apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2026-04-23 00:17:25 +08:00
Code Issues Actions 19 Packages Projects Releases Wiki Activity

[Optimize] optimize mask_quant & swiglu (#6222)

Browse Source 
* optimize mask_quant op speed up 1.5

* fix calculate sequence

* add fused

* rm log

* push kernel code

* add ut

* accuracy ok

* add ue8m0

* add ut

* add merge develop

* rm ut of mask_per_token_quant
This commit is contained in:
fxyfxy777
2026-02-02 13:52:38 +08:00
committed by GitHub
parent 25656455ee
commit 2ada119a38
7 changed files with 555 additions and 452 deletions
Download Patch File Download Diff File Expand all files Collapse all files
custom_ops/setup_ops.py
+1 -1
View File
@@ -293,7 +293,7 @@ elif paddle.is_compiled_with_cuda():
"gpu_ops/step_system_cache.cu",
"gpu_ops/cpp_extensions.cc",
"gpu_ops/share_external_data.cu",
"gpu_ops/per_token_quant_fp8.cu",
"gpu_ops/fused_mask_swiglu_fp8_quant_kernel.cu",
"gpu_ops/update_split_fuse_input.cu",
"gpu_ops/text_image_index_out.cu",
"gpu_ops/text_image_gather_scatter.cu",