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

Updated logistic regression example

Browse source 
Signed-off-by: Fabian Sauter <sauter.fabian@mailbox.org>
This commit is contained in:
Fabian Sauter 2022-06-24 10:10:11 +02:00
parent 7d16b73d14
commit 4b9b6607d0
7 changed files with 142 additions and 124 deletions
Download patch file Download diff file Expand all files Collapse all files

4
examples/logistic_regression/src/CMakeLists.txt Normal file
View file

@ -0,0 +1,4 @@
cmake_minimum_required(VERSION 3.15)
add_executable(kompute_logistic_regression main.cpp)
target_link_libraries(kompute_logistic_regression PRIVATE shader kompute::kompute)

72
examples/logistic_regression/src/Main.cpp
View file

@ -1,72 +0,0 @@
#include <iostream>
#include <memory>
#include <vector>
#include "kompute/Kompute.hpp"
int
main()
{
#if KOMPUTE_ENABLE_SPDLOG
spdlog::set_level(
static_cast<spdlog::level::level_enum>(SPDLOG_ACTIVE_LEVEL));
#endif
uint32_t ITERATIONS = 100;
float learningRate = 0.1;
kp::Manager mgr;
auto xI = mgr.tensor({ 0, 1, 1, 1, 1 });
auto xJ = mgr.tensor({ 0, 0, 0, 1, 1 });
auto y = mgr.tensor({ 0, 0, 0, 1, 1 });
auto wIn = mgr.tensor({ 0.001, 0.001 });
auto wOutI = mgr.tensor({ 0, 0, 0, 0, 0 });
auto wOutJ = mgr.tensor({ 0, 0, 0, 0, 0 });
auto bIn = mgr.tensor({ 0 });
auto bOut = mgr.tensor({ 0, 0, 0, 0, 0 });
auto lOut = mgr.tensor({ 0, 0, 0, 0, 0 });
std::vector<std::shared_ptr<kp::Tensor>> params = { xI, xJ, y,
wIn, wOutI, wOutJ,
bIn, bOut, lOut };
std::vector<uint32_t> spirv(
(uint32_t*)kp::shader_data::shaders_glsl_logisticregression_comp_spv,
(uint32_t*)(kp::shader_data::shaders_glsl_logisticregression_comp_spv +
kp::shader_data::
shaders_glsl_logisticregression_comp_spv_len));
std::shared_ptr<kp::Algorithm> algo = mgr.algorithm(
params, spirv, kp::Workgroup({ 5 }), std::vector<float>({ 5.0 }));
mgr.sequence()->eval<kp::OpTensorSyncDevice>(params);
std::shared_ptr<kp::Sequence> sq =
mgr.sequence()
->record<kp::OpTensorSyncDevice>({ wIn, bIn })
->record<kp::OpAlgoDispatch>(algo)
->record<kp::OpTensorSyncLocal>({ wOutI, wOutJ, bOut, lOut });
// Iterate across all expected iterations
for (size_t i = 0; i < ITERATIONS; i++) {
sq->eval();
for (size_t j = 0; j < bOut->size(); j++) {
wIn->data()[0] -= learningRate * wOutI->data()[j];
wIn->data()[1] -= learningRate * wOutJ->data()[j];
bIn->data()[0] -= learningRate * bOut->data()[j];
}
}
std::cout << "RESULTS" << std::endl;
std::cout << "w1: " << wIn->data()[0] << std::endl;
std::cout << "w2: " << wIn->data()[1] << std::endl;
std::cout << "b: " << bIn->data()[0] << std::endl;
}

66
examples/logistic_regression/src/main.cpp Normal file
View file

@ -0,0 +1,66 @@
#include <iostream>
#include <memory>
#include <vector>
#include "kompute/Tensor.hpp"
#include "my_shader.hpp"
#include <kompute/Kompute.hpp>
int
main()
{
uint32_t ITERATIONS = 100;
float learningRate = 0.1;
kp::Manager mgr;
std::shared_ptr<kp::TensorT<float>> xI = mgr.tensor({ 0, 1, 1, 1, 1 });
std::shared_ptr<kp::TensorT<float>> xJ = mgr.tensor({ 0, 0, 0, 1, 1 });
std::shared_ptr<kp::TensorT<float>> y = mgr.tensor({ 0, 0, 0, 1, 1 });
std::shared_ptr<kp::TensorT<float>> wIn = mgr.tensor({ 0.001, 0.001 });
std::shared_ptr<kp::TensorT<float>> wOutI = mgr.tensor({ 0, 0, 0, 0, 0 });
std::shared_ptr<kp::TensorT<float>> wOutJ = mgr.tensor({ 0, 0, 0, 0, 0 });
std::shared_ptr<kp::TensorT<float>> bIn = mgr.tensor({ 0 });
std::shared_ptr<kp::TensorT<float>> bOut = mgr.tensor({ 0, 0, 0, 0, 0 });
std::shared_ptr<kp::TensorT<float>> lOut = mgr.tensor({ 0, 0, 0, 0, 0 });
const std::vector<std::shared_ptr<kp::Tensor>> params = {
xI, xJ, y, wIn, wOutI, wOutJ, bIn, bOut, lOut
};
const std::vector<uint32_t> shader = std::vector<uint32_t>(
shader::MY_SHADER_COMP_SPV.begin(), shader::MY_SHADER_COMP_SPV.end());
std::shared_ptr<kp::Algorithm> algo = mgr.algorithm(
params, shader, kp::Workgroup({ 5 }), std::vector<float>({ 5.0 }));
mgr.sequence()->eval<kp::OpTensorSyncDevice>(params);
std::shared_ptr<kp::Sequence> sq =
mgr.sequence()
->record<kp::OpTensorSyncDevice>({ wIn, bIn })
->record<kp::OpAlgoDispatch>(algo)
->record<kp::OpTensorSyncLocal>({ wOutI, wOutJ, bOut, lOut });
// Iterate across all expected iterations
for (size_t i = 0; i < ITERATIONS; i++) {
sq->eval();
for (size_t j = 0; j < bOut->size(); j++) {
wIn->data()[0] -= learningRate * wOutI->data()[j];
wIn->data()[1] -= learningRate * wOutJ->data()[j];
bIn->data()[0] -= learningRate * bOut->data()[j];
}
}
std::cout << "RESULTS" << std::endl;
std::cout << "w1: " << wIn->data()[0] << std::endl;
std::cout << "w2: " << wIn->data()[1] << std::endl;
std::cout << "b: " << bIn->data()[0] << std::endl;
}