Проверил метки на символы не из словаря.

.gitignore

@@ -33,3 +33,4 @@ paddleocr.egg-info/
 /deploy/android_demo/app/cache/
 test_tipc/web/models/
 test_tipc/web/node_modules/
+paddle/

scripts/dataset_validation.py

@@ -0,0 +1,44 @@
+import os
+import cv2
+
+def rename_files(directory_path: str):
+    extensions = set()
+
+    for file in os.listdir(directory_path):
+        filename, extension = file.split(".")
+        extensions.add(extension)
+        old_filename = os.path.join(directory_path, file)
+        new_filename = os.path.join(directory_path, f"{filename}.jpg")
+        if old_filename != new_filename:
+            print(f"{old_filename} -> {new_filename}")
+            os.rename(old_filename, new_filename)
+
+def check_images(directory_path: str):
+    for file in os.listdir(directory_path):
+        try:
+            img = cv2.imread(os.path.join(directory_path, file))
+            if img is None:
+                print(file)
+        except Exception:
+            print(file)
+
+def check_labels(dir_path: str):
+    for filename in os.listdir(dir_path):
+        name, extension = filename.split(".")
+        if len(name) != 17:
+            print(filename)
+
+def check_symbols(dir_path: str):
+    with open(os.path.join(dir_path, "dict.txt"), "r") as dict_file:
+        dict_content = dict_file.readlines()
+        dict_chars = set(char.strip() for char in dict_content)
+        for filename in os.listdir(os.path.join(dir_path, "images")):
+            label, extension = filename.split(".")
+            if any([char not in dict_chars for char in label]):
+                print(filename)
+
+# rename_files("train_data/images/")
+# check_images("train_data/images/")
+# check_labels("train_data/images/")
+check_symbols("train_data/")
+