Commits · c9dddda4ef6eaef2eaf36d4a97335239f8e00ce7 · ModelZoo / ResNet50_tensorflow

"docs/vscode:/vscode.git/clone" did not exist on "ff15ba1b5ae8c589332d21212295436c3b68c4fa"

03 Apr, 2018 32 commits
- Minor fixes to third party code. · c9dddda4
  Zhichao Lu authored Apr 02, 2018
```
- Fix externally broken test input_test.py
- Add Cython to object detection REQUIRED_PACKAGES

PiperOrigin-RevId: 191338837
```
  c9dddda4
- Mscoco download script minor fix. · 5444724e
  Zhichao Lu authored Apr 01, 2018
```
PiperOrigin-RevId: 191248237
```
  5444724e
- Identity context manager to help setting up conditional with statements. · bc99a72e
  Zhichao Lu authored Mar 30, 2018
```
PiperOrigin-RevId: 191112757
```
  bc99a72e
- Add object detection gwsq to gwsq config. · 24ca4146
  Zhichao Lu authored Mar 29, 2018
```
PiperOrigin-RevId: 191014662
```
  24ca4146
- Add support to single channel images in visualization tool. · 8da696bf
  Zhichao Lu authored Mar 29, 2018
```
PiperOrigin-RevId: 190996725
```
  8da696bf
- Utility to write config files to disk. · 642e1005
  Zhichao Lu authored Mar 29, 2018
```
PiperOrigin-RevId: 190972737
```
  642e1005
- Updating model to infer when labels are padded, rather than relying solely on... · e9f85e83
  Zhichao Lu authored Mar 29, 2018
```
Updating model to infer when labels are padded, rather than relying solely on the mode. This is necessary for evaluating on train data.

PiperOrigin-RevId: 190960173
```
  e9f85e83
- Use keyword arguments in model builder for ssd feature extractor. This was... · 312b09e2
  Zhichao Lu authored Mar 29, 2018
```
Use keyword arguments in model builder for ssd feature extractor. This was incorreclty setting reuse argument.

PiperOrigin-RevId: 190940462
```
  312b09e2
- Internal code review config change. · 5bbe41c4
  Zhichao Lu authored Mar 28, 2018
```
PiperOrigin-RevId: 190790653
```
  5bbe41c4
- Removing flag defaults for train steps and eval steps. If not provided, will... · 056bea45
  Zhichao Lu authored Mar 27, 2018
```
Removing flag defaults for train steps and eval steps. If not provided, will revert to the values specified in the config.

PiperOrigin-RevId: 190706470
```
  056bea45
- Add PNASNet-based Faster R-CNN object detection. · 85b1bdf0
  Zhichao Lu authored Mar 27, 2018
```
PiperOrigin-RevId: 190705850
```
  85b1bdf0
- Provide option to perform in-place batch norm updates for ssd feature extractors. · d2c5bfac
  Zhichao Lu authored Mar 27, 2018
```
PiperOrigin-RevId: 190688309
```
  d2c5bfac
- Use mobilenet_v1_arg_scope to pass the weight_decay and batch norm params. · 3956d90e
  Zhichao Lu authored Mar 27, 2018
```
PiperOrigin-RevId: 190682119
```
  3956d90e
- Add warning if shuffle=false but num_readers > 1. The latter would cause the... · d724a08b
  Zhichao Lu authored Mar 27, 2018
```
Add warning if shuffle=false but num_readers > 1. The latter would cause the input data stream to still be shuffled.

PiperOrigin-RevId: 190678792
```
  d724a08b
- Improved shape inference for reshape · 952dd997
  Zhichao Lu authored Mar 27, 2018
```
PiperOrigin-RevId: 190651873
```
  952dd997
- Write groundtruth weights from input pipeline into model. · bb43ed96
  Zhichao Lu authored Mar 27, 2018
```
PiperOrigin-RevId: 190636417
```
  bb43ed96
- Update Installation.md · 45069b91
  Zhichao Lu authored Mar 26, 2018
```
Cython is must installed package for building cocoapi. else the build will not
initiate.

PiperOrigin-RevId: 190553060
```
  45069b91
- Tfmini requires positive values for the dimensions. Therefore, we need to test... · 2398e6a5
  Zhichao Lu authored Mar 26, 2018
```
Tfmini requires positive values for the dimensions. Therefore, we need to test for no paddings in both width and height direction.

PiperOrigin-RevId: 190520529
```
  2398e6a5
- Use get_data_files_path to access test data. · 4a91d110
  Zhichao Lu authored Mar 26, 2018
```
PiperOrigin-RevId: 190505306
```
  4a91d110
- Switch line orders in trainer so that restore_map is called after moving... · 93b8168a
