Commits · a12cec09073704c9f7598dab89b2ee4dc4e39e8c · ModelZoo / ResNet50_tensorflow

12 Mar, 2020 1 commit
- Internal change · a12cec09
  Hongkun Yu authored Mar 11, 2020
```
PiperOrigin-RevId: 300477605
```
  a12cec09
11 Mar, 2020 1 commit
- Add TimeHistory callback to BERT. · b39958cf
  Will Cromar authored Mar 11, 2020
```
PiperOrigin-RevId: 300433601
```
  b39958cf
10 Mar, 2020 2 commits
- Internal change · 08f45dc4
  A. Unique TensorFlower authored Mar 10, 2020
```
PiperOrigin-RevId: 300203487
```
  08f45dc4
- Internal change · 6e910b1f
  A. Unique TensorFlower authored Mar 09, 2020
```
PiperOrigin-RevId: 300009545
```
  6e910b1f
09 Mar, 2020 1 commit
- Rename Attention to MultiheadAttention · f3a29bdd
  Hongkun Yu authored Mar 09, 2020
```
PiperOrigin-RevId: 299901483
```
  f3a29bdd
07 Mar, 2020 1 commit
- Add TimeHistory callback to BERT. · 7d86c317
  Hongkun Yu authored Mar 07, 2020
```
PiperOrigin-RevId: 299594839
```
  7d86c317
06 Mar, 2020 1 commit

Temporarily disable explicit allreduce in BERT SQuAD · 11ccb99e

Zongwei Zhou authored Mar 05, 2020

In BERT SQuAD, disable explicit allreduce for now to keep the original clip_by_global_norm math. With explicit allreduce, the gradients before allreduce are scaled so even if we move clip_by_global_norm before allreduce (as in TF1 and pre-TF 2.2) it will operate on scaled gradients, the math will be changed. So with explicit allreduce, it is better to move clip_by_global_norm to after allreduce.

