Add Cuda iniitialization in Classifier

CMakeLists.txt

@@ -51,10 +51,15 @@ endif (CMAKE_BUILD_TYPE STREQUAL "Debug")
 
 
 if (CODE_COVERAGE)
+get_property(LANGUAGES GLOBAL PROPERTY ENABLED_LANGUAGES)
+message("ALL LANGUAGES: ${LANGUAGES}")
+foreach(LANG ${LANGUAGES})
+    message("${LANG} compiler is \"${CMAKE_${LANG}_COMPILER_ID}\"")
+endforeach()
   enable_testing()
-  include(CodeCoverage)
+  #include(CodeCoverage)
-  MESSAGE("Code coverage enabled")
+  #MESSAGE("Code coverage enabled")
-  SET(GCC_COVERAGE_LINK_FLAGS " ${GCC_COVERAGE_LINK_FLAGS} -lgcov --coverage")
+  #SET(GCC_COVERAGE_LINK_FLAGS " ${GCC_COVERAGE_LINK_FLAGS} -lgcov --coverage")
 endif (CODE_COVERAGE)
 
 if (ENABLE_CLANG_TIDY)

bayesnet/classifiers/Classifier.cc

@@ -9,7 +9,15 @@
 #include "Classifier.h"
 
 namespace bayesnet {
-    Classifier::Classifier(Network model) : model(model), m(0), n(0), metrics(Metrics()), fitted(false) {}
+    Classifier::Classifier(Network model) : model(model), m(0), n(0), metrics(Metrics()), fitted(false), device(torch::kCPU)
+    {
+        if (torch::cuda::is_available()) {
+            device = torch::Device(torch::kCUDA);
+            std::cout << "CUDA is available! Using GPU." << std::endl;
+        } else {
+            std::cout << "CUDA is not available. Using CPU." << std::endl;
+        }
+    }
     const std::string CLASSIFIER_NOT_FITTED = "Classifier has not been fitted";
     Classifier& Classifier::build(const std::vector<std::string>& features, const std::string& className, std::map<std::string, std::vector<int>>& states, const torch::Tensor& weights, const Smoothing_t smoothing)
     {
@@ -31,7 +39,7 @@ namespace bayesnet {
     {
         try {
             auto yresized = torch::transpose(ytmp.view({ ytmp.size(0), 1 }), 0, 1);
-            dataset = torch::cat({ dataset, yresized }, 0);
+            dataset = torch::cat({ dataset, yresized }, 0).to(device);
         }
         catch (const std::exception& e) {
             std::stringstream oss;
@@ -50,7 +58,7 @@ namespace bayesnet {
     {
         dataset = X;
         buildDataset(y);
-        const torch::Tensor weights = torch::full({ dataset.size(1) }, 1.0 / dataset.size(1), torch::kDouble);
+        const torch::Tensor weights = torch::full({ dataset.size(1) }, 1.0 / dataset.size(1), torch::kDouble).to(device);
         return build(features, className, states, weights, smoothing);
     }
     // X is nxm where n is the number of features and m the number of samples

bayesnet/classifiers/Classifier.h

@@ -38,6 +38,7 @@ namespace bayesnet {
         std::string dump_cpt() const override;
         void setHyperparameters(const nlohmann::json& hyperparameters) override; //For classifiers that don't have hyperparameters
     protected:
+        torch::Device device;
         bool fitted;
         unsigned int m, n; // m: number of samples, n: number of features
         Network model;

bayesnet/network/Node.cc

@@ -97,7 +97,7 @@ namespace bayesnet {
         dimensions.push_back(numStates);
         transform(parents.begin(), parents.end(), back_inserter(dimensions), [](const auto& parent) { return parent->getNumStates(); });
         // Create a tensor of zeros with the dimensions of the CPT
-        cpTable = torch::zeros(dimensions, torch::kDouble) + smoothing;
+        cpTable = torch::zeros(dimensions, torch::kDouble).to(device) + smoothing;
         // Fill table with counts
         auto pos = find(features.begin(), features.end(), name);
         if (pos == features.end()) {

sample/sample.cc

@@ -7,6 +7,7 @@
 #include <ArffFiles.hpp>
 #include <CPPFImdlp.h>
 #include <bayesnet/ensembles/BoostAODE.h>
+#include <torch/torch.h>
 
 std::vector<mdlp::labels_t> discretizeDataset(std::vector<mdlp::samples_t>& X, mdlp::labels_t& y)
 {
@@ -19,7 +20,8 @@ std::vector<mdlp::labels_t> discretizeDataset(std::vector<mdlp::samples_t>& X, m
     }
     return Xd;
 }
-tuple<torch::Tensor, torch::Tensor, std::vector<std::string>, std::string, map<std::string, std::vector<int>>> loadDataset(const std::string& name, bool class_last)
+
+tuple<torch::Tensor, torch::Tensor, std::vector<std::string>, std::string, map<std::string, std::vector<int>>> loadDataset(const std::string& name, bool class_last, torch::Device device)
 {
     auto handler = ArffFiles();
     handler.load(name, class_last);
@@ -34,16 +36,16 @@ tuple<torch::Tensor, torch::Tensor, std::vector<std::string>, std::string, map<s
     torch::Tensor Xd;
     auto states = map<std::string, std::vector<int>>();
     auto Xr = discretizeDataset(X, y);
-    Xd = torch::zeros({ static_cast<int>(Xr.size()), static_cast<int>(Xr[0].size()) }, torch::kInt32);
+    Xd = torch::zeros({ static_cast<int>(Xr.size()), static_cast<int>(Xr[0].size()) }, torch::kInt32).to(device);
     for (int i = 0; i < features.size(); ++i) {
         states[features[i]] = std::vector<int>(*max_element(Xr[i].begin(), Xr[i].end()) + 1);
         auto item = states.at(features[i]);
         iota(begin(item), end(item), 0);
-        Xd.index_put_({ i, "..." }, torch::tensor(Xr[i], torch::kInt32));
+        Xd.index_put_({ i, "..." }, torch::tensor(Xr[i], torch::kInt32).to(device));
     }
     states[className] = std::vector<int>(*max_element(y.begin(), y.end()) + 1);
     iota(begin(states.at(className)), end(states.at(className)), 0);
-    return { Xd, torch::tensor(y, torch::kInt32), features, className, states };
+    return { Xd, torch::tensor(y, torch::kInt32).to(device), features, className, states };
 }
 
 int main(int argc, char* argv[])
@@ -53,16 +55,22 @@ int main(int argc, char* argv[])
         return 1;
     }
     std::string file_name = argv[1];
+    torch::Device device(torch::kCPU);
+    if (torch::cuda::is_available()) {
+        device = torch::Device(torch::kCUDA);
+        std::cout << "CUDA is available! Using GPU." << std::endl;
+    } else {
+        std::cout << "CUDA is not available. Using CPU." << std::endl;
+    }
     torch::Tensor X, y;
     std::vector<std::string> features;
     std::string className;
     map<std::string, std::vector<int>> states;
     auto clf = bayesnet::BoostAODE(false); // false for not using voting in predict
     std::cout << "Library version: " << clf.getVersion() << std::endl;
-    tie(X, y, features, className, states) = loadDataset(file_name, true);
+    tie(X, y, features, className, states) = loadDataset(file_name, true, device);
     clf.fit(X, y, features, className, states, bayesnet::Smoothing_t::LAPLACE);
     auto score = clf.score(X, y);
     std::cout << "File: " << file_name << " Model: BoostAODE score: " << score << std::endl;
     return 0;
-}
+}
-