Migrate conv2d layer to Tensor

include/backend.hpp

@@ -40,6 +40,18 @@ class Backend {
         const size_t           input_size,
         const size_t           output_size
     ) = 0;
+
+    virtual CUDANet::Tensor& conv2d(
+        const CUDANet::Tensor& weights,
+        const CUDANet::Tensor& biases,
+        const CUDANet::Tensor& input,
+        CUDANet::Tensor& output,
+        const CUDANet::Shape in_shape,
+        const CUDANet::Shape padding_shape,
+        const CUDANet::Shape kernel_shape,
+        const CUDANet::Shape stride_shape,
+        const CUDANet::Shape out_shape
+    ) = 0;
 };
 
 }  // namespace CUDANet

include/backend/cuda.cuh

@@ -36,6 +36,18 @@ class CUDA : public Backend {
         const size_t           input_size,
         const size_t           output_size
     ) override;
+
+    CUDANet::Tensor& conv2d(
+        const CUDANet::Tensor& weights,
+        const CUDANet::Tensor& biases,
+        const CUDANet::Tensor& input,
+        CUDANet::Tensor& output,
+        const CUDANet::Shape in_shape,
+        const CUDANet::Shape padding_shape,
+        const CUDANet::Shape kernel_shape,
+        const CUDANet::Shape stride_shape,
+        const CUDANet::Shape out_shape
+    ) override;
 };
 
 }  // namespace CUDANet::Backend

include/kernels/convolution.cuh

@@ -1,39 +1,20 @@
-#ifndef CUDANET_CONVOLUTION_H
+#pragma once
-#define CUDANET_CONVOLUTION_H
 
 #include <cuda_runtime.h>
 #include "layer.hpp"
 
 namespace CUDANet::Kernels {
 
-/**
- * @brief Convolution kernel
- *
- * @param d_input Device pointer to the input matrix
- * @param d_kernel Device pointer to the convolution kernel
- * @param d_bias Device pointer to the bias
- * @param d_output Device pointer to the output matrix
- * @param inputSize Width and height of the input matrix
- * @param nChannels Number of channels in the input matrix
- * @param kernelSize Width and height of the convolution kernel
- * @param stride Convolution stride
- * @param nFilters Number of output filters
- * @param outputSize Width and height of the output matrix
- */
 __global__ void convolution(
     const float* __restrict__ d_input,
     const float* __restrict__ d_kernel,
     const float* __restrict__ d_bias,
     float* __restrict__ d_output,
-    const shape2d inputSize,
+    const Shape input_shape,
-    const int nChannels,
+    const Shape padding_shape,
-    const shape2d paddingSize,
+    const Shape kernel_shape,
-    const shape2d kernelSize,
+    const Shape stride_shape,
-    const shape2d stride,
+    const Shape output_shape
-    const int nFilters,
-    const shape2d outputSize
 );
 
 }  // namespace CUDANet::Kernels
-
-#endif  // CUDANET_CONVOLUTION_H

include/layers/conv2d.hpp

@@ -1,5 +1,4 @@
-#ifndef CUDANET_CONV_LAYER_H
+#pragma once
-#define CUDANET_CONV_LAYER_H
 
 #include <vector>
 
@@ -12,149 +11,52 @@ namespace CUDANet::Layers {
  * @brief 2D convolutional layer
  *
  */
-class Conv2d : public WeightedLayer, public TwoDLayer {
+class Conv2d : public Layer {
   public:
-    /**
-     * @brief Construct a new Conv 2d layer
-     *
-     * @param inputSize Width and height of the input matrix
-     * @param inputChannels Number of channels in the input matrix
-     * @param kernelSize Width and height of the convolution kernel
-     * @param stride Convolution stride
-     * @param numFilters Number of output filters
-     * @param paddingSize Padding size
-     * @param activationType Activation function type ('RELU', 'SIGMOID',
-     * 'SOFTMAX' or 'NONE')
-     */
     Conv2d(
-        shape2d        inputSize,
+        CUDANet::Shape    input_shape,
-        int            inputChannels,
+        CUDANet::Shape    kernel_shape,
-        shape2d        kernelSize,
+        CUDANet::Shape    stride_shape,
-        shape2d        stride,
+        CUDANet::Shape    padding_shape,
-        int            numFilters,
+        CUDANet::Backend* backend
-        shape2d        paddingSize,
-        ActivationType activationType
     );
 
-    /**
+    ~Conv2d() {};
-     * @brief Destroy the Conv 2d object
-     *
-     */
-    ~Conv2d();
 
