Mathis/CUDANet
1
0
Fork 0
mirror of https://github.com/lordmathis/CUDANet.git synced 2025-11-07 02:04:26 +00:00
Code Issues Packages Projects Releases Wiki Activity

Refactor model test

Browse Source
This commit is contained in:
Matúš Námešný
2024-04-14 00:05:32 +02:00
parent 457e0f669f
commit 53c976733b
2 changed files with 67 additions and 54 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
src/layers/conv2d.cu
View File

@@ -1,6 +1,3 @@
#include <iostream>
#include <string>
#include "activation.cuh" #include "activation.cuh"
#include "conv2d.cuh" #include "conv2d.cuh"
#include "convolution.cuh" #include "convolution.cuh"

118
test/model/test_model.cu
View File

@@ -5,60 +5,78 @@
#include "max_pooling.cuh" #include "max_pooling.cuh"
#include "model.hpp" #include "model.hpp"
TEST(Model, TestModelPredict) { class ModelTest : public ::testing::Test {
int inputSize = 6; protected:
int inputChannels = 2; CUDANet::Model* commonTestSetup(
int outputSize = 3; int inputSize = 6,
int inputChannels = 2,
int outputSize = 3,
int kernelSize = 3; int kernelSize = 3,
int stride = 1; int stride = 1,
int numFilters = 2; int numFilters = 2,
int poolingSize = 2; int poolingSize = 2,
int poolingStride = 2; int poolingStride = 2
) {
CUDANet::Model model(inputSize, inputChannels, outputSize); CUDANet::Model *model = new CUDANet::Model(inputSize, inputChannels, outputSize);
// Conv2d // Conv2d
CUDANet::Layers::Conv2d conv2d( CUDANet::Layers::Conv2d *conv2d = new CUDANet::Layers::Conv2d(
inputSize, inputChannels, kernelSize, stride, numFilters, CUDANet::Layers::Padding::VALID, inputSize, inputChannels, kernelSize, stride, numFilters, CUDANet::Layers::Padding::VALID,
CUDANet::Layers::ActivationType::NONE CUDANet::Layers::ActivationType::NONE
); );
// weights 6*6*2*2 // weights 6*6*2*2
std::vector<float> conv2dWeights = { std::vector<float> conv2dWeights = {
0.18313f, 0.53363f, 0.39527f, 0.27575f, 0.3433f, 0.41746f, 0.18313f, 0.53363f, 0.39527f, 0.27575f, 0.3433f, 0.41746f,
0.16831f, 0.61693f, 0.54599f, 0.99692f, 0.77127f, 0.25146f, 0.16831f, 0.61693f, 0.54599f, 0.99692f, 0.77127f, 0.25146f,
0.4206f, 0.16291f, 0.93484f, 0.79765f, 0.74982f, 0.78336f, 0.4206f, 0.16291f, 0.93484f, 0.79765f, 0.74982f, 0.78336f,
0.6386f, 0.87744f, 0.33587f, 0.9691f, 0.68437f, 0.65098f, 0.6386f, 0.87744f, 0.33587f, 0.9691f, 0.68437f, 0.65098f,
0.48153f, 0.97546f, 0.8026f, 0.36689f, 0.98152f, 0.37351f, 0.48153f, 0.97546f, 0.8026f, 0.36689f, 0.98152f, 0.37351f,
0.68407f, 0.2684f, 0.2855f, 0.76195f, 0.67828f, 0.603f 0.68407f, 0.2684f, 0.2855f, 0.76195f, 0.67828f, 0.603f
}; };
conv2d.setWeights(conv2dWeights.data()); conv2d->setWeights(conv2dWeights.data());
model.addLayer("conv2d", &conv2d); model->addLayer("conv2d", conv2d);
// maxpool2d // maxpool2d
CUDANet::Layers::MaxPooling2D maxpool2d( CUDANet::Layers::MaxPooling2D *maxpool2d = new CUDANet::Layers::MaxPooling2D(
inputSize - kernelSize + 1, numFilters, poolingSize, poolingStride, CUDANet::Layers::ActivationType::RELU inputSize - kernelSize + 1, numFilters, poolingSize, poolingStride, CUDANet::Layers::ActivationType::RELU
); );
model.addLayer("maxpool2d", &maxpool2d); model->addLayer("maxpool2d", maxpool2d);
// dense // dense
CUDANet::Layers::Dense dense( CUDANet::Layers::Dense *dense = new CUDANet::Layers::Dense(
8, 3, CUDANet::Layers::ActivationType::SOFTMAX 8, 3, CUDANet::Layers::ActivationType::SOFTMAX
); );
// dense weights 18*6 // dense weights 18*6
std::vector<float> denseWeights = { std::vector<float> denseWeights = {
0.36032f, 0.33115f, 0.02948f, 0.36032f, 0.33115f, 0.02948f,
0.09802f, 0.45072f, 0.56266f, 0.09802f, 0.45072f, 0.56266f,
0.43514f, 0.80946f, 0.43439f, 0.43514f, 0.80946f, 0.43439f,
0.90916f, 0.08605f, 0.07473f, 0.90916f, 0.08605f, 0.07473f,
0.94788f, 0.66168f, 0.34927f, 0.94788f, 0.66168f, 0.34927f,
0.09464f, 0.61963f, 0.73775f, 0.09464f, 0.61963f, 0.73775f,
0.51559f, 0.81916f, 0.64915f, 0.51559f, 0.81916f, 0.64915f,
0.03934f, 0.87608f, 0.68364f, 0.03934f, 0.87608f, 0.68364f,
}; };
dense.setWeights(denseWeights.data()); dense->setWeights(denseWeights.data());
model.addLayer("dense", &dense); model->addLayer("dense", dense);
return model;
}
void commonTestTeardown(float* d_input) {
cudaDeviceReset();
}
cudaError_t cudaStatus;
};
TEST_F(ModelTest, TestModelPredict) {
int outputSize = 3;
CUDANet::Model *model = commonTestSetup();
// input 6*6*2 // input 6*6*2
std::vector<float> input = { std::vector<float> input = {
@@ -78,7 +96,7 @@ TEST(Model, TestModelPredict) {
std::vector<float> expected = {2e-05f, 0.00021f, 0.99977f}; std::vector<float> expected = {2e-05f, 0.00021f, 0.99977f};
// predict // predict
const float* output = model.predict(input.data()); const float* output = model->predict(input.data());
float sum = 0.0f; float sum = 0.0f;
for (int i = 0; i < outputSize; ++i) { for (int i = 0; i < outputSize; ++i) {
@@ -87,6 +105,4 @@ TEST(Model, TestModelPredict) {
} }
EXPECT_NEAR(sum, 1.0f, 1e-5f); EXPECT_NEAR(sum, 1.0f, 1e-5f);
cudaDeviceReset();
} }