mirrors/icefall
1
0
Fork 0
mirror of https://github.com/k2-fsa/icefall.git synced 2025-12-11 06:55:27 +00:00
Code Issues Packages Projects Releases Wiki Activity

Revert "Make sub-module dropped out independently."

Browse Source 
This reverts commit 3ff3f440ee6d2a367cc3cc45e40f8eb69d122861.
This commit is contained in:
Daniel Povey 2022-11-11 21:36:36 +08:00
parent 742bcaa340
commit f7aff4f507
1 changed files with 18 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
egs/librispeech/ASR/pruned_transducer_stateless7/zipformer.py
View File

@ -487,6 +487,11 @@ class ZipformerEncoderLayer(nn.Module):
# dropout rate for submodules that interact with time.
dynamic_dropout = self.get_dynamic_dropout_rate()
# multi-headed self-attention module
# TODO: make the various attention-using models be dropped
# out independently.
use_self_attn = (random.random() > dynamic_dropout)
if torch.jit.is_scripting() or use_self_attn:
# attn_weights: (num_heads, batch_size, seq_len, seq_len)
attn_weights = self.self_attn_weights(
src,
@ -495,22 +500,22 @@ class ZipformerEncoderLayer(nn.Module):
key_padding_mask=src_key_padding_mask,
)
if torch.jit.is_scripting() or random.random() > dynamic_dropout:
if torch.jit.is_scripting() or use_self_attn:
src = src + self.self_attn1(
src, attn_weights)
# convolution module
if torch.jit.is_scripting() or random.random() > dynamic_dropout:
if torch.jit.is_scripting() or use_self_attn:
src = src + self.nonlin_attention_module(src,
attn_weights[0:1])
src = src + self.feed_forward2(src)
# pooling module
if torch.jit.is_scripting() or random.random() > dynamic_dropout:
if torch.jit.is_scripting() or use_self_attn:
src = src + self.attention_squeeze1(src, attn_weights[1:2])
if torch.jit.is_scripting() or random.random() > dynamic_dropout:
if torch.jit.is_scripting() or use_self_attn:
src = src + self.self_attn2(
src, attn_weights)
@ -520,9 +525,10 @@ class ZipformerEncoderLayer(nn.Module):
src = src + self.feed_forward3(src)
# pooling module
if torch.jit.is_scripting() or random.random() > dynamic_dropout:
if torch.jit.is_scripting() or use_self_attn:
src = src + self.attention_squeeze2(src, attn_weights[2:3])
src = self.norm_final(self.balancer(src))
delta = src - src_orig