apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2026-04-23 17:11:21 +08:00
Code Issues Actions 23 Packages Projects Releases Wiki Activity

[BugFix] Fix token_penalty kernel (#6069)

Browse Source 
* fix token_penalty kernel

* try to fix xpu

* fix xpu

* fix unit test
This commit is contained in:
freeliuzc
2026-01-28 12:03:05 +08:00
committed by GitHub
parent 85db063da6
commit ce06c6dfb3
13 changed files with 320 additions and 246 deletions
Download Patch File Download Diff File Expand all files Collapse all files
fastdeploy/model_executor/layers/sample/meta_data.py
+2
View File
@@ -43,6 +43,8 @@ class SamplingMetadata:
step_idx: paddle.Tensor
top_p: paddle.Tensor
# only GPU used
bad_words_token_len: Optional[paddle.Tensor] = None
top_k: Optional[paddle.Tensor] = None
top_k_list: Optional[list] = None
min_p: Optional[paddle.Tensor] = None