apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-05 16:48:03 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

fa3_rope (#4190)

Browse Source
This commit is contained in:
xiaoxiaohehe001
2025-09-21 22:04:59 +08:00
committed by GitHub
parent 5223065d59
commit 9f1882d9a8
4 changed files with 17 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
custom_ops/gpu_ops/cpp_extensions.cc
View File

@@ -107,7 +107,8 @@ std::vector<paddle::Tensor> GQARopeWriteCacheKernel(
const paddle::optional<paddle::Tensor> &cache_v_zp,
const paddle::optional<paddle::Tensor> &kv_signal_data,
const int kv_token_num, const int max_seq_len,
const std::string &cache_quant_type);
const std::string &cache_quant_type,
const bool rope_3d);
std::vector<paddle::Tensor>
PreCacheLenConcat(const paddle::Tensor &seq_lens_decoder,