def test_yolov2_tiny_2(self):
     from dlpy.applications import Tiny_YoloV2
     anchors=[]
     anchors.append(1)
     anchors.append(1)
     anchors.append(1)
     anchors.append(1)
     model = Tiny_YoloV2(self.s, anchors, predictions_per_grid=2, max_label_per_image=3, max_boxes=1)
     model.print_summary()
Exemple #2
0
 def test_yolov2_tiny(self):
     from dlpy.applications import Tiny_YoloV2
     anchors = []
     anchors.append(1)
     anchors.append(1)
     anchors.append(1)
     anchors.append(1)
     model = Tiny_YoloV2(self.s, anchors, predictions_per_grid=2)
     model.print_summary()
Exemple #3
0
    def test_evaluate_obj_det(self):

        if self.data_dir is None:
            unittest.TestCase.skipTest(
                self, "DLPY_DATA_DIR is not set in the environment variables")

        caslib, path = caslibify(self.s,
                                 path=self.data_dir +
                                 'evaluate_obj_det_det.sashdat',
                                 task='load')

        self.s.table.loadtable(caslib=caslib,
                               casout={
                                   'name': 'evaluate_obj_det_det',
                                   'replace': True
                               },
                               path=path)

        self.s.table.loadtable(caslib=caslib,
                               casout={
                                   'name': 'evaluate_obj_det_gt',
                                   'replace': True
                               },
                               path='evaluate_obj_det_gt.sashdat')
        yolo_anchors = (5.9838598901098905, 3.4326923076923075,
                        2.184993862520458, 1.9841448445171848,
                        1.0261752136752136, 1.2277777777777779)
        yolo_model = Tiny_YoloV2(self.s,
                                 grid_number=17,
                                 scale=1.0 / 255,
                                 n_classes=1,
                                 height=544,
                                 width=544,
                                 predictions_per_grid=3,
                                 anchors=yolo_anchors,
                                 max_boxes=100,
                                 coord_type='yolo',
                                 max_label_per_image=100,
                                 class_scale=1.0,
                                 coord_scale=2.0,
                                 prediction_not_a_object_scale=1,
                                 object_scale=5,
                                 detection_threshold=0.05,
                                 iou_threshold=0.2)

        metrics = yolo_model.evaluate_object_detection(
            ground_truth='evaluate_obj_det_gt',
            coord_type='yolo',
            detection_data='evaluate_obj_det_det',
            iou_thresholds=0.5)