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

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
+3
View File
@@ -232,6 +232,9 @@ class Envs:
SGLANG_ROCM_FUSED_DECODE_MLA = EnvBool(False)
SGLANG_ROCM_DISABLE_LINEARQUANT = EnvBool(False)
# NPU
SGLANG_NPU_DISABLE_ACL_FORMAT_WEIGHT = EnvBool(False)
# Quantization
SGLANG_INT4_WEIGHT = EnvBool(False)
SGLANG_CPU_QUANTIZATION = EnvBool(False)