Commits · 68f301f7dedb1a82e1b3b0959d552d898c0df75d · ModelZoo / ResNet50_tensorflow

11 Dec, 2020 5 commits
- [keras_cv] Support custom policy for `AutoAugment`. · 68f301f7
  Hye Soo Yang authored Dec 11, 2020
```
PiperOrigin-RevId: 347069098
```
  68f301f7
- [keras_cv] Add more AutoAugment policies (svhn, reduced imagenet, reduced cifar10). · 590201bb
  Hye Soo Yang authored Dec 11, 2020
```
PiperOrigin-RevId: 347020444
```
  590201bb
- Internal change · 20453ae1
  Pengchong Jin authored Dec 11, 2020
```
PiperOrigin-RevId: 347013275
```
  20453ae1
- [Orbit] The global_step variable is required and cannot be None. · 55331222
  Hongkun Yu authored Dec 10, 2020
```
We use kwargs for all arguments and enforce them should be kwargs.
Reorder args.

PiperOrigin-RevId: 346908658
```
  55331222
- Provide PerClass IoU · 8de71c4b
  Zhenyu Tan authored Dec 10, 2020
```
PiperOrigin-RevId: 346895957
```
  8de71c4b
10 Dec, 2020 2 commits
- Move MobileBertMaskedLM to official/nlp/modeling/layers/mobile_bert_layers.py · be324167
  Chen Chen authored Dec 10, 2020
```
PiperOrigin-RevId: 346826537
```
  be324167
- Internal change · 03d1cda8
  A. Unique TensorFlower authored Dec 10, 2020
```
PiperOrigin-RevId: 346790357
```
  03d1cda8
09 Dec, 2020 4 commits
- Consolidate the self attention mask layer. · 5fc4c351
  Hongkun Yu authored Dec 09, 2020
```
Ask users to use keras_nlp.

PiperOrigin-RevId: 346621226
```
  5fc4c351
- Internal change · 74438979
  Poorva Potdar authored Dec 09, 2020
```
PiperOrigin-RevId: 346570482
```
  74438979
- Add train.md to describe how to run official/nlp/train.py in open-source world. · 4c6e23c8
  Chen Chen authored Dec 08, 2020
```
PiperOrigin-RevId: 346486072
```
  4c6e23c8
- Internal change · 5ee7a462
  Yeqing Li authored Dec 08, 2020
```
PiperOrigin-RevId: 346436382
```
  5ee7a462
08 Dec, 2020 4 commits
- Internal change · a99dc30d
  Chen Chen authored Dec 08, 2020
```
PiperOrigin-RevId: 346397839
```
  a99dc30d
- Internal change · b86dff8c
  A. Unique TensorFlower authored Dec 08, 2020
```
PiperOrigin-RevId: 346346141
```
  b86dff8c
- Remove files depending on tf-text because the nightly package has version problems. · 6d451a79
  Hongkun Yu authored Dec 07, 2020
```
PiperOrigin-RevId: 346217285
```
  6d451a79
- Update locations for NCF and Transformer, since the data has been moved to placer · 852d3ef6
  Pankaj Kanwar authored Dec 07, 2020
```
with the right replications.

PiperOrigin-RevId: 346213547
```
  852d3ef6
07 Dec, 2020 2 commits
- Internal change · 84dc1fd6
  A. Unique TensorFlower authored Dec 07, 2020
```
PiperOrigin-RevId: 346129282
```
  84dc1fd6
- Internal change · fd270433
  A. Unique TensorFlower authored Dec 07, 2020
```
PiperOrigin-RevId: 346088379
```
  fd270433
06 Dec, 2020 1 commit
- Internal change. · 8ffa9448
  Liangzhe Yuan authored Dec 06, 2020
```
PiperOrigin-RevId: 345961168
```
  8ffa9448
05 Dec, 2020 2 commits

Internal change · 09157a72
Yeqing Li authored Dec 04, 2020
```
PiperOrigin-RevId: 345796945
```
09157a72

Context R-CNN Updates: Added capabilities to use multi-headed or multi-layered... · 30b1c958

Sara Beery authored Dec 04, 2020

Context R-CNN Updates: Added capabilities to use multi-headed or multi-layered attention, to place the attention heads pre- or post-second stage feature extraction, and to work with embedded features in the context feature back from pre- or post-second stage feature extraction. Added an option for RPN feature map crops to be piped through to model outputs.

