-
Notifications
You must be signed in to change notification settings - Fork 4
/
lr.cpp
329 lines (313 loc) · 9.75 KB
/
lr.cpp
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
#include "mpi.h"
#include "omp.h"
#include "lr.h"
#include <string.h>
#include <stdlib.h>
#include <time.h>
#include <iomanip>
LR::LR()
{}
LR::~LR()
{}
float LR::sigmoid(float x)
{
double sgm = 1/(1+exp(-(double)x));
return (float)sgm;
}
vector<string> LR::splitline(string &line)
{
vector<string> tmp_vec;
size_t beg = 0, end = 0;
string split_tag = "\t";
while((end = line.find_first_of(split_tag,beg)) != string::npos)
{
if(end > beg)
{
string index_str = line.substr(beg, end - beg);
tmp_vec.push_back(index_str);
}
beg = end + 1;
}
if(beg < line.size())
{
string index_end = line.substr(beg);
tmp_vec.push_back(index_end);
}
return tmp_vec;
}
int LR::get_feature_num(string sample_filename, vector<int>& label, vec_vec& feature_matrix,int myid,int numprocs)
{
//cout<<"Calculate feature number......"<<endl;
ifstream fin(sample_filename.c_str());
if(!fin)cerr<<"open error get feature number..."<<sample_filename<<endl;
string line, splittag = ":";
int max_index = 0;
vector<string> feature_index;
sparse_feature sf;
vec key_val;
int i = 0;
while(getline(fin,line)){
// if(i%10000 == 0)cout<<"for pthreadid: "<<myid<<"\t"<<i<<endl;
i++;
key_val.clear();
feature_index.clear();
feature_index = splitline(line);
int y = atoi(feature_index[0].c_str());
//cout<<y<<endl;
label.push_back(y);
for(int i = 1; i < feature_index.size(); i++){
int index = 0, beg = 0, end = 0;
while((end = feature_index[i].find_first_of(":",beg)) != string::npos){
if(end > beg){
string indexstr = feature_index[i].substr(beg,end-beg);
// cout<<indexstr<<"??i = "<<i<<endl;
index = atoi(indexstr.c_str());
// cout<<index<<endl;
// if(index == 0){cout<<i<<line<<endl;return 1;}
if (index > max_index)
max_index = index;
// n cout<<index<<" ";
sf.id_index = index - 1;
}
//beg += 1; //this code must remain,it makes me crazy two days!!!
beg = end + 1;
}
if(beg < feature_index[i].size()){
string indexend = feature_index[i].substr(beg);
int value = atoi(indexend.c_str());
sf.id_val = value;
}
key_val.push_back(sf);
//cout<<"--------------------"<<endl;
//if (index > max_index)
// max_index = index;
}
feature_matrix.push_back(key_val);
}
fin.close();
//int send = 1;
//for(int tid = 1; tid < numprocs; tid++)
// MPI_Send(&send,1,MPI_INTEGER,tid,1,MPI_COMM_WORLD);
cout<<"maxindex = "<<max_index<<endl;
return max_index;
}
void load_one_sample(string sample_filename)
{}
void LR::init_theta(vector<float>& theta, vector<float> &delta_theta,int feature_size)
{
float init_theta = 0.0;
for(size_t i = 0; i < feature_size; i++){
theta.push_back(init_theta);
delta_theta.push_back(init_theta);
}
}
void LR::train(string filename,vector<float>& theta, vector<float> &delta_theta, vector<int>& label,vector<vector<sparse_feature> >& feature_matrix,int myid,int numprocs)
{
size_t step = 0;
string sample_line;
vector<int> index;
//cout<<"traing start......"<<endl;
while(step < 1)
{
int count = 0;
size_t size = feature_matrix.size();
int start = myid * (size / numprocs);
int stop = (myid+1) * (size / numprocs);
if(myid == numprocs - 1) stop = size;
for(size_t i = start; i < stop; i++)
{
// if(i % 1000000 == 0)cout<<i<<endl;
float val,x = 0.0, y = 0.0;
//#pragma omp parallel num_threads(16)
//{
for(size_t j = 0; j < feature_matrix[i].size(); j++){
int index = feature_matrix[i][j].id_index;
float val = feature_matrix[i][j].id_val;
x += theta[index]*val;
}
//}
y = sigmoid(x);
//#pragma omp parallel num_threads(16)
//{
for(size_t j = 0; j < feature_matrix[i].size(); j++)
{
delta_theta[feature_matrix[i][j].id_index] += y-label[i];
}
//}
//#pragma omp parallel num_threads(16)
//{//
// cout<<i<<endl;
if(i%10000 == 0){
for(size_t j = 0; j < theta.size(); j++)
theta[j] -= 0.1*delta_theta[j]/1000;
delta_theta.clear();
}
//}
// cout<<"--------------------------------"<<endl;
}
step++;
//cout<<"step "<<step<<endl;
}
}
void LR::savemodel(vector<float> &theta,int myid){
ofstream fout("train.model");
fout.setf(ios::fixed,ios::floatfield);
fout.precision(7);
//cout<<"model size = "<<theta.size()<<endl;
