apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-06 09:07:10 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

FDTensor support GPU device (#190)

Browse Source 
* fdtensor support GPU

* TRT backend support GPU FDTensor

* FDHostAllocator add FASTDEPLOY_DECL

* fix FDTensor Data

* fix FDTensor dtype

Co-authored-by: Jason <jiangjiajun@baidu.com>
This commit is contained in:
heliqi
2022-09-08 03:53:08 -05:00
committed by GitHub
parent bc8e9e4dae
commit 4d1f264d01
17 changed files with 432 additions and 153 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
csrc/fastdeploy/backends/paddle/paddle_backend.h
View File

@@ -44,8 +44,11 @@ struct PaddleBackendOption {
std::vector<std::string> delete_pass_names = {};
};
// convert FD device to paddle place type
paddle_infer::PlaceType ConvertFDDeviceToPlace(Device device);
// Share memory buffer with paddle_infer::Tensor from fastdeploy::FDTensor
void ShareTensorFromCpu(paddle_infer::Tensor* tensor, FDTensor& fd_tensor);
void ShareTensorFromFDTensor(paddle_infer::Tensor* tensor, FDTensor& fd_tensor);
// Copy memory data from paddle_infer::Tensor to fastdeploy::FDTensor
void CopyTensorToCpu(std::unique_ptr<paddle_infer::Tensor>& tensor,
@@ -72,6 +75,8 @@ class PaddleBackend : public BaseBackend {
TensorInfo GetInputInfo(int index);
TensorInfo GetOutputInfo(int index);
std::vector<TensorInfo> GetInputInfo();
std::vector<TensorInfo> GetOutputInfo();
private:
paddle_infer::Config config_;