chenchenghao/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 13 Packages Projects Releases Wiki Activity

[sgl-kernel Code Clean] Remove useless lightning_attention kernel (#13819)

Browse Source
This commit is contained in:
Xiaoyu Zhang
2025-11-24 18:26:25 +08:00
committed by GitHub
parent aeac622058
commit ecefc7904f
10 changed files with 0 additions and 1754 deletions
Download Patch File Download Diff File Expand all files Collapse all files
sgl-kernel/csrc/common_extension.cc
-4
View File
@@ -50,10 +50,6 @@ TORCH_LIBRARY_FRAGMENT(sgl_kernel, m) {
/*
* From csrc/attention
*/
m.def(
"lightning_attention_decode(Tensor q, Tensor k, Tensor v, Tensor past_kv, Tensor slope, Tensor! output, Tensor! "
"new_kv) -> ()");
m.impl("lightning_attention_decode", torch::kCUDA, &lightning_attention_decode);
m.def("merge_state(Tensor v_a, Tensor s_a, Tensor v_b, Tensor s_b, Tensor! v_merged, Tensor! s_merged) -> ()");
m.impl("merge_state", torch::kCUDA, &merge_state);
m.def("merge_state_v2(Tensor v_a, Tensor s_a, Tensor v_b, Tensor s_b, Tensor! v_merged, Tensor! s_merged) -> ()");