-    /**
+    CUDANet::Tensor& forward(const CUDANet::Tensor& input) override;
-     * @brief Forward pass of the convolutional layer
-     *
-     * @param d_input Device pointer to the input matrix
-     * @return Device pointer to the output matrix
-     */
-    float* forward(const float* d_input);
 
-    /**
+    CUDANet::Shape input_shape() override;
-     * @brief Set the weights of the convolutional layer
-     *
-     * @param weights_input Pointer to the weights
-     */
-    void setWeights(const float* weights_input);
 
-    /**
+    CUDANet::Shape output_shape() override;
-     * @brief Get the weights of the convolutional layer
-     *
-     * @return std::vector<float>
-     */
-    std::vector<float> getWeights();
 
-    /**
+    size_t input_size() override;
-     * @brief Set the biases of the convolutional layer
-     *
-     * @param biases_input Pointer to the biases
-     */
-    void setBiases(const float* biases_input);
 
-    /**
+    size_t output_size();
-     * @brief Get the biases of the convolutional layer
-     *
-     * @return std::vector<float>
-     */
-    std::vector<float> getBiases();
 
-    /**
+    void set_weights(void* input) override;
-     * @brief Get output size
-     *
-     * @return int output size
-     */
-    int getOutputSize();
 
-    /**
+    CUDANet::Tensor& get_weights() override;
-     * @brief Get input size
-     *
-     * @return int input size
-     */
-    int getInputSize();
 
-    /**
+    void set_biases(void* input) override;
-     * @brief Get the padding size of the layer
-     *
-     * @return int
-     */
-    shape2d getPaddingSize() {
-        return paddingSize;
-    }
 
-    shape2d getOutputDims();
+    CUDANet::Tensor& get_biases() override;
+
+    CUDANet::Shape get_padding_shape();
 
   private:
-    // Inputs
+    CUDANet::Backend* backend;
-    shape2d inputSize;
-    int     inputChannels;
 
-    // Outputs
+    CUDANet::Shape in_shape;
-    shape2d outputSize;
+    CUDANet::Shape out_shape;
 
-    // Kernel
+    CUDANet::Shape kernel_shape;
-    shape2d kernelSize;
+    CUDANet::Shape stride_shape;
-    shape2d stride;
+    CUDANet::Shape padding_shape;
-    shape2d paddingSize;
-    int     numFilters;
 
-    // Kernels
+    CUDANet::Tensor weights;
-    std::vector<float> weights;
+    CUDANet::Tensor biases;
-    std::vector<float> biases;
 
-    float* forwardCPU(const float* input);
+    CUDANet::Tensor output;
-
-// Cuda
-#ifdef USE_CUDA
-    float* d_output;
-    float* d_weights;
-    float* d_biases;
-
-    float* forwardCUDA(const float* d_input);
-    void   initCUDA();
-    void   delCUDA();
-
-    /**
-     * @brief Copy weights and biases to the device
-     *
-     */
-    void toCuda();
-#endif
-
-    Activation* activation;
-
-    /**
-     * @brief Initialize weights of the convolutional layer with zeros
-     *
-     */
-    void initializeWeights();
-
-    /**
-     * @brief Initialize biases of the convolutional layer with zeros
-     *
-     */
-    void initializeBiases();
 };
 
 }  // namespace CUDANet::Layers
-
-#endif  // CUDANET_CONV_LAYER_H

include/layers/dense.hpp

@@ -14,7 +14,7 @@ namespace CUDANet::Layers {
 class Dense : public Layer {
   public:
 
-    Dense(CUDANet::Backend *backend, CUDANet::Shape input_shape, CUDANet::Shape output_shape);
+    Dense(CUDANet::Shape input_shape, CUDANet::Shape output_shape, CUDANet::Backend *backend);
 
     ~Dense();
 

src/backends/cuda/kernels/convolution.cu

@@ -9,52 +9,50 @@ __global__ void Kernels::convolution(
     const float* __restrict__ d_kernel,
     const float* __restrict__ d_bias,
     float* __restrict__ d_output,
-    const shape2d inputSize,
+    const Shape input_shape,
-    const int   nChannels,
+    const Shape padding_shape,
-    const shape2d paddingSize,
+    const Shape kernel_shape,
-    const shape2d kernelSize,
+    const Shape stride_shape,
-    const shape2d stride,
+    const Shape output_shape
-    const int   nFilters,
-    const shape2d outputSize
 ) {
     int j = blockDim.x * blockIdx.x + threadIdx.x;
     int i = blockDim.y * blockIdx.y + threadIdx.y;
     int f = blockDim.z * blockIdx.z + threadIdx.z;
 
