Internal change

PiperOrigin-RevId: 285844156

official/vision/detection/configs/base_config.py

@@ -30,10 +30,11 @@ REGULARIZATION_VAR_REGEX = r'.*(kernel|weight):0$'
 BASE_CFG = {
     'model_dir': '',
     'use_tpu': True,
+    'strategy_type': 'tpu',
     'isolate_session_state': False,
     'train': {
         'iterations_per_loop': 100,
-        'train_batch_size': 64,
+        'batch_size': 64,
         'total_steps': 22500,
         'num_cores_per_replica': None,
         'input_partition_dims': None,
@@ -57,13 +58,13 @@ BASE_CFG = {
         'frozen_variable_prefix': RESNET_FROZEN_VAR_PREFIX,
         'train_file_pattern': '',
         'train_dataset_type': 'tfrecord',
-        'transpose_input': True,
+        'transpose_input': False,
         'regularization_variable_regex': REGULARIZATION_VAR_REGEX,
         'l2_weight_decay': 0.0001,
         'gradient_clip_norm': 0.0,
     },
     'eval': {
-        'eval_batch_size': 8,
+        'batch_size': 8,
         'eval_samples': 5000,
         'min_eval_interval': 180,
         'eval_timeout': None,

official/vision/detection/configs/maskrcnn_config.py

@@ -34,6 +34,7 @@ MASKRCNN_CFG.override({
     'maskrcnn_parser': {
         'use_bfloat16': True,
         'output_size': [1024, 1024],
+        'num_channels': 3,
         'rpn_match_threshold': 0.7,
         'rpn_unmatched_threshold': 0.3,
         'rpn_batch_size_per_im': 256,

official/vision/detection/dataloader/maskrcnn_parser.py

@@ -275,6 +275,10 @@ class Parser(object):
     if self._use_bfloat16:
       image = tf.cast(image, dtype=tf.bfloat16)
 
+    inputs = {
+        'image': image,
+        'image_info': image_info,
+    }
     # Packs labels for model_fn outputs.
     labels = {
         'anchor_boxes': input_anchor.multilevel_boxes,
@@ -282,15 +286,16 @@ class Parser(object):
         'rpn_score_targets': rpn_score_targets,
         'rpn_box_targets': rpn_box_targets,
     }
-    labels['gt_boxes'] = input_utils.pad_to_fixed_size(
-        boxes, self._max_num_instances, -1)
-    labels['gt_classes'] = input_utils.pad_to_fixed_size(
+    inputs['gt_boxes'] = input_utils.pad_to_fixed_size(boxes,
+                                                       self._max_num_instances,
+                                                       -1)
+    inputs['gt_classes'] = input_utils.pad_to_fixed_size(
         classes, self._max_num_instances, -1)
     if self._include_mask:
-      labels['gt_masks'] = input_utils.pad_to_fixed_size(
+      inputs['gt_masks'] = input_utils.pad_to_fixed_size(
           masks, self._max_num_instances, -1)
 
-    return image, labels
+    return inputs, labels
 
   def _parse_eval_data(self, data):
     """Parses data for evaluation."""
@@ -348,11 +353,7 @@ class Parser(object):
         self._anchor_size,
         (image_height, image_width))
 
-    labels = {
-        'source_id': dataloader_utils.process_source_id(data['source_id']),
-        'anchor_boxes': input_anchor.multilevel_boxes,
-        'image_info': image_info,
-    }
+    labels = {}
 
     if self._mode == ModeKeys.PREDICT_WITH_GT:
       # Converts boxes from normalized coordinates to pixel coordinates.
@@ -372,6 +373,11 @@ class Parser(object):
           groundtruths['source_id'])
       groundtruths = dataloader_utils.pad_groundtruths_to_fixed_size(
           groundtruths, self._max_num_instances)
+      # TODO(yeqing):  Remove the `groundtrtuh` layer key (no longer needed).
       labels['groundtruths'] = groundtruths
+    inputs = {
+        'image': image,
+        'image_info': image_info,
+    }
 
-    return image, labels
+    return inputs, labels

official/vision/detection/modeling/base_model.py

@@ -99,6 +99,7 @@ class Model(object):
         params.train.learning_rate)
 
     self._frozen_variable_prefix = params.train.frozen_variable_prefix
+    self._l2_weight_decay = params.train.l2_weight_decay
 
     # Checkpoint restoration.
     self._checkpoint = params.train.checkpoint.as_dict()

official/vision/detection/modeling/losses.py

@@ -147,6 +147,7 @@ class RpnBoxLoss(object):
   """Region Proposal Network box regression loss function."""
 
   def __init__(self, params):
+    self._delta = params.huber_loss_delta
     self._huber_loss = tf.keras.losses.Huber(
         delta=params.huber_loss_delta, reduction=tf.keras.losses.Reduction.SUM)
 
@@ -212,7 +213,7 @@ class FastrcnnClassLoss(object):
       a scalar tensor representing total class loss.
     """
     with tf.name_scope('fast_rcnn_loss'):
-      _, _, _, num_classes = class_outputs.get_shape().as_list()
+      _, _, num_classes = class_outputs.get_shape().as_list()
       class_targets = tf.cast(class_targets, dtype=tf.int32)
       class_targets_one_hot = tf.one_hot(class_targets, num_classes)
       return self._fast_rcnn_class_loss(class_outputs, class_targets_one_hot)
@@ -320,9 +321,6 @@ class FastrcnnBoxLoss(object):
 class MaskrcnnLoss(object):
   """Mask R-CNN instance segmentation mask loss function."""
 
-  def __init__(self):
-    raise ValueError('Not TF 2.0 ready.')
-
   def __call__(self, mask_outputs, mask_targets, select_class_targets):
     """Computes the mask loss of Mask-RCNN.
 

official/vision/detection/modeling/retinanet_model.py

@@ -56,7 +56,6 @@ class RetinanetModel(base_model.Model):
     self._generate_detections_fn = postprocess_ops.MultilevelDetectionGenerator(
         params.postprocess)
 
-    self._l2_weight_decay = params.train.l2_weight_decay
     self._transpose_input = params.train.transpose_input
     assert not self._transpose_input, 'Transpose input is not supportted.'
     # Input layer.
@@ -134,6 +133,7 @@ class RetinanetModel(base_model.Model):
     return self._keras_model
 
   def post_processing(self, labels, outputs):
+    # TODO(yeqing): Moves the output related part into build_outputs.
     required_output_fields = ['cls_outputs', 'box_outputs']
     for field in required_output_fields:
       if field not in outputs: