PLSSVM - Parallel Least Squares Support Vector Machine  2.0.0
A Least Squares Support Vector Machine implementation using different backends.
Public Member Functions | Public Attributes | List of all members
plssvm::detail::cmd::parser_train Class Reference

Class for encapsulating all necessary parameters for training; normally provided through command line arguments. More...

#include <parser_train.hpp>

Collaboration diagram for plssvm::detail::cmd::parser_train:
[legend]

Public Member Functions

 parser_train (int argc, char **argv)
 Parse the command line arguments argv using cxxopts and set the training parameters accordingly. More...
 

Public Attributes

plssvm::parameter csvm_params {}
 Other base C-SVM parameters.
 
default_value< double > epsilon { default_init<double>{ 0.001 } }
 The error tolerance parameter for the CG algorithm.
 
default_value< std::size_t > max_iter { default_init<std::size_t>{ 0 } }
 The maximum number of iterations in the CG algorithm.
 
backend_type backend { backend_type::automatic }
 The used backend: automatic (depending on the specified target_platforms), OpenMP, CUDA, HIP, OpenCL, or SYCL.
 
target_platform target { target_platform::automatic }
 The target platform: automatic (depending on the used backend), CPUs or GPUs from NVIDIA, AMD or Intel.
 
sycl::kernel_invocation_type sycl_kernel_invocation_type { sycl::kernel_invocation_type::automatic }
 The kernel invocation type when using SYCL as backend.
 
sycl::implementation_type sycl_implementation_type { sycl::implementation_type::automatic }
 The SYCL implementation to use with –backend=sycl.
 
bool strings_as_labels { false }
 true if std::string should be used as label type instead of the default type ìnt.
 
bool float_as_real_type { false }
 true if float should be used as real type instead of the default type double.
 
std::string input_filename {}
 The name of the data/test file to parse.
 
std::string model_filename {}
 The name of the model file to write the learned support vectors to/to parse the saved model from.
 
std::string performance_tracking_filename {}
 If performance tracking has been enabled, provides the name of the file where the performance tracking results are saved to. If the filename is empty, the results are dumped to stdout instead.
 

Detailed Description

Class for encapsulating all necessary parameters for training; normally provided through command line arguments.

Constructor & Destructor Documentation

◆ parser_train()

plssvm::detail::cmd::parser_train::parser_train ( int  argc,
char **  argv 
)

Parse the command line arguments argv using cxxopts and set the training parameters accordingly.

If no model filename is given, uses the input filename and appends a ".model". The model file is than saved in the current working directory.

Parameters
[in]argcthe number of passed command line arguments
[in]argvthe command line arguments

The documentation for this class was generated from the following file: