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
change
sglang
Commits
71d41212
"vscode:/vscode.git/clone" did not exist on "7e8af32b1de5ae1d27cfb24b753623a7a2d284f9"
Unverified
Commit
71d41212
authored
Oct 24, 2025
by
Qiaolin Yu
Committed by
GitHub
Oct 24, 2025
Browse files
Fix dpsk-r1-fp4 launching crash (#12063)
parent
b9fb74f3
Changes
2
Hide whitespace changes
Inline
Side-by-side
Showing
2 changed files
with
11 additions
and
20 deletions
+11
-20
python/sglang/srt/layers/quantization/mxfp4.py
python/sglang/srt/layers/quantization/mxfp4.py
+5
-7
python/sglang/srt/layers/quantization/unquant.py
python/sglang/srt/layers/quantization/unquant.py
+6
-13
No files found.
python/sglang/srt/layers/quantization/mxfp4.py
View file @
71d41212
...
...
@@ -587,13 +587,11 @@ class Mxfp4MoEMethod(FusedMoEMethodBase):
self
,
layer
:
torch
.
nn
.
Module
,
moe_runner_config
:
MoeRunnerConfig
):
self
.
moe_runner_config
=
moe_runner_config
backend
=
get_moe_runner_backend
()
if
backend
.
is_auto
():
backend
=
(
MoeRunnerBackend
.
TRITON_KERNELS
if
self
.
use_triton_kernels
else
MoeRunnerBackend
.
TRITON
)
backend
=
(
MoeRunnerBackend
.
TRITON_KERNELS
if
self
.
use_triton_kernels
else
MoeRunnerBackend
.
TRITON
)
self
.
runner
=
MoeRunner
(
backend
,
moe_runner_config
)
def
apply
(
...
...
python/sglang/srt/layers/quantization/unquant.py
View file @
71d41212
...
...
@@ -8,12 +8,7 @@ from torch.nn.parameter import Parameter
from
sglang.srt.custom_op
import
CustomOp
from
sglang.srt.layers.amx_utils
import
_amx_process_weight_after_loading
from
sglang.srt.layers.moe
import
(
MoeRunner
,
MoeRunnerBackend
,
MoeRunnerConfig
,
get_moe_runner_backend
,
)
from
sglang.srt.layers.moe
import
MoeRunner
,
MoeRunnerBackend
,
MoeRunnerConfig
from
sglang.srt.layers.moe.moe_runner.triton
import
TritonMoeQuantInfo
from
sglang.srt.layers.quantization.base_config
import
(
FusedMoEMethodBase
,
...
...
@@ -225,13 +220,11 @@ class UnquantizedFusedMoEMethod(FusedMoEMethodBase, CustomOp):
self
,
layer
:
torch
.
nn
.
Module
,
moe_runner_config
:
MoeRunnerConfig
):
self
.
moe_runner_config
=
moe_runner_config
backend
=
get_moe_runner_backend
()
if
backend
.
is_auto
():
backend
=
(
MoeRunnerBackend
.
TRITON_KERNELS
if
self
.
use_triton_kernels
else
MoeRunnerBackend
.
TRITON
)
backend
=
(
MoeRunnerBackend
.
TRITON_KERNELS
if
self
.
use_triton_kernels
else
MoeRunnerBackend
.
TRITON
)
self
.
runner
=
MoeRunner
(
backend
,
moe_runner_config
)
def
apply
(
...
...
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