fanasina/y_project
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

debug : nan, it's the learning rate too high,s try to find optimum: 0.001

Browse Source
This commit is contained in:
fanasina
2025-12-13 05:00:23 +01:00
parent 32207170f6
commit 3d54639d4c
18 changed files with 834 additions and 148 deletions
Download Patch File Download Diff File Expand all files Collapse all files
y_network_neural_network_/include/y_net_neur_net/y_nnn_screen_manager.h
+1
View File
@@ -61,6 +61,7 @@ struct arg_bash{
pthread_mutex_t *mut_bash_var;
pthread_cond_t *cond_bash_var;
int go_on;
int weight_net_print;
};
struct arg_bash *create_arg_bash();
y_network_neural_network_/src/y_net_neur_net/y_nnn_manager.c
+15 -8
View File
@@ -129,10 +129,11 @@ void* runBashPrint(void *arg){
while( check_go_on_bash(bash_arg) && (new_bash_exist(bash_arg)) && check_go_on_print_params(pprint) && !is_ending(qlStatus)){
if(/*(qlStatus->nb_episodes %125 == 0) &&*/ pprint->printed){
//pthread_mutex_lock(&(pprint->mut_printed));
pthread_mutex_lock(&(car->mut_coord));
bash_print_vehicle_n_path(car, pprint->scale_x, pprint->scale_y,bash_arg);
pthread_mutex_unlock(&(car->mut_coord));
if(bash_arg->weight_net_print == 0){
pthread_mutex_lock(&(car->mut_coord));
bash_print_vehicle_n_path(car, pprint->scale_x, pprint->scale_y,bash_arg);
pthread_mutex_unlock(&(car->mut_coord));
}
//pthread_mutex_unlock(&(pprint->mut_printed));
////printf("%s ",pprint->string_space);
len_buf=sprintf(buf,"%s ",pprint->string_space);
@@ -150,7 +151,7 @@ void* runBashPrint(void *arg){
}
////printf("\n< %5.2f > ( %s ) \n", car->direction, action_name[qlStatus->action % COUNT_ACTION]);
len_buf=sprintf(buf,"\n< %5.2f > ( %s ) \n", car->direction, action_name[qlStatus->action % COUNT_ACTION]);
len_buf=sprintf(buf,"\n< %5.2f ( %s ) \n", car->direction, action_name[qlStatus->action % COUNT_ACTION]);
BASH_WRITE_IF_EXIST(bash_arg, buf, len_buf)
//print_weight_in_neurons_TYPE_FLOAT(net_main, "net_main_wei");
//PRINT_ATTRIBUTE_TENS_IN_ALL_LAYERS(TYPE_FLOAT, net_main, weight_in, "net_main_we_in");
@@ -158,9 +159,15 @@ void* runBashPrint(void *arg){
//PRINT_ATTRIBUTE_TENS_IN_ALL_LAYERS(TYPE_FLOAT, net_target, output, "net_target_out");
//PRINT_ATTRIBUTE_TENS_IN_ALL_LAYERS(TYPE_FLOAT, net_main, input, "net_main_input");
////printf(" action : %d , factor : %f nb_episodes : %ld \n",qlStatus->action,rlAgent->qlearnParams->exploration_factor, rlAgent->status->nb_episodes);
BASH_PRINT_ATTRIBUTE_TENS_IN_ALL_LAYERS(TYPE_FLOAT, bash_arg, net_main, weight_in, "net_main_we_in");
len_buf=sprintf(buf," action : %d , factor : %f nb_episodes : %ld \n",qlStatus->action,rlAgent->qlearnParams->exploration_factor, rlAgent->status->nb_episodes);
if(bash_arg->weight_net_print){
BASH_PRINT_ATTRIBUTE_TENS_IN_ALL_LAYERS(TYPE_FLOAT, bash_arg, net_target, weight_in, "net_main_we_in", false);
}
len_buf=sprintf(buf," action : %d , learning_rate: %f, factor : %f nb_episodes : %ld \n",
qlStatus->action,
rlAgent->qlearnParams->learning_rate,
rlAgent->qlearnParams->exploration_factor,
rlAgent->status->nb_episodes);
BASH_WRITE_IF_EXIST(bash_arg, buf, len_buf)
FOR_LIST_FORM_BEGIN(TYPE_L_INT, qlStatus->progress_best_cumul){