-    if (i >= outputSize.first || j >= outputSize.second || f >= nFilters) {
+    if (i >= output_shape[0] || j >= output_shape[1] || f >= output_shape[2]) {
         return;
     }
 
     float sum = 0.0f;
 
     // Iterate over kernel and input matrix
-    for (int c = 0; c < nChannels; c++) {
+    for (int c = 0; c < input_shape[2]; c++) {
-        for (int k = 0; k < kernelSize.first; k++) {
+        for (int k = 0; k < kernel_shape[0]; k++) {
-            for (int l = 0; l < kernelSize.second; l++) {
+            for (int l = 0; l < kernel_shape[1]; l++) {
                 // if i, j is in the padding region
-                if (i * stride.first + k < paddingSize.first ||
+                if (i * stride_shape[0] + k < padding_shape[0] ||
-                    i * stride.first + k >=
+                    i * stride_shape[0] + k >=
-                        (inputSize.first + paddingSize.first) ||
+                        (input_shape[0] + padding_shape[0]) ||
-                    j * stride.second + l < paddingSize.second ||
+                    j * stride_shape[1] + l < padding_shape[1] ||
-                    j * stride.second + l >=
+                    j * stride_shape[1] + l >=
-                        (inputSize.second + paddingSize.second)) {
+                        (input_shape[1] + padding_shape[1])) {
                     continue;
                 }
 
                 int kernelIndex =
-                    f * kernelSize.first * kernelSize.second * nChannels +
+                    f * kernel_shape[0] * kernel_shape[1] * input_shape[2] +
-                    c * kernelSize.first * kernelSize.second +
+                    c * kernel_shape[0] * kernel_shape[1] +
-                    k * kernelSize.second + l;
+                    k * kernel_shape[1] + l;
-                int inputIndex = c * inputSize.first * inputSize.second +
+                int inputIndex = c * input_shape[0] * input_shape[1] +
-                                 (i * stride.first + k - paddingSize.first) *
+                                 (i * stride_shape[0] + k - padding_shape[0]) *
-                                     inputSize.second +
+                                     input_shape[1] +
-                                 (j * stride.second + l - paddingSize.second);
+                                 (j * stride_shape[1] + l - padding_shape[1]);
 
                 sum += d_kernel[kernelIndex] * d_input[inputIndex];
             }
         }
     }
 
-    d_output[f * outputSize.first * outputSize.second + i * outputSize.second + j] =
+    d_output[f * output_shape[0] * output_shape[1] + i * output_shape[1] + j] =
         sum + d_bias[f];
 }

src/backends/cuda/layer_ops.cu

@@ -1,5 +1,6 @@
 #include "backend/cuda.cuh"
 #include "kernels/activation_functions.cuh"
+#include "kernels/convolution.cuh"
 #include "kernels/matmul.cuh"
 #include "utils/cuda_helper.cuh"
 
@@ -57,7 +58,7 @@ CUDANet::Tensor& CUDA::dense(
     const CUDANet::Tensor& weights,
     const CUDANet::Tensor& biases,
     const CUDANet::Tensor& input,
-    CUDANet::Tensor& output,
+    CUDANet::Tensor&       output,
     const size_t           input_size,
     const size_t           output_size
 ) {
@@ -80,3 +81,32 @@ CUDANet::Tensor& CUDA::dense(
 
     return output;
 }
+
+CUDANet::Tensor& CUDA::conv2d(
+    const CUDANet::Tensor& weights,
+    const CUDANet::Tensor& biases,
+    const CUDANet::Tensor& input,
+    CUDANet::Tensor&       output,
+    const CUDANet::Shape   in_shape,
+    const CUDANet::Shape   padding_shape,
+    const CUDANet::Shape   kernel_shape,
+    const CUDANet::Shape   stride_shape,
+    const CUDANet::Shape   out_shape
+) {
+    dim3 block(8, 8, 8);
+    dim3 grid(
+        (out_shape[0] + block.x - 1) / block.x,
+        (out_shape[1] + block.y - 1) / block.y,
+        (out_shape[3] + block.z - 1) / block.z
+    );
+
+    Kernels::convolution<<<grid, block>>>(
+        input.data<float>(), weights.data<float>(), biases.data<float>(),
+        output.data<float>(), in_shape, padding_shape, kernel_shape,
+        stride_shape, out_shape
+    );
+    CUDA_CHECK(cudaGetLastError());
+    CUDA_CHECK(cudaDeviceSynchronize());
+
+    return output;
+}

src/backends/cuda/layers/conv2d.cu

@@ -49,25 +49,5 @@ void Conv2d::toCuda() {
 
 float* Conv2d::forwardCUDA(const float* d_input) {
     // Convolve
-    dim3 block(8, 8, 8);
-    dim3 grid(
-        (outputSize.first + block.x - 1) / block.x,
-        (outputSize.second + block.y - 1) / block.y,
-        (numFilters + block.z - 1) / block.z
-    );
 