//for(int i = 0; i < theta.size(); i++){
int k = theta.size();
cout<<"k = "<<k<<endl;
if(myid == 0)
for(int i = 0; i < k; i++){
fout<<theta[i]<<endl;
}
fout.close();
}
void LR::predict(string trainfile, vector<float>& theta)
{
cout<<"predic start-----------------------------------"<<endl;
ifstream fin(trainfile.c_str());
string train_line;
vector<float> predict_result;
vector<string> predict_feature;
float x;
vector<int> preindex;
vector<float> preval;
while(getline(fin, train_line))
{
x = 0.0;
predict_feature.clear();
predict_feature = splitline(train_line);
preindex.clear();
preval.clear();
for(size_t j = 0; j < predict_feature.size(); j++)
{
int beg = 0, end = 0;
while((end = predict_feature[j].find_first_of(":",beg)) != string::npos)
{
if(end > beg)
{
string string_sub = predict_feature[j].substr(beg, end - beg);
int k = atoi(string_sub.c_str());
preindex.push_back(k-1);
}
beg += 1;
}
string string_end = predict_feature[j].substr(beg);
int t = atoi(string_end.c_str());
preval.push_back(t);
//x += theta[k-1];
}
for(size_t j = 0; j < preindex.size(); j++)
{
//cout<<preindex[j]<<":"<<preval[j]<<endl;
x += theta[preindex[j]]*preval[j];
}
float y = sigmoid(x);
predict_result.push_back(y);
}
for(size_t j = 0; j < predict_result.size(); j++)
{
cout<<"predict result:"<<endl;
cout<<predict_result[j]<<endl;
}
}
struct A{ int a;char b; short c;char d;};
struct _feature{
int id_index;
float id_val;
};
int main(int argc,char* argv[])
{
/* MPI_Init(&argc, &argv);
int my_rank;
MPI_Comm_rank(MPI_COMM_WORLD, &my_rank);
int finalize_retcode = MPI_Finalize();
if(0 == my_rank) fprintf(stderr, "Process, return_code\n");
fprintf(stderr, "%i, %i\n", my_rank, finalize_retcode);*/
//cout<<sizeof(A)<<endl;
int myid, numprocs,ans = 0;
MPI_Status status;
MPI_Init(&argc,&argv);
MPI_Comm_rank(MPI_COMM_WORLD,&myid);
MPI_Comm_size(MPI_COMM_WORLD,&numprocs);
clock_t time1 ,time2;
LR lr;
int feature_num = 0;
//string train_file = "train_feature";
string train_file = "sampling";
// string predict_file = "predict_file";
// train_file = argv[1];
/* struct _feature matrix;
MPI_Datatype myvalue;
MPI_Datatype old_types[2];
MPI_Aint indices[2];
int blocklens[2];
blocklens[0] = 1;
blocklens[1] = 1;
old_types[0] = MPI_INT;
old_types[1] = MPI_FLOAT;
MPI_Address(&matrix,&indices[0]);
MPI_Address(&matrix,&indices[1]);
indices[1] -= indices[0];
indices[0] = 0;
MPI_Type_struct(2,blocklens,indices,old_types,&myvalue);
MPI_Type_commit(&myvalue);
*/
time1 = clock();
//if(myid == 0){
feature_num = lr.get_feature_num(train_file, lr.label,lr.feature_matrix,myid,numprocs);
// int size = lr.feature_matrix.size();
/* MPI_Datatype f_m;
MPI_Type_vector(1,1,size,myvalue,&f_m);
MPI_Type_commit(&f_m);
MPI_Datatype f_matrix;
MPI_Type_vector(size,1,size,myvalue,&f_matrix);
MPI_Type_commit(&f_matrix);
MPI_Bcast(&lr.feature_matrix,size*size,f_matrix,0,MPI_COMM_WORLD);
ans = 1;
*/
//for(int otherid = 1; otherid < numprocs; otherid++)
// MPI_Send(&ans,1,MPI_INT,otherid,1,MPI_COMM_WORLD);
//}
time2 = clock();
/* for(size_t i = 0; i < 4; i++)
{
cout<<lr.feature_matrix[i].size()<<endl;
for(size_t j = 0; j < lr.feature_matrix[i].size(); j++)
{
cout<<lr.feature_matrix[i][j].id_index<<":"<<lr.feature_matrix[i][j].id_val<<" ";
if(lr.feature_matrix[i][j].id_index == -1)
{
cout<<i<<"-"<<j<<endl;
return 1;
}
cout<<endl;
}
}*/
//cout<<"---------------"<<feature_num<<endl;
lr.init_theta(lr.theta,lr.delta_theta,feature_num);
//cout<< time2 - time1 <<endl;
//if(myid != 0){
// MPI_Recv(&ans,1,MPI_INT,0,1,MPI_COMM_WORLD,&status);
//if(ans == 1){
//lr.delta_theta(lr.theta.size(),0.0);
lr.train(train_file,lr.theta, lr.delta_theta,lr.label, lr.feature_matrix,myid,numprocs);
lr.savemodel(lr.theta, myid);
// }
//}else{
// lr.train(train_file,lr.theta, lr.label, lr.feature_matrix,myid,numprocs);
//}
//lr.theta.clear();
//lr.delta_theta.clear();
//lr.feature_matrix.clear();
clock_t time3 = clock();
//cout<<time3 - time2<<endl;
//lr.label.clear();
//lr.feature_matrix.clear();
//lr.predict(predict_file, lr.theta);
int ret_code = MPI_Finalize();
if(0 == myid) fprintf(stderr,"process,return_code");
fprintf(stderr,"%i,%i",myid, ret_code);
return 0;
}