fix c++ version

CMakeLists.txt

@@ -0,0 +1,101 @@
+cmake_minimum_required(VERSION 3.15 FATAL_ERROR)
+
+project(difacto VERSION 0.1 LANGUAGES C CXX)
+
+# 设置C++标准
+set(CMAKE_CXX_STANDARD 11)
+set(CMAKE_CXX_STANDARD_REQUIRED ON)
+set(CMAKE_CXX_EXTENSIONS OFF)
+
+# 定义依赖路径
+set(DEPS_PATH ${CMAKE_CURRENT_SOURCE_DIR}/deps)
+
+# 配置选项
+option(USE_CITY "Build with CITY HASH support" ON)
+option(USE_LZ4 "Build with LZ4 support" ON)
+option(NO_REVERSE_ID "Disable reverse feature ID" OFF)
+
+# 添加路径定义
+include_directories(
+  ${CMAKE_CURRENT_SOURCE_DIR}/src
+  ${CMAKE_CURRENT_SOURCE_DIR}/include
+  ${CMAKE_CURRENT_SOURCE_DIR}/dmlc-core/include
+  ${CMAKE_CURRENT_SOURCE_DIR}/ps-lite/include
+  ${CMAKE_CURRENT_SOURCE_DIR}/dmlc-core/src
+  ${DEPS_PATH}/include
+)
+
+# 定义全局编译选项
+set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11 -fopenmp -fPIC -O3 -ggdb -Wall -finline-functions -DDMLC_LOG_FATAL_THROW=0")
+
+if(NO_REVERSE_ID)
+  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DREVERSE_FEATURE_ID=0")
+endif()
+
+# 添加dmlc-core子项目
+add_subdirectory(${CMAKE_CURRENT_SOURCE_DIR}/dmlc-core)
+
+# 添加ps-lite子项目
+add_subdirectory(${CMAKE_CURRENT_SOURCE_DIR}/ps-lite)
+
+# 编译difacto库
+file(GLOB_RECURSE DIFACTO_SRCS
+  src/loss/loss.cc
+  src/updater.cc
+  src/sgd/sgd_updater.cc
+  src/sgd/sgd_learner.cc
+  src/learner.cc
+  src/bcd/bcd_learner.cc
+  src/lbfgs/lbfgs_learner.cc
+  src/store/store.cc
+  src/tracker/tracker.cc
+  src/reporter/reporter.cc
+  src/data/localizer.cc
+  src/reader/batch_reader.cc
+)
+
+add_library(difacto STATIC ${DIFACTO_SRCS})
+
+# City Hash 依赖处理
+if(USE_CITY)
+  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DDIFACTO_USE_CITY=1")
+  target_link_libraries(difacto ${DEPS_PATH}/lib/libcityhash.a)
+endif()
+
+# LZ4 依赖处理
+if(USE_LZ4)
+  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DDIFACTO_USE_LZ4=1")
+  target_link_libraries(difacto ${DEPS_PATH}/lib/liblz4.a)
+endif()
+
+# 添加可执行文件
+add_executable(difacto_main src/main.cc)
+set_target_properties(difacto_main PROPERTIES OUTPUT_NAME "difacto")
+target_link_libraries(difacto_main difacto dmlc pslite ${CMAKE_THREAD_LIBS_INIT})
+
+# 添加测试目标
+file(GLOB_RECURSE TEST_SRCS
+  tests/cpp/main.cc
+  tests/cpp/spmv_test.cc
+  tests/cpp/spmm_test.cc
+  tests/cpp/spmt_test.cc
+  tests/cpp/compressed_row_block_test.cc
+  tests/cpp/find_position_test.cc
+  tests/cpp/localizer_test.cc
+  tests/cpp/kv_match_test.cc
+  tests/cpp/kv_union_test.cc
+  tests/cpp/sgd_learner_test.cc
+  tests/cpp/bcd_learner_test.cc
+  tests/cpp/lbfgs_learner_test.cc
+  tests/cpp/lbfgs_twoloop_test.cc
+  tests/cpp/batch_reader_test.cc
+  tests/cpp/fm_loss_test.cc
+  tests/cpp/logit_loss_delta_test.cc
+  tests/cpp/data_store_test.cc
+)
+
+add_executable(difacto_tests ${TEST_SRCS})
+target_link_libraries(difacto_tests difacto dmlc pslite gtest gtest_main pthread ${CMAKE_THREAD_LIBS_INIT})
+
+# 如果需要单元测试，在这里添加
+# add_subdirectory(tests) 

