Code

Simplify transformer tests#100

Code

Merged

michakraus merged 58 commits intomainJuliaGNI/GeometricMachineLearning.jl:mainfrom simplify_transformer_testsJuliaGNI/GeometricMachineLearning.jl:simplify_transformer_testsCopy head branch name to clipboardDec 20, 2023

+401-791

Commits

Commits on Dec 13, 2023

• Remove old array tests.

  benedict-96

  committed
  ba96b35

  Copy full SHA for ba96b35

• Removed the block matrices. This is now done by the LinearSympNetLayers.

  benedict-96

  committed
  0fb9156

  Copy full SHA for 0fb9156

• Hope this fixed the TODO list.

  benedict-96

  committed
  dd4e8a5

  Copy full SHA for dd4e8a5

• We are not using those arrays anymore.

  benedict-96

  committed
  35c0d48

  Copy full SHA for 35c0d48

• Factored out sampling of arrays.

  benedict-96

  committed
  1f1e30d

  Copy full SHA for 1f1e30d

• Got rid of some some symplectic arrays and included the supertype for horizontal components as an extra file.

  benedict-96

  committed
  51c7c76

  Copy full SHA for 51c7c76

• Updated docs.

  benedict-96

  committed
  3876be5

  Copy full SHA for 3876be5

• New supertype for the horizontal components.

  benedict-96

  committed
  48300b2

  Copy full SHA for 48300b2

• Removed tests for various symplectic arrays that are no longer part of GML.

  benedict-96

  committed
  969a6cb

  Copy full SHA for 969a6cb

• Fixed docs for optimize_for_one_epoch

  benedict-96

  committed
  5d58358

  Copy full SHA for 5d58358

Commits on Dec 14, 2023

• Added descriptions for some functions in the docs.

  benedict-96

  committed
  b43fb39

  Copy full SHA for b43fb39

• Added eltype to AbstractCache.

  benedict-96

  committed
  d032f3e

  Copy full SHA for d032f3e

• Removed comments that seem to have become unnecessary.

  benedict-96

  committed
  81ac240

  Copy full SHA for 81ac240

• Removed symplectic part (again).

  benedict-96

  committed
  9002452

  Copy full SHA for 9002452

• Refactored all the transformer tests and gave each of them a name.

  benedict-96

  committed
  4c82474

  Copy full SHA for 4c82474

Commits on Dec 19, 2023

• Renamed file to something more descriptive.

  benedict-96

  committed
  35b3db6

  Copy full SHA for 35b3db6

• Renamed file to something meaningful and made the test more readable.

  benedict-96

  committed
  66f3279

  Copy full SHA for 66f3279

• Merge pull request #97 from JuliaGNI/put_sampling_of_arrays_into_extra_test_file

  Show description for e8ba5c1
  michakraus

  authored
  e8ba5c1

  Copy full SHA for e8ba5c1

• Added descriptions and a wrapper for using the custom loss function with NeuralNetwork.

  benedict-96

  committed
  257aeee

  Copy full SHA for 257aeee

• Added descriptions and a wrapper for using the custom loss function with NeuralNetwork.

  benedict-96

  committed
  9ca60ed

  Copy full SHA for 9ca60ed

• Added documentation.

  benedict-96

  committed
  fced8d8

  Copy full SHA for fced8d8

• Added constructor for optimizer if input arguments are flipped.

  benedict-96

  committed
  d95268d

  Copy full SHA for d95268d

• Added a comment saying that the constructor can be called with DataLoader as input.

  benedict-96

  committed
  2abed87

  Copy full SHA for 2abed87

• Added default for number of epochs.

  benedict-96

  committed
  0f62b8f

  Copy full SHA for 0f62b8f

• Combined matrix and tensor routines into one. Added another loss for autoencoder qp data.

  benedict-96

  committed
  3ab42f2

  Copy full SHA for 3ab42f2

• Commented out a section that is probably not needed.

  benedict-96

  committed
  5d22855

  Copy full SHA for 5d22855

• Test data loader for qp data.

  benedict-96

  committed
  86485ad

  Copy full SHA for 86485ad

• Renamed and added tests.

  benedict-96

  committed
  cac1f18

  Copy full SHA for cac1f18

• Adjusted symplectic matrix.

