Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Add LayoutLMv2 + LayoutXLM #12604

Merged
merged 114 commits into from Aug 30, 2021
Merged

Add LayoutLMv2 + LayoutXLM #12604

merged 114 commits into from Aug 30, 2021

Commits on Aug 30, 2021

  1. First commit

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    4744ab8 View commit details
    Browse the repository at this point in the history

  2. Make style

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    0a3a7e4 View commit details
    Browse the repository at this point in the history

  3. Fix dummy objects

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    39da694 View commit details
    Browse the repository at this point in the history

  4. Add Detectron2 config

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    5b18c28 View commit details
    Browse the repository at this point in the history

  5. Add LayoutLMv2 pooler

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    5348460 View commit details
    Browse the repository at this point in the history

  6. More improvements, add documentation

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    9be733e View commit details
    Browse the repository at this point in the history

  7. More improvements

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    3814c54 View commit details
    Browse the repository at this point in the history

  8. Add model tests

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    125ada5 View commit details
    Browse the repository at this point in the history

  9. Add clarification regarding image input

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    76a5a0f View commit details
    Browse the repository at this point in the history

  10. Improve integration test

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    480ebe1 View commit details
    Browse the repository at this point in the history

  11. Fix bug

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    5b2f585 View commit details
    Browse the repository at this point in the history

  12. Fix another bug

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    060a684 View commit details
    Browse the repository at this point in the history

  13. Fix another bug

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    5e61df4 View commit details
    Browse the repository at this point in the history

  14. Fix another bug

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    e731f67 View commit details
    Browse the repository at this point in the history

  15. More improvements

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    d0ca865 View commit details
    Browse the repository at this point in the history

  16. Make more tests pass

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    604dd9b View commit details
    Browse the repository at this point in the history

  17. Make more tests pass

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    b4c172e View commit details
    Browse the repository at this point in the history

  18. Improve integration test

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    7fb70b5 View commit details
    Browse the repository at this point in the history

  19. Remove gradient checkpointing and add head masking

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    e6c1318 View commit details
    Browse the repository at this point in the history

  20. Add integration test

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    aaef300 View commit details
    Browse the repository at this point in the history

  21. Add LayoutLMv2ForSequenceClassification to the tests

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    b470d03 View commit details
    Browse the repository at this point in the history

  22. Add LayoutLMv2ForQuestionAnswering

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    dfe5ea7 View commit details
    Browse the repository at this point in the history

  23. More improvements

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    59c1cf6 View commit details
    Browse the repository at this point in the history

  24. More improvements

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    33ffd98 View commit details
    Browse the repository at this point in the history

  25. Small improvements

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    aa15dbf View commit details
    Browse the repository at this point in the history

  26. Fix _LazyModule

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    28b576a View commit details
    Browse the repository at this point in the history

  27. Fix fast tokenizer

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    6229e02 View commit details
    Browse the repository at this point in the history

  28. Move sync_batch_norm to a separate method

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    d9ff738 View commit details
    Browse the repository at this point in the history

  29. Replace dummies by requires_backends

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    c681657 View commit details
    Browse the repository at this point in the history

  30. Move calculation of visual bounding boxes to separate method + update… 

    … README
    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    fa97538 View commit details
    Browse the repository at this point in the history

  31. Add models to main init

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    ba0bc0e View commit details
    Browse the repository at this point in the history

  32. First draft

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    cd67bfa View commit details
    Browse the repository at this point in the history

  33. More improvements

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    287abfa View commit details
    Browse the repository at this point in the history

  34. More improvements

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    8c0948f View commit details
    Browse the repository at this point in the history

  35. More improvements

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    88be5de View commit details
    Browse the repository at this point in the history

  36. More improvements

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    373811f View commit details
    Browse the repository at this point in the history

  37. More improvements

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    48c53c0 View commit details
    Browse the repository at this point in the history

  38. Remove is_split_into_words

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    b92db14 View commit details
    Browse the repository at this point in the history

  39. More improvements

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    fcb505a View commit details
    Browse the repository at this point in the history

  40. Simply tesseract - no use of pandas anymore

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    86bb3ab View commit details
    Browse the repository at this point in the history

