chenchenghao/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 11 Packages Projects Releases Wiki Activity

Reorg moe code (#2563)

Browse Source
This commit is contained in:
Ke Bao
2024-12-24 01:10:22 +08:00
committed by GitHub
parent 23e5e50fd5
commit e835a50021
88 changed files with 338 additions and 344 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
python/sglang/srt/layers/quantization/__init__.py
View File

@@ -60,8 +60,8 @@ def fp8_get_quant_method(self, layer, prefix):
is_layer_skipped,
)
from sglang.srt.layers.fused_moe_triton.layer import FusedMoE
from sglang.srt.layers.linear import UnquantizedLinearMethod
from sglang.srt.layers.moe.fused_moe_triton.layer import FusedMoE
from sglang.srt.layers.quantization.fp8 import Fp8LinearMethod, Fp8MoEMethod
if isinstance(layer, LinearBase):
@@ -80,7 +80,7 @@ def gptq_get_quant_method(self, layer, prefix):
GPTQMarlinMoEMethod,
)
from sglang.srt.layers.fused_moe_triton.layer import FusedMoE
from sglang.srt.layers.moe.fused_moe_triton.layer import FusedMoE
if isinstance(layer, LinearBase):
return GPTQMarlinLinearMethod(self)
@@ -96,7 +96,7 @@ def awq_get_quant_method(self, layer, prefix):
AWQMoEMethod,
)
from sglang.srt.layers.fused_moe_triton.layer import FusedMoE
from sglang.srt.layers.moe.fused_moe_triton.layer import FusedMoE
if isinstance(layer, LinearBase):
return AWQMarlinLinearMethod(self)