Update new_dataset_script.py

new_dataset_script.py

@@ -133,15 +133,16 @@ class NewDataset(datasets.GeneratorBasedBuilder):
         # Shuffle the list for random split
         random.seed(42)  # Set a random seed for reproducibility
         random.shuffle(all_image_filenames)
+
+        num_files = len(all_image_filenames)
+        train_split_end = int(num_files * 0.7)
+        val_split_end = train_split_end + int(num_files * 0.15)
+        
+        train_files = all_image_filenames[:train_split_end]
+        val_files = all_image_filenames[train_split_end:val_split_end]
+        test_files = all_image_filenames[val_split_end:]
     
-        if self.config.name == "full":
-                # Use all the data for the full configuration
-                train_files, val_files, test_files = self._split_files(all_image_filenames)
-        elif self.config.name == "small":
-                # Use only a subset of the data for the small configuration
-                small_subset = all_image_filenames[:1000]  # For example, use only the first 100 images
-                train_files, val_files, test_files = self._split_files(small_subset)
-            
+