[Performance] Optimze the performance of Qwen25VL (#15640)

Co-authored-by: Xinyuan Tong <115166877+JustinTong0323@users.noreply.github.com>
This commit is contained in:
Siyuan Chen
2026-01-02 23:15:36 -08:00
committed by GitHub
co-authored by Xinyuan Tong
parent 5b4f790200
commit 9a414b164c
3 changed files with 20 additions and 16 deletions
@@ -16,6 +16,7 @@ from sglang.srt.server_args import get_global_server_args
from sglang.srt.utils import (
cpu_has_amx_support,
get_bool_env_var,
get_compiler_backend,
is_cpu,
is_cuda,
is_hip,
@@ -2748,6 +2749,7 @@ def rotate_half(x):
return torch.cat((-x2, x1), dim=-1)
@torch.compile(dynamic=True, backend=get_compiler_backend())
def apply_rotary_pos_emb_native(
q: torch.Tensor,
k: torch.Tensor,