def test_boolean_mask(self):
corners = tf.constant(
[4 * [0.0], 4 * [1.0], 4 * [2.0], 4 * [3.0], 4 * [4.0]])
indicator = tf.constant([True, False, True, False, True], tf.bool)
expected_subset = [4 * [0.0], 4 * [2.0], 4 * [4.0]]
boxes = box_list.BoxList(corners)
subset = box_list_ops.boolean_mask(boxes, indicator)
with self.test_session() as sess:
subset_output = sess.run(subset.get())
self.assertAllClose(subset_output, expected_subset)
def test_boolean_mask_with_field(self):
corners = tf.constant(
[4 * [0.0], 4 * [1.0], 4 * [2.0], 4 * [3.0], 4 * [4.0]])
indicator = tf.constant([True, False, True, False, True], tf.bool)
weights = tf.constant([[.1], [.3], [.5], [.7], [.9]], tf.float32)
expected_subset = [4 * [0.0], 4 * [2.0], 4 * [4.0]]
expected_weights = [[.1], [.5], [.9]]
boxes = box_list.BoxList(corners)
boxes.add_field('weights', weights)
subset = box_list_ops.boolean_mask(boxes, indicator, ['weights'])
with self.test_session() as sess:
subset_output, weights_output = sess.run(
[subset.get(), subset.get_field('weights')])
self.assertAllClose(subset_output, expected_subset)
self.assertAllClose(weights_output, expected_weights)
def compute_loss():
sampled_boxlist = box_list_ops.boolean_mask(
detection_boxlist, sampled_indices)
sampled_padded_boxlist = box_list_ops.pad_or_clip_box_list(
sampled_boxlist, num_boxes=self.batch_size)
detection_boxes = sampled_padded_boxlist.get()
detection_transcriptions = sampled_padded_boxlist.get_field(
fields.BoxListFields.transcription)
# detection_transcriptions = tf.Print(detection_transcriptions, [detection_transcriptions], message="These are the subsampled GTs transcr.", summarize=99999)
detection_scores = sampled_padded_boxlist.get_field(
fields.BoxListFields.scores)
num_detections = tf.minimum(sampled_boxlist.num_boxes(),
self.batch_size)
transcriptions_dict, eval_metric_ops = self._predict_lstm(
rpn_features_to_crop, detection_boxes,
detection_transcriptions, detection_scores, num_detections)
return [
self.loss(transcriptions_dict),
(transcriptions_dict, eval_metric_ops)
]