-    CUDANet::Utils::clear(d_output, outputSize.first * outputSize.second * numFilters);
-
-    Kernels::convolution<<<grid, block>>>(
-        d_input, d_weights, d_biases, d_output, inputSize, inputChannels,
-        paddingSize, kernelSize, stride, numFilters, outputSize
-    );
-    CUDA_CHECK(cudaGetLastError());
-
-    // Apply activation
-    activation->activate(d_output);
-
-    CUDA_CHECK(cudaDeviceSynchronize());
-
-    return d_output;
 }

src/layers/conv2d.cpp

@@ -1,111 +1,136 @@
-#include <stdexcept>
-#include <vector>
-
-#include "activation.hpp"
 #include "conv2d.hpp"
+
+#include <format>
+#include <stdexcept>
+
 #include "layer.hpp"
+#include "tensor.hpp"
 
 using namespace CUDANet::Layers;
 
 Conv2d::Conv2d(
-    shape2d        inputSize,
+    CUDANet::Shape    input_shape,
-    int            inputChannels,
+    CUDANet::Shape    kernel_shape,
-    shape2d        kernelSize,
+    CUDANet::Shape    stride_shape,
-    shape2d        stride,
+    CUDANet::Shape    padding_shape,
-    int            numFilters,
+    CUDANet::Backend* backend
-    shape2d        paddingSize,
-    ActivationType activationType
 )
-    : inputSize(inputSize),
+    : in_shape(input_shape),
-      inputChannels(inputChannels),
+      kernel_shape(kernel_shape),
-      kernelSize(kernelSize),
+      stride_shape(stride_shape),
-      stride(stride),
+      padding_shape(padding_shape),
-      numFilters(numFilters),
+      backend(backend) {
-      paddingSize(paddingSize) {
+    if (in_shape.size() != 3) {
-    outputSize = {
+        throw std::runtime_error(
-        (inputSize.first - kernelSize.first + 2 * paddingSize.first) /
+            std::format(
-                stride.first +
+                "Invalid input shape. Expected 3 dims, got {}", in_shape
-            1,
+            )
-        (inputSize.second - kernelSize.second + 2 * paddingSize.second) /
+        );
-                stride.second +
+    }
-            1
-    };
 
-    activation = new Activation(
+    if (kernel_shape.size() != 3) {
-        activationType, outputSize.first * outputSize.second * numFilters
+        throw std::runtime_error(
+            std::format(
+                "Invalid kernel shape. Expected 3 dims, got {}", kernel_shape
+            )
+        );
+    }
+
+    if (stride_shape.size() != 2) {
+        throw std::runtime_error(
+            std::format(
+                "Invalid stride shape. Expected 2 dims, got {}", stride_shape
+            )
+        );
+    }
+
+    if (padding_shape.size() != 2) {
+        throw std::runtime_error(
+            std::format(
+                "Invalid padding shape. Expected 2 dims, got {}", padding_shape
+            )
+        );
+    }
+
+    size_t out_h = (in_shape[0] - kernel_shape[0] + 2 * padding_shape[0]) /
+                       stride_shape[0] +
+                   1;
+    size_t out_w = (in_shape[1] - kernel_shape[1] + 2 * padding_shape[1]) /
+                       stride_shape[1] +
+                   1;
+    out_shape.resize(3);
+    out_shape[0] = out_h;
+    out_shape[1] = out_w;
+    out_shape[2] = kernel_shape[2];
+    output = CUDANet::Tensor(
+        Shape{out_shape[0] * out_shape[1] * out_shape[3]},
+        CUDANet::DType::FLOAT32, backend
     );
 
-    weights.resize(
+    weights = CUDANet::Tensor(
-        kernelSize.first * kernelSize.second * inputChannels * numFilters
+        Shape{
+            kernel_shape[0] * kernel_shape[1] * kernel_shape[2] * in_shape[2]
+        },
+        CUDANet::DType::FLOAT32, backend
+    );
+    biases = CUDANet::Tensor(
+        Shape{kernel_shape[2]}, CUDANet::DType::FLOAT32, backend
     );
-    initializeWeights();
 
-    biases.resize(numFilters);
+    weights.zero();
-    initializeBiases();
+    biases.zero();
-
-#ifdef USE_CUDA
-    initCUDA();
-    toCuda();
-#endif
 }
 
