Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in
Toggle navigation
Menu
Open sidebar
OpenDAS
vllm_cscc
Commits
e0734387
Unverified
Commit
e0734387
authored
Feb 28, 2025
by
Jee Jee Li
Committed by
GitHub
Feb 28, 2025
Browse files
[Bugfix] Fix MoeWNA16Method activation (#14024)
parent
f58f8b5c
Changes
1
Show whitespace changes
Inline
Side-by-side
Showing
1 changed file
with
2 additions
and
1 deletion
+2
-1
vllm/model_executor/layers/quantization/moe_wna16.py
vllm/model_executor/layers/quantization/moe_wna16.py
+2
-1
No files found.
vllm/model_executor/layers/quantization/moe_wna16.py
View file @
e0734387
...
@@ -293,9 +293,10 @@ class MoeWNA16Method(FusedMoEMethodBase):
...
@@ -293,9 +293,10 @@ class MoeWNA16Method(FusedMoEMethodBase):
custom_routing_function
:
Optional
[
Callable
]
=
None
,
custom_routing_function
:
Optional
[
Callable
]
=
None
,
scoring_func
:
str
=
"softmax"
,
scoring_func
:
str
=
"softmax"
,
e_score_correction_bias
:
Optional
[
torch
.
Tensor
]
=
None
,
e_score_correction_bias
:
Optional
[
torch
.
Tensor
]
=
None
,
activation
:
str
=
"silu"
,
)
->
torch
.
Tensor
:
)
->
torch
.
Tensor
:
from
vllm.model_executor.layers.fused_moe
import
fused_experts
from
vllm.model_executor.layers.fused_moe
import
fused_experts
assert
activation
==
"silu"
,
"Only SiLU activation is supported."
topk_weights
,
topk_ids
=
FusedMoE
.
select_experts
(
topk_weights
,
topk_ids
=
FusedMoE
.
select_experts
(
hidden_states
=
x
,
hidden_states
=
x
,
router_logits
=
router_logits
,
router_logits
=
router_logits
,
...
...
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