Code

Customizable scorer#118

Code

Merged

AKuederle merged 30 commits intomainmad-lab-fau/tpcp:mainfrom customizable_scorermad-lab-fau/tpcp:customizable_scorerCopy head branch name to clipboardAug 23, 2024

Commits

Commits on Aug 23, 2024

• Refactoring of scoring to allow Configurable aggregator

  AKuederle

  committed
  dcf6425

  Copy full SHA for dcf6425

• Made everything configuragble aggregators. This means all aggregators are now instances not classes

  AKuederle

  committed
  81d6ea5

  Copy full SHA for 81d6ea5

• Reworked example text

  AKuederle

  committed
  15429b8

  Copy full SHA for 15429b8

• Typo

  AKuederle

  committed
  ceed06e

  Copy full SHA for ceed06e

• Changelog and some more tests

  AKuederle

  committed
  85062f9

  Copy full SHA for 85062f9

• Formatting

  AKuederle

  committed
  a82a890

  Copy full SHA for a82a890

• Made it possible to add default aggs to doc

  AKuederle

  committed
  937f526

  Copy full SHA for 937f526

• FIxed stupid bug

  AKuederle

  committed
  1e75897

  Copy full SHA for 1e75897

• Added 3.12 to testing

  AKuederle

  committed
  0bb5abb

  Copy full SHA for 0bb5abb

• opefully fixed tests for 3.12

  AKuederle

  committed
  aa54d2d

  Copy full SHA for aa54d2d

• Now correctly assert mock

  AKuederle

  committed
  d0aa3b0

  Copy full SHA for d0aa3b0

• Got sidetracked implemented new global caching for pipelines

  AKuederle

  committed
  77a3bcf

  Copy full SHA for 77a3bcf

• Reworked custom scorere example yet again and added new Macro Scorer

  AKuederle

  committed
  a334a7c

  Copy full SHA for a334a7c

• Massive refactor of global cahce to make it more robust in parallel envs

  AKuederle

  committed
  277c5b3

  Copy full SHA for 277c5b3

• Implemented more caching tests based on recent changes

  AKuederle

  committed
  f1396cc

  Copy full SHA for f1396cc

• Fixed some remaining bugs with caching refactor

  AKuederle

  committed
  3e320be

  Copy full SHA for 3e320be

• Delayed import of torch and tensorflow. This should massivly increase import speed of tpcp, when torch and tensorflow are installed, but not used in the context of tpcp

  AKuederle

  committed
  41ecc3b

  Copy full SHA for 41ecc3b

• Linting

  AKuederle

  committed
  1a6b744

  Copy full SHA for 1a6b744

• Added new remove func for global parallel backend

  AKuederle

  committed
  9cc19c9

  Copy full SHA for 9cc19c9

• Added info about the torch/tensorflow import fix

  AKuederle

  committed
  f7477f6

  Copy full SHA for f7477f6

• Removed redundant type annot

  AKuederle

  committed
  dd973a7

  Copy full SHA for dd973a7

• Make sure to wait for process pool to shut down before running next test

  AKuederle

  committed
  76200e2

  Copy full SHA for 76200e2

• Added file with test class

  AKuederle

  committed
  7be6869

  Copy full SHA for 7be6869

• Readded fix for parallel tests

  AKuederle

  committed
  092485d

  Copy full SHA for 092485d

• Allow hashing of more nested objects

  AKuederle

  committed
  9db105e

  Copy full SHA for 9db105e

• Added some black magic, to allow the caching of classes defined in __main__

  AKuederle

  committed
  9de7300

  Copy full SHA for 9de7300

• Fixed recusion error in repr or aggregators

  AKuederle

  committed
  45c923b

  Copy full SHA for 45c923b

• Updated custom scorer example

  AKuederle

  committed
  b0496c6

  Copy full SHA for b0496c6

• Added newest changes

  AKuederle

  committed
  c030ce5

  Copy full SHA for c030ce5

• Added MacroFloatAgg to docs

  AKuederle

  committed
  31bb6a3

  Copy full SHA for 31bb6a3