EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

feat: use sgl-kernel 0.0.3 in sglang (#3179)

Browse Source
This commit is contained in:
Yineng Zhang
2025-01-27 21:39:52 +08:00
committed by GitHub
parent 8a96f74988
commit 2f79f58873
6 changed files with 20 additions and 24 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
python/sglang/srt/models/deepseek_v2.py
View File

@@ -56,12 +56,12 @@ from sglang.srt.layers.vocab_parallel_embedding import (
from sglang.srt.managers.schedule_batch import global_server_args_dict
from sglang.srt.model_executor.forward_batch_info import ForwardBatch
from sglang.srt.model_loader.weight_utils import default_weight_loader
from sglang.srt.utils import is_flashinfer_available, is_hip
from sglang.srt.utils import is_cuda_available, is_hip
is_hip_ = is_hip()
if is_flashinfer_available():
from flashinfer import bmm_fp8
if is_cuda_available():
from sgl_kernel import bmm_fp8
class DeepseekV2MLP(nn.Module):

6
python/sglang/srt/models/minicpm3.py
View File

@@ -40,10 +40,10 @@ from sglang.srt.layers.vocab_parallel_embedding import (
from sglang.srt.managers.schedule_batch import global_server_args_dict
from sglang.srt.model_executor.forward_batch_info import ForwardBatch
from sglang.srt.model_loader.weight_utils import default_weight_loader
from sglang.srt.utils import is_flashinfer_available
from sglang.srt.utils import is_cuda_available
if is_flashinfer_available():
from flashinfer import bmm_fp8
if is_cuda_available():
from sgl_kernel import bmm_fp8
class MiniCPM3MLP(nn.Module):