LabeliaLabs/distributed-learning-contributivity

Issues

• Use Pipfile and Pipfile.lock instead of requirements.txt

  #365 opened 3 years ago by HeytemBou
  0

• Approximation of amounts_per_partner

  #341 opened 3 years ago by kat-leen
  4

• Unable to run single partner learning approach

  #351 opened 3 years ago by HeytemBou
  1

• Clarify in documentation the use of save_path in Scenario and Experiment

  #329 opened 3 years ago by bowni
  0

• Add automated test for F-MNIST dataset

  #357 opened 3 years ago by HeytemBou
  0

• Tests need an update

  #261 opened 3 years ago by arthurPignet
  0

• Add end to end test for other multi partner learning approach

  #227 opened 3 years ago by RomainGoussault
  0

• AIS_Kriging_S contrib method fails

  #302 opened 4 years ago by RomainGoussault
  1

• S-model can only be used with 10-labels datasets

  #347 opened 4 years ago by arthurPignet
  0

• Add a artificial metric of training time

  #340 opened 4 years ago by arthurPignet
  0

• Refactor to global batches + local batches (vs. "mini-batch" + "gradient_updates_per_pass")

  #339 opened 4 years ago by bowni
  0

• Does the library benefit from GPU ?

  #250 opened 4 years ago by arthurPignet
  8

• re-build the model reset the optimizer

  #307 opened 4 years ago by arthurPignet
  1

• remove val_loss and val_accuracy evaluation can speed up the keras' .fit()

  #309 opened 4 years ago by arthurPignet
  0

• Scenario parameters dataset and dataset_name should be merged

  #317 opened 4 years ago by arthurPignet
  0

• Experiment documentation

  #292 opened 4 years ago by RomainGoussault
  0

• Clarify in README and/or documentation when results are saved

  #331 opened 4 years ago by bowni
  0

• problem solved

  #332 opened 4 years ago by lamnd09
  0

• Add multiclass performance metrics in results

  #330 opened 4 years ago by jeromechambost
  0

• What happens when we clip batch_size ?

  #232 opened 4 years ago by RomainGoussault
  2

• Need to update to Tf v2.4.1

  #327 opened 4 years ago by arthurPignet
  0

• Aggregation weighting parameter not taken into account

  #305 opened 4 years ago by RomainGoussault
  0

• The MNIST model seems to deep/complexe

  #324 opened 4 years ago by arthurPignet
  0

• Scenario/mpl log show inconsistent syntax

  #323 opened 4 years ago by arthurPignet
  0

• Add make build & make tests

  #312 opened 4 years ago by natct10
  0

• Add 'both' option for validation dataset and test data set.

  #297 opened 4 years ago by arthurPignet
  0

• IS_reg_S default to Shapley when partners_count < 4

  #303 opened 4 years ago by RomainGoussault
  0

• Federated step by step methods should ouput normalized score

  #293 opened 4 years ago by RomainGoussault
  1

• Switch to github action

  #296 opened 4 years ago by natct10
  2

• Need a parameter to choose if the val/test sets would be local or global

  #287 opened 4 years ago by arthurPignet
  0

• Create changelog

  #291 opened 4 years ago by natct10
  0

• Add an Experiment object to gather scenarios and run them in sequence

  #273 opened 4 years ago by bowni
  3

• init scenario logs are quite ugly for aggregator and mpl method

  #285 opened 4 years ago by arthurPignet
  0

• Logs are not shown when running test

  #254 opened 4 years ago by RomainGoussault
  4

• The repo is way too heavy!

  #258 opened 4 years ago by natct10
  8

• ESC50 preprocessing is time consuming, the preprocessed data should be stored in cache

  #286 opened 4 years ago by arthurPignet
  0

• Cifar model converges really slowly

  #282 opened 4 years ago by arthurPignet
  0

• Early stopping doesn't work any more

  #278 opened 4 years ago by Thomas-Galtier
  0

• Titanic example fails if minibatch contains only one class

  #257 opened 4 years ago by RomainGoussault
  2

• Add in documentation that the /experiments folder where results are saved is relative to the current working dir

  #272 opened 4 years ago by bowni
  0

• make the library label agnostic

  #252 opened 4 years ago by arthurPignet
  2

• Dynamique change of mplc parameters during the training

  #242 opened 4 years ago by arthurPignet
  0

• Refacto of the multipartnerlearning file

  #251 opened 4 years ago by arthurPignet
  0

• The shuffle option

  #259 opened 4 years ago by arthurPignet
  0

• Update flake8 config

  #234 opened 4 years ago by natct10
  1

• Conda package

  #229 opened 4 years ago by natct10
  2

• Partial corruption

  #240 opened 4 years ago by Thomas-Galtier
  0

• Update TF version

  #245 opened 4 years ago by natct10
  1

• Gaussian noise dataset for corrupted partner

  #237 opened 4 years ago by arthurPignet
  1

• Remove config_quick_debug run from Travis

  #228 opened 4 years ago by RomainGoussault
  0