[FIX] use time_warping_para to constraint control width, add logic to disable cache [ADD] num_control_points

DeepSpeech.py

@@ -427,7 +427,8 @@ def train():
             FLAGS.augmentation_spec_dropout_keeprate < 1 or
             FLAGS.augmentation_freq_and_time_masking or
             FLAGS.augmentation_pitch_and_tempo_scaling or
-            FLAGS.augmentation_speed_up_std > 0):
+            FLAGS.augmentation_speed_up_std > 0 or
+            FLAGS.augmentation_sparse_warp):
         do_cache_dataset = False
 
     # Create training and validation datasets

util/feeding.py

@@ -48,7 +48,8 @@ def samples_to_mfccs(samples, sample_rate, train_phase=False):
                                               time_warping_para=FLAGS.augmentation_sparse_warp_time_warping_para,
                                               interpolation_order=FLAGS.augmentation_sparse_warp_interpolation_order,
                                               regularization_weight=FLAGS.augmentation_sparse_warp_regularization_weight,
-                                              num_boundary_points=FLAGS.augmentation_sparse_warp_num_boundary_points)
+                                              num_boundary_points=FLAGS.augmentation_sparse_warp_num_boundary_points,
+                                              num_control_points=FLAGS.augmentation_sparse_warp_num_control_points)
 
         if FLAGS.augmentation_freq_and_time_masking:
             spectrogram = augment_freq_time_mask(spectrogram,

util/flags.py

@@ -30,6 +30,7 @@ def create_flags():
     f.DEFINE_float('augmentation_spec_dropout_keeprate', 1, 'keep rate of dropout augmentation on spectrogram (if 1, no dropout will be performed on spectrogram)')
 
     f.DEFINE_boolean('augmentation_sparse_warp', False, 'whether to use spectrogram sparse warp')
+    f.DEFINE_integer('augmentation_sparse_warp_num_control_points', 1, 'specify number of control points')
     f.DEFINE_integer('augmentation_sparse_warp_time_warping_para', 80, 'time_warping_para')
     f.DEFINE_integer('augmentation_sparse_warp_interpolation_order', 2, 'sparse_warp_interpolation_order')
     f.DEFINE_float('augmentation_sparse_warp_regularization_weight', 0.0, 'sparse_warp_regularization_weight')

util/spectrogram_augmentations.py

@@ -68,7 +68,7 @@ def augment_dropout(spectrogram,
     return tf.nn.dropout(spectrogram, rate=1-keep_prob)
 
 
-def augment_sparse_warp(spectrogram, time_warping_para=80, interpolation_order=2, regularization_weight=0.0, num_boundary_points=1):
+def augment_sparse_warp(spectrogram, time_warping_para=80, interpolation_order=2, regularization_weight=0.0, num_boundary_points=1, num_control_points=1):
     """Reference: https://arxiv.org/pdf/1904.08779.pdf
     Args:
         spectrogram: `[batch, time, frequency]` float `Tensor`
@@ -77,6 +77,7 @@ def augment_sparse_warp(spectrogram, time_warping_para=80, interpolation_order=2
         regularization_weight: used to put into `sparse_image_warp`
         num_boundary_points: used to put into `sparse_image_warp`,
                             default=1 means boundary points on 4 corners of the image
+        num_control_points: number of control points
     Returns:
         warped_spectrogram: `[batch, time, frequency]` float `Tensor` with same
             type as input image.
@@ -92,25 +93,26 @@ def augment_sparse_warp(spectrogram, time_warping_para=80, interpolation_order=2
     time_warping_para = tf.math.minimum(
         time_warping_para, tf.math.subtract(tf.math.floordiv(tau, 2), 1))
 
-    mid_tau = tf.math.floordiv(tau, 2)
     mid_freq = tf.math.floordiv(freq_size, 2)
     left_mid_point = [0, mid_freq]
     right_mid_point = [tau, mid_freq]
 
-    # dest control point must between (W, tau-W), which means the first and last W interval of the spectrogram won't be warped
-    random_dest_time_point = tfv1.random_uniform(
-        [], time_warping_para, tau - time_warping_para, tf.int32)
-    source_control_point_locations = tf.cast([[
-        left_mid_point,
-        [mid_tau, mid_freq], # source control point must start from the center of the image
-        right_mid_point
-    ]], tf.float32)
+    random_source_times = [tfv1.random_uniform( # generate source points `t` of time axis between (W, tau-W)
+        [], time_warping_para, tf.math.subtract(tau, time_warping_para), tf.int32) for i in range(num_control_points)]
+    random_dest_times = [tfv1.random_uniform( # generate dest points `t'` of time axis between (t-W, t+W)
+        [], tf.math.subtract(source_time, time_warping_para), tf.math.add(source_time, time_warping_para), tf.int32) for source_time in random_source_times]
 
-    dest_control_point_locations = tf.cast([[
-        left_mid_point,
-        [random_dest_time_point, mid_freq],
-        right_mid_point
-    ]], tf.float32)
+    source_control_point_locations = tf.cast([
+        [left_mid_point] +
+        [[source_time, mid_freq] for source_time in random_source_times] +
+        [right_mid_point]
+    ], tf.float32)
+
+    dest_control_point_locations = tf.cast([
+        [left_mid_point] +
+        [[dest_time, mid_freq] for dest_time in random_dest_times] +
+        [right_mid_point]
+    ], tf.float32)
 
     warped_spectrogram, _ = sparse_image_warp(spectrogram,
                                               source_control_point_locations=source_control_point_locations,