clean code

rapidsai · Jun 26, 2023 · 5a9cb6c · 5a9cb6c
1 parent b60b3e4
commit 5a9cb6c
Show file tree

Hide file tree

Showing 4 changed files with 3 additions and 107 deletions.
diff --git a/cpp/include/cuml/linear_model/qn_mg.hpp b/cpp/include/cuml/linear_model/qn_mg.hpp
@@ -1,16 +1,8 @@
 #include <raft/core/comms.hpp>
-// #include <raft/core/handle.hpp>
-// #include <raft/core/device_mdarray.hpp>
-// #include <raft/util/cudart_utils.hpp>
-// #include <raft/comms/std_comms.hpp>
 #include <cuml/common/logger.hpp>
-#include <cuml/linear_model/qn.h> // to use qn_params
-
+#include <cuml/linear_model/qn.h> 
 #include <cuda_runtime.h>
 
-// #include <vector>
-#include <iostream>
-
 namespace ML {
 namespace GLM {
 namespace opg {
@@ -35,42 +27,6 @@ void qnFit(const raft::handle_t &handle,
            int rank,
            int n_ranks);
 
-/**
- * @brief Fit a GLM using quasi newton methods.
- *
- * @param cuml_handle   reference to raft::handle_t object
- * @param params        model parameters
- * @param X             device pointer to a contiguous feature matrix of dimension [N, D]
- * @param X_col_major   true if X is stored column-major
- * @param y             device pointer to label vector of length N
- * @param N             number of examples
- * @param D             number of features
- * @param C             number of outputs (number of classes or `1` for regression)
- * @param w0            device pointer of size (D + (fit_intercept ? 1 : 0)) * C with initial point,
- *                      overwritten by final result.
- * @param f             host pointer holding the final objective value
- * @param num_iters     host pointer holding the actual number of iterations taken
- * @param sample_weight device pointer to sample weight vector of length n_rows (nullptr
-   for uniform weights)
- * @param svr_eps       epsilon parameter for svr
- */
-/*
-template <typename T, typename I = int>
-void qnFit(const raft::handle_t& cuml_handle,
-           const qn_params& params,
-           float* X,
-           bool X_col_major,
-           float* y,
-           int N,
-           int D,
-           int C,
-           float* w0,
-           float* f,
-           int* num_iters,
-           float* sample_weight = nullptr,
-           T svr_eps        = 0);
-           */
-
 };  // namespace opg
 };  // namespace GLM
 };  // namespace ML

diff --git a/cpp/src/glm/qn/glm_base_mg.cuh b/cpp/src/glm/qn/glm_base_mg.cuh
@@ -3,7 +3,6 @@
 #include <raft/core/handle.hpp>
 #include <raft/linalg/multiply.cuh>
 
-// the following are for PCA
 #include "glm/qn/glm_base.cuh"
 #include "glm/qn/glm_logistic.cuh"
 #include "glm/qn/qn_util.cuh"
@@ -47,8 +46,6 @@ inline void linearBwdMG(const raft::handle_t& handle,
     const SimpleDenseMat<T>* X_simple_p = (const SimpleDenseMat<T>*)(&X);
     Gweights.assign_gemm(handle, 1.0 / n_samples, dZ, false, X, false, beta / n_ranks, stream);
 
-    // Stats::opg::mean(handle, mu_data, input_data, input_desc, streams, n_streams);
-
     raft::stats::mean(Gbias.data, dZ.data, dZ.m, dZ.n, false, true, stream);
     T bias_factor = 1.0 * dZ.n / n_samples;
     raft::linalg::multiplyScalar(Gbias.data, Gbias.data, bias_factor, dZ.m, stream);
@@ -95,21 +92,8 @@ struct GLMWithDataMG : ML::GLM::detail::GLMWithData<T, GLMObjective> {
         raft::update_host(&reg_host, dev_scalar, 1, stream);
 
