

FitBenchmarking v0.2.0 is the result of 10 months of work, and we've introduced a lot of changes to the FitBenchmarking package.

This release now requires Python 3.7.1 or greater.

New Features

  • Support for different cost functions was added; see below
  • Improved support for different jacobians; see below
  • Better handling of conflicting options
  • More descriptions of the methods in docs
  • Box constraints supported
  • Licences of the software used documented
  • Minimizers are collected into types
  • Second derivatives (Hessians) supported, and can be compared
  • Added an option to select algorithms by type
  • Results directory can be set from the command line
  • A maximum runtime can be selected for each solver

Parsers

  • Mantid's Crystal Field objective is supported
  • Native parser has been refactored
  • IVP parser has been added

Fitting Software

  • iminuit>2.0 is supported
  • levmar supported
  • Matlab supported (basic, and curve_fitting, statistics and optimization toolboxes)
  • scipy_go supported
  • gradient_free_optimizers supported

Cost functions

  • NLLS and WeightedNLLSs cost functions replace the defaults
  • HellingerNLLS cost function
  • Poisson cost function added
  • Multiple cost functions can be compared at once

Jacobians

  • numdifftools Jacobians have been added
  • Analytic Jacobians added for NIST and SIF file formats
  • Solver default Jacobians have been added
  • Multiple Jacobians can be compared at once

Tables

  • the generation of these has been refactored
  • Failed fits are highlighted in the tables
  • Customizable colourmaps in the tables
  • Header row and leading column of results table is frozen
  • Problem descriptions displayed in fitting report
  • Accuracy and runtime displayed in fitting report
  • Problem summary page has been added

Data sets

  • more SASView examples added
  • Data Assimilation (IVP) examples added
  • CrystalField examples added
  • Global optimization SIF file examples added

Breaking Changes

  • Support is dropped for iminuit<2.0
  • lm-scipy-no-jac option under scipy_ls in [MINIMIZERS] has been removed (this is now available under the solver default jacobian)
  • results_dir has moved from [PLOTTING] into [OUTPUT] in the options file
  • use_errors has been removed as an option, and this functionality is controlled by the choice of cost function
  • The Jacobian method SciPyFD is no longer supported; this has been renamed to scipy
  • colour_scale option in [PLOTTING] is no longer supported; this is replaced by the colour_map

Scientific Software - Peer-reviewed - Python
Published by tyronerees about 4 years ago

FitBenchmarking - FitBenchmarking v0.1.5

FitBenchmarking 0.1.5 has no new features compared to 0.1.4

Changes:

  • Support iminuit 2.0 (and drop support for iminuit 1.x)
  • Update Zenodo metadata file

Scientific Software - Peer-reviewed - Python
Published by tyronerees over 4 years ago

FitBenchmarking - FitBenchmarking v0.1.5 (release candidate 1)

FitBenchmarking 0.1.5 has no new features compared to 0.1.4

Changes:

  • Support iminuit 2.0 (and drop support for iminuit 1.x)
  • Update Zenodo metadata file

Scientific Software - Peer-reviewed - Python
Published by tyronerees over 4 years ago

FitBenchmarking - FitBenchmarking v0.1.4

FitBenchmarking 0.1.4 has no new features compared to 0.1.3

Changes: - Fix a bug in the zenodo json file - Fix documentation errors - Update version of urllib needed

Scientific Software - Peer-reviewed - Python
Published by tyronerees over 4 years ago

FitBenchmarking - FitBenchmarking v0.1.4 (release candidate 2)

FitBenchmarking 0.1.4 has no new features compared to 0.1.3

Changes: - Fix a bug in the zenodo json file - Fix documentation errors - Update version of urllib needed

Scientific Software - Peer-reviewed - Python
Published by tyronerees over 4 years ago

FitBenchmarking - FitBenchmarking v0.1.4 (release candidate 1)

FitBenchmarking 0.1.4 has no new features compared to 0.1.3

Changes:

Fix a bug in the zenodo json file

Scientific Software - Peer-reviewed - Python
Published by tyronerees almost 5 years ago

FitBenchmarking - FitBenchmarking v0.1.3

FitBenchmarking 0.1.3 has no new features compared to 0.1.2

Changes:

  • Fix a bug where the labels in the plots were inacccurate
  • Add a json file for zenodo metadata

Scientific Software - Peer-reviewed - Python
Published by tyronerees almost 5 years ago

FitBenchmarking - FitBenchmarking 0.1.2

FitBenchmarking 0.1.2 has no new features compared to 0.1.1

Changes:

  • Fix a bug in the CI, where results were not being sent to coveralls
  • Update the README file to point to places of interest in the docs

Scientific Software - Peer-reviewed - Python
Published by tyronerees almost 5 years ago

FitBenchmarking - FitBenchmarking 0.1.1

FitBenchmarking 0.1.1 Release Notes

FitBenchmarking 0.1.1 has no new features compared to 0.1.0

Changes: - Fix a bug where SASView problems were sometimes in the wrong order - Updated test results to reflect changes in the current versions of software we support - Update the paper based on comments suggested by the JOSS reviewers

Scientific Software - Peer-reviewed - Python
Published by tyronerees almost 5 years ago

FitBenchmarking - First Release!

FitBenchmarking is an open source tool for comparing different minimizers/fitting frameworks. FitBenchmarking is cross platform and we support Windows, Linux and Mac OS.

Fitting a mathematical model to data is a fundamental task across all scientific disciplines. (At least) three groups of people have an interest in fitting software:

  • Scientists, who want to know what is the best algorithm for fitting their model to data they might encounter, on their specific hardware;
  • Scientific software developers, who want to know what is the state-of-the-art in fitting algorithms and implementations, what they should recommend as their default solver, and if they should implement a new method in their software; and
  • Mathematicians and numerical software developers, who want to understand the types of problems on which current algorithms do not perform well, and to have a route to expose newly developed methods to users.

Representatives of each of these communities have got together to build FitBenchmarking. We hope this tool will help foster fruitful interactions and collaborations across the disciplines.

This is our first release and we are actively working to expand the offerings of FitBenchmarking! If you have any feature requests or would like to be involved as a contributer please create an issue or pull request.

Scientific Software - Peer-reviewed - Python
Published by AndrewLister-STFC about 5 years ago

FitBenchmarking - Release candidate for version 0.1.0

Scientific Software - Peer-reviewed - Python
Published by AndrewLister-STFC about 5 years ago

FitBenchmarking - Release candidate for version 0.1.0

Scientific Software - Peer-reviewed - Python
Published by AndrewLister-STFC about 5 years ago

FitBenchmarking - Release candidate for version 0.1.0

Scientific Software - Peer-reviewed - Python
Published by AndrewLister-STFC about 5 years ago

FitBenchmarking - Second pre-release version of FitBenchmarking

This includes features to make it more friendly as a pip-install. In particular, this version - prints the options used into the log - notifiesthe user where the results are stored - allows fitbenchmarking to run without extra - makes data downloadable from the docs - only uses minimally installed software as default

Scientific Software - Peer-reviewed - Python
Published by tyronerees over 5 years ago

FitBenchmarking - First pre-release

This is the first pre-release version of FitBenchmarking

Scientific Software - Peer-reviewed - Python
Published by tyronerees over 5 years ago