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

Migrated default base type of tensor to float

Browse source
This commit is contained in:
Alejandro Saucedo 2020-08-31 16:01:27 +01:00
parent 00c66f347b
commit 08a1300450
13 changed files with 255 additions and 249 deletions
Download patch file Download diff file Expand all files Collapse all files

8
src/include/kompute/Tensor.hpp
View file

@ -42,7 +42,7 @@ class Tensor
* @param data Vector of data that will be used by the tensor
* @param tensorType Type for the tensor which is of type TensorTypes
*/
Tensor(std::vector<uint32_t> data,
Tensor(std::vector<float> data,
TensorTypes tensorType = TensorTypes::eDevice);
/**
@ -70,7 +70,7 @@ class Tensor
*
* @return Vector of elements representing the data in the tensor.
*/
std::vector<uint32_t> data();
std::vector<float> data();
/**
* Returns the size/magnitude of the Tensor, which will be the total number
* of elements across all dimensions
@ -103,7 +103,7 @@ class Tensor
* Sets / resets the vector data of the tensor. This function does not
* perform any copies into GPU memory and is only performed on the host.
*/
void setData(const std::vector<uint32_t>& data);
void setData(const std::vector<float>& data);
/**
* Records a copy from the memory of the tensor provided to the current
@ -163,7 +163,7 @@ class Tensor
bool mFreeMemory;
// -------------- ALWAYS OWNED RESOURCES
std::vector<uint32_t> mData;
std::vector<float> mData;
TensorTypes mTensorType = TensorTypes::eDevice;