Commits · e91779b873a4e0261cd9a5a41297d4ffd93eb53e · ModelZoo / ResNet50_tensorflow

01 May, 2020 1 commit
- Internal change · e0d33132
  Tianqi Liu authored Apr 30, 2020
```
PiperOrigin-RevId: 309331866
```
  e0d33132
26 Apr, 2020 1 commit
- Internal change · 4b53cd23
  Chen Chen authored Apr 26, 2020
```
PiperOrigin-RevId: 308473864
```
  4b53cd23
25 Apr, 2020 1 commit
- Make BertPretrainer to accept embedding_table explicitly · f7852565
  Sergey Mironov authored Apr 16, 2020
  
  f7852565
23 Apr, 2020 1 commit
- Internal change · e3ae4b5e
  André Susano Pinto authored Apr 23, 2020
```
PiperOrigin-RevId: 307998953
```
  e3ae4b5e
22 Apr, 2020 2 commits
- Add model metrics to the training summary. · 22f10c60
  Chen Chen authored Apr 22, 2020
```
PiperOrigin-RevId: 307914609
```
  22f10c60
- Internal change · 04165099
  Le Hou authored Apr 21, 2020
```
PiperOrigin-RevId: 307712303
```
  04165099
21 Apr, 2020 2 commits
- Internal change · 330b34fe
  Hongkun Yu authored Apr 21, 2020
```
PiperOrigin-RevId: 307689094
```
  330b34fe
- Test that training=True activates dropout in a reusable SavedModel for BERT. · 5e539a3d
  A. Unique TensorFlower authored Apr 21, 2020
```
PiperOrigin-RevId: 307633374
```
  5e539a3d
20 Apr, 2020 2 commits
- In BERT's export to TF Hub, fix shape propagation for seq_length. · b3377b09
  A. Unique TensorFlower authored Apr 20, 2020
```
PiperOrigin-RevId: 307425903
```
  b3377b09
- Internal change · e289c8c7
  A. Unique TensorFlower authored Apr 20, 2020
```
PiperOrigin-RevId: 307416709
```
  e289c8c7
19 Apr, 2020 1 commit
- Internal Change · 4e434726
  Chen Chen authored Apr 18, 2020
```
PiperOrigin-RevId: 307233807
```
  4e434726
17 Apr, 2020 3 commits
- Remove hooks. · 2ddd716e
  Hongkun Yu authored Apr 16, 2020
```
PiperOrigin-RevId: 306994199
```
  2ddd716e
- Internal change · 01c4ee81
  Chen Chen authored Apr 16, 2020
```
PiperOrigin-RevId: 306986791
```
  01c4ee81
- Thread custom callbacks through BERT classifier. This allows injection of... · b523f160
  Jeremiah Harmsen authored Apr 16, 2020
```
Thread custom callbacks through BERT classifier.  This allows injection of custom callbacks for library users of run_classifier.

PiperOrigin-RevId: 306966758
```
  b523f160
16 Apr, 2020 1 commit

Add calling of custom callback on_epoch_begin() and on_epoch_end() methods. · 31da2245

Jeremiah Harmsen authored Apr 16, 2020

Minor refactoring to (1) encapsulate callbacks in a CallbackList and (2) allow evaluation results to be threaded to on_epoch_end callback.

PiperOrigin-RevId: 306827406

31da2245

15 Apr, 2020 3 commits
- Add flag for do_lower_case in export_tfhub.py · 31711f66
  Elizabeth Kemp authored Apr 15, 2020
```
PiperOrigin-RevId: 306669872
```
  31711f66
- Internal change · 89599a23
  Le Hou authored Apr 15, 2020
```
PiperOrigin-RevId: 306664983
```
  89599a23
- Remove extraneous evaluation & export which happens on the final epoch. · 8931d298
  Jeremiah Harmsen authored Apr 15, 2020
```
Preserves current behavior for submodels of exporting duplicates (e.g., for 100 total steps, my_submodel_step_100.ckpt and my_submodel.ckpt are equivalent and both written after the final epoch).

PiperOrigin-RevId: 306629202
```
  8931d298
13 Apr, 2020 1 commit
- Allow to not use next sentence labels in pretraining. · 3db5347d
  Chen Chen authored Apr 13, 2020
```
PiperOrigin-RevId: 306324960
```
  3db5347d
10 Apr, 2020 1 commit
- Add end learning rate (end_lr) FLAG for linear learning rate decay. · 22de957a
  A. Unique TensorFlower authored Apr 10, 2020
```
PiperOrigin-RevId: 305948522
```
  22de957a
08 Apr, 2020 2 commits

Rename arguments of EncoderScaffold: · db6aca44

Chen Chen authored Apr 08, 2020

num_output_classes -> pooled_output_dim
classification_layer_initializer -> pooler_layer_initializer

PiperOrigin-RevId: 305550415

db6aca44

Internal change · 91495d13
Chen Chen authored Apr 08, 2020
```
PiperOrigin-RevId: 305530702
```
91495d13

07 Apr, 2020 1 commit
- Internal change · 2ff0a3fe
  A. Unique TensorFlower authored Apr 07, 2020
```
PiperOrigin-RevId: 305332859
```
  2ff0a3fe
31 Mar, 2020 1 commit
- Internal change · ad34b621
  A. Unique TensorFlower authored Mar 30, 2020
```
PiperOrigin-RevId: 303871250
```
  ad34b621
27 Mar, 2020 2 commits
- Support Lamb optimizer within BERT. · da5860f2
  A. Unique TensorFlower authored Mar 27, 2020
```
PiperOrigin-RevId: 303356961
```
  da5860f2
- Internal change · 0265f59c
  A. Unique TensorFlower authored Mar 26, 2020
```
PiperOrigin-RevId: 303225001
```
  0265f59c
26 Mar, 2020 1 commit
- Move modeling_training_utils.py to bert folder · 6ef71d49
  Hongkun Yu authored Mar 25, 2020
```
PiperOrigin-RevId: 303050809
```
  6ef71d49
25 Mar, 2020 1 commit
- Internal change · 38486725
  Hongkun Yu authored Mar 25, 2020
```
PiperOrigin-RevId: 302977474
```
  38486725
24 Mar, 2020 1 commit
- Internal change · 1ec383c8
  A. Unique TensorFlower authored Mar 23, 2020
```
PiperOrigin-RevId: 302552996
```
  1ec383c8
19 Mar, 2020 1 commit
- Internal change · 4ae295c3
  A. Unique TensorFlower authored Mar 18, 2020
```
PiperOrigin-RevId: 301741984
```
  4ae295c3
17 Mar, 2020 3 commits
- Adds scale_loss to run_customized_training_loop, which should be the correct treatment. · bf6d6f6f
  Hongkun Yu authored Mar 17, 2020
```
PiperOrigin-RevId: 301441181
```
  bf6d6f6f
- Uses performance.configure_optimizer to reduce duplicate code. · 69837027
  Hongkun Yu authored Mar 16, 2020
```
PiperOrigin-RevId: 301308623
```
  69837027
- Internal change · 2d7223ad
  Ken Franko authored Mar 16, 2020
```
PiperOrigin-RevId: 301277545
```
  2d7223ad
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
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 3 commits
- Internal change · 2565629c
  Hongkun Yu authored Mar 05, 2020
```
PiperOrigin-RevId: 299169021
```
  2565629c
- 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