minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

[Kernel] Hotfix: update sgl-kernel imports of relocated fp8_kernel (RFC #29630 #30784) (#31089)

Browse Source 
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
Xiaoyu Zhang
2026-07-14 08:41:23 +08:00
committed by GitHub
co-authored by Claude Opus 4.8
parent cfe4eefabb
commit 4c997310f5
3 changed files with 6 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files
sgl-kernel/benchmark/bench_fp8_blockwise_gemm.py
+1 -1
View File
@@ -20,7 +20,7 @@ except ImportError:
vllm_scaled_mm = None
VLLM_AVAILABLE = False
from sglang.srt.layers.quantization.fp8_kernel import (
from sglang.kernels.ops.quantization.fp8_kernel import (
w8a8_block_fp8_matmul_triton as w8a8_block_fp8_matmul,
)
sgl-kernel/benchmark/bench_per_token_group_quant_8bit.py
+3 -3
View File
@@ -8,13 +8,13 @@ import torch
import triton
from sgl_kernel.test_utils import create_per_token_group_quant_test_data
from sglang.srt.layers.quantization.fp8_kernel import (
from sglang.kernels.ops.quantization.fp8_kernel import (
create_per_token_group_quant_fp8_output_scale,
)
from sglang.srt.layers.quantization.fp8_kernel import (
from sglang.kernels.ops.quantization.fp8_kernel import (
per_token_group_quant_8bit as triton_per_token_group_quant_8bit,
)
from sglang.srt.layers.quantization.fp8_kernel import (
from sglang.kernels.ops.quantization.fp8_kernel import (
sglang_per_token_group_quant_8bit,
)
from sglang.srt.utils import is_hip