PiperOrigin-RevId: 345777219

30b1c958

04 Dec, 2020 10 commits
- Internal change · e7dfe641
  A. Unique TensorFlower authored Dec 04, 2020
```
PiperOrigin-RevId: 345773699
```
  e7dfe641
- Internal change · ef45f4a3
  Abdullah Rashwan authored Dec 04, 2020
```
PiperOrigin-RevId: 345761622
```
  ef45f4a3
- Internal change · 0fcb6dc8
  A. Unique TensorFlower authored Dec 04, 2020
```
PiperOrigin-RevId: 345761218
```
  0fcb6dc8
- Internal change · 53155a0d
  A. Unique TensorFlower authored Dec 04, 2020
```
PiperOrigin-RevId: 345712582
```
  53155a0d
- Piping image id lists for context features for attention visualization · f6b4cbcd
  Sara Beery authored Dec 04, 2020
```
DEFAULT_VALUE_OK=setting prefetch batches default to 2 as using -1 can cause memory issues for some models such as context r-cnn that use sequence examples.

PiperOrigin-RevId: 345682268
```
  f6b4cbcd
- Add checkpoint type 'full' to SSD meta arch. · 78f0e355
  Vighnesh Birodkar authored Dec 04, 2020
```
PiperOrigin-RevId: 345620862
```
  78f0e355
- Internal change · fd6b24c1
  A. Unique TensorFlower authored Dec 03, 2020
```
PiperOrigin-RevId: 345606908
```
  fd6b24c1
- Internal change · 3c61d591
  A. Unique TensorFlower authored Dec 03, 2020
```
PiperOrigin-RevId: 345588520
```
  3c61d591
- Support for config override of sample_from_dataset_weights. · d933decc
  Yoni Ben-Meshulam authored Dec 03, 2020
```
PiperOrigin-RevId: 345586686
```
  d933decc
- Improve testing and documentation for weighted dataset sampling. · b4499b90
  Yoni Ben-Meshulam authored Dec 03, 2020
```
PiperOrigin-RevId: 345586199
```
  b4499b90
03 Dec, 2020 6 commits

Updating Context R-CNN dataset tools to fix a bug in context feature bank... · 101282ad

Sara Beery authored Dec 03, 2020

Updating Context R-CNN dataset tools to fix a bug in context feature bank building logic when the number of context examples in the time horizon is greater than the specified maximum, and adding capabilities to track and save the number of embeddings stored per image.

PiperOrigin-RevId: 345548385

101282ad

Internal change · 1e205552
A. Unique TensorFlower authored Dec 03, 2020
```
PiperOrigin-RevId: 345529749
```
1e205552
Internal change · 495fbc4a
A. Unique TensorFlower authored Dec 03, 2020
```
PiperOrigin-RevId: 345500920
```
495fbc4a
Internal change · 609d31c4
A. Unique TensorFlower authored Dec 03, 2020
```
PiperOrigin-RevId: 345493076
```
609d31c4

Refactored the keypoint postprocessing logics to reduce duplicated codes and · ca0eb4a6

Yu-hui Chen authored Dec 03, 2020

added a new operating mode for single-instance prediction.

1) Refactored the _postprocess_keypoints_for_class_and_image function such that
it can be reused by single/multi class keypoint tasks.
2) Removed the "mod" operator to make the model compatible with WASM.

PiperOrigin-RevId: 345468250

ca0eb4a6

Internal change · e3f8ea22
Hongkun Yu authored Dec 02, 2020
```
PiperOrigin-RevId: 345383199
```
e3f8ea22

02 Dec, 2020 3 commits
- Internal change · abcb643f
  A. Unique TensorFlower authored Dec 02, 2020
```
PiperOrigin-RevId: 345267761
```
  abcb643f
- Fix the case when output_range is set and attention_mask is None. · bd7535b2
  A. Unique TensorFlower authored Dec 02, 2020
```
PiperOrigin-RevId: 345261173
```
  bd7535b2
- Update op handler (#9521) · c97c6208
  pyoung2778 authored Dec 01, 2020
  
  c97c6208
01 Dec, 2020 1 commit

Avoid explicit tf.unstack in... · f7b4c6de

Vivek Rathod authored Dec 01, 2020

Avoid  explicit tf.unstack in `convert_strided_predictions_to_normalized_boxes`. This enables exporting a saved model with dynamic batch size.

PiperOrigin-RevId: 345055072

f7b4c6de