GitLab

PiperOrigin-RevId: 273966871

PiperOrigin-RevId: 273861263

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

* 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

PiperOrigin-RevId: 273795511

PiperOrigin-RevId: 273653001

PiperOrigin-RevId: 273562498

PiperOrigin-RevId: 273527676

Bert fp16 perf improvements, do the matmul in intermediate later in fp16, and also remove explicit casting to fp32 for layerNorm.

PiperOrigin-RevId: 273379063

PiperOrigin-RevId: 273371605

PiperOrigin-RevId: 273358759

PiperOrigin-RevId: 273233857

PiperOrigin-RevId: 273066504

PiperOrigin-RevId: 272934570

PiperOrigin-RevId: 272915002

PiperOrigin-RevId: 272777104

PiperOrigin-RevId: 272700366

As the model code is subject to a major change, we do not release hub module at this moment.

PiperOrigin-RevId: 272688279

* 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

PiperOrigin-RevId: 272280612

PiperOrigin-RevId: 272121528

PiperOrigin-RevId: 272077584

PiperOrigin-RevId: 272043067

Adds a swish activation without customized gradients.

PiperOrigin-RevId: 272029817

PiperOrigin-RevId: 271873759

Support variable reshape to make TF1 checkpoint compatible with a Bert without reshape in einsum layers.

PiperOrigin-RevId: 271613961

PiperOrigin-RevId: 271611082

PiperOrigin-RevId: 271422100

The reference implementation can be found at
https://github.com/tensorflow/minigo

This fork was originally created to experiment with performance upgrades
for MLPerf, but since MLPerf work is focused in the original repo,
this fork's existence only serves to confuse.

PiperOrigin-RevId: 270926016

PiperOrigin-RevId: 270817869

PiperOrigin-RevId: 270791796

PiperOrigin-RevId: 270749832

Update readme

PiperOrigin-RevId: 270346317

Add specification for what dataset link to use on the website.