y_network_neural_network_/src/y_net_neur_net/y_nnn_screen_manager.c
+1
View File
@@ -116,6 +116,7 @@ struct arg_bash *create_arg_bash(){
pthread_cond_init(b_arg->cond_bash_var,NULL);
b_arg->go_on=1;
b_arg->weight_net_print=0;
b_arg->thread_launch=NULL;
b_arg->thread_run_newbash=NULL;
y_network_neural_network_/test/is_good.c
+51 -48
View File
@@ -336,7 +336,7 @@ HIDE_TEST(_first_learn_vehicle_50__9){
size_t nb_prod_thread = 2;
size_t nb_calc_thread = 4;
float learning_rate = 0.001 /*0.01*/ /* 0.001*/;
struct networks_qlearning *nnetworks = create_nework_qlearning(
struct networks_qlearning *nnetworks = create_network_qlearning(
pconf,
randomize, minR, maxR, randomRange,
nb_prod_thread, nb_calc_thread,
@@ -364,12 +364,12 @@ struct status_qlearning *qlstatus = create_status_qlearning ();
20/*long int nb_training_before_update_weight_in_target*/,
10000/*size_t number_episodes*/
);
#if 0
UPDATE_ATTRIBUTE_NEURONE_IN_ALL_LAYERS(TYPE_FLOAT, nnetworks->main_net, d_f_act , df );
UPDATE_ATTRIBUTE_NEURONE_IN_ALL_LAYERS(TYPE_FLOAT, nnetworks->main_net, f_act, f );
UPDATE_ATTRIBUTE_NEURONE_IN_ALL_LAYERS(TYPE_FLOAT, nnetworks->target_net, d_f_act , df );
UPDATE_ATTRIBUTE_NEURONE_IN_ALL_LAYERS(TYPE_FLOAT, nnetworks->target_net, f_act , f );
#endif
struct print_params *pprint = create_print_params(
12/*float scale_x*/,12 /*float scale_y*/,
dly/*struct delay_params * dly_p*/
@@ -529,7 +529,7 @@ HIDE_TEST(_first_learn_vehicle_50__10){
size_t nb_prod_thread = 2;
size_t nb_calc_thread = 4;
float learning_rate = 0.00001 /* 0.001*/;
struct networks_qlearning *nnetworks = create_nework_qlearning(
struct networks_qlearning *nnetworks = create_network_qlearning(
pconf,
randomize, minR, maxR, randomRange,
nb_prod_thread, nb_calc_thread,
@@ -614,33 +614,13 @@ struct status_qlearning *qlstatus = create_status_qlearning ();
}
#endif
TEST(_first_learn_vehicle_50__11_9){
HIDE_TEST(_first_learn_vehicle_50__11_9){
size_t nb_block = 7;
size_t dim= 2;
struct blocks * path = create_blocks(nb_block, dim);
#if 1
copy_coordinate(path->lower_bound_block[0], (float[]){0,0});
copy_coordinate(path->upper_bound_block[0], (float[]){100,250});
copy_coordinate(path->lower_bound_block[1], (float[]){100,0});
copy_coordinate(path->upper_bound_block[1], (float[]){250,80});
copy_coordinate(path->lower_bound_block[2], (float[]){250,0});
copy_coordinate(path->upper_bound_block[2], (float[]){360,140});
copy_coordinate(path->lower_bound_block[3], (float[]){360,70});
copy_coordinate(path->upper_bound_block[3], (float[]){600,140});
copy_coordinate(path->lower_bound_block[4], (float[]){600,90});
copy_coordinate(path->upper_bound_block[4], (float[]){720,300});
copy_coordinate(path->lower_bound_block[5], (float[]){300,300});
copy_coordinate(path->upper_bound_block[5], (float[]){720,350});
copy_coordinate(path->lower_bound_block[6], (float[]){0,250});
copy_coordinate(path->upper_bound_block[6], (float[]){410,300});
#else
copy_coordinate(path->lower_bound_block[0], (float[]){0,0});
copy_coordinate(path->upper_bound_block[0], (float[]){150,250});
copy_coordinate(path->lower_bound_block[1], (float[]){150,0});