-Conv2d::~Conv2d() {
+Conv2d::~Conv2d() {}
-#ifdef USE_CUDA
+
-    delCUDA();
+CUDANet::Tensor& Conv2d::forward(const CUDANet::Tensor& input) {
-#endif
+    output.zero();
-    delete activation;
+    backend->conv2d(
+        weights,
+        biases,
+        input,
+        output,
+        in_shape,
+        padding_shape,
+        kernel_shape,
+        stride_shape,
+        out_shape
+    );
+    return output;
 }
 
-void Conv2d::initializeWeights() {
+CUDANet::Shape Conv2d::input_shape() {
-    std::fill(weights.begin(), weights.end(), 0.0f);
+    return in_shape;
 }
 
-void Conv2d::initializeBiases() {
+CUDANet::Shape Conv2d::output_shape() {
-    std::fill(biases.begin(), biases.end(), 0.0f);
+    return out_shape;
 }
 
-void Conv2d::setWeights(const float* weights_input) {
+size_t Conv2d::input_size() {
-    std::copy(weights_input, weights_input + weights.size(), weights.begin());
+    return sizeof(float) * in_shape[0] * in_shape[1] * in_shape[2];
-#ifdef USE_CUDA
-    toCuda();
-#endif
 }
 
-std::vector<float> Conv2d::getWeights() {
+size_t Conv2d::output_size() {
+    return sizeof(float) * out_shape[0] * out_shape[1] * out_shape[2];
+}
+
+void Conv2d::set_weights(void* input) {
+    weights.set_data<float>(static_cast<float*>(input));
+}
+
+CUDANet::Tensor& Conv2d::get_weights() {
     return weights;
 }
 
-void Conv2d::setBiases(const float* biases_input) {
+void Conv2d::set_biases(void* input) {
-    std::copy(biases_input, biases_input + biases.size(), biases.begin());
+    biases.set_data<float>(static_cast<float*>(input));
-#ifdef USE_CUDA
-    toCuda();
-#endif
 }
 
-std::vector<float> Conv2d::getBiases() {
+CUDANet::Tensor& Conv2d::get_biases() {
     return biases;
 }
 
-float* Conv2d::forwardCPU(const float* input) {
+CUDANet::Shape Conv2d::get_padding_shape() {
-    throw std::logic_error("Not implemented");
+    return padding_shape;
-}
-
-float* Conv2d::forward(const float* input) {
-#ifdef USE_CUDA
-    return forwardCUDA(input);
-#else
-    return forwardCPU(input);
-#endif
-}
-
-int Conv2d::getOutputSize() {
-    return outputSize.first * outputSize.second * numFilters;
-}
-
-int Conv2d::getInputSize() {
-    return inputSize.first * inputSize.second * inputChannels;
-}
-
-shape2d Conv2d::getOutputDims() {
-    return outputSize;
 }

src/layers/dense.cpp

@@ -5,34 +5,30 @@
 
 using namespace CUDANet::Layers;
 
-Dense::Dense(CUDANet::Backend* backend, CUDANet::Shape in, CUDANet::Shape out)
+Dense::Dense(CUDANet::Shape in, CUDANet::Shape out, CUDANet::Backend* backend)
     : backend(backend),
       in_shape(in),
-      out_shape(out),
+      out_shape(out) {
-      weights(
-          CUDANet::Tensor(Shape{in[0] * out[0]}, CUDANet::DType::FLOAT32, backend)
-      ),
-      biases(CUDANet::Tensor(Shape{out[0]}, CUDANet::DType::FLOAT32, backend)),
-      output(CUDANet::Tensor(Shape{out[0]}, CUDANet::DType::FLOAT32, backend)) {
-    // Allocate memory for weights and biases
 
     if (in.size() != 1) {
         throw std::runtime_error(
-            std::format("Invalid shape. Expected [1], got {}", in)
+            std::format("Invalid shape. Expected [1], got {}", in_shape)
         );
     }
 
     if (out.size() != 1) {
         throw std::runtime_error(
-            std::format("Invalid shape. Expected [1], got {}", out)
+            std::format("Invalid shape. Expected [1], got {}", out_shape)
         );
     }
 
-    auto input_len  = in[0];
+    weights = CUDANet::Tensor(Shape{in[0] * out[0]}, CUDANet::DType::FLOAT32, backend);
-    auto output_len = out[0];
+    biases = CUDANet::Tensor(Shape{out[0]}, CUDANet::DType::FLOAT32, backend);
+    output = CUDANet::Tensor(Shape{out[0]}, CUDANet::DType::FLOAT32, backend);
 
     weights.zero();
     biases.zero();
+    output.zero();
 }
 
 Dense::~Dense() {}