  41. Add LayoutLMv2Processor

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    0ae53ff View commit details
    Browse the repository at this point in the history

  42. Update is_pytesseract_available

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    1adbaf8 View commit details
    Browse the repository at this point in the history

  43. Fix bugs

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    d5cf7c2 View commit details
    Browse the repository at this point in the history

  44. Improve feature extractor

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    0382104 View commit details
    Browse the repository at this point in the history

  45. Fix bug

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    d06248d View commit details
    Browse the repository at this point in the history

  46. Add print statement

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    075590b View commit details
    Browse the repository at this point in the history

  47. Add truncation of bounding boxes

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    b6b277e View commit details
    Browse the repository at this point in the history

  48. Add tests for LayoutLMv2FeatureExtractor and LayoutLMv2Tokenizer

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    258060a View commit details
    Browse the repository at this point in the history

  49. Improve tokenizer tests

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    2a166ca View commit details
    Browse the repository at this point in the history

  50. Make more tokenizer tests pass

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    ab3b0ef View commit details
    Browse the repository at this point in the history

  51. Make more tests pass, add integration tests

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    214b491 View commit details
    Browse the repository at this point in the history

  52. Finish integration tests

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    0ae6e3b View commit details
    Browse the repository at this point in the history

  53. More improvements

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    ea84ad6 View commit details
    Browse the repository at this point in the history

  54. More improvements - update API of the tokenizer

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    bba6100 View commit details
    Browse the repository at this point in the history

  55. More improvements

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    ebc2541 View commit details
    Browse the repository at this point in the history

  56. Remove support for VQA training

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    93d93b7 View commit details
    Browse the repository at this point in the history

  57. Remove some files

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    0b4c97b View commit details
    Browse the repository at this point in the history

  58. Improve feature extractor

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    5a24365 View commit details
    Browse the repository at this point in the history

  59. Improve documentation and one more tokenizer test

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    f04672c View commit details
    Browse the repository at this point in the history

  60. Make quality and small docs improvements

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    98ca2a2 View commit details
    Browse the repository at this point in the history

  61. Add batched tests for LayoutLMv2Processor, remove fast tokenizer

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    7804d69 View commit details
    Browse the repository at this point in the history

  62. Add truncation of labels

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    0ea905b View commit details
    Browse the repository at this point in the history

  63. Apply suggestions from code review

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    8db4e13 View commit details
    Browse the repository at this point in the history

  64. Improve processor tests

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    0e7d10e View commit details
    Browse the repository at this point in the history

  65. Fix failing tests and add suggestion from code review

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    4bccc97 View commit details
    Browse the repository at this point in the history

  66. Fix tokenizer test

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    fd12133 View commit details
    Browse the repository at this point in the history

  67. Add detectron2 CI job

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    23d0570 View commit details
    Browse the repository at this point in the history

  68. Simplify CI job

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    40c1b6d View commit details
    Browse the repository at this point in the history

  69. Comment out non-detectron2 jobs and specify number of processes

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    124dd86 View commit details
    Browse the repository at this point in the history

  70. Add pip install torchvision

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    c59bffe View commit details
    Browse the repository at this point in the history

  71. Add durations to see which tests are slow

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    c299ff0 View commit details
    Browse the repository at this point in the history

  72. Fix tokenizer test and make model tests smaller

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    f7ea2fe View commit details
    Browse the repository at this point in the history

  73. Frist draft

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    da85fbc View commit details
    Browse the repository at this point in the history

  74. Use setattr

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    0401e4d View commit details
    Browse the repository at this point in the history

  75. Possible fix

    @LysandreJik @NielsRogge
    LysandreJik authored and NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    2e43af8 View commit details
    Browse the repository at this point in the history

  76. Proposal with configuration

    @LysandreJik @NielsRogge
    LysandreJik authored and NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    e6d6efc View commit details
    Browse the repository at this point in the history

  77. First draft of fast tokenizer

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    546bfb9 View commit details
    Browse the repository at this point in the history

  78. More improvements

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    507d724 View commit details
    Browse the repository at this point in the history

  79. Enable fast tokenizer tests

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    4101b29 View commit details
    Browse the repository at this point in the history

