mirror/darknet
1
0
Fork 0
mirror of https://github.com/pjreddie/darknet.git synced 2023-08-10 21:13:14 +03:00
Code Issues Projects Releases Wiki Activity

Minor Python and C API improvement

Browse Source
This commit is contained in:
AlexeyAB
2019-02-06 14:38:12 +03:00
parent b76f1c0006
commit c50b0e0c8a
9 changed files with 15 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
build/darknet/x64/darknet.py
View File

@ -125,7 +125,7 @@ lib.network_width.restype = c_int
lib.network_height.argtypes = [c_void_p] lib.network_height.argtypes = [c_void_p]
lib.network_height.restype = c_int lib.network_height.restype = c_int
predict = lib.network_predict predict = lib.network_predict_ptr
predict.argtypes = [c_void_p, POINTER(c_float)] predict.argtypes = [c_void_p, POINTER(c_float)]
predict.restype = POINTER(c_float) predict.restype = POINTER(c_float)
@ -151,7 +151,7 @@ free_detections.argtypes = [POINTER(DETECTION), c_int]
free_ptrs = lib.free_ptrs free_ptrs = lib.free_ptrs
free_ptrs.argtypes = [POINTER(c_void_p), c_int] free_ptrs.argtypes = [POINTER(c_void_p), c_int]
network_predict = lib.network_predict network_predict = lib.network_predict_ptr
network_predict.argtypes = [c_void_p, POINTER(c_float)] network_predict.argtypes = [c_void_p, POINTER(c_float)]
reset_rnn = lib.reset_rnn reset_rnn = lib.reset_rnn

4
build/darknet/x64/darknet_video.py
View File

@ -60,7 +60,7 @@ lib.network_width.restype = c_int
lib.network_height.argtypes = [c_void_p] lib.network_height.argtypes = [c_void_p]
lib.network_height.restype = c_int lib.network_height.restype = c_int
predict = lib.network_predict predict = lib.network_predict_ptr
predict.argtypes = [c_void_p, POINTER(c_float)] predict.argtypes = [c_void_p, POINTER(c_float)]
predict.restype = POINTER(c_float) predict.restype = POINTER(c_float)
@ -88,7 +88,7 @@ free_detections.argtypes = [POINTER(DETECTION), c_int]
free_ptrs = lib.free_ptrs free_ptrs = lib.free_ptrs
free_ptrs.argtypes = [POINTER(c_void_p), c_int] free_ptrs.argtypes = [POINTER(c_void_p), c_int]
network_predict = lib.network_predict network_predict = lib.network_predict_ptr
network_predict.argtypes = [c_void_p, POINTER(c_float)] network_predict.argtypes = [c_void_p, POINTER(c_float)]
reset_rnn = lib.reset_rnn reset_rnn = lib.reset_rnn

2
darknet.py
View File

@ -151,7 +151,7 @@ free_detections.argtypes = [POINTER(DETECTION), c_int]
free_ptrs = lib.free_ptrs free_ptrs = lib.free_ptrs
free_ptrs.argtypes = [POINTER(c_void_p), c_int] free_ptrs.argtypes = [POINTER(c_void_p), c_int]
network_predict = lib.network_predict network_predict = lib.network_predict_ptr
network_predict.argtypes = [c_void_p, POINTER(c_float)] network_predict.argtypes = [c_void_p, POINTER(c_float)]
reset_rnn = lib.reset_rnn reset_rnn = lib.reset_rnn

4
darknet_video.py
View File

@ -60,7 +60,7 @@ lib.network_width.restype = c_int
lib.network_height.argtypes = [c_void_p] lib.network_height.argtypes = [c_void_p]
lib.network_height.restype = c_int lib.network_height.restype = c_int
predict = lib.network_predict predict = lib.network_predict_ptr
predict.argtypes = [c_void_p, POINTER(c_float)] predict.argtypes = [c_void_p, POINTER(c_float)]
predict.restype = POINTER(c_float) predict.restype = POINTER(c_float)
@ -88,7 +88,7 @@ free_detections.argtypes = [POINTER(DETECTION), c_int]
free_ptrs = lib.free_ptrs free_ptrs = lib.free_ptrs
free_ptrs.argtypes = [POINTER(c_void_p), c_int] free_ptrs.argtypes = [POINTER(c_void_p), c_int]
network_predict = lib.network_predict network_predict = lib.network_predict_ptr
network_predict.argtypes = [c_void_p, POINTER(c_float)] network_predict.argtypes = [c_void_p, POINTER(c_float)]
reset_rnn = lib.reset_rnn reset_rnn = lib.reset_rnn

