Internal change

PiperOrigin-RevId: 371847148

official/vision/beta/configs/experiments/maskrcnn/coco_spinenet143_casrcnn_tpu.yaml

+# Expect to reach: box mAP: 51.6%, mask mAP: 44.5% on COCO
+runtime:
+  distribution_strategy: 'tpu'
+  mixed_precision_dtype: 'bfloat16'
+task:
+  init_checkpoint: null
+  train_data:
+    global_batch_size: 256
+    parser:
+      aug_rand_hflip: true
+      aug_scale_min: 0.1
+      aug_scale_max: 2.5
+  losses:
+    l2_weight_decay: 0.00004
+  model:
+    anchor:
+      anchor_size: 4.0
+      num_scales: 3
+    min_level: 3
+    max_level: 7
+    input_size: [1280, 1280, 3]
+    backbone:
+      spinenet:
+        stochastic_depth_drop_rate: 0.2
+        model_id: '143'
+      type: 'spinenet'
+    decoder:
+      type: 'identity'
+    detection_head:
+      cascade_class_ensemble: true
+      class_agnostic_bbox_pred: true
+    rpn_head:
+      num_convs: 2
+      num_filters: 256
+    roi_sampler:
+      cascade_iou_thresholds: [0.7]
+      foreground_iou_threshold: 0.6
+    norm_activation:
+      norm_epsilon: 0.001
+      norm_momentum: 0.99
+      use_sync_bn: true
+      activation: 'swish'
+    detection_generator:
+      pre_nms_top_k: 1000
+trainer:
+  train_steps: 162050
+  optimizer_config:
+    learning_rate:
+      type: 'stepwise'
+      stepwise:
+        boundaries: [148160, 157420]
+        values: [0.32, 0.032, 0.0032]
+    warmup:
+      type: 'linear'
+      linear:
+        warmup_steps: 2000
+        warmup_learning_rate: 0.0067

official/vision/beta/configs/experiments/maskrcnn/coco_spinenet49_casrcnn_tpu.yaml

+runtime:
+  distribution_strategy: 'tpu'
+  mixed_precision_dtype: 'bfloat16'
+task:
+  init_checkpoint: null
+  train_data:
+    global_batch_size: 256
+    parser:
+      aug_rand_hflip: true
+      aug_scale_min: 0.1
+      aug_scale_max: 2.0
+  losses:
+    l2_weight_decay: 0.00004
+  model:
+    anchor:
+      anchor_size: 3.0
+      num_scales: 3
+    min_level: 3
+    max_level: 7
+    input_size: [640, 640, 3]
+    backbone:
+      spinenet:
+        stochastic_depth_drop_rate: 0.2
+        model_id: '49'
+      type: 'spinenet'
+    decoder:
+      type: 'identity'
+    detection_head:
+      cascade_class_ensemble: true
+      class_agnostic_bbox_pred: true
+    rpn_head:
+      num_convs: 2
+      num_filters: 256
+    roi_sampler:
+      cascade_iou_thresholds: [0.7]
+      foreground_iou_threshold: 0.6
+    norm_activation:
+      norm_epsilon: 0.001
+      norm_momentum: 0.99
+      use_sync_bn: true
+      activation: 'swish'
+    detection_generator:
+      pre_nms_top_k: 1000
+trainer:
+  train_steps: 231000
+  optimizer_config:
+    learning_rate:
+      type: 'stepwise'
+      stepwise:
+        boundaries: [219450, 226380]
+        values: [0.28, 0.028, 0.0028]
+    warmup:
+      type: 'linear'
+      linear:
+        warmup_steps: 2000
+        warmup_learning_rate: 0.0067

official/vision/beta/configs/experiments/maskrcnn/coco_spinenet96_casrcnn_tpu.yaml

