Mathis/CUDANet
1
0
Fork 0
mirror of https://github.com/lordmathis/CUDANet.git synced 2025-11-06 01:34:22 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add Kernels namespace

Browse Source
This commit is contained in:
Matúš Námešný
2024-03-11 21:04:23 +01:00
parent e0178e2d5c
commit d2ab78fbc7
18 changed files with 188 additions and 186 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
test/kernels/test_activations.cu
View File

@@ -25,7 +25,7 @@ TEST(ActivationsTest, SigmoidSanityCheck) {
cudaStatus = cudaMemcpy(d_input, input, sizeof(float) * 3, cudaMemcpyHostToDevice);
EXPECT_EQ(cudaStatus, cudaSuccess);
sigmoid_kernel<<<1, 3>>>(d_input, d_output, 3);
Kernels::sigmoid<<<1, 3>>>(d_input, d_output, 3);
cudaStatus = cudaDeviceSynchronize();
EXPECT_EQ(cudaStatus, cudaSuccess);

4
test/kernels/test_padding.cu
View File

@@ -3,7 +3,7 @@
#include <iostream>
#include "padding.cuh"
#include "convolution.cuh"
TEST(PaddingTest, SimplePaddingTest) {
cudaError_t cudaStatus;
@@ -51,7 +51,7 @@ TEST(PaddingTest, SimplePaddingTest) {
int THREADS_PER_BLOCK = 64;
int BLOCKS = paddedSize / THREADS_PER_BLOCK + 1;
pad_matrix_kernel<<<BLOCKS, THREADS_PER_BLOCK>>>(
Kernels::padding<<<BLOCKS, THREADS_PER_BLOCK>>>(
d_input, d_padded, w, h, n, p
);
cudaStatus = cudaDeviceSynchronize();

12
test/layers/test_conv2d.cu
View File

@@ -12,9 +12,9 @@ class Conv2dTest : public ::testing::Test {
int inputChannels,
int kernelSize,
int stride,
Padding padding,
Layers::Padding padding,
int numFilters,
Activation activation,
Layers::Activation activation,
std::vector<float>& input,
float* kernels,
float*& d_input,
@@ -65,9 +65,9 @@ TEST_F(Conv2dTest, SimpleTest) {
int inputChannels = 1;
int kernelSize = 2;
int stride = 1;
Padding padding = VALID;
Layers::Padding padding = Layers::Padding::VALID;
int numFilters = 1;
Activation activation = LINEAR;
Layers::Activation activation = Layers::Activation::NONE;
std::vector<float> input = {1.0f, 2.0f, 3.0f, 4.0f, 5.0f, 6.0f,
7.0f, 8.0f, 9.0f, 10.0f, 11.0f, 12.0f,
@@ -114,9 +114,9 @@ TEST_F(Conv2dTest, ComplexTest) {
int inputChannels = 3;
int kernelSize = 3;
int stride = 1;
Padding padding = SAME;
Layers::Padding padding = Layers::Padding::SAME;
int numFilters = 2;
Activation activation = LINEAR;
Layers::Activation activation = Layers::Activation::NONE;
// clang-format off
std::vector<float> input = {

19
test/layers/test_dense.cu
View File

@@ -16,7 +16,7 @@ class DenseLayerTest : public ::testing::Test {
float* biases,
float*& d_input,
float*& d_output,
Activation activation
Layers::Activation activation
) {
// Create Dense layer
Layers::Dense denseLayer(inputSize, outputSize, activation);
@@ -57,7 +57,9 @@ TEST_F(DenseLayerTest, Init) {
int inputSize = i;
int outputSize = j;
Layers::Dense denseLayer(inputSize, outputSize, SIGMOID);
Layers::Dense denseLayer(
inputSize, outputSize, Layers::Activation::SIGMOID
);
}
}
}
@@ -76,7 +78,9 @@ TEST_F(DenseLayerTest, setWeights) {
};
// clang-format on
Layers::Dense denseLayer(inputSize, outputSize, SIGMOID);
Layers::Dense denseLayer(
inputSize, outputSize, Layers::Activation::SIGMOID
);
denseLayer.setWeights(weights.data());
}
@@ -102,7 +106,7 @@ TEST_F(DenseLayerTest, ForwardUnitWeightMatrixLinear) {
Layers::Dense denseLayer = commonTestSetup(
inputSize, outputSize, input, weights.data(), biases.data(), d_input,
d_output, LINEAR
d_output, Layers::Activation::NONE
);
denseLayer.forward(d_input, d_output);
@@ -142,7 +146,8 @@ TEST_F(DenseLayerTest, ForwardRandomWeightMatrixRelu) {
float* d_output;
Layers::Dense denseLayer = commonTestSetup(
inputSize, outputSize, input, weights.data(), biases.data(), d_input, d_output, RELU
inputSize, outputSize, input, weights.data(), biases.data(), d_input,
d_output, Layers::Activation::RELU
);
denseLayer.forward(d_input, d_output);
@@ -186,8 +191,8 @@ TEST_F(DenseLayerTest, ForwardRandomWeightMatrixSigmoid) {
float* d_output;
Layers::Dense denseLayer = commonTestSetup(
inputSize, outputSize, input, weights.data(), biases.data(), d_input, d_output,
SIGMOID
inputSize, outputSize, input, weights.data(), biases.data(), d_input,
d_output, Layers::Activation::SIGMOID
);
denseLayer.forward(d_input, d_output);