Fix loading strict dataset

BabyLM.py

@@ -234,39 +234,9 @@ class BabyLM(datasets.GeneratorBasedBuilder):
         Returns data for different splits 
         """
 
-
-        if "fr_lang_small" in self.config.name:
-            train_data_dir = "FR"
-        elif "de_lang_small" in self.config.name:
-            train_data_dir = "DE"
-        elif "zh_lang_small" in self.config.name:
-            train_data_dir = "ZH"
-        elif "ja_lang_small" in self.config.name:
-            train_data_dir = "JA"
-        elif "es_lang_small" in self.config.name:
-            train_data_dir = "SP"
-        elif "nld_lang_small" in self.config.name:
-            train_data_dir = "DUT"
-        elif "po_lang_small" in self.config.name:
-            train_data_dir = "PORT"
-        elif "ron_lang_small" in self.config.name:
-            train_data_dir = "ROM"
-        elif "cat_lang_small" in self.config.name:
-            train_data_dir = "CAT"
-        elif "sem_strict_small" in self.config.name:
-            train_data_dir = "SEM"
-        elif "sem_zh_small" in self.config.name:
-            train_data_dir = "SEM-ZH"
-        elif "strict_small" in self.config.name: #default settings – English
-            train_data_dir = "10M"
-        else: 
-            train_data_dir = "100M"
-
-
         folder = 'original_tagged' if 'original' in self.config.name else 'clean_tagged' #
         folder = folder + '_gold' if 'gold' in self.config.name else folder #gold tags for french, german, japanese and english
 
-
         #modified urls to download
 
         urls_to_download = {
@@ -319,10 +289,14 @@ class BabyLM(datasets.GeneratorBasedBuilder):
             urls_to_download["train"] = [f"{folder}/SEM/{fn}" for fn in sem_en_filenames]
             urls_to_download["dev"] = [f"{folder}/dev/{fn}" for fn in en_filenames]
             urls_to_download["test"] = [f"{folder}/test/{fn}" for fn in en_filenames]
-        else:
+        elif 'strict_small' in self.config.name:
             urls_to_download["train"] = [f"{folder}/10M/{fn}" for fn in en_filenames]
             urls_to_download["dev"] = [f"{folder}/dev/{fn}" for fn in en_filenames]
             urls_to_download["test"] = [f"{folder}/test/{fn}" for fn in en_filenames]
+        else:
+            urls_to_download["train"] = [f"{folder}/100M/{fn}" for fn in en_filenames]
+            urls_to_download["dev"] = [f"{folder}/dev/{fn}" for fn in en_filenames]
+            urls_to_download["test"] = [f"{folder}/test/{fn}" for fn in en_filenames]
 
         
         downloaded_files = dl_manager.download_and_extract(urls_to_download)