- 12 10月, 2019 3 次提交
-
-
由 Yeqing Li 提交于
PiperOrigin-RevId: 274241934
-
由 A. Unique TensorFlower 提交于
Change summary directory and model checkpoint directory so that training via Keras Compile/Fit() and custom training loop is consistent. PiperOrigin-RevId: 274202793
-
由 Hongkun Yu 提交于
PiperOrigin-RevId: 274201399
-
- 11 10月, 2019 11 次提交
-
-
-
由 Tao 提交于
-
由 Hongkun Yu 提交于
PiperOrigin-RevId: 274090672
-
由 Reed Wanderman-Milne 提交于
PiperOrigin-RevId: 274090348
-
由 A. Unique TensorFlower 提交于
change benchmark's log verbosity to logging.INFO. it seems to me that DEBUG map to ---v=1 internally, which is way to verbose for the purpose of benchmarking. PiperOrigin-RevId: 274040907
-
由 Yeqing Li 提交于
PiperOrigin-RevId: 274035928
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 274028786
-
由 Yeqing Li 提交于
PiperOrigin-RevId: 274023277
-
由 Hongkun Yu 提交于
PiperOrigin-RevId: 274015143
-
由 Yeqing Li 提交于
PiperOrigin-RevId: 274010788
-
由 Navid Lambert-Shirzad 提交于
-
- 10 10月, 2019 5 次提交
-
-
由 Hongkun Yu 提交于
PiperOrigin-RevId: 273966871
-
由 Hongkun Yu 提交于
PiperOrigin-RevId: 273861263
-
由 Reed Wanderman-Milne 提交于
Instead of needing to ensure variables are float32, casting inputs to float32, etc, instead dtype="float32" is passed to the layer constructor, which will do all that logic automatically. The only difference is the output of LayerNorm is now float32 instead of float16, so an extra cast is needed elsewhere. PiperOrigin-RevId: 273833286
-
由 Pooya Davoodi 提交于
* Updating python API to use CombinedNonMaxSuppresion TF operator 1. Adds a unit test to test post_processing python API 2. Currently sets clip_window to None as the kernel uses the default clip_window of [0,0,1,1] 3. Added use_static_shapes to the API. In old API if use_static_shapes is true, then it pads/clips outputs to max_total_size, if specified. If not specified, it pads to num_classes*max_size_per_class. If use_static_shapes is false, it always pads/clips to max_total_size. Update unit test to account for clipped bouding boxes Changed the name to CombinedNonMaxSuppression based on feedback from Google Added additional parameters to combinedNMS python function. They are currently unused and required for networks like FasterRCNN and MaskRCNN * Delete selected_indices from API Because it was removed from CombinedNMS recently in the PR. * Improve doc of function combined_non_max_suppression * Enable CombinedNonMaxSuppression for first_stage_nms * fix bug * Ensure agnostic_nms is not used with combined_nms Remove redundant arguments from combined_nms * Fix pylint * Add checks for unsupported args * Fix pylint * Move combined_non_max_suppression to batch_multiclass_non_max_suppression Also rename combined_nms to use_combined_nms * Delete combined_nms for first_stage_nms because it does not work * Revert "Delete combined_nms for first_stage_nms because it does not work" This reverts commit 2a3cc5145f17cee630a67ddedd20e90c2920fa9f. * Use nmsed_additional_fields.get to avoid error * Merge combined_non_max_suppression with main nms function * Rename combined_nms for first stage nms * Improve docs * Use assertListEqual for numpy arrays * Fix pylint errors * End comments with period
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 273795511
-
- 09 10月, 2019 2 次提交
-
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 273653001
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 273562498
-
- 08 10月, 2019 4 次提交
-
-
由 George Karpenkov 提交于
PiperOrigin-RevId: 273527676
-
由 A. Unique TensorFlower 提交于
Bert fp16 perf improvements, do the matmul in intermediate later in fp16, and also remove explicit casting to fp32 for layerNorm. PiperOrigin-RevId: 273379063
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 273371605
-
由 Jing Li 提交于
PiperOrigin-RevId: 273358759
-
- 07 10月, 2019 1 次提交
-
-
由 Jing Li 提交于
PiperOrigin-RevId: 273233857
-
- 06 10月, 2019 1 次提交
-
-
由 Jing Li 提交于
PiperOrigin-RevId: 273066504
-
- 05 10月, 2019 2 次提交
-
-
由 Hongkun Yu 提交于
PiperOrigin-RevId: 272934570
-
由 Jing Li 提交于
PiperOrigin-RevId: 272915002
-
- 04 10月, 2019 3 次提交
-
-
由 Hongkun Yu 提交于
PiperOrigin-RevId: 272777104
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 272700366
-
由 Hongkun Yu 提交于
As the model code is subject to a major change, we do not release hub module at this moment. PiperOrigin-RevId: 272688279
-
- 03 10月, 2019 2 次提交
-
-
由 Sergey Mironov 提交于
-
由 Guangda Lai 提交于
* Fix TensorRT test output and add int8 result. * Fix comments * Removing footnote declaration as well * Remove TF-TRT examples and point users to new example repo * Update urls * Fix review comments and links
-
- 02 10月, 2019 1 次提交
-
-
由 Hongkun Yu 提交于
PiperOrigin-RevId: 272280612
-
- 01 10月, 2019 4 次提交
-
-
由 David Chen 提交于
PiperOrigin-RevId: 272121528
-
由 George Karpenkov 提交于
PiperOrigin-RevId: 272077584
-
由 Hongkun Yu 提交于
PiperOrigin-RevId: 272043067
-
由 Hongkun Yu 提交于
Adds a swish activation without customized gradients. PiperOrigin-RevId: 272029817
-
- 30 9月, 2019 1 次提交
-
-
由 Hongkun Yu 提交于
PiperOrigin-RevId: 271873759
-