obj atr eval script now writes a json of obj predictions

object_attribute_classifier_cached_features/eval.py

@@ -76,18 +76,21 @@ def create_batch_generator(region_ids_json):
 class eval_mgr():
     def __init__(
             self, 
-            scores_dirname, 
+            attribute_scores_dirname, 
+            object_scores_dirname,
             inv_object_labels_dict,
             vis_dirname,
             genome_region_dir):
 
-        self.scores_dirname = scores_dirname
+        self.attribute_scores_dirname = attribute_scores_dirname
+        self.object_scores_dirname = object_scores_dirname
         self.inv_object_labels_dict = inv_object_labels_dict
         self.vis_dirname = vis_dirname
         self.genome_region_dir = genome_region_dir
         self.epsilon = 0.00001
         self.num_iter = 0.0
         self.object_accuracy = 0.0
+        self.obj_pred_json = dict()
         self.precision = np.zeros([11], np.float32)
         self.recall = np.zeros([11], np.float32)
         self.fall_out = np.zeros([11], np.float32)
@@ -133,20 +136,33 @@ class eval_mgr():
             eval_vars_dict['attribute_prob'],
             labels['attributes'])
 
-        pred_obj_labels = self.get_top_k_labels(eval_vars_dict['object_prob'],5)[0:10]
-        gt_obj_labels = self.get_gt_labels(labels['objects'])[0:10]
+        pred_obj_labels = self.get_top_k_labels(eval_vars_dict['object_prob'],5)
+        gt_obj_labels = self.get_gt_labels(labels['objects'])
         region_paths = self.get_region_paths(image_ids, region_ids)
 
+        for i, region_id in enumerate(region_ids):
+            self.obj_pred_json[region_id] = {
+                'gt': gt_obj_labels[i],
+                'pred': pred_obj_labels[i],
+            }
+
         if constants.visualize_object_predictions:
             self.save_image_pred(
-                pred_obj_labels,
-                gt_obj_labels,
+                pred_obj_labels[0:10],
+                gt_obj_labels[0:10],
                 region_ids,
                 region_paths)
 
         if iter%500 == 0:
             self.write_scores()
 
+            filename = os.path.join(
+                self.object_scores_dirname,
+                'object_predictions.json')
+
+            with open(filename,'w') as file:
+                ujson.dump(self.obj_pred_json,file,indent=4)
+
     def save_image_pred(
             self,
             pred_obj_labels,
@@ -217,13 +233,13 @@ class eval_mgr():
     def write_scores(self):
         for i in xrange(10):
             filename = os.path.join(
-                self.scores_dirname,
+                self.attribute_scores_dirname,
                 'scores_' + str(i) + '.json')
             with open(filename, 'w') as file:
                 ujson.dump(self.scores_dict[i], file, indent=4)
 
             filename = os.path.join(
-                self.scores_dirname,
+                self.attribute_scores_dirname,
                 'labels_' + str(i) + '.json')
             with open(filename, 'w') as file:
                 ujson.dump(self.labels_dict[i], file, indent=4)
@@ -434,6 +450,7 @@ if __name__=='__main__':
     print 'Creating evaluator...'
     evaluator = eval_mgr(
         constants.region_attribute_scores_dirname,
+        constants.region_object_scores_dirname,
         inv_object_labels_dict,
         constants.region_pred_vis_dirname,
         constants.image_dir)