v0.1.9

EasyHybrid v0.1.9

Diff since v0.1.8

Merged pull requests:

• new file structure (#241) (@lazarusA)
• print_banner (#243) (@lazarusA)
• fixes edit links (#244) (@lazarusA)
• v0.1.8 to v0.1.9 (#245) (@lazarusA)

v0.1.8

EasyHybrid v0.1.8

Diff since v0.1.7

Merged pull requests:

• Ba/r_nse_loss (#179) (@BernhardAhrens)
• zenodo badge (#188) (@lazarusA)
• hybrid_model config save + tune targets (#189) (@BernhardAhrens)
• CHANGELOG update (#192) (@BernhardAhrens)
• Update PrettyTables version in Project.toml to include 2.4.0 and 3.1.2 (#193) (@BernhardAhrens)
• harmonise format (#194) (@lazarusA)
• contributor's guide (#195) (@lazarusA)
• Continuation: use Lux.Training.single_train_step! (#198) (@BernhardAhrens)
• GenericHybrid with DimensionalData and extra_loss function (#200) (@BernhardAhrens)
• update show io (#201) (@lazarusA)
• coverage (#202) (@lazarusA)
• cov badge (#203) (@lazarusA)
• [wip] KGE and fix for build_chains (#206) (@BernhardAhrens)
• column_labels is not available in 2.4.0 (#208) (@lazarusA)
• per target (#210) (@lazarusA)
• add backwards compatibility PrettyTables (#211) (@lazarusA)
• losses tests (#212) (@lazarusA)
• more loss tests (#213) (@lazarusA)
• fix train show and tests (#214) (@lazarusA)
• wrap tuples tests (#215) (@lazarusA)
• [WIP] LSTM in GenericHybridModel (#217) (@BernhardAhrens)
• update macOS arch (#219) (@lazarusA)
• mention DD (#220) (@lazarusA)
• docs compat (#221) (@lazarusA)
• delete (#222) (@lazarusA)
• overview of tutorials for link to all (#223) (@BernhardAhrens)
• Input / output of LSTM clearer (#224) (@BernhardAhrens)
• updates for DVpress 0.3 (#225) (@lazarusA)
• CompatHelper: bump compat for DimensionalData to 0.30, (keep existing compat) (#227) (@github-actions[bot])
• finish book chapter move (#228) (@lazarusA)
• stars (#229) (@lazarusA)
• setup yaml (#231) (@lazarusA)
• be smooth (#232) (@lazarusA)
• soft deprecation () save_ps (#234) (@BernhardAhrens)
• update dependencies package.json (#239) (@lazarusA)
• v0.1.7 to v0.1.8 (#240) (@lazarusA)

Closed issues:

• Prior for mechanistic parameters via penalty in loss function (#158)
• allow / incorporate physical constraints (priors) in the loss (#173)
• switch to TrainState and/or single_train_step! (#177)
• Kinda deprecate () in train(hybridModel, data, ()) (#187)
• Loss on latent variables/outputs (#199)
• move book chapter to docs (#209)
• setup YAML config workflow (#230)

v0.1.7

EasyHybrid v0.1.7

Diff since v0.1.6

Merged pull requests:

• CompatHelper: bump compat for MLJ to 0.22, (keep existing compat) (#178) (@github-actions[bot])
• update CI, explicit imports and formatter (#183) (@lazarusA)
• more badges (#184) (@lazarusA)
• patch release (#185) (@lazarusA)
• fix compat (#186) (@lazarusA)

Closed issues:

• remove Flux dependency (#180)

v0.1.6

EasyHybrid v0.1.6

Diff since v0.1.5

Merged pull requests:

• generic loss (#163) (@lazarusA)
• best_model_.jld2 wo underscore (#165) (@BernhardAhrens)
• Cleanup research docu and move CUE (#166) (@BernhardAhrens)
• get overview back (#167) (@BernhardAhrens)
• losses (#175) (@lazarusA)
• bump (#176) (@lazarusA)

Closed issues:

• do a more generic loss function (#162)

v0.1.5

EasyHybrid v0.1.5

Diff since v0.1.4

Merged pull requests:

• [wip] starting point for folds (#153) (@BernhardAhrens)
• tests contructHybridModel (#154) (@BernhardAhrens)
• quick fix. did not like the file names (?) (#155) (@BernhardAhrens)
• Ba/fix hyperopt plots (#156) (@BernhardAhrens)
• changelog (#159) (@BernhardAhrens)
• 0.1.5 (#161) (@BernhardAhrens)

Closed issues:

• hyper parameter optim save error (#148)

v0.1.4

EasyHybrid v0.1.4

Diff since v0.1.3

Merged pull requests:

• custom output folder (#134) (@lazarusA)
• forward run of hybrid model, parallel script (#135) (@BernhardAhrens)
• fix (#137) (@BernhardAhrens)
• dispatch on DD (#140) (@lazarusA)
• move to SOC_PBM own repo (#143) (@BernhardAhrens)
• fixes abspath io (#144) (@lazarusA)
• fix prepare data, and add k-folds support (#146) (@lazarusA)
• forward model run still goes into EasyHybrid (#147) (@BernhardAhrens)
• minimal slurm (#149) (@lazarusA)
• changelog (#150) (@BernhardAhrens)
• 0.1.4 (#151) (@BernhardAhrens)
• add name to best model (#152) (@lazarusA)

Closed issues:

• spatial inputs and automatic differentiation (#139)
• prepare data + splits for DD (#145)

v0.1.3

EasyHybrid v0.1.3

Diff since v0.1.2

Merged pull requests:

• move project CUE_CN to own repo (#128) (@BernhardAhrens)
• fix prepare_data for when you input a Dataframe with NaNs or missing (#129) (@BernhardAhrens)
• subsample for trainboard fixed (#130) (@BernhardAhrens)
• example for Ellis summerschool (#131) (@BernhardAhrens)
• legend for quantiles in neural parameters (#132) (@BernhardAhrens)
• changelog and bump (#133) (@BernhardAhrens)

v0.1.2

EasyHybrid v0.1.2

Diff since v0.1.1

Merged pull requests:

• Hyperparameter tuning with Hyperopt.jl (#109) (@BernhardAhrens)
• stable badge (#121) (@lazarusA)
• CompatHelper: bump compat for JLD2 to 0.6, (keep existing compat) (#122) (@github-actions[bot])
• monthly downloads badge (#124) (@lazarusA)
• update projects (#125) (@BernhardAhrens)
• heading and Hyperopt reference (#126) (@BernhardAhrens)
• changelog and bump (#127) (@lazarusA)

Closed issues:

• train, validation, test split and hyperparameter tuning (#57)

v0.1.1

EasyHybrid v0.1.1

Diff since v0.1.0

Merged pull requests:

• target and update readme (#119) (@lazarusA)
• bump (#120) (@lazarusA)

v0.1.0

EasyHybrid v0.1.0