         // apply linearFwd, getLossAndDz, linearBwd
-        /*
-          inline void loss_grad(T* loss_val,
-                        Mat& G,
-                        const Mat& W,
-                        const SimpleMat<T>& Xb,
-                        const Vec& yb,
-                        Mat& Zb,
-                        cudaStream_t stream,
-                        bool initGradZero = true)
-        */
-        // call linearFwd, linearBwd, getLossAndDz
         ML::GLM::detail::linearFwd(lossFunc->handle, *(this->Z), *(this->X), W);                  // linear part: forward pass
 
-        //raft::interruptible::synchronize(stream);
-        // raft::comms::comms_t const& communicator = raft::resource::get_comms(handle);   
         raft::comms::comms_t const& communicator = raft::resource::get_comms(*(this->handle_p));   
 
         lossFunc->getLossAndDZ(dev_scalar, *(this->Z), *(this->y), stream);  // loss specific part
@@ -126,7 +110,6 @@ struct GLMWithDataMG : ML::GLM::detail::GLMWithData<T, GLMObjective> {
         raft::resource::sync_stream(*(this->handle_p));    
         raft::interruptible::synchronize(stream);
 
-        //linearBwd(lossFunc->handle, G, *(this->X), *(this->Z), false);    // linear part: backward pass
         linearBwdMG(lossFunc->handle, G, *(this->X), *(this->Z), false, n_samples, n_ranks);    // linear part: backward pass
         raft::interruptible::synchronize(stream);
 

diff --git a/cpp/src/glm/qn_mg.cu b/cpp/src/glm/qn_mg.cu
@@ -1,16 +1,8 @@
 #include <raft/core/comms.hpp>
 #include <raft/util/cudart_utils.hpp>
 #include <raft/core/handle.hpp>
-
 #include <cuml/common/logger.hpp>
-
-//#include "glm/qn/glm_base.cuh"
-//#include "glm/qn/glm_logistic.cuh"
-//#include "glm/qn/qn_util.cuh"
-//#include "glm/qn/qn_solvers.cuh"
-//#include "glm/qn/glm_regularizer.cuh"
-
-#include <cuml/linear_model/qn.h> // to use qn_params
+#include <cuml/linear_model/qn.h> 
 #include "qn/simple_mat/dense.hpp"
 #include "qn/qn_util.cuh"
 #include "qn/glm_logistic.cuh"
@@ -19,8 +11,6 @@
 #include "qn/glm_base_mg.cuh"
 
 #include <cuda_runtime.h>
-#include <iostream>
-
 
 namespace ML {
 namespace GLM {
@@ -127,37 +117,3 @@ void qnFit(const raft::handle_t &handle,
 };  // namespace OPG
 };  // namespace GLM
 };  // namespace ML
-
-// #include <raft/core/device_mdarray.hpp>
-//#include <iostream>
-//#include <vector>
-
-//#include <cuml/linear_model/qn_mg.hpp>
-
-/*
-namespace ML {
-namespace GLM {
-
-
-};  // namespace GLM
-};  // namespace ML
-*/
-
-/*
-template <typename T>
-void qnFit(const raft::handle_t& handle,
-           const qn_params& pams,
-           T* X_data,
-           bool X_col_major,
-           T* y_data,
-           int N,
-           int D,
-           int C,
-           T* w0_data,
-           T* f,
-           int* num_iters,
-           T* sample_weight = nullptr,
-           T svr_eps        = 0)
-{
-}
-*/
diff --git a/python/cuml/linear_model/logistic_regression_mg.pyx b/python/cuml/linear_model/logistic_regression_mg.pyx
@@ -148,6 +148,7 @@ class LogisticRegressionMG(LogisticRegression):
                 coef_size, dtype=self.dtype, order='C')
 
     def fit(self, X, y, rank, n_ranks, n_samples, n_classes, convert_dtype=False) -> "LogisticRegressionMG":
+        assert (n_classes == 2):
 
         cdef handle_t* handle_ = <handle_t*><size_t>self.handle.getHandle()