chore: upgrade sgl-kernel 0.0.8.post3 (#5342)

2025-04-13 00:45:59 -07:00
parent c1270aabc5
commit f58b929a51
3 changed files with 3 additions and 3 deletions
--- a/python/sglang/srt/layers/quantization/fp8_kernel.py
+++ b/python/sglang/srt/layers/quantization/fp8_kernel.py
@@ -41,7 +41,7 @@ fp8_type_ = torch.float8_e4m3fnuz if _is_hip else torch.float8_e4m3fn

 _is_cuda = is_cuda()
 if _is_cuda:
-    import deep_gemm  # `pip install "sgl-kernel>=0.0.4.post3"`
+    import deep_gemm
    from sgl_kernel import sgl_per_token_group_quant_fp8, sgl_per_token_quant_fp8

    sm_version = get_device_sm()