Commits · 0375a63b548b77720c175152ba43c896b03f71b8 · ModelZoo / ResNet50_tensorflow

"router/vscode:/vscode.git/clone" did not exist on "6e127dcc9616a0e39efd95a3e08324138e8c4df7"

12 Jul, 2021 2 commits
- Fix the issue when both numerator and denominator are 0, the dice score is... · 0375a63b
  Fan Yang authored Jul 12, 2021
```
Fix the issue when both numerator and denominator are 0, the dice score is incorrectly computed as 1, which should be 0.

PiperOrigin-RevId: 384257731
```
  0375a63b
- Internal change. · 9c377959
  Fan Yang authored Jul 12, 2021
```
PiperOrigin-RevId: 384257276
```
  9c377959
10 Jul, 2021 2 commits
- Internal change · adc12a16
  A. Unique TensorFlower authored Jul 10, 2021
```
PiperOrigin-RevId: 384018258
```
  adc12a16
- [efficient] Fix is_short_seq order so that we can also apply feature transform... · 7aab3475
  Frederick Liu authored Jul 09, 2021
```
[efficient] Fix is_short_seq order so that we can also apply feature transform and apply softmax afterwards.

PiperOrigin-RevId: 383967806
```
  7aab3475
09 Jul, 2021 4 commits

Refactor decoder factory to allow registering other decoders. · 9a052f52
Fan Yang authored Jul 09, 2021
```
PiperOrigin-RevId: 383944185
```
9a052f52
Internal change · c6afac2c
Dan Kondratyuk authored Jul 09, 2021
```
PiperOrigin-RevId: 383861528
```
c6afac2c
Internal change · 1437baae
Dan Kondratyuk authored Jul 09, 2021
```
PiperOrigin-RevId: 383856955
```
1437baae

With float16, always use LossScaleOptimizer. · be3575f5

Reed Wanderman-Milne authored Jul 08, 2021

Before, it was too easy to accidentally forget to set runtime.loss_scale, which had to always be done if mixed precision is used, otherwise the model would converge to worse accuracy. Now, all that needs to be done to use mixed precision is to set runtime.mixed_precision_dtype=float16.

