Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in / Register
Toggle navigation
Menu
Open sidebar
OpenDAS
FastMoE
Commits
96ed8976
Commit
96ed8976
authored
Feb 25, 2021
by
Jiezhong Qiu
Browse files
put dropout after relu
although order of dp and relu doesn't matter
parent
03b2a725
Changes
1
Hide whitespace changes
Inline
Side-by-side
Showing
1 changed file
with
1 addition
and
1 deletion
+1
-1
examples/transformer-xl/mem_transformer.py
examples/transformer-xl/mem_transformer.py
+1
-1
No files found.
examples/transformer-xl/mem_transformer.py
View file @
96ed8976
...
@@ -380,8 +380,8 @@ from fmoe import FMoETransformerMLP
...
@@ -380,8 +380,8 @@ from fmoe import FMoETransformerMLP
class
CustomizedMoEPositionwiseFF
(
FMoETransformerMLP
):
class
CustomizedMoEPositionwiseFF
(
FMoETransformerMLP
):
def
__init__
(
self
,
d_model
,
d_inner
,
dropout
,
pre_lnorm
=
False
,
moe_num_expert
=
64
,
moe_top_k
=
2
):
def
__init__
(
self
,
d_model
,
d_inner
,
dropout
,
pre_lnorm
=
False
,
moe_num_expert
=
64
,
moe_top_k
=
2
):
activation
=
nn
.
Sequential
(
activation
=
nn
.
Sequential
(
nn
.
Dropout
(
dropout
),
nn
.
ReLU
()
nn
.
ReLU
()
nn
.
Dropout
(
dropout
),
)
)
super
().
__init__
(
num_expert
=
moe_num_expert
,
d_model
=
d_model
,
d_hidden
=
d_inner
,
top_k
=
moe_top_k
,
super
().
__init__
(
num_expert
=
moe_num_expert
,
d_model
=
d_model
,
d_hidden
=
d_inner
,
top_k
=
moe_top_k
,
do_lnorm
=
True
,
pre_lnorm
=
pre_lnorm
,
activation
=
activation
,
dropout
=
dropout
)
do_lnorm
=
True
,
pre_lnorm
=
pre_lnorm
,
activation
=
activation
,
dropout
=
dropout
)
...
...
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment