apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-12-24 13:28:13 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

[Excutor] Increase buffer size to prevent address corruption; add forward metadata debug tool (#3404)

Browse Source 
* 修复buffer申请不够大，增加打印forwardmetadata的工具

* fix mistake

* Make CPU tensor in CPUPlace

* Add test about forward_meta_str and Add unitest_requirement

---------

Co-authored-by: RAM <gstian5555@outlook.com>
This commit is contained in:
Jundong Liu
2025-08-18 16:14:09 +08:00
committed by GitHub
parent 5585cf7aa5
commit ea4a3b479c
5 changed files with 146 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
custom_ops/gpu_ops/append_attn/get_block_shape_and_split_kv_block.cu
View File

@@ -289,7 +289,7 @@ std::vector<paddle::Tensor> GetBlockShapeAndSplitKVBlock(
kv_tile_ids_per_batch =
GetEmptyTensor({0}, paddle::DataType::INT32, seq_lens_encoder.place());
kv_num_blocks_x_cpu =
GetEmptyTensor({0}, paddle::DataType::INT32, seq_lens_encoder.place());
GetEmptyTensor({0}, paddle::DataType::INT32, paddle::CPUPlace());
}
if (max_just_dec_len_this_time > 0) {