README.md

@@ -21,7 +21,7 @@ and train FM with 2-dimension on it.
 git clone --recursive https://github.com/dmlc/difacto
 cd difacto; git submodule update --init; make -j8
 ./tools/download.sh gisette
-build/difacto data_in=data/gisette_scale val_data=data/gisette_scale.t lr=.02 V_dim=2 V_lr=.001
+build/difacto data_in=data/gisette_scale val_data=data/gisette_scale.t lr=.02 V_dim=2 V_lr=.001 batch_size=256
 ```
 
 ### History
@@ -35,3 +35,20 @@ Origins from
 
 Mu Li, Ziqi Liu, Alex Smola, and Yu-Xiang Wang.
 DiFacto — Distributed Factorization Machines. In WSDM, 2016
+
+## Build
+
+### 使用CMake构建（推荐）
+
+```bash
+# 创建编译目录
+mkdir -p build
+cd build
+
+# 配置和编译项目
+cmake ..
+make -j4
+
+# 运行示例
+./difacto task=train learner=sgd batch_size=100 data_in=/path/to/data V_dim=10
+```

example/criteo_sgd.conf

@@ -1,19 +1,40 @@
 # data
-data_in = data/criteo_kaggle/criteo_train.rec
-data_val = data/criteo_kaggle/criteo_val.rec
-data_format = rec
+# data_in = /workspace/public_data/criteo/dac/train.txt
+# data_val = /workspace/public_data/criteo/dac/test.txt
+data_in = ./data/criteo.train_sample
+data_val = ./data/criteo.train_sample
+# data_in = ./data/criteo.train_sample_libsvm
+# data_val = ./data/criteo.train_sample_libsvm
+data_format = criteo
+# data_format = libsvm
+
+# 日志设置
+# log_level = 1
+# report_interval = 60  # 每60秒报告一次进度
 
 # learner
 task = train
 learner = sgd
-max_num_epochs = 10
-batch_size = 10000
+max_num_epochs = 5
+batch_size = 1000
 
-# linear term
-l1 = 10
-l2 = 10
+# # linear term
+# l1 = 10
+# l2 = 10
 
 # embedding term
-V_dim = 10
+V_dim = 8
 V_threshold = 10
-V_l2 = 10
+# V_l2 = 10
+
+# loss
+loss = fm
+
+# learning rate
+lr = 0.01
+lr_beta = 1
+V_lr = 0.01
+V_lr_beta = 1
+
+# number of jobs per epoch
+num_jobs_per_epoch = 10

example/gisette_sgd.conf

@@ -0,0 +1,20 @@
+# data
+data_in = data/gisette_scale
+data_val = data/gisette_scale.t
+
+# learner
+task = train
+learner = sgd
+max_num_epochs = 2
+batch_size = 512
+lr = 0.02 
+V_lr=.001
+
+# linear term
+l1 = 10
+l2 = 10
+
+# embedding term
+V_dim = 10
+V_threshold = 10
+V_l2 = 10

include/difacto/store.h

@@ -6,6 +6,7 @@
 #include <memory>
 #include <vector>
 #include <string>
+#include <functional>
 #include "./base.h"
 #include "dmlc/io.h"
 #include "dmlc/parameter.h"

src/reader/adfea_parser.h

@@ -10,7 +10,7 @@
 #include "difacto/base.h"
 #include "data/row_block.h"
 #include "data/parser.h"
-#include "data/strtonum.h"
+#include "dmlc/strtonum.h"
 namespace difacto {
 
 /**
@@ -33,9 +33,9 @@ class AdfeaParser : public dmlc::data::ParserImpl<feaid_t> {
   }
   bool ParseNext(
       std::vector<dmlc::data::RowBlockContainer<feaid_t> > *data) override {
-    using dmlc::data::isspace;
-    using dmlc::data::isdigit;
-    using dmlc::data::strtoull;
+    using dmlc::isspace;
+    using dmlc::isdigit;
+    using dmlc::strtoull;
 
     dmlc::InputSplit::Blob chunk;
 

src/reader/criteo_parser.h

@@ -6,14 +6,13 @@
 #ifndef DIFACTO_READER_CRITEO_PARSER_H_
 #define DIFACTO_READER_CRITEO_PARSER_H_
 #include <limits>
-#if DIFACTO_USE_CITY
-#include <city.h>
-#endif  // DIFACTO_USE_CITY
+#include <string>
+#include <functional>
 #include <vector>
 #include "difacto/base.h"
 #include "data/row_block.h"
 #include "data/parser.h"
-#include "data/strtonum.h"
+#include "dmlc/strtonum.h"
 namespace difacto {
 
 /**
@@ -93,12 +92,9 @@ class CriteoParser : public dmlc::data::ParserImpl<feaid_t> {
 
  private:
   inline feaid_t Hash(const char* p, size_t len) {
-#if DIFACTO_USE_CITY
-    return CityHash64(p, len);
-#else
-    LOG(FATAL) << "compile with USE_CITY=1";
-    return 0;
-#endif  // DIFACTO_USE_CITY
+    std::string str(p, len);
+    std::hash<std::string> hasher;
+    return hasher(str);
   }
 
   // implement strchr