simplify the control logic for using shared experts fusion (#5504)

This commit is contained in:
Xiaoyu Zhang
2025-04-20 04:17:35 +08:00
committed by GitHub
parent bf86c5e990
commit d58e354472
16 changed files with 69 additions and 54 deletions

View File

@@ -81,7 +81,6 @@ global_server_args_dict = {
"moe_dense_tp_size": ServerArgs.moe_dense_tp_size,
"chunked_prefill_size": ServerArgs.chunked_prefill_size,
"n_share_experts_fusion": ServerArgs.n_share_experts_fusion,
"disable_shared_experts_fusion": ServerArgs.disable_shared_experts_fusion,
"disable_chunked_prefix_cache": ServerArgs.disable_chunked_prefix_cache,
}