@@ -657,9 +637,30 @@ TEST(_first_learn_vehicle_50__11_9){
copy_coordinate(path->upper_bound_block[6], (float[]){410,300});
#else
copy_coordinate(path->lower_bound_block[0], (float[]){0,0});
copy_coordinate(path->upper_bound_block[0], (float[]){100,250});
copy_coordinate(path->lower_bound_block[1], (float[]){100,0});
copy_coordinate(path->upper_bound_block[1], (float[]){250,80});
copy_coordinate(path->lower_bound_block[2], (float[]){250,0});
copy_coordinate(path->upper_bound_block[2], (float[]){360,140});
copy_coordinate(path->lower_bound_block[3], (float[]){360,70});
copy_coordinate(path->upper_bound_block[3], (float[]){600,140});
copy_coordinate(path->lower_bound_block[4], (float[]){600,90});
copy_coordinate(path->upper_bound_block[4], (float[]){720,300});
copy_coordinate(path->lower_bound_block[5], (float[]){300,300});
copy_coordinate(path->upper_bound_block[5], (float[]){720,350});
copy_coordinate(path->lower_bound_block[6], (float[]){0,250});
copy_coordinate(path->upper_bound_block[6], (float[]){410,300});
///////////////////////////////////////
#if 0
copy_coordinate(path->lower_bound_block[0], (float[]){0,3});
copy_coordinate(path->upper_bound_block[0], (float[]){4,7});
copy_coordinate(path->lower_bound_block[1], (float[]){1,0});
@@ -675,6 +676,7 @@ TEST(_first_learn_vehicle_50__11_9){
copy_coordinate(path->lower_bound_block[6], (float[]){1,7});
copy_coordinate(path->upper_bound_block[6], (float[]){8,9.75});
#endif
#endif
@@ -683,15 +685,15 @@ TEST(_first_learn_vehicle_50__11_9){
struct vehicle *car = create_vehicle(path);
config_layers *pconf = create_config_layers_from_OneD(4,(size_t[]){3,24,24,3}); /* 3 input , 3 target; 2 hidden layer with 24 neurons each */
//config_layers *pconf = create_config_layers_from_OneD(4,(size_t[]){3,14,14,3}); /* 3 input , 3 target; 2 hidden layer with 24 neurons each */
//config_layers *pconf = create_config_layers_from_OneD(4,(size_t[]){3,4,4,3}); /* 3 input , 3 target; 2 hidden layer with 14 neurons each */
bool randomize=true;
float minR = -0.5, maxR = 0.5;
int randomRange = 500;
size_t nb_prod_thread = 2;
size_t nb_calc_thread = 4;
float learning_rate = 0.001; /* 0.00001*/ /* 0.001*/;
struct networks_qlearning *nnetworks = create_nework_qlearning(
float learning_rate = 0.01; /* 0.00001*/ /* 0.001*/;
struct networks_qlearning *nnetworks = create_network_qlearning(
pconf,
randomize, minR, maxR, randomRange,
nb_prod_thread, nb_calc_thread,
@@ -771,7 +773,7 @@ struct status_qlearning *qlstatus = create_status_qlearning ();
}
#if 1
HIDE_TEST(_first_learn_vehicle_50__11){
TEST(_first_learn_vehicle_50__11){
size_t nb_block = 7;
size_t dim= 2;
struct blocks * path = create_blocks(nb_block, dim);
@@ -873,15 +875,16 @@ copy_coordinate(path->lower_bound_block[0], (float[]){0,0});
struct vehicle *car = create_vehicle(path);
config_layers *pconf = create_config_layers_from_OneD(4,(size_t[]){3,24,24,3}); /* 3 input , 3 target; 2 hidden layer with 24 neurons each */
//config_layers *pconf = create_config_layers_from_OneD(4,(size_t[]){3,14,14,3}); /* 3 input , 3 target; 2 hidden layer with 24 neurons each */
