Skip to content

GitLab

Switch branch/tag
  1. 03 Aug, 2020 5 commits
  3. 01 Aug, 2020 2 commits
  5. 31 Jul, 2020 8 commits
  7. 30 Jul, 2020 11 commits
    • A. Unique TensorFlower's avatar
      Internal change · 93248443
      A. Unique TensorFlower authored 
      PiperOrigin-RevId: 324084377
      93248443
    • A. Unique TensorFlower's avatar
      Internal change · 2e785497
      A. Unique TensorFlower authored 
      PiperOrigin-RevId: 324084377
      2e785497
    • xinliupitt's avatar
      func name change · 673231ff
      xinliupitt authored
      673231ff
    • xinliupitt's avatar
      transformer test case · 72efa854
      xinliupitt authored
      72efa854
    • xinliupitt's avatar
      get_config and doc · 1c5dca7f
      xinliupitt authored
      1c5dca7f
    • A. Unique TensorFlower's avatar
      Internal change · fe04a5ba
      A. Unique TensorFlower authored 
      PiperOrigin-RevId: 323948101
      fe04a5ba
    • A. Unique TensorFlower's avatar
      Internal change · 5366f605
      A. Unique TensorFlower authored 
      PiperOrigin-RevId: 323948101
      5366f605
    • Tomer Kaftan's avatar
      Pre-emptively disable the KerasTensors refactoring for the detection models in... · 071bb243
      Tomer Kaftan authored 
      Pre-emptively disable the KerasTensors refactoring for the detection models in tensorflow models/official/vision/detection, because they rely on several unsupported things that will stop working entirely when the refactoring goes live.

Specifically:
* The custom layers implement `__call__` instead of `call` and rely on manually enter the keras backend graph
* The vision models try to use `tf.while_loop` as Keras op layers during functional API construction, which is unsupported.

Updating the models to avoid this would subtly change the variable names and break the pre-existing tf1-style name-based checkpoints, so for now we will just disable the KerasTensors refactoring for these models.

PiperOrigin-RevId: 323937426
      071bb243
    • Tomer Kaftan's avatar
      Pre-emptively disable the KerasTensors refactoring for the detection models in... · ccc60760
      Tomer Kaftan authored 
      Pre-emptively disable the KerasTensors refactoring for the detection models in tensorflow models/official/vision/detection, because they rely on several unsupported things that will stop working entirely when the refactoring goes live.

Specifically:
* The custom layers implement `__call__` instead of `call` and rely on manually enter the keras backend graph
* The vision models try to use `tf.while_loop` as Keras op layers during functional API construction, which is unsupported.

Updating the models to avoid this would subtly change the variable names and break the pre-existing tf1-style name-based checkpoints, so for now we will just disable the KerasTensors refactoring for these models.

PiperOrigin-RevId: 323937426
      ccc60760
    • Hongkun Yu's avatar
      Remove backward compatibility hack. · 52504780
      Hongkun Yu authored 
      PiperOrigin-RevId: 323927843
      52504780
    • Hongkun Yu's avatar
      Remove backward compatibility hack. · de8c6ace
      Hongkun Yu authored 
      PiperOrigin-RevId: 323927843
      de8c6ace
  9. 29 Jul, 2020 11 commits
  11. 28 Jul, 2020 3 commits