Commits · bc43944064000e413ca23ab681ea19e05699dfd9 · ModelZoo / ResNet50_tensorflow

"...matte-anything_pytorch.git" did not exist on "bea945787e15a4ef79c73e2d196e0da6d9cc17cb"

25 Mar, 2021 4 commits
- Clarify which part of the StandardEvaluator functions is in eager or graph mode. · bc439440
  Ruoxin Sang authored Mar 25, 2021
```
PiperOrigin-RevId: 365133146
```
  bc439440
- Open source the second half of multi-task library · e773b9b3
  A. Unique TensorFlower authored Mar 25, 2021
```
PiperOrigin-RevId: 365085378
```
  e773b9b3
- Internal change · 983837ff
  A. Unique TensorFlower authored Mar 25, 2021
```
PiperOrigin-RevId: 365082527
```
  983837ff
- Adds unit tests for using `cls_head` in `BertClassifier`. · a49c7332
  Jeremiah Liu authored Mar 24, 2021
```
PiperOrigin-RevId: 364952108
```
  a49c7332
24 Mar, 2021 5 commits
- Internal change · 44c2b999
  A. Unique TensorFlower authored Mar 23, 2021
```
PiperOrigin-RevId: 364736401
```
  44c2b999
- Add seed to DataConfig. Add deterministic tests for input reader. · 7f66d532
  Hongkun Yu authored Mar 23, 2021
```
PiperOrigin-RevId: 364734183
```
  7f66d532
- Merge pull request #9670 from ChristopherBrix:nlp_data_tagging_quadratic · 5ae21ff1
  A. Unique TensorFlower authored Mar 23, 2021
```
PiperOrigin-RevId: 364726354
```
  5ae21ff1
- Deprecate `network.Classification` from `BertClassifier`. · f87d6696
  Jeremiah Liu authored Mar 23, 2021
```
PiperOrigin-RevId: 364722225
```
  f87d6696
- Internal change · 0032b254
  A. Unique TensorFlower authored Mar 23, 2021
```
PiperOrigin-RevId: 364722131
```
  0032b254
23 Mar, 2021 8 commits
- Open source bert pretrain dataloader with dynamic sequence handling. · b25dec2b
  Hongkun Yu authored Mar 23, 2021
```
PiperOrigin-RevId: 364634581
```
  b25dec2b
- Internal change · 692da0ae
  Le Hou authored Mar 23, 2021
```
PiperOrigin-RevId: 364625895
```
  692da0ae
- Internal change · 9920d1b9
  A. Unique TensorFlower authored Mar 23, 2021
```
PiperOrigin-RevId: 364620438
```
  9920d1b9
- Parallelize the batching of the Shakespeare benchmark. · 0ffd6ab0
  Jay Shi authored Mar 23, 2021
```
PiperOrigin-RevId: 364541922
```
  0ffd6ab0
- Internal change · d5afcc72
  Yeqing Li authored Mar 22, 2021
```
PiperOrigin-RevId: 364488384
```
  d5afcc72
- Internal change · 4ddca747
  Yeqing Li authored Mar 22, 2021
```
PiperOrigin-RevId: 364483380
```
  4ddca747
- Internal change · b8a5c570
  Fan Yang authored Mar 22, 2021
```
PiperOrigin-RevId: 364458994
```
  b8a5c570
- Internal change · 44f942b4
  Yeqing Li authored Mar 22, 2021
```
PiperOrigin-RevId: 364443016
```
  44f942b4
22 Mar, 2021 7 commits

Internal change · 7bf19211
Yeqing Li authored Mar 22, 2021
```
PiperOrigin-RevId: 364433023
```
7bf19211
Internal change · b83ef168
A. Unique TensorFlower authored Mar 22, 2021
```
PiperOrigin-RevId: 364395434
```
b83ef168
Internal change · 8cb5ac1e
Poorva Potdar authored Mar 22, 2021
```
PiperOrigin-RevId: 364378436
```
8cb5ac1e

Add a new option "recreate_iterator_for_each_eval" into StandardEvaluator. · 0e6f8848

A. Unique TensorFlower authored Mar 22, 2021

In current implementation of StandardEvaluator, the iterator of evaluate dataset will go back (actually create a new iterator) to the beginning when evaluate() is called every time.

In some case, the iterator need to go ahead forever. So add the new option "recreate_iterator_for_each_eval" to control this behavior.