+runtime:
+  distribution_strategy: 'tpu'
+  mixed_precision_dtype: 'bfloat16'
+task:
+  init_checkpoint: null
+  train_data:
+    global_batch_size: 256
+    parser:
+      aug_rand_hflip: true
+      aug_scale_min: 0.1
+      aug_scale_max: 2.0
+  losses:
+    l2_weight_decay: 0.00004
+  model:
+    anchor:
+      anchor_size: 3.0
+      num_scales: 3
+    min_level: 3
+    max_level: 7
+    input_size: [1024, 1024, 3]
+    backbone:
+      spinenet:
+        stochastic_depth_drop_rate: 0.2
+        model_id: '96'
+      type: 'spinenet'
+    decoder:
+      type: 'identity'
+    detection_head:
+      cascade_class_ensemble: true
+      class_agnostic_bbox_pred: true
+    rpn_head:
+      num_convs: 2
+      num_filters: 256
+    roi_sampler:
+      cascade_iou_thresholds: [0.7]
+      foreground_iou_threshold: 0.6
+    norm_activation:
+      norm_epsilon: 0.001
+      norm_momentum: 0.99
+      use_sync_bn: true
+      activation: 'swish'
+    detection_generator:
+      pre_nms_top_k: 1000
+trainer:
+  train_steps: 231000
+  optimizer_config:
+    learning_rate:
+      type: 'stepwise'
+      stepwise:
+        boundaries: [219450, 226380]
+        values: [0.32, 0.032, 0.0032]
+    warmup:
+      type: 'linear'
+      linear:
+        warmup_steps: 2000
+        warmup_learning_rate: 0.0067

official/vision/beta/modeling/factory.py

@@ -114,7 +114,28 @@ def build_maskrcnn(
       use_sync_bn=norm_activation_config.use_sync_bn,
       norm_momentum=norm_activation_config.norm_momentum,
       norm_epsilon=norm_activation_config.norm_epsilon,
-      kernel_regularizer=l2_regularizer)
+      kernel_regularizer=l2_regularizer,
+      name='detection_head')
+  if roi_sampler_config.cascade_iou_thresholds:
+    detection_head_cascade = [detection_head]
+    for cascade_num in range(len(roi_sampler_config.cascade_iou_thresholds)):
+      detection_head = instance_heads.DetectionHead(
+          num_classes=model_config.num_classes,
+          num_convs=detection_head_config.num_convs,
+          num_filters=detection_head_config.num_filters,
+          use_separable_conv=detection_head_config.use_separable_conv,
+          num_fcs=detection_head_config.num_fcs,
+          fc_dims=detection_head_config.fc_dims,
+          class_agnostic_bbox_pred=detection_head_config
+          .class_agnostic_bbox_pred,
+          activation=norm_activation_config.activation,
+          use_sync_bn=norm_activation_config.use_sync_bn,
+          norm_momentum=norm_activation_config.norm_momentum,
+          norm_epsilon=norm_activation_config.norm_epsilon,
+          kernel_regularizer=l2_regularizer,
+          name='detection_head_{}'.format(cascade_num + 1))
+      detection_head_cascade.append(detection_head)
+    detection_head = detection_head_cascade
 
   roi_generator_obj = roi_generator.MultilevelROIGenerator(
       pre_nms_top_k=roi_generator_config.pre_nms_top_k,

official/vision/beta/modeling/maskrcnn_model.py

@@ -31,7 +31,8 @@ class MaskRCNNModel(tf.keras.Model):
                backbone: tf.keras.Model,
                decoder: tf.keras.Model,
                rpn_head: tf.keras.layers.Layer,
-               detection_head: tf.keras.layers.Layer,
+               detection_head: Union[tf.keras.layers.Layer,
+                                     List[tf.keras.layers.Layer]],
                roi_generator: tf.keras.layers.Layer,
                roi_sampler: Union[tf.keras.layers.Layer,
                                   List[tf.keras.layers.Layer]],
@@ -54,7 +55,7 @@ class MaskRCNNModel(tf.keras.Model):
       backbone: `tf.keras.Model`, the backbone network.
       decoder: `tf.keras.Model`, the decoder network.
       rpn_head: the RPN head.
-      detection_head: the detection head.
+      detection_head: the detection head or a list of heads.
       roi_generator: the ROI generator.
       roi_sampler: a single ROI sampler or a list of ROI samplers for cascade
         detection heads.
@@ -104,7 +105,10 @@ class MaskRCNNModel(tf.keras.Model):
     self.backbone = backbone
     self.decoder = decoder
     self.rpn_head = rpn_head
-    self.detection_head = detection_head
+    if not isinstance(detection_head, (list, tuple)):
+      self.detection_head = [detection_head]
+    else:
+      self.detection_head = detection_head
     self.roi_generator = roi_generator
     if not isinstance(roi_sampler, (list, tuple)):
       self.roi_sampler = [roi_sampler]
@@ -191,7 +195,7 @@ class MaskRCNNModel(tf.keras.Model):
            gt_classes=gt_classes,
            training=training,
            model_outputs=model_outputs,
-           layer_num=cascade_num,
+           cascade_num=cascade_num,
            regression_weights=regression_weights)
       all_class_outputs.append(class_outputs)
 
