chenchenghao/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 13 Packages Projects Releases Wiki Activity

Clean up server_args.py (#7037)

Browse Source
This commit is contained in:
Lianmin Zheng
2025-06-10 05:34:29 -07:00
committed by GitHub
parent 019851d099
commit 6406408a70
7 changed files with 394 additions and 331 deletions
Download Patch File Download Diff File Expand all files Collapse all files
python/sglang/srt/managers/schedule_batch.py
+7 -6
View File
@@ -72,32 +72,33 @@ INIT_INCREMENTAL_DETOKENIZATION_OFFSET = 5
GLOBAL_SERVER_ARGS_KEYS = [
"attention_backend",
"mm_attention_backend",
"debug_tensor_dump_inject",
"debug_tensor_dump_output_folder",
"chunked_prefill_size",
"deepep_mode",
"device",
"disable_chunked_prefix_cache",
"disable_radix_cache",
"enable_deepep_moe",
"enable_dp_attention",
"enable_two_batch_overlap",
"enable_dp_lm_head",
"enable_deepep_moe",
"deepep_mode",
"enable_ep_moe",
"moe_dense_tp_size",
"ep_dispatch_algorithm",
"deepep_config",
"ep_num_redundant_experts",
"enable_nan_detection",
"flashinfer_mla_disable_ragged",
"max_micro_batch_size",
"moe_dense_tp_size",
"ep_dispatch_algorithm",
"disable_shared_experts_fusion",
"sampling_backend",
"speculative_accept_threshold_acc",
"speculative_accept_threshold_single",
"torchao_config",
"triton_attention_reduce_in_fp32",
"ep_num_redundant_experts",
"mm_attention_backend",
"num_reserved_decode_tokens",
]
# Put some global args for easy access