In some case, the iterator creation is time consuming. For example, careting an iterator of distributed dataset for multi worker in eager mode. So it's better to make a repeat dataset instead of recreating an iterator.

If "recreate_iterator_for_each_eval" is True, then the iterator will go back; Otherwise, keep going ahead.

PiperOrigin-RevId: 364356169

0e6f8848

Update layer documentation. · e27323d0
Jeremiah Liu authored Mar 22, 2021
```
PiperOrigin-RevId: 364328905
```
e27323d0
Implements `GaussianProcessClassificationHead`. · ac83d2b5
Jeremiah Liu authored Mar 21, 2021
```
PiperOrigin-RevId: 364226289
```
ac83d2b5
Minor readability improvements. · a6bbd3fa
Le Hou authored Mar 21, 2021
```
PiperOrigin-RevId: 364219111
```
a6bbd3fa

20 Mar, 2021 3 commits
- Adds spectral normalization and Gaussian process layers. · 1e9cbdce
  Jeremiah Liu authored Mar 20, 2021
```
PiperOrigin-RevId: 364109398
```
  1e9cbdce
- remove calls to metrics.result() inside the train_step · 1a8f129b
  A. Unique TensorFlower authored Mar 20, 2021
```
PiperOrigin-RevId: 364096705
```
  1a8f129b
- Internal change · 2b91303a
  A. Unique TensorFlower authored Mar 19, 2021
```
PiperOrigin-RevId: 364028490
```
  2b91303a
19 Mar, 2021 1 commit
- [core] Only use runtime_options for training. · 13bed58a
  Frederick Liu authored Mar 18, 2021
```
PiperOrigin-RevId: 363782489
```
  13bed58a
18 Mar, 2021 4 commits
- Fixes typos and use string for endpoints key. · 9abd85f2
  Yeqing Li authored Mar 18, 2021
```
PiperOrigin-RevId: 363736548
```
  9abd85f2
- Internal change · dc04b2b0
  Dana Movshovitz-Attias authored Mar 18, 2021
```
PiperOrigin-RevId: 363673488
```
  dc04b2b0
- [core] Add tpu_enable_xla_dynamic_padder to RuntimeConfig. · 124e3918
  Frederick Liu authored Mar 18, 2021
```
PiperOrigin-RevId: 363595377
```
  124e3918
- Add a unit test to demonstrate how to load an encoder through yaml and checkpoint. · 5e5f12f3
  Hongkun Yu authored Mar 17, 2021
```
PiperOrigin-RevId: 363593694
```
  5e5f12f3
17 Mar, 2021 2 commits
- Internal change · 9557e02a
  Chen Chen authored Mar 17, 2021
```
PiperOrigin-RevId: 363529035
```
  9557e02a
- Update TODOs for {Bert,Sentencepiece}Tokenizer.get_config(): · e4388f88
  A. Unique TensorFlower authored Mar 17, 2021
```
both are tracked in the same bug.

PiperOrigin-RevId: 363431367
```
  e4388f88
16 Mar, 2021 5 commits
- [nlp][translation] Consistently use float32 for layer norm for both encoder and decoder. · 45fd6cee
  Frederick Liu authored Mar 16, 2021
```
PiperOrigin-RevId: 363233691
```
  45fd6cee
- Support arbitrary dimensions in stochastic depth. · 17cec5c0
  Dan Kondratyuk authored Mar 16, 2021
```
PiperOrigin-RevId: 363188927
```
  17cec5c0
- [nlp][translation] Remove seq2seq model _dtype argument and break transformer utils dependency. · 8aae5ece
  Frederick Liu authored Mar 16, 2021
```
PiperOrigin-RevId: 363121226
```
  8aae5ece
- Internal change · cf7742e0
  A. Unique TensorFlower authored Mar 15, 2021
```
PiperOrigin-RevId: 363083122
```
  cf7742e0
- Fix AdamWeightDecay: self.gradient_clip_norm should be used. · 76f1113e
  Hongkun Yu authored Mar 15, 2021
```
PiperOrigin-RevId: 363069319
```
  76f1113e
15 Mar, 2021 1 commit
- Internal change · 3dcc078a
  A. Unique TensorFlower authored Mar 15, 2021
```
PiperOrigin-RevId: 362993979
```
  3dcc078a