- 12 11月, 2020 4 次提交
-
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 341913626
-
由 Chen Chen 提交于
PiperOrigin-RevId: 341913299
-
由 Hongkun Yu 提交于
When loss is NaN, the weights should also be NaN. PiperOrigin-RevId: 341886095
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 341852981
-
- 11 11月, 2020 7 次提交
-
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 341713387
-
由 Chen Chen 提交于
PiperOrigin-RevId: 341709524
-
由 Chen Chen 提交于
PiperOrigin-RevId: 341708068
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 341693911
-
由 Pengchong Jin 提交于
PiperOrigin-RevId: 341687182
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 341682938
-
由 Allen Wang 提交于
PiperOrigin-RevId: 341642152
-
- 10 11月, 2020 3 次提交
-
-
由 A. Unique TensorFlower 提交于
PiperOrigin-RevId: 341573525
-
由 Pengchong Jin 提交于
PiperOrigin-RevId: 341443730
-
由 Poorva Potdar 提交于
PiperOrigin-RevId: 341433786
-
- 09 11月, 2020 1 次提交
-
-
由 Ruoxin Sang 提交于
PiperOrigin-RevId: 341329653
-
- 07 11月, 2020 6 次提交
-
-
由 Vighnesh Birodkar 提交于
PiperOrigin-RevId: 341188685
-
由 Vivek Rathod 提交于
Make label_id_offset an argument so it can be turned off when eval_util.result_dict_for_batched_example is applied on the outputs of an exported SavedModel. Exported saved models already apply label offset. PiperOrigin-RevId: 341176846
-
由 Vivek Rathod 提交于
Run inference under distribution strategy, gather outputs locally and evaluate the results with coco tools on cpu. PiperOrigin-RevId: 341162083
-
由 Poorva Potdar 提交于
PiperOrigin-RevId: 341123614
-
由 thunderfyc 提交于
* Rename sequence_projection to seq_flow_lite * Rename sequence_projection to seq_flow_lite
-
由 Le Hou 提交于
PiperOrigin-RevId: 341074795
-
- 06 11月, 2020 2 次提交
-
-
由 Pengchong Jin 提交于
PiperOrigin-RevId: 340908345
-
由 Allen Wang 提交于
PiperOrigin-RevId: 340897846
-
- 05 11月, 2020 2 次提交
-
-
由 Hongkun Yu 提交于
PiperOrigin-RevId: 340748908
-
由 Hongkun Yu 提交于
PiperOrigin-RevId: 340745997
-
- 04 11月, 2020 3 次提交
-
-
由 Hongkun Yu 提交于
PiperOrigin-RevId: 340580428
-
由 Pankaj Kanwar 提交于
PiperOrigin-RevId: 340527173
-
由 Hongkun Yu 提交于
PiperOrigin-RevId: 340481326
-
- 03 11月, 2020 3 次提交
-
-
由 Hongkun Yu 提交于
PiperOrigin-RevId: 340385514
-
由 Jonathan Huang 提交于
PiperOrigin-RevId: 340366780
-
由 Jonathan Huang 提交于
Fixes two bugs when handling verified_neg_classes and not_exhaustive_classes fields for LVIS evaluation: (1) Before, if one of these fields was empty, the input pipeline would default to an all-ones representation; this CL turns out off this behavior for the LVIS-specific classes. (2) Labels are now 1-indexed coming out of the _prepare_groundtruth_for_eval function in model_lib (as they should be). PiperOrigin-RevId: 340357845
-
- 31 10月, 2020 3 次提交
-
-
由 Pankaj Kanwar 提交于
PiperOrigin-RevId: 339927293
-
由 Abdullah Rashwan 提交于
PiperOrigin-RevId: 339917189
-
由 Le Hou 提交于
PiperOrigin-RevId: 339901099
-
- 30 10月, 2020 3 次提交
-
-
由 Hongkun Yu 提交于
PiperOrigin-RevId: 339805681
-
由 Hongkun Yu 提交于
PiperOrigin-RevId: 339804825
-
由 Abdullah Rashwan 提交于
PiperOrigin-RevId: 339773662
-
- 29 10月, 2020 3 次提交
-
-
由 Jonathan Huang 提交于
Plumb LVIS specific fields (e.g. `neg_category_ids`, `not_exhaustive_category_ids`) through input pipelines. PiperOrigin-RevId: 339614575
-
由 Abdullah Rashwan 提交于
PiperOrigin-RevId: 339598839
-
由 Yin Cui 提交于
PiperOrigin-RevId: 339565966
-