Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[cif] remove concat after in cif #1763

Merged
merged 1 commit into from
Mar 20, 2023
+1 −25
Merged

[cif] remove concat after in cif #1763

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
13 changes: 1 addition & 12 deletions wenet/cif/cif_decoder.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -48,11 +48,6 @@ class BaseDecoder(nn.Module):
use_output_layer: whether to use output layer
pos_enc_class: PositionalEncoding or ScaledPositionalEncoding
normalize_before: whether to use layer_norm before the first block
concat_after: whether to concat attention layer's input and output
if True, additional linear will be applied.
i.e. x -> x + linear(concat(x, att(x)))
if False, no additional linear will be applied.
i.e. x -> x + att(x)
"""

def __init__(
Expand Down Expand Up @@ -177,7 +172,6 @@ def __init__(
use_output_layer: bool = True,
pos_enc_class=PositionalEncoding,
normalize_before: bool = True,
concat_after: bool = False,
embeds_id: int = -1,
):
assert check_argument_types()
Expand Down Expand Up @@ -205,8 +199,7 @@ def __init__(
PositionwiseFeedForward(attention_dim, linear_units,
dropout_rate),
dropout_rate,
normalize_before,
concat_after)
normalize_before)
for _ in range(num_blocks)
])

Expand Down Expand Up @@ -294,7 +287,6 @@ def __init__(
use_output_layer: bool = True,
pos_enc_class=PositionalEncoding,
normalize_before: bool = True,
concat_after: bool = False,
att_layer_num: int = 6,
kernel_size: int = 21,
sanm_shfit: int = 0
Expand Down Expand Up @@ -355,7 +347,6 @@ def __init__(
dropout_rate),
dropout_rate,
normalize_before,
concat_after,
) for _ in range(att_layer_num)
])
if num_blocks - att_layer_num <= 0:
Expand All @@ -374,7 +365,6 @@ def __init__(
dropout_rate),
dropout_rate,
normalize_before,
concat_after,
) for _ in range(num_blocks - att_layer_num)
])
self.decoders3 = torch.nn.ModuleList([
Expand All @@ -386,7 +376,6 @@ def __init__(
dropout_rate),
dropout_rate,
normalize_before,
concat_after,
) for _ in range(1)
])

Expand Down
13 changes: 0 additions & 13 deletions wenet/cif/decoder_layer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -34,11 +34,6 @@ class DecoderLayerSANM(nn.Module):
dropout_rate (float): Dropout rate.
normalize_before (bool): Whether to use layer_norm before the first
block.
concat_after (bool): Whether to concat attention layer's input and
output.
if True, additional linear will be applied.
i.e. x -> x + linear(concat(x, att(x)))
if False, no additional linear will be applied. i.e. x -> x + att(x)
"""

def __init__(
Expand All @@ -49,7 +44,6 @@ def __init__(
feed_forward: nn.Module,
dropout_rate: float,
normalize_before: bool = True,
concat_after: bool = False,
):
"""Construct an DecoderLayer object."""
super(DecoderLayerSANM, self).__init__()
Expand All @@ -64,13 +58,6 @@ def __init__(
self.norm3 = nn.LayerNorm(size, eps=1e-12)
self.dropout = nn.Dropout(dropout_rate)
self.normalize_before = normalize_before
self.concat_after = concat_after
if self.concat_after:
self.concat_linear1 = nn.Linear(size + size, size)
self.concat_linear2 = nn.Linear(size + size, size)
else:
self.concat_linear1 = nn.Identity()
self.concat_linear2 = nn.Identity()

def forward(
self,
Expand Down