Update csv2arrow.py

Tencent · Jun 14, 2024 · 6602373 · 6602373
1 parent 28ff0e2
commit 6602373
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/hydit/data_loader/csv2arrow.py b/hydit/data_loader/csv2arrow.py
@@ -19,7 +19,7 @@ def parse_data(data):
 
         with open(img_path, "rb") as fp:
             image = fp.read()
-            md5 = hashlib.md5(fp.read()).hexdigest()
+            md5 = hashlib.md5(image).hexdigest()
 
         with Image.open(img_path) as f:
             width, height = f.size
@@ -85,4 +85,4 @@ def make_arrow(csv_root, dataset_root, start_id=0, end_id=-1):
     csv_root = sys.argv[1]
     output_arrow_data_path = sys.argv[2]
     pool = Pool(500)
-    make_arrow(csv_root, output_arrow_data_path)
+    make_arrow(csv_root, output_arrow_data_path)