@@ -266,7 +270,7 @@ class MaskRCNNModel(tf.keras.Model):
     return model_outputs
 
   def _run_frcnn_head(self, features, rois, gt_boxes, gt_classes, training,
-                      model_outputs, layer_num, regression_weights):
+                      model_outputs, cascade_num, regression_weights):
     """Runs the frcnn head that does both class and box prediction.
 
     Args:
@@ -279,7 +283,7 @@ class MaskRCNNModel(tf.keras.Model):
         classes. It is padded with -1s to indicate the invalid classes.
       training: `bool`, if model is training or being evaluated.
       model_outputs: `dict`, used for storing outputs used for eval and losses.
-      layer_num: `int`, the current frcnn layer in the cascade.
+      cascade_num: `int`, the current frcnn layer in the cascade.
       regression_weights: `list`, weights used for l1 loss in bounding box
         regression.
 
@@ -305,7 +309,7 @@ class MaskRCNNModel(tf.keras.Model):
     if training and gt_boxes is not None:
       rois = tf.stop_gradient(rois)
 
-      current_roi_sampler = self.roi_sampler[layer_num]
+      current_roi_sampler = self.roi_sampler[cascade_num]
       rois, matched_gt_boxes, matched_gt_classes, matched_gt_indices = (
           current_roi_sampler(rois, gt_boxes, gt_classes))
       # Create bounding box training targets.
@@ -317,10 +321,11 @@ class MaskRCNNModel(tf.keras.Model):
               tf.expand_dims(tf.equal(matched_gt_classes, 0), axis=-1),
               [1, 1, 4]), tf.zeros_like(box_targets), box_targets)
       model_outputs.update({
-          'class_targets_{}'.format(layer_num)
-          if layer_num else 'class_targets':
+          'class_targets_{}'.format(cascade_num)
+          if cascade_num else 'class_targets':
               matched_gt_classes,
-          'box_targets_{}'.format(layer_num) if layer_num else 'box_targets':
+          'box_targets_{}'.format(cascade_num)
+          if cascade_num else 'box_targets':
               box_targets,
       })
 
@@ -328,12 +333,14 @@ class MaskRCNNModel(tf.keras.Model):
     roi_features = self.roi_aligner(features, rois)
 
     # Run frcnn head to get class and bbox predictions.
-    class_outputs, box_outputs = self.detection_head(roi_features)
+    current_detection_head = self.detection_head[cascade_num]
+    class_outputs, box_outputs = current_detection_head(roi_features)
 
     model_outputs.update({
-        'class_outputs_{}'.format(layer_num) if layer_num else 'class_outputs':
+        'class_outputs_{}'.format(cascade_num)
+        if cascade_num else 'class_outputs':
             class_outputs,
-        'box_outputs_{}'.format(layer_num) if layer_num else 'box_outputs':
+        'box_outputs_{}'.format(cascade_num) if cascade_num else 'box_outputs':
             box_outputs,
     })
     return (class_outputs, box_outputs, model_outputs, matched_gt_boxes,

official/vision/beta/modeling/maskrcnn_model_test.py

@@ -373,7 +373,7 @@ class MaskRCNNModelTest(parameterized.TestCase, tf.test.TestCase):
         backbone=backbone,
         decoder=decoder,
         rpn_head=rpn_head,
-        detection_head=detection_head)
+        detection_head=[detection_head])
     if include_mask:
       expect_checkpoint_items['mask_head'] = mask_head
     self.assertAllEqual(expect_checkpoint_items, model.checkpoint_items)