  80. Make more tests pass

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    a582226 View commit details
    Browse the repository at this point in the history

  81. Make more tests pass

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    67cca2f View commit details
    Browse the repository at this point in the history

  82. More improvements

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    2379176 View commit details
    Browse the repository at this point in the history

  83. Addd padding to fast tokenizer

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    c8151e7 View commit details
    Browse the repository at this point in the history

  84. Mkae more tests pass

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    d6ea661 View commit details
    Browse the repository at this point in the history

  85. Make more tests pass

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    b0c7eca View commit details
    Browse the repository at this point in the history

  86. Make all tests pass for fast tokenizer

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    7613c27 View commit details
    Browse the repository at this point in the history

  87. Make fast tokenizer support overflowing boxes and labels

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    38934e9 View commit details
    Browse the repository at this point in the history

  88. Add support for overflowing_labels to slow tokenizer

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    066a9ec View commit details
    Browse the repository at this point in the history

  89. Add support for fast tokenizer to the processor

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    4446c8a View commit details
    Browse the repository at this point in the history

  90. Update processor tests for both slow and fast tokenizers

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    42ebf01 View commit details
    Browse the repository at this point in the history

  91. Add head models to model mappings

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    5082dbd View commit details
    Browse the repository at this point in the history

  92. Make style & quality

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    b703ea2 View commit details
    Browse the repository at this point in the history

  93. Remove Detectron2 config file

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    b22011d View commit details
    Browse the repository at this point in the history

  94. Add configurable option to label all subwords

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    beb6f69 View commit details
    Browse the repository at this point in the history

  95. Fix test

    @LysandreJik @NielsRogge
    LysandreJik authored and NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    be1eaa1 View commit details
    Browse the repository at this point in the history

  96. Skip visual segment embeddings in test

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    659bd94 View commit details
    Browse the repository at this point in the history

  97. Use ResNet-18 backbone in tests instead of ResNet-101

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    66b5cbe View commit details
    Browse the repository at this point in the history

  98. Proposal

    @LysandreJik @NielsRogge
    LysandreJik authored and NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    ba5a44f View commit details
    Browse the repository at this point in the history

  99. Re-enable all jobs on CI

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    a8e6997 View commit details
    Browse the repository at this point in the history

  100. Fix installation of tesseract

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    3ab8384 View commit details
    Browse the repository at this point in the history

  101. Fix failing test

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    c417b04 View commit details
    Browse the repository at this point in the history

  102. Fix index table

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    84b33b8 View commit details
    Browse the repository at this point in the history

  103. Add LayoutXLM doc page, first draft of code examples

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    6142f97 View commit details
    Browse the repository at this point in the history

  104. Improve documentation a lot

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    2c5d412 View commit details
    Browse the repository at this point in the history

  105. Update expected boxes for Tesseract 4.0.0 beta

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    e86b4cf View commit details
    Browse the repository at this point in the history

  106. Use offsets to create labels instead of checking if they start with ##

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    9eb372b View commit details
    Browse the repository at this point in the history

  107. Update expected boxes for Tesseract 4.1.1

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    e5c5f61 View commit details
    Browse the repository at this point in the history

  108. Fix conflict

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    5114199 View commit details
    Browse the repository at this point in the history

  109. Make variable names cleaner, add docstring, add link to notebooks

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    d429492 View commit details
    Browse the repository at this point in the history

  110. Revert "Fix conflict" 

    This reverts commit a9b46ce.
    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    2e0a4f9 View commit details
    Browse the repository at this point in the history

  111. Revert to make integration test pass

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    748308f View commit details
    Browse the repository at this point in the history

  112. Apply suggestions from @LysandreJik's review

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    e273e77 View commit details
    Browse the repository at this point in the history

  113. Address @patrickvonplaten's comments

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    a72f080 View commit details
    Browse the repository at this point in the history

  114. Remove fixtures DocVQA in favor of dataset on the hub

    @NielsRogge
    NielsRogge committed Aug 30, 2021
    Configuration menu
    Copy the full SHA
    2391ca5 View commit details
    Browse the repository at this point in the history