optimize per token group quant fp8 (#3490)

This commit is contained in:
Xiaoyu Zhang
2025-02-11 22:19:05 +08:00
committed by GitHub
parent fdf04a1426
commit bb418ced80
8 changed files with 509 additions and 0 deletions

View File

@@ -100,6 +100,7 @@ sources = [
"src/sgl-kernel/csrc/fused_add_rms_norm_kernel.cu",
"src/sgl-kernel/csrc/eagle_utils.cu",
"src/sgl-kernel/csrc/speculative_sampling.cu",
"src/sgl-kernel/csrc/per_token_group_quant_fp8.cu",
"3rdparty/flashinfer/csrc/activation.cu",
"3rdparty/flashinfer/csrc/bmm_fp8.cu",
"3rdparty/flashinfer/csrc/norm.cu",