2
include/darknet.h
View File

@ -745,6 +745,7 @@ LIB_API void do_nms_obj(detection *dets, int total, int classes, float thresh);
// network.h // network.h
LIB_API float *network_predict(network net, float *input); LIB_API float *network_predict(network net, float *input);
LIB_API float *network_predict_ptr(network *net, float *input);
LIB_API detection *get_network_boxes(network *net, int w, int h, float thresh, float hier, int *map, int relative, int *num, int letter); LIB_API detection *get_network_boxes(network *net, int w, int h, float thresh, float hier, int *map, int relative, int *num, int letter);
LIB_API void free_detections(detection *dets, int n); LIB_API void free_detections(detection *dets, int n);
LIB_API void fuse_conv_batchnorm(network net); LIB_API void fuse_conv_batchnorm(network net);
@ -766,6 +767,7 @@ LIB_API void optimize_picture(network *net, image orig, int max_layer, float sca
// image.h // image.h
LIB_API image resize_image(image im, int w, int h); LIB_API image resize_image(image im, int w, int h);
LIB_API void copy_image_from_bytes(image im, char *pdata);
LIB_API image letterbox_image(image im, int w, int h); LIB_API image letterbox_image(image im, int w, int h);
LIB_API void rgbgr_image(image im); LIB_API void rgbgr_image(image im);
LIB_API image make_image(int w, int h, int c); LIB_API image make_image(int w, int h, int c);

5
src/detector.c
View File

@ -247,6 +247,9 @@ void train_detector(char *datacfg, char *cfgfile, char *weightfile, int *gpus, i
if (mean_average_precision > 0) { if (mean_average_precision > 0) {
printf("\n Last accuracy mAP@0.5 = %2.2f %% ", mean_average_precision*100); printf("\n Last accuracy mAP@0.5 = %2.2f %% ", mean_average_precision*100);
} }
int calc_map_for_each = 4 * train_images_num / (net.batch * net.subdivisions); // calculate mAP for each 4 Epochs
if (calc_map) printf("\n (next mAP calculation at %d iterations) ", (iter_map + calc_map_for_each));
if (net.cudnn_half) { if (net.cudnn_half) {
if (i < net.burn_in * 3) printf("\n Tensor Cores are disabled until the first %d iterations are reached.", 3 * net.burn_in); if (i < net.burn_in * 3) printf("\n Tensor Cores are disabled until the first %d iterations are reached.", 3 * net.burn_in);
else printf("\n Tensor Cores are used."); else printf("\n Tensor Cores are used.");
@ -254,8 +257,6 @@ void train_detector(char *datacfg, char *cfgfile, char *weightfile, int *gpus, i
printf("\n %d: %f, %f avg loss, %f rate, %lf seconds, %d images\n", get_current_batch(net), loss, avg_loss, get_current_rate(net), (what_time_is_it_now() - time), i*imgs); printf("\n %d: %f, %f avg loss, %f rate, %lf seconds, %d images\n", get_current_batch(net), loss, avg_loss, get_current_rate(net), (what_time_is_it_now() - time), i*imgs);
int draw_precision = 0; int draw_precision = 0;
int calc_map_for_each = 4 * train_images_num / (net.batch * net.subdivisions); // calculate mAP for each 4 Epochs
if (calc_map) printf(" Next mAP calculation at %d iterations \n", (iter_map + calc_map_for_each));
if (calc_map && (i >= (iter_map + calc_map_for_each) || i == net.max_batches) && i >= net.burn_in && i >= 1000) { if (calc_map && (i >= (iter_map + calc_map_for_each) || i == net.max_batches) && i >= net.burn_in && i >= 1000) {
if (l.random) { if (l.random) {
printf("Resizing to initial size: %d x %d \n", init_w, init_h); printf("Resizing to initial size: %d x %d \n", init_w, init_h);

2
src/image.h
View File

@ -33,7 +33,7 @@ image random_crop_image(image im, int w, int h);
image random_augment_image(image im, float angle, float aspect, int low, int high, int size); image random_augment_image(image im, float angle, float aspect, int low, int high, int size);
void random_distort_image(image im, float hue, float saturation, float exposure); void random_distort_image(image im, float hue, float saturation, float exposure);
//LIB_API image resize_image(image im, int w, int h); //LIB_API image resize_image(image im, int w, int h);
LIB_API void copy_image_from_bytes(image im, char *pdata); //LIB_API void copy_image_from_bytes(image im, char *pdata);
void fill_image(image m, float s); void fill_image(image m, float s);
void letterbox_image_into(image im, int w, int h, image boxed); void letterbox_image_into(image im, int w, int h, image boxed);
//LIB_API image letterbox_image(image im, int w, int h); //LIB_API image letterbox_image(image im, int w, int h);

1
src/network.c
View File

@ -738,6 +738,7 @@ float *network_predict_image(network *net, image im)
{ {
//image imr = letterbox_image(im, net->w, net->h); //image imr = letterbox_image(im, net->w, net->h);
float *p; float *p;
if(net->batch != 1) set_batch_network(net, 1);
if (im.w == net->w && im.h == net->h) { if (im.w == net->w && im.h == net->h) {
// Input image is the same size as our net, predict on that image // Input image is the same size as our net, predict on that image
p = network_predict(*net, im.data); p = network_predict(*net, im.data);

2
src/network.h
View File

@ -122,7 +122,7 @@ float train_network_datum(network net, float *x, float *y);
matrix network_predict_data(network net, data test); matrix network_predict_data(network net, data test);
//LIB_API float *network_predict(network net, float *input); //LIB_API float *network_predict(network net, float *input);
LIB_API float *network_predict_ptr(network *net, float *input); //LIB_API float *network_predict_ptr(network *net, float *input);
float network_accuracy(network net, data d); float network_accuracy(network net, data d);
float *network_accuracies(network net, data d, int n); float *network_accuracies(network net, data d, int n);
float network_accuracy_multi(network net, data d, int n); float network_accuracy_multi(network net, data d, int n);