PiperOrigin-RevId: 383767033

be3575f5

08 Jul, 2021 3 commits
- Internal change · 9499011f
  A. Unique TensorFlower authored Jul 08, 2021
```
PiperOrigin-RevId: 383679175
```
  9499011f
- Internal change · ee5693ce
  Hongkun Yu authored Jul 08, 2021
```
PiperOrigin-RevId: 383667471
```
  ee5693ce
- Fix typo in TransformerEncoderBlock docstring · a977b4e1
  Philip Pham authored Jul 08, 2021
```
PiperOrigin-RevId: 383660517
```
  a977b4e1
07 Jul, 2021 2 commits
- Internal change · f3022ed8
  Abdullah Rashwan authored Jul 07, 2021
```
PiperOrigin-RevId: 383505893
```
  f3022ed8
- Internal change · 10b7712e
  Hongkun Yu authored Jul 07, 2021
```
PiperOrigin-RevId: 383505455
```
  10b7712e
05 Jul, 2021 1 commit
- [core] Remove adafactor factory test since it's not implemented which causes external test to fail. · f504a75b
  Frederick Liu authored Jul 04, 2021
```
PiperOrigin-RevId: 383066915
```
  f504a75b
04 Jul, 2021 1 commit
- [nlp] Support norm_first option for Bert/Bigbird/Kernel encoders. · 097a8296
  Frederick Liu authored Jul 03, 2021
```
PiperOrigin-RevId: 382960239
```
  097a8296
03 Jul, 2021 1 commit
- Internal change · c2582c3e
  Frederick Liu authored Jul 02, 2021
```
PiperOrigin-RevId: 382846192
```
  c2582c3e
02 Jul, 2021 2 commits
- Remove private Keras API usage. · ef21dabb
  Dan Kondratyuk authored Jul 02, 2021
```
PiperOrigin-RevId: 382833629
```
  ef21dabb
- Internal change · c1e086ef
  A. Unique TensorFlower authored Jul 01, 2021
```
PiperOrigin-RevId: 382655843
```
  c1e086ef
01 Jul, 2021 7 commits
- Change activation/add/upsampling/identity to a keras layer form for the spinenet_mobile. · 7b0dde4d
  Jaehong Kim authored Jul 01, 2021
```
Added use_keras_upsampling_2d flags to use keras upsampling layer instead of optimized custom spatial_transform_ops.

PiperOrigin-RevId: 382626602
```
  7b0dde4d
- Revert causal padding until new checkpoints are trained. · b40047b6
  Dan Kondratyuk authored Jul 01, 2021
```
PiperOrigin-RevId: 382606601
```
  b40047b6
- Internal change · cf3e3d30
  A. Unique TensorFlower authored Jul 01, 2021
```
PiperOrigin-RevId: 382571340
```
  cf3e3d30
- Internal change · d4256e47
  Abdullah Rashwan authored Jul 01, 2021
```
PiperOrigin-RevId: 382561014
```
  d4256e47
- Merge pull request #10081 from Atze00:padding_causal_convolution · eb8dda4d
  A. Unique TensorFlower authored Jul 01, 2021
```
PiperOrigin-RevId: 382547358
```
  eb8dda4d
- Internal change · 683b1caa
  Abdullah Rashwan authored Jul 01, 2021
```
PiperOrigin-RevId: 382462213
```
  683b1caa
- Internal change · eeff8d87
  Abdullah Rashwan authored Jun 30, 2021
```
PiperOrigin-RevId: 382422745
```
  eeff8d87
30 Jun, 2021 4 commits
- Internal change · f24c95ae
  Xianzhi Du authored Jun 30, 2021
```
PiperOrigin-RevId: 382408293
```
  f24c95ae
- Add placeholder for Adafactor optimizer. · edc65712
  A. Unique TensorFlower authored Jun 30, 2021
```
PiperOrigin-RevId: 382396878
```
  edc65712
- Internal change · d7333c86
  Xianzhi Du authored Jun 30, 2021
```
PiperOrigin-RevId: 382382393
```
  d7333c86
- Internal change · 2e9e0478
  Abdullah Rashwan authored Jun 29, 2021
```
PiperOrigin-RevId: 382240268
```
  2e9e0478
29 Jun, 2021 2 commits
- Add an identity decoder field to the volumetric_models decoder config. · 1624201f
  A. Unique TensorFlower authored Jun 29, 2021
```
PiperOrigin-RevId: 382051540
```
  1624201f
- Internal change · 97ae894b
  A. Unique TensorFlower authored Jun 29, 2021
```
PiperOrigin-RevId: 382051212
```
  97ae894b
28 Jun, 2021 1 commit
- Add activation function parameters to backbone. · 1b8ac0cb
  Dan Kondratyuk authored Jun 28, 2021
```
PiperOrigin-RevId: 381908843
```
  1b8ac0cb
25 Jun, 2021 5 commits
- Internal change · 3d62e895
  A. Unique TensorFlower authored Jun 25, 2021
```
PiperOrigin-RevId: 381540013
```
  3d62e895
- Internal change · c8a91782
  Fan Yang authored Jun 25, 2021
```
PiperOrigin-RevId: 381516130
```
  c8a91782
- Internal change · 6e5cbee1
  A. Unique TensorFlower authored Jun 24, 2021
```
PiperOrigin-RevId: 381396116
```
  6e5cbee1
- Add functionality to create fake data and use that in train_test to show how an e2e training works. · fd505f48
  Fan Yang authored Jun 24, 2021
```
PiperOrigin-RevId: 381368435
```
  fd505f48
- Internal change · c73c0127
  A. Unique TensorFlower authored Jun 24, 2021
```
PiperOrigin-RevId: 381367878
```
  c73c0127
24 Jun, 2021 3 commits
- Adds a feature to process a dictionary for `input_paths` in `DataConfig`, to... · 88b2a354
  A. Unique TensorFlower authored Jun 24, 2021
```
Adds a feature to process a dictionary for `input_paths` in `DataConfig`, to allow combining multiple datasets using a user defined combine_fn.

PiperOrigin-RevId: 381363688
```
  88b2a354
- small change · c703ef58
  Atze00 authored Jun 24, 2021
  
  c703ef58
- Add model export config to RetinaNet, which optionally cast model outputs to... · 50ebc683
  A. Unique TensorFlower authored Jun 24, 2021
```
Add model export config to RetinaNet, which optionally cast model outputs to floats, and normalize output box coordinates.

PiperOrigin-RevId: 381334851
```
  50ebc683