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

[NPU][1/N] NPU basic functions refactor and new modelslim quant type (#13359)

Browse Source
This commit is contained in:
Even Zhou
2025-12-04 16:15:31 +08:00
committed by GitHub
parent d6c490192d
commit 894c0dc57c
43 changed files with 2500 additions and 2058 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/srt/ascend/test_ascend_deepep.py
+1 -1
View File
@@ -35,7 +35,7 @@ class TestAscendDeepEP(CustomTestCase):
"--attention-backend",
"ascend",
"--quantization",
"w8a8_int8",
"modelslim",
"--mem-fraction-static",
0.8,
"--disable-radix-cache",
test/srt/ascend/test_ascend_deepseek_mtp.py
+1 -1
View File
@@ -33,7 +33,7 @@ class TestAscendDeepSeekMTP(CustomTestCase):
"--attention-backend",
"ascend",
"--quantization",
"w8a8_int8",
"modelslim",
"--mem-fraction-static",
0.8,
"--disable-radix-cache",
test/srt/ascend/test_ascend_mla_fia_w8a8int8.py
+1 -1
View File
@@ -38,7 +38,7 @@ class TestAscendMlaW8A8Int8(CustomTestCase):
"--attention-backend",
"ascend",
"--quantization",
"w8a8_int8",
"modelslim",
"--tp-size",
2,
"--disable-radix-cache",
test/srt/ascend/test_ascend_mla_w8a8int8.py
+1 -1
View File
@@ -37,7 +37,7 @@ class TestAscendMlaW8A8Int8(CustomTestCase):
"--attention-backend",
"ascend",
"--quantization",
"w8a8_int8",
"modelslim",
"--tp-size",
4,
"--disable-radix-cache",