  benedict-96

  committed
  4528053

  Copy full SHA for 4528053

• Renamed file to something more descriptive.

  benedict-96

  committed
  6ed5c62

  Copy full SHA for 6ed5c62

• Renamed file to something meaningful and made the test more readable.

  benedict-96

  committed
  f4ddd3f

  Copy full SHA for f4ddd3f

• Added descriptions and a wrapper for using the custom loss function with NeuralNetwork.

  benedict-96

  committed
  92ac2fe

  Copy full SHA for 92ac2fe

• Added descriptions and a wrapper for using the custom loss function with NeuralNetwork.

  benedict-96

  committed
  cb19f5b

  Copy full SHA for cb19f5b

• Added documentation.

  benedict-96

  committed
  cb7725f

  Copy full SHA for cb7725f

• Added constructor for optimizer if input arguments are flipped.

  benedict-96

  committed
  90bbaa7

  Copy full SHA for 90bbaa7

• Added a comment saying that the constructor can be called with DataLoader as input.

  benedict-96

  committed
  137d640

  Copy full SHA for 137d640

• Added default for number of epochs.

  benedict-96

  committed
  b57d148

  Copy full SHA for b57d148

• Combined matrix and tensor routines into one. Added another loss for autoencoder qp data.

  benedict-96

  committed
  3100b0b

  Copy full SHA for 3100b0b

• Commented out a section that is probably not needed.

  benedict-96

  committed
  718174f

  Copy full SHA for 718174f

• Test data loader for qp data.

  benedict-96

  committed
  370ee05

  Copy full SHA for 370ee05

• Renamed and added tests.

  benedict-96

  committed
  14d6ae2

  Copy full SHA for 14d6ae2

• Adjusted symplectic matrix.

  benedict-96

  committed
  9edec7b

  Copy full SHA for 9edec7b

• dim was missing for specifying default.

  benedict-96

  committed
  2acdd24

  Copy full SHA for 2acdd24

• dl has no field data.

  benedict-96

  committed
  f8b8fff

  Copy full SHA for f8b8fff

• Fixed typos.

  benedict-96

  committed
  8c4c73b

  Copy full SHA for 8c4c73b

• Merge branch 'increase_data_loader_test_coverage' of https://github.com/JuliaGNI/GeometricMachineLearning.jl into increase_data_loader_test_coverage

  benedict-96

  committed
  fc92faa

  Copy full SHA for fc92faa

• Removed method that appeared twice for some reason.

  benedict-96

  committed
  d072594

  Copy full SHA for d072594

• Forgot to commit before.

  benedict-96

  committed
  d071839

  Copy full SHA for d071839

Commits on Dec 20, 2023

• CompatHelper: bump compat for GPUArrays to 10, (keep existing compat)

  

  CompatHelper Julia

  committed
  2bb3f29

  Copy full SHA for 2bb3f29

• Merge pull request #102 from JuliaGNI/compathelper/new_version/2023-12-20-00-35-42-820-02318887484

  Show description for 8696833
  michakraus

  authored
  8696833

  Copy full SHA for 8696833

• Merge pull request #101 from JuliaGNI/increase_data_loader_test_coverage

  Show description for 9b56781
  michakraus

  authored
  9b56781

  Copy full SHA for 9b56781

• Fixed docs for optimize_for_one_epoch

  benedict-96

  committed
  d160bcf

  Copy full SHA for d160bcf

• Added descriptions for some functions in the docs.

  benedict-96

  committed
  b35e5e2

  Copy full SHA for b35e5e2

• Added eltype to AbstractCache.

  benedict-96

  committed
  8fe4832

  Copy full SHA for 8fe4832

• Removed comments that seem to have become unnecessary.

  benedict-96

  committed
  ebc5f71

  Copy full SHA for ebc5f71

• Resolved merge conflict with array_tests.jl (i.e. added vectorization test).

  benedict-96

  committed
  3c967e8

  Copy full SHA for 3c967e8

• Added sampling test for custom arrays (resolved conflict).

  benedict-96

  committed
  da77a96

  Copy full SHA for da77a96

• Resolved merge conflicts (for some reason appeared again).

  benedict-96

  committed
  dd343b9

  Copy full SHA for dd343b9