PiperOrigin-RevId: 299278082

11ccb99e

05 Mar, 2020 6 commits
- Internal change · 2565629c
  Hongkun Yu authored Mar 05, 2020
```
PiperOrigin-RevId: 299169021
```
  2565629c
- Remove force_v2_in_keras_compile. experimental_run_tf_function is no-op now. · d3d7f15f
  Hongkun Yu authored Mar 05, 2020
```
PiperOrigin-RevId: 299160422
```
  d3d7f15f
- Internal change · 12271d7c
  A. Unique TensorFlower authored Mar 05, 2020
```
PiperOrigin-RevId: 299149155
```
  12271d7c
- Add README files in nlp/modeling folder. · ab6d40ca
  Jing Li authored Mar 05, 2020
```
PiperOrigin-RevId: 299126470
```
  ab6d40ca
- Internal change · cf01596c
  Zongwei Zhou authored Mar 04, 2020
```
PiperOrigin-RevId: 299007295
```
  cf01596c
- Internal change · 7a3d6c4c
  Chen Chen authored Mar 04, 2020
```
PiperOrigin-RevId: 298986381
```
  7a3d6c4c
04 Mar, 2020 2 commits
- Creates a models folder to host all keras.Model end points. · 0066ae22
  Hongkun Yu authored Mar 04, 2020
```
PiperOrigin-RevId: 298905114
```
  0066ae22
- Internal change · 238321c0
  A. Unique TensorFlower authored Mar 04, 2020
```
PiperOrigin-RevId: 298817372
```
  238321c0
03 Mar, 2020 2 commits
- Attributes->Arguments. Be consistent with keras style. · 651677f5
  Hongkun Yu authored Mar 03, 2020
```
PiperOrigin-RevId: 298692558
```
  651677f5
- Remove @tf.function(experimental_compile=True) annotation · add0620a
  George Karpenkov authored Mar 02, 2020
```
Removed with plans to re-add later once the feature stabilizes more.

PiperOrigin-RevId: 298486867
```
  add0620a
02 Mar, 2020 1 commit
- Add TimeHistory callback to BERT. · 533d1e6b
  Will Cromar authored Mar 02, 2020
```
PiperOrigin-RevId: 298466825
```
  533d1e6b
01 Mar, 2020 2 commits
- Add a README file in ALBERT folder. · fae6ca34
  Chen Chen authored Mar 01, 2020
```
PiperOrigin-RevId: 298227701
```
  fae6ca34
- Use hub_module_url in BERT readme file. · 6a76ce5b
  Chen Chen authored Feb 29, 2020
```
PiperOrigin-RevId: 298146392
```
  6a76ce5b
26 Feb, 2020 6 commits
- Test the Transformer layer with dynamic sequence · 2f9f2479
  George Karpenkov authored Feb 26, 2020
```
PiperOrigin-RevId: 297412889
```
  2f9f2479
- Internal change · ce83a9db
  Hongkun Yu authored Feb 26, 2020
```
PiperOrigin-RevId: 297383836
```
  ce83a9db
- Add a class-based decorator for only adding tf.function in TF2 mode · f2cf0cf4
  George Karpenkov authored Feb 26, 2020
```
PiperOrigin-RevId: 297366405
```
  f2cf0cf4
- Do not recreate tf.function on different calls to the same layer · e0c2c302
  George Karpenkov authored Feb 26, 2020
```
PiperOrigin-RevId: 297366158
```
  e0c2c302
- Internal change · 7b9365dd
  Hongkun Yu authored Feb 25, 2020
```
PiperOrigin-RevId: 297247401
```
  7b9365dd
- Clean up doc string. · 7c0e458b
  Hongkun Yu authored Feb 25, 2020
```
PiperOrigin-RevId: 297226166
```
  7c0e458b
25 Feb, 2020 4 commits
- Internal change · 2781377d
  Zongwei Zhou authored Feb 25, 2020
```
PiperOrigin-RevId: 297222995
```
  2781377d
- Convert numpy arrays to floats using float constructor, not .item(). · ed2e3bc3
  Reed Wanderman-Milne authored Feb 25, 2020
```
I plan on changing Keras to return Python floats instead of Numpy scalars, and this will break if .item() is called. Also, using the float constructor is cleaner, as it clearly communicates that a Python float is desired.

PiperOrigin-RevId: 297192289
```
  ed2e3bc3
- Only apply tf.function(experimental_compile=True) in eager mode · fe1fa6d6
  George Karpenkov authored Feb 25, 2020
```
Application in graph mode still leads to some crashes.

PiperOrigin-RevId: 297144398
```
  fe1fa6d6
- Creates modeling/performance.py to include mix prediction related stuff · fb35d6be
  Hongkun Yu authored Feb 24, 2020
```
PiperOrigin-RevId: 297002741
```
  fb35d6be
24 Feb, 2020 1 commit
- Branch a run_squad.py for albert. · 02d78796
  Chen Chen authored Feb 24, 2020
```
PiperOrigin-RevId: 296933982
```
  02d78796
23 Feb, 2020 2 commits
- Internal change · 89dd9a4e
  Zongwei Zhou authored Feb 23, 2020
```
PiperOrigin-RevId: 296767846
```
  89dd9a4e
- tf.keras.backend.sparse_categorical_crossentropy -> tf.keras.losses.sparse_categorical_crossentropy · c25d7711
  Hongkun Yu authored Feb 23, 2020
```
PiperOrigin-RevId: 296755118
```
  c25d7711
22 Feb, 2020 2 commits
- Internal change · 6f0e3a0b
  Hongkun Yu authored Feb 21, 2020
```
PiperOrigin-RevId: 296561671
```
  6f0e3a0b
- Internal change · e185b94c
  Hongkun Yu authored Feb 21, 2020
```
PiperOrigin-RevId: 296550109
```
  e185b94c
21 Feb, 2020 3 commits
- Enable XLA compilation using `@tf.function(experimental_compile=True) for transformer layer. · 393c1399
  George Karpenkov authored Feb 21, 2020
```
To debug the tf.function this API can be used: https://www.tensorflow.org/api_docs/python/tf/config/experimental_run_functions_eagerly

PiperOrigin-RevId: 296458870
```
  393c1399
- Fix tranformer path. · 867f0c47
  Hongkun Yu authored Feb 20, 2020
```
PiperOrigin-RevId: 296366426
```
  867f0c47
- Internal Change · e90a0b16
  Chen Chen authored Feb 20, 2020
```
PiperOrigin-RevId: 296325607
```
  e90a0b16
20 Feb, 2020 1 commit

Move data pre-processing related files (classifier_data_lib.py,... · c0ece9cc

Chen Chen authored Feb 20, 2020

Move data pre-processing related files (classifier_data_lib.py, create_finetuning_data.py, create_pretraining_data.py, squad_lib.py, squad_lib_sp.py) to data folder.

PiperOrigin-RevId: 296254023

c0ece9cc

18 Feb, 2020 1 commit
- Internal change · 2926ba69
  A. Unique TensorFlower authored Feb 18, 2020
```
PiperOrigin-RevId: 295757618
```
  2926ba69