  Zhichao Lu authored Mar 26, 2018
```
Switch line orders in trainer so that restore_map is called after moving average variables are created.

Moving averages are now properly loaded during fine-tuning, instead of being recreated.

PiperOrigin-RevId: 190496046
```
  93b8168a
- Expose read block length via input reader proto so it can be tuned. · ed877c00
  Zhichao Lu authored Mar 23, 2018
```
PiperOrigin-RevId: 190292280
```
  ed877c00
- Change to fit Python 3, where filter() works as generator. · 2a43dca3
  Zhichao Lu authored Mar 22, 2018
```
PiperOrigin-RevId: 190115628
```
  2a43dca3
- Remove unused leakr_badwords.dic · 88f116f8
  Zhichao Lu authored Mar 22, 2018
```
PiperOrigin-RevId: 190098000
```
  88f116f8
- Make box_list_ops_test.py work with the C API enabled. · f74ae662
  Zhichao Lu authored Mar 22, 2018
```
The C API improves static shape inference, making more errors caught at graph construction time instead of runtime.

PiperOrigin-RevId: 190095450
```
  f74ae662
- Correct the return value from ProcessBatch in the case of an exception. · 8fdf31d6
  Zhichao Lu authored Mar 22, 2018
```
PiperOrigin-RevId: 190068159
```
  8fdf31d6
- Update submission check. · 60f170e7
  Zhichao Lu authored Mar 21, 2018
```
PiperOrigin-RevId: 190002150
```
  60f170e7
- Minor change to remove redundant check for finetune_checkpoint_type. · bda5126e
  Zhichao Lu authored Mar 21, 2018
```
PiperOrigin-RevId: 189997094
```
  bda5126e
- Update running_pets instructions to link to running locally. · 2672cdbd
  Zhichao Lu authored Mar 21, 2018
```
PiperOrigin-RevId: 189933390
```
  2672cdbd
- Provide groundtruth masks and keypoints in the evaluator. · d7f255f6
  Zhichao Lu authored Mar 21, 2018
```
PiperOrigin-RevId: 189929413
```
  d7f255f6
- Apply fix for image summary in eval metrics, so that eval on train works appropriately. · e46021a7
  Zhichao Lu authored Mar 20, 2018
```
PiperOrigin-RevId: 189815553
```
  e46021a7
- Updating transform_input_data to resize original image. This is necessary for... · 7007d9e3
  Zhichao Lu authored Mar 20, 2018
```
Updating transform_input_data to resize original image. This is necessary for result_dict_for_single_example(), since it expects the input image and groundtruth masks to be of the same spatial dimensions.

PiperOrigin-RevId: 189786443
```
  7007d9e3
- Enabling both train and eval image summaries. Note that eval summaries are not... · 6f1756bc
  Zhichao Lu authored Mar 19, 2018
```
Enabling both train and eval image summaries. Note that eval summaries are not created tf-learn environment. To get them to show up, added the summary image into the eval_metric_ops.

PiperOrigin-RevId: 189658259
```
  6f1756bc
23 Mar, 2018 5 commits
- Merge pull request #3669 from sm176357/patch-1 · 1d38a225
  Yukun Zhu authored Mar 22, 2018
```
syntax error : download_and_convert_VOC2012.sh
```
  1d38a225
- Merge pull request #3709 from YknZhu/master · f78281e9
  Yukun Zhu authored Mar 22, 2018
```
Add mobilenetv2 support in deeplab
```
  f78281e9
- Merge pull request #3686 from hsm207/patch-2 · d4cd02b9
  Yukun Zhu authored Mar 22, 2018
```
Fix DeprecationWarning
```
  d4cd02b9
- Merge pull request #3693 from hsm207/patch-3 · 51ea4925
  Yukun Zhu authored Mar 22, 2018
```
Update comment on folder structure
```
  51ea4925
- PiperOrigin-RevId: 190154671 · 7a2f1a3e
  Liang-Chieh Chen authored Mar 22, 2018
  
  7a2f1a3e
22 Mar, 2018 3 commits
- Merge pull request #3691 from tensorflow/readme-run-instructions · 67d65c69
  Neal Wu authored Mar 22, 2018
```
Updated README run instructions
```
  67d65c69
- Merge pull request #3694 from tensorflow/resnet-l2-loss · f478547e
  Neal Wu authored Mar 22, 2018
```
Add a tf.summary.scalar call for L2 loss in resnet
```
  f478547e
- Merge pull request #3702 from cclauss/from-six.moves-import-xrange-yet-again · 6daba3b4
  Neal Wu authored Mar 22, 2018
```
from six.moves import xrange (en masse) YET AGAIN
```
  6daba3b4