//config_layers *pconf = create_config_layers_from_OneD(3,(size_t[]){3,24,3}); /* 3 input , 3 target; 2 hidden layer with 24 neurons each */
bool randomize=true;
float minR = -0.5, maxR = 0.5;
//float minR = 0, maxR = 1;
int randomRange = 500;
size_t nb_prod_thread = 2;
size_t nb_calc_thread = 4;
float learning_rate = 0; /* 0.000001*/ /* 0.001*/;
struct networks_qlearning *nnetworks = create_nework_qlearning(
float learning_rate = 0.0007 /*0.001*//* 0.0001*/; /* 0.000001*/ /* 0.001*/;
struct networks_qlearning *nnetworks = create_network_qlearning(
pconf,
randomize, minR, maxR, randomRange,
nb_prod_thread, nb_calc_thread,
@@ -894,8 +897,8 @@ EXTRACT_FILE_TO_TENSOR_ATTRIBUTE_NNEURONS(TYPE_FLOAT, nnetworks->target_net, wei
//EXTRACT_FILE_TO_TENSOR_ATTRIBUTE_NNEURONS(TYPE_FLOAT, nnetworks->main_net, weight_in, ".ff_main_20250508_17h50m56s_26300.txt");
//EXTRACT_FILE_TO_TENSOR_ATTRIBUTE_NNEURONS(TYPE_FLOAT, nnetworks->target_net, weight_in, ".ff_target_20250508_17h50m56s_26300.txt");
EXTRACT_FILE_TO_TENSOR_ATTRIBUTE_NNEURONS(TYPE_FLOAT, nnetworks->main_net, weight_in, ".ff_main_.symlink");
EXTRACT_FILE_TO_TENSOR_ATTRIBUTE_NNEURONS(TYPE_FLOAT, nnetworks->target_net, weight_in, ".ff_target_.symlink");
///EXTRACT_FILE_TO_TENSOR_ATTRIBUTE_NNEURONS(TYPE_FLOAT, nnetworks->main_net, weight_in, ".ff_main_.symlink");
///EXTRACT_FILE_TO_TENSOR_ATTRIBUTE_NNEURONS(TYPE_FLOAT, nnetworks->target_net, weight_in, ".ff_target_.symlink");
/*
EXTRACT_FILE_TO_TENSOR_ATTRIBUTE_NNEURONS(TYPE_FLOAT, nnetworks->main_net, weight_in, ".ff_main_20250508_23h02m40s_29000.txt");
EXTRACT_FILE_TO_TENSOR_ATTRIBUTE_NNEURONS(TYPE_FLOAT, nnetworks->target_net, weight_in, ".ff_target_20250508_23h02m40s_29000.txt");
@@ -910,7 +913,7 @@ struct status_qlearning *qlstatus = create_status_qlearning ();
0.95/*float gamma*/,
learning_rate,
0 /* (not used!)float discount_factor*/,
0.0001/*0.99*/ /*float exploration_factor*/,
1.0/*0.99*//*0.0001*//*0.99*/ /*float exploration_factor*/,
20/*long int nb_training_before_update_weight_in_target*/,
10000/*size_t number_episodes*/
);
@@ -1086,7 +1089,7 @@ copy_coordinate(path->lower_bound_block[0], (float[]){0,0});
size_t nb_prod_thread = 2;
size_t nb_calc_thread = 4;
float learning_rate = 0.0000001 /* 0.001*/;
struct networks_qlearning *nnetworks = create_nework_qlearning(
struct networks_qlearning *nnetworks = create_network_qlearning(
pconf,
randomize, minR, maxR, randomRange,
nb_prod_thread, nb_calc_thread,
@@ -1233,7 +1236,7 @@ HIDE_TEST(__first_learn_vehicle13){
size_t nb_prod_thread = 2;
size_t nb_calc_thread = 4;
float learning_rate = 0.1;
struct networks_qlearning *nnetworks = create_nework_qlearning(
struct networks_qlearning *nnetworks = create_network_qlearning(
pconf,
randomize, minR, maxR, randomRange,
nb_prod_thread, nb_calc_thread,
@@ -1371,7 +1374,7 @@ HIDE_TEST(first_learn_vehicle){
size_t nb_prod_thread = 2;
size_t nb_calc_thread = 4;
float learning_rate = 0.1;
struct networks_qlearning *nnetworks = create_nework_qlearning(
struct networks_qlearning *nnetworks = create_network_qlearning(
pconf,
randomize, minR, maxR, randomRange,
nb_prod_thread, nb_calc_thread,
@@ -1482,7 +1485,7 @@ HIDE_TEST(first_learn_vehicle){
size_t nb_prod_thread = 2;
