update

README.md

@@ -53,7 +53,7 @@ python face-cluster-by-infomap
 | GCN-V + GCN-E | 92.56 | 83.74 | 87.93 |
 | Infomap(ours)(k=50,min_sim=0.58) | 95.50 | 92.51 | 93.98 |
 
-![avatar](./image/evaluate.png)
+![](image/evaluate.png)
 
 ## Results on YouTube-Faces
 

data/README.md

@@ -16,6 +16,8 @@ The data directory is constucted as follows:
 - `features` currently supports binary file.
 - `labels` supports plain text where each line indicates a label corresponding to the feature file.
 
+```
+
 ### Feature Extraction
 To experiment with your own face pictures, it is required to extracted face features from the pictures.
 

face-cluster-by-infomap.py

@@ -231,11 +231,12 @@ def cluster_by_infomap(nbrs, dists, pred_label_path, save_result=False):
             for idx in range(idx_len):
                 of.write(str(idx2label[idx]) + '\n')
 
-    pred_labels = intdict2ndarray(idx2label)
-    true_lb2idxs, true_idx2lb = read_meta(label_path)
-    gt_labels = intdict2ndarray(true_idx2lb)
-    for metric in metrics:
-        evaluate(gt_labels, pred_labels, metric)
+    if label_path is not None:
+        pred_labels = intdict2ndarray(idx2label)
+        true_lb2idxs, true_idx2lb = read_meta(label_path)
+        gt_labels = intdict2ndarray(true_idx2lb)
+        for metric in metrics:
+            evaluate(gt_labels, pred_labels, metric)
 
 
 def get_dist_nbr(feature_path, k=80, knn_method='faiss-cpu'):