sglang/sgl-kernel/python/sgl_kernel/attention.py

import torch


def lightning_attention_decode(q, k, v, past_kv, slope, output, new_kv):
    torch.ops.sgl_kernel.lightning_attention_decode(
        q, k, v, past_kv, slope, output, new_kv
    )
Reorganize python source files in sgl-kernel with multiple files (#4027) 2025-03-03 06:36:40 -08:00			`import torch`


			`def lightning_attention_decode(q, k, v, past_kv, slope, output, new_kv):`
Rename files in sgl kernel to avoid nested folder structure (#4213) Co-authored-by: zhyncs <me@zhyncs.com> 2025-03-08 22:54:51 -08:00			`torch.ops.sgl_kernel.lightning_attention_decode(`
Reorganize python source files in sgl-kernel with multiple files (#4027) 2025-03-03 06:36:40 -08:00			`q, k, v, past_kv, slope, output, new_kv`
			`)`