thek0tyara/llama-cpp-turboquant
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 4

Python implementation

Browse source
This commit is contained in:
Alejandro Saucedo 2021-02-27 19:37:31 +00:00
parent 198fb46eb6
commit 4c4d073b90
18 changed files with 3172 additions and 3349 deletions
Download patch file Download diff file Expand all files Collapse all files

5
src/include/kompute/Manager.hpp
View file

@ -7,8 +7,6 @@
#include "kompute/Sequence.hpp"
#include "kompute/operations/OpTensorSyncDevice.hpp"
#define KP_DEFAULT_SESSION "DEFAULT"
namespace kp {
@ -83,8 +81,7 @@ class Manager
*/
std::shared_ptr<Tensor> tensor(
const std::vector<float>& data,
Tensor::TensorTypes tensorType = Tensor::TensorTypes::eDevice,
bool syncDataToGPU = true);
Tensor::TensorTypes tensorType = Tensor::TensorTypes::eDevice);
std::shared_ptr<Algorithm> algorithm(
const std::vector<std::shared_ptr<Tensor>>& tensors = {},