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
dd38ba3a
Unverified
Commit
dd38ba3a
authored
Dec 04, 2025
by
Jee Jee Li
Committed by
GitHub
Dec 04, 2025
Browse files
[Bugfix] Fix adapter_enabled IMA (#29977)
Signed-off-by:
Jee Jee Li
<
pandaleefree@gmail.com
>
parent
5f91cdda
Changes
1
Show whitespace changes
Inline
Side-by-side
Showing
1 changed file
with
6 additions
and
2 deletions
+6
-2
vllm/lora/ops/triton_ops/fused_moe_lora_op.py
vllm/lora/ops/triton_ops/fused_moe_lora_op.py
+6
-2
No files found.
vllm/lora/ops/triton_ops/fused_moe_lora_op.py
View file @
dd38ba3a
...
...
@@ -96,10 +96,14 @@ def _fused_moe_lora_kernel(
slice_id
=
tl
.
program_id
(
axis
=
1
)
lora_idx
=
tl
.
program_id
(
axis
=
2
)
lora_id
=
tl
.
load
(
lora_ids
+
lora_idx
)
moe_enabled
=
tl
.
load
(
adapter_enabled
+
lora_id
)
if
lora_id
==
-
1
or
moe_enabled
==
0
:
if
lora_id
==
-
1
:
# Early exit for the no-lora case.
return
moe_enabled
=
tl
.
load
(
adapter_enabled
+
lora_id
)
if
moe_enabled
==
0
:
# Early exit for the no moe lora case.
return
max_loras
=
tl
.
num_programs
(
axis
=
2
)
grid_k
=
tl
.
cdiv
(
K
,
BLOCK_SIZE_K
*
SPLIT_K
)
...
...
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