[NPU] fix model llava-onevision-qwen2-7b-ov torch compiles error in npu case (#26573)

2026-05-30 17:10:58 +08:00
parent 714bcd84e2
commit 90eb894564
1 changed files with 4 additions and 2 deletions
@@ -1,10 +1,12 @@
 import torch

 from sglang.srt.sampling.penaltylib.orchestrator import _BatchedPenalizer
-from sglang.srt.utils import get_compiler_backend
+from sglang.srt.utils import get_compiler_backend, is_npu
+
+_is_npu = is_npu()


-@torch.compile(dynamic=True, backend=get_compiler_backend())
+@torch.compile(dynamic=True, backend=get_compiler_backend(), disable=_is_npu)
 def apply_scaling_penalties(logits, scaling_penalties):
    logits[:] = torch.where(
        logits < 0,