|
|
@ -320,6 +320,9 @@ class DataGenerator(object):
|
|
|
|
if flatten:
|
|
|
|
if flatten:
|
|
|
|
padded_audio = padded_audio.flatten()
|
|
|
|
padded_audio = padded_audio.flatten()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# Stride size for conv0 is (3, 2)
|
|
|
|
|
|
|
|
# Stride size for conv1 to convN is (1, 2)
|
|
|
|
|
|
|
|
# Same as the network, hard-coded here
|
|
|
|
padded_instance = [padded_audio, text]
|
|
|
|
padded_instance = [padded_audio, text]
|
|
|
|
padded_conv0_h = (padded_audio.shape[0] - 1) // 2 + 1
|
|
|
|
padded_conv0_h = (padded_audio.shape[0] - 1) // 2 + 1
|
|
|
|
padded_conv0_w = (padded_audio.shape[1] - 1) // 3 + 1
|
|
|
|
padded_conv0_w = (padded_audio.shape[1] - 1) // 3 + 1
|
|
|
@ -327,6 +330,8 @@ class DataGenerator(object):
|
|
|
|
padded_instance += [
|
|
|
|
padded_instance += [
|
|
|
|
[0], # sequence offset, always 0
|
|
|
|
[0], # sequence offset, always 0
|
|
|
|
[valid_w], # valid sequence length
|
|
|
|
[valid_w], # valid sequence length
|
|
|
|
|
|
|
|
# Index ranges for channel, height and width
|
|
|
|
|
|
|
|
# Please refer scale_sub_region layer to see details
|
|
|
|
[1, 32, 1, padded_conv0_h, valid_w + 1, padded_conv0_w]
|
|
|
|
[1, 32, 1, padded_conv0_h, valid_w + 1, padded_conv0_w]
|
|
|
|
]
|
|
|
|
]
|
|
|
|
pre_padded_h = padded_conv0_h
|
|
|
|
pre_padded_h = padded_conv0_h
|
|
|
|