- 24 Jul, 2020 1 commit
-
-
Hongkun Yu authored
PiperOrigin-RevId: 322960082
-
- 20 Jul, 2020 1 commit
-
-
Hongkun Yu authored
Modify the unit test of encoder scaffold to demonstrate that a subclass embedding network can be used. PiperOrigin-RevId: 322083775
-
- 16 Jul, 2020 1 commit
-
-
Hongkun Yu authored
PiperOrigin-RevId: 321591514
-
- 20 Jun, 2020 1 commit
-
-
A. Unique TensorFlower authored
Contributed by mickeystroller PiperOrigin-RevId: 317411747
-
- 16 Jun, 2020 1 commit
-
-
Hongkun Yu authored
PiperOrigin-RevId: 316593329
-
- 10 Jun, 2020 1 commit
-
-
A. Unique TensorFlower authored
PiperOrigin-RevId: 315789209
-
- 03 Jun, 2020 1 commit
-
-
Maxim Neumann authored
PiperOrigin-RevId: 314486753
-
- 21 May, 2020 2 commits
-
-
A. Unique TensorFlower authored
PiperOrigin-RevId: 312751112
-
Hongkun Yu authored
Transformer Encoder: when embedding width differs from hidden size, add a projection to hidden size. PiperOrigin-RevId: 312708922
-
- 14 May, 2020 2 commits
-
-
Chen Chen authored
PiperOrigin-RevId: 311597242
-
Jeremiah Harmsen authored
Add network and BERT model to perform per-token classification (e.g., for named entity recognition tasks). PiperOrigin-RevId: 311480326
-
- 13 May, 2020 2 commits
-
-
Chen Chen authored
PiperOrigin-RevId: 311428193
-
Scott Zhu authored
The intention of this change is to reduce the code complexity within Keras class, especially for Network, which currently contains logic for both subclass Model and functional Model. After this change, the subclass model and functional model become individual class and become self contained. 1. Model is now the base class for subclass model. It doesn't contains network structure management, and the topology will be created within __init__ and __call__, which is for user to implement. It also contains compile/fit/eval/predict, which is the basic functionality for model training. 2. Functional is created based on existing Network class. It extends the Model, which allows it leverage compile/fit/eval/predict. In addition, it also take input/output as init parameter and manage the network topology. 3. Sequential model is now a subclass of Functional, since it will use Functional's method to manage it topology (layer stacking). Model(input, output) will create a Functional under the hood, and behave the same way as before. PiperOrigin-RevId: 311232972
-
- 09 May, 2020 1 commit
-
-
A. Unique TensorFlower authored
Add an interface in nlp.modeling.networks.encoder_scaffold.EncoderScaffold and nlp.modeling.networks.transformer_encoder.TransformerEncoder, to get the pooler Dense layer reference. PiperOrigin-RevId: 310675150
-
- 08 May, 2020 1 commit
-
-
A. Unique TensorFlower authored
Add interfaces in nlp.modeling.networks.encoder_scaffold.EncoderScaffold, allowing it to output all hidden layer references, and all intermediate output data references. PiperOrigin-RevId: 310509202
-
- 05 May, 2020 1 commit
-
-
Hongkun Yu authored
PiperOrigin-RevId: 310032518
-
- 25 Apr, 2020 1 commit
-
-
Sergey Mironov authored
-
- 20 Apr, 2020 1 commit
-
-
Chen Chen authored
PiperOrigin-RevId: 307500045
-
- 08 Apr, 2020 1 commit
-
-
Chen Chen authored
num_output_classes -> pooled_output_dim classification_layer_initializer -> pooler_layer_initializer PiperOrigin-RevId: 305550415
-
- 12 Mar, 2020 1 commit
-
-
Hongkun Yu authored
PiperOrigin-RevId: 300477605
-
- 05 Mar, 2020 2 commits
-
-
Hongkun Yu authored
PiperOrigin-RevId: 299169021
-
Jing Li authored
PiperOrigin-RevId: 299126470
-
- 04 Mar, 2020 1 commit
-
-
Hongkun Yu authored
PiperOrigin-RevId: 298905114
-
- 03 Mar, 2020 1 commit
-
-
Hongkun Yu authored
PiperOrigin-RevId: 298692558
-
- 13 Feb, 2020 1 commit
-
-
Chen Chen authored
PiperOrigin-RevId: 294813724
-
- 12 Feb, 2020 1 commit
-
-
Hongkun Yu authored
PiperOrigin-RevId: 294588219
-
- 08 Feb, 2020 1 commit
-
-
Zongwei Zhou authored
PiperOrigin-RevId: 293958491
-
- 05 Feb, 2020 1 commit
-
-
Hongkun Yu authored
Adds return_all_encoder_outputs option to TransformerEncoder to build a network returning all encoder layers outputs. This is useful for Transformer Seq2Seq. PiperOrigin-RevId: 293395112
-
- 04 Feb, 2020 1 commit
-
-
Hongkun Yu authored
PiperOrigin-RevId: 293081282
-
- 21 Jan, 2020 1 commit
-
-
Hongkun Yu authored
Keras: "manual" shape inference is only required if the layer is dynamic (otherwise we use TF's static shape inference capabilities) PiperOrigin-RevId: 290821518
-
- 14 Jan, 2020 1 commit
-
-
A. Unique TensorFlower authored
PiperOrigin-RevId: 289743636
-
- 08 Jan, 2020 1 commit
-
-
A. Unique TensorFlower authored
PiperOrigin-RevId: 288774884
-
- 07 Jan, 2020 1 commit
-
-
Hongkun Yu authored
PiperOrigin-RevId: 288554071
-
- 19 Dec, 2019 1 commit
-
-
A. Unique TensorFlower authored
PiperOrigin-RevId: 286477560
-
- 06 Dec, 2019 1 commit
-
-
Chen Chen authored
PiperOrigin-RevId: 284203261
-
- 05 Dec, 2019 1 commit
-
-
Chen Chen authored
PiperOrigin-RevId: 283885197
-
- 13 Nov, 2019 1 commit
-
-
A. Unique TensorFlower authored
PiperOrigin-RevId: 280123567
-
- 11 Nov, 2019 1 commit
-
-
Hongkun Yu authored
- Update classifier example. - Add new converted checkpoints. - Update benchmark, PiperOrigin-RevId: 279762797
-