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

Remove --modelopt-config in server_args (#2758)

Browse Source
This commit is contained in:
Lianmin Zheng
2025-01-06 16:35:45 -08:00
committed by GitHub
parent b01febdca0
commit 9dec582dab
2 changed files with 0 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
python/sglang/srt/model_executor/model_runner.py
View File

@@ -150,7 +150,6 @@ class ModelRunner:
"enable_nan_detection": server_args.enable_nan_detection, "enable_nan_detection": server_args.enable_nan_detection,
"enable_dp_attention": server_args.enable_dp_attention, "enable_dp_attention": server_args.enable_dp_attention,
"enable_ep_moe": server_args.enable_ep_moe, "enable_ep_moe": server_args.enable_ep_moe,
"modelopt_config": server_args.modelopt_config,
} }
) )

7
python/sglang/srt/server_args.py
View File

@@ -149,7 +149,6 @@ class ServerArgs:
torch_compile_max_bs: int = 32 torch_compile_max_bs: int = 32
cuda_graph_max_bs: Optional[int] = None cuda_graph_max_bs: Optional[int] = None
torchao_config: str = "" torchao_config: str = ""
modelopt_config: str = ""
enable_nan_detection: bool = False enable_nan_detection: bool = False
enable_p2p_check: bool = False enable_p2p_check: bool = False
triton_attention_reduce_in_fp32: bool = False triton_attention_reduce_in_fp32: bool = False
@@ -810,12 +809,6 @@ class ServerArgs:
default=ServerArgs.torchao_config, default=ServerArgs.torchao_config,
help="Optimize the model with torchao. Experimental feature. Current choices are: int8dq, int8wo, int4wo-<group_size>, fp8wo, fp8dq-per_tensor, fp8dq-per_row", help="Optimize the model with torchao. Experimental feature. Current choices are: int8dq, int8wo, int4wo-<group_size>, fp8wo, fp8dq-per_tensor, fp8dq-per_row",
) )
parser.add_argument(
"--modelopt-config",
type=str,
default=ServerArgs.modelopt_config,
help="Optimize the model with nvidia-modelopt. Experimental feature. Current choices are: fp8",
)
parser.add_argument( parser.add_argument(
"--enable-nan-detection", "--enable-nan-detection",
action="store_true", action="store_true",