Skip to content
Snippets Groups Projects

Repository graph

You can move around the graph by using the arrow keys.
Select Git revision
  • 313-allow-non-monotonic-window-lead-times-in-helpers-statistics-py-ahead_names-definition
  • 414-include-crps-analysis-and-other-ens-verif-methods-or-plots
  • IntelliO3-ts
  • Kleinert_et_al_2022_Representing
  • develop protected
  • develop_IntelliO3-ts
  • enxhi_issue460_remove_TOAR-I_access
  • falco_issue271_feat_add-cdc-database-datahandler
  • felix_PaperVersion
  • felix_TF2_merge_test_sectorial_skill_scores
  • felix_issue267-create-wrf-chem-data-handler
  • felix_issue287_tech-wrf-datahandler-should-inherit-from-singlestationdatahandler
  • felix_issue303-wrf-dh-major-minor
  • felix_issue319-toar-statistics4wrf-dh
  • felix_issue400_feat-include-tf-probability
  • felix_issue411-feat_include-postprocessing-for-bnns
  • felix_issue412-create-ens-predictions-for-bnns
  • felix_issue420-datahandler-with-multiple-stats-per-variable
  • felix_issue_365-tech-set-ci-cd-artifact-expire-time
  • lukas_issue363_feat_calculate-toar-metrics-on-hourly-forecasts
  • v2.4.0 protected
  • v2.3.0 protected
  • v2.2.0 protected
  • v2.1.0 protected
  • Kleinert_etal_2022_initial_submission
  • v2.0.0 protected
  • v1.5.0 protected
  • v1.4.0 protected
  • v1.3.0 protected
  • v1.2.1 protected
  • v1.2.0 protected
  • v1.1.0 protected
  • IntelliO3-ts-v1.0_R1-submit
  • v1.0.0 protected
  • v0.12.2 protected
  • v0.12.1 protected
  • v0.12.0 protected
  • v0.11.0 protected
  • v0.10.0 protected
  • IntelliO3-ts-v1.0_initial-submit
40 results
Created with Raphaël 2.2.015Oct141312118765427Sep241413987632131Aug30272625242322212019181716141312119643227Jul26232221201918161587542130Jun2928232221181728May27262522212019181764329Apr282721201918161514131211109876131Mar302926252322181716151211109854326Feb2524232219draft uncertainiy bootsdraft uncertainiy bootsblock_length and evaluate_competitors can be set from outsidecan calculate block mseMerge branch 'lukas_issue332_feat_report-error-metrics-for-all-competitors' into 'issue333_feat_test-set-sample-uncertainty'add helpersremove does not work properly stillfix issue with remove items if length is of 0fixed another problem with missing station forecastasmake_predict_function() changed to non-private version in training:102, abstract_model_class:134+ disabled __compare_keras_optimizers and rewrite empty lists with Noneadded weighted skill score mean to tablesanother adjustment for cases that no observation but a competitor is availablechange naming of skill score reportssmall fix if no model_list is availalso store competitive skill scoresobs dim is now derived from class attributerecent develop branch changes put in data_handler_single_station manually. May has overwritten something. Certainly {self.target_dim: helpers.to_list(self.target_var)}) # ToDo: is it right to expand this dim??experiment_setup.py: if no experiment_name is given (None), only the experiment_date is chosen as a nameerrors of external competitors are now added tooreports now errors for all competitors tootf_upgrade_tf2 run on whole directory. Only training.py and abstract_model_class.py directly affected by the script. Other problems were mostly import keras related and solved with import tensorflow.keras as keras. Not possible for advanced_paddings.py. For details see: https://gitlab.jsc.fz-juelich.de/esde/machine-learning/mlair/-/issues/331Merge branch 'lukas_issue330_bug_calculate-bootstraps' into 'develop'manually loaded CUDA/10.1.105 added to mlt_modules_juwels.shmanually loaded CUDA/11.0.2 added to mlt_modules_juwels.shmax_number_multiprocessing=7 added, np.linalg.det(first_param) < 1e-5 in legendre_reduction addedlazy preprocessed data only saved when not foundcalculate_legendre_reduction loops correctly over all split intervals nowpre_processing.py when station raised an error no _res_file is returned in f_proc, but None. This is catched within validate_station before it is read there againcalculate_legendre_reduction removed old dask codecalculate_legendre_reduction refactored to a splitted dask computation using client.submit and client.gather functionalitycsv file is now created only when results are availablestore results from bootstrapping in csv fileonly reset bootstrap_skill_scores on iteration 0Merge branch 'lukas_issue326_refac_do-not-stop-if-filter-plots-fail' into 'develop'station sample selection only in main functionupdate U-netupdate U-netadd U-Net incl LSTM in lowest leveladd U-Net incl LSTM in lowest levelInclude a LSTM layer after conv