perf #70: introducing changes proposed by Lucashsmello

- See this [PR](adambielski/siamese-triplet#54)
  where the original author of the PR proposes a faster implementation
  for the triplet selector

src/adambielski_lib/utils.py

@@ -162,19 +162,18 @@ def get_triplets(self, embeddings, labels):
             anchor_positives = list(combinations(label_indices, 2))  # All anchor-positive pairs
             anchor_positives = np.array(anchor_positives)
 
-            ap_distances = distance_matrix[anchor_positives[:, 0], anchor_positives[:, 1]]
-            for anchor_positive, ap_distance in zip(anchor_positives, ap_distances):
-                loss_values = ap_distance - distance_matrix[torch.LongTensor(np.array([anchor_positive[0]])), torch.LongTensor(negative_indices)] + self.margin
-                loss_values = loss_values.data.cpu().numpy()
-                hard_negative = self.negative_selection_fn(loss_values)
+            ap_distances = distance_matrix[anchor_positives[:, 0], anchor_positives[:, 1]] + self.margin
+            idxs = np.ix_(anchor_positives[:, 0], negative_indices)
+            loss_values = ap_distances.unsqueeze(dim=1) - distance_matrix[idxs]
+            loss_values = loss_values.data.cpu().numpy()
+            for i, loss_val in enumerate(loss_values):
+                hard_negative = self.negative_selection_fn(loss_val)
                 if hard_negative is not None:
                     hard_negative = negative_indices[hard_negative]
-                    triplets.append([anchor_positive[0], anchor_positive[1], hard_negative])
+                    triplets.append([anchor_positives[i][0], anchor_positives[i][1], hard_negative])
 
         if len(triplets) == 0:
-            triplets.append([anchor_positive[0], anchor_positive[1], negative_indices[0]])
-
-        triplets = np.array(triplets)
+            triplets.append([anchor_positives[-1][0], anchor_positives[-1][1], negative_indices[0]])
 
         return torch.LongTensor(triplets)