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

Replace torch.jit.script with torch.compile in get_masked_input_and_mask to fix benchmark underreporting (#8733)

Browse Source
This commit is contained in:
YyWangCS
2025-08-04 10:02:51 +08:00
committed by GitHub
parent b102353f8f
commit f024795e57
1 changed files with 8 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
python/sglang/srt/layers/vocab_parallel_embedding.py
View File

@@ -26,7 +26,12 @@ from sglang.srt.layers.quantization.base_config import (
method_has_implemented_embedding,
)
from sglang.srt.layers.quantization.unquant import UnquantizedEmbeddingMethod
from sglang.srt.utils import cpu_has_amx_support, is_cpu, set_weight_attrs
from sglang.srt.utils import (
cpu_has_amx_support,
get_compiler_backend,
is_cpu,
set_weight_attrs,
)
DEFAULT_VOCAB_PADDING_SIZE = 64
@@ -117,7 +122,7 @@ class VocabParallelEmbeddingShardIndices:
assert self.num_added_elements <= self.num_added_elements_padded
@torch.jit.script
@torch.compile(dynamic=True, backend=get_compiler_backend())
def get_masked_input_and_mask(
input_: torch.Tensor,
org_vocab_start_index: int,
@@ -126,7 +131,7 @@ def get_masked_input_and_mask(
added_vocab_start_index: int,
added_vocab_end_index: int,
) -> Tuple[torch.Tensor, torch.Tensor]:
# torch.jit.script will fuse all of the pointwise ops below
# torch.compile will fuse all of the pointwise ops below
# into a single kernel, making it very fast
org_vocab_mask = (input_ >= org_vocab_start_index) & (input_ < org_vocab_end_index)
added_vocab_mask = (input_ >= added_vocab_start_index) & (