size_t nb_calc_thread = 4;
float learning_rate = 0.1;
struct networks_qlearning *nnetworks = create_nework_qlearning(
struct networks_qlearning *nnetworks = create_network_qlearning(
pconf,
randomize, minR, maxR, randomRange,
nb_prod_thread, nb_calc_thread,
@@ -2006,7 +2009,7 @@ HIDE_TEST(first_learn_vehicle_rev50_8){
size_t nb_prod_thread = 2;
size_t nb_calc_thread = 4;
float learning_rate = 0.001; // 0.00001 /*0.001*/;
struct networks_qlearning *nnetworks = create_nework_qlearning(
struct networks_qlearning *nnetworks = create_network_qlearning(
pconf,
randomize, minR, maxR, randomRange,
nb_prod_thread, nb_calc_thread,
@@ -2170,7 +2173,7 @@ HIDE_TEST(first_learn_vehicle_50__9){
size_t nb_prod_thread = 2;
size_t nb_calc_thread = 4;
float learning_rate = 0.00001 /* 0.001*/;
struct networks_qlearning *nnetworks = create_nework_qlearning(
struct networks_qlearning *nnetworks = create_network_qlearning(
pconf,
randomize, minR, maxR, randomRange,
nb_prod_thread, nb_calc_thread,
@@ -2340,7 +2343,7 @@ HIDE_TEST(first_learn_vehicle_50__10){
size_t nb_prod_thread = 2;
size_t nb_calc_thread = 4;
float learning_rate = 0.00001 /* 0.001*/;
struct networks_qlearning *nnetworks = create_nework_qlearning(
struct networks_qlearning *nnetworks = create_network_qlearning(
pconf,
randomize, minR, maxR, randomRange,
nb_prod_thread, nb_calc_thread,
@@ -2512,7 +2515,7 @@ copy_coordinate(path->lower_bound_block[0], (float[]){0,0});
size_t nb_prod_thread = 2;
size_t nb_calc_thread = 4;
float learning_rate = 0; /* 0.000001*/ /* 0.001*/;
struct networks_qlearning *nnetworks = create_nework_qlearning(
struct networks_qlearning *nnetworks = create_network_qlearning(
pconf,
randomize, minR, maxR, randomRange,
nb_prod_thread, nb_calc_thread,
@@ -2691,7 +2694,7 @@ copy_coordinate(path->lower_bound_block[0], (float[]){0,0});
size_t nb_prod_thread = 2;
size_t nb_calc_thread = 4;
float learning_rate = 0.0000001 /* 0.001*/;
struct networks_qlearning *nnetworks = create_nework_qlearning(
struct networks_qlearning *nnetworks = create_network_qlearning(
pconf,
randomize, minR, maxR, randomRange,
nb_prod_thread, nb_calc_thread,
@@ -2813,7 +2816,7 @@ HIDE_TEST(_first_learn_vehicle13){
size_t nb_prod_thread = 2;
size_t nb_calc_thread = 4;
float learning_rate = 0.1;
struct networks_qlearning *nnetworks = create_nework_qlearning(
struct networks_qlearning *nnetworks = create_network_qlearning(
pconf,
randomize, minR, maxR, randomRange,
nb_prod_thread, nb_calc_thread,
@@ -2951,7 +2954,7 @@ HIDE_TEST(first_learn_vehicle){
size_t nb_prod_thread = 2;
size_t nb_calc_thread = 4;
float learning_rate = 0.1;
struct networks_qlearning *nnetworks = create_nework_qlearning(
struct networks_qlearning *nnetworks = create_network_qlearning(
pconf,
randomize, minR, maxR, randomRange,
nb_prod_thread, nb_calc_thread,
@@ -3062,7 +3065,7 @@ HIDE_TEST(first_learn_vehicle){
size_t nb_prod_thread = 2;
size_t nb_calc_thread = 4;
float learning_rate = 0.1;
struct networks_qlearning *nnetworks = create_nework_qlearning(
struct networks_qlearning *nnetworks = create_network_qlearning(
pconf,
randomize, minR, maxR, randomRange,
nb_prod_thread, nb_calc_thread,