chenchenghao/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 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

8
sgl-kernel/include/sgl_kernel_ops.h
View File

@@ -103,14 +103,6 @@ void mscclpp_allreduce(fptr_t _context, torch::Tensor& inp, torch::Tensor& out,
/*
* From csrc/attention
*/
void lightning_attention_decode(
const torch::Tensor& q,
const torch::Tensor& k,
const torch::Tensor& v,
const torch::Tensor& past_kv,
const torch::Tensor& slope,
torch::Tensor output,
torch::Tensor new_kv);
void merge_state(
at::Tensor v_a, at::Tensor s_a, at::Tensor v_b, at::Tensor s_b, at::Tensor v_merged, at::Tensor s_merged);
void merge_state_v2(