Update to run hailo demos

Author: jrrodri

abraia/__init__.py

@@ -2,7 +2,7 @@
 from dotenv import load_dotenv
 load_dotenv()
 
-__version__ = '0.26.0'
+__version__ = '0.26.1'
 
 from . import config
 from .client import Abraia, APIError

abraia/demo.py

@@ -15,6 +15,7 @@
 
 DEMOS = {
     'apple': {
+        'model': 'multiple/models/yolov8n-seg.onnx',
         'src': '5479199-hd_1920_1080_25fps.mp4',
         'labels': ['apple']
     },

abraia/hailo/toolbox.py

@@ -1321,44 +1321,32 @@ def _get_results(self, bindings):
         #     mh, mw, _ = self.get_input_shape()
         #     detections = []
         #     for det in infer_results:
-        #         print(det)
         #         if det.score < self.score_threshold: break
         #         xmin, ymin, xmax, ymax = map_box_to_orig([det.x_min * mw, det.y_min * mh, det.x_max * mw, det.y_max * mh], (oh, ow), (mh, mw))
         #         detection = {'label': self.labels[det.class_id] if self.labels else str(det.class_id), 'score': float(det.score), 'box': [xmin, ymin, xmax - xmin, ymax - ymin], 'class_id': det.class_id}
         #         if self.task == 'segment':
         #             mask = resize_mask_to_unpadded_box(det.mask, [xmin, ymin, xmax, ymax], [det.x_min * mw, det.y_min * mh, det.x_max * mw, det.y_max * mh])
         #             if mask is not None: detection['mask'] = mask
         #         detections.append(detection)
-        #     print(detections)
         #     return detections
 
         def _process_nms_results(self, result, image):
             infer_results = result if isinstance(result, list) else [result]
             img_height, img_width = image.shape[:2]
             size = max(img_height, img_width)
             padding_length = int(abs(img_height - img_width) / 2)
-
             detections = []
             for det in infer_results:
-                if det.score < self.score_threshold:
-                    break
-
+                if det.score < self.score_threshold: break
                 box_on_input_image, box_on_padded_image = convert_box_from_normalized(
                     [det.x_min, det.y_min, det.x_max, det.y_max], size, padding_length, img_height, img_width)
-                
                 xmin, ymin, xmax, ymax = box_on_input_image
-                detection = {
-                    'label': self.labels[det.class_id] if self.labels else str(det.class_id),
-                    'score': float(det.score),
-                    'box': [xmin, ymin, xmax - xmin, ymax - ymin],
-                    'class_id': det.class_id
-                }
-
+                detection = {'label': self.labels[det.class_id] if self.labels else str(det.class_id),
+                    'score': float(det.score), 'box': [xmin, ymin, xmax - xmin, ymax - ymin], 'class_id': det.class_id}
                 if self.task == 'segment':
                     mask = resize_mask_to_unpadded_box(det.mask, box_on_input_image, box_on_padded_image)
                     if mask is not None:
                         detection['mask'] = mask
-                
                 detections.append(detection)
             return detections
 

abraia/inference/ops.py

@@ -237,12 +237,11 @@ def softmax(x, axis=-1):
 
 
 def count_objects(results):
-    counts, colors = {}, {}
+    counts = {}
     for result in results:
-        label, color = result['label'], result['color']
+        label= result['label']
         counts[label] = counts.get(label, 0) + 1
-        colors[label] = color
-    objects = [{'label': label, 'count': counts[label], 'color': colors[label]} for label in counts.keys()]
+    objects = [{'label': label, 'count': counts[label]} for label in counts.keys()]
     return objects