diff --git a/tensorflow_datasets/datasets/patch_camelyon/patch_camelyon_dataset_builder.py b/tensorflow_datasets/datasets/patch_camelyon/patch_camelyon_dataset_builder.py
index 67ddc43f2b0..9785722db0b 100644
--- a/tensorflow_datasets/datasets/patch_camelyon/patch_camelyon_dataset_builder.py
+++ b/tensorflow_datasets/datasets/patch_camelyon/patch_camelyon_dataset_builder.py
@@ -54,13 +54,13 @@ def _split_generators(self, dl_manager):
     paths = dl_manager.download_and_extract(resources)
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST, gen_kwargs=dict(split='test', paths=paths)
+            name=tfds.Split.TEST, gen_kwargs=dict(split='test', paths=paths)  # pyrefly: ignore[missing-attribute]
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN, gen_kwargs=dict(split='train', paths=paths)
+            name=tfds.Split.TRAIN, gen_kwargs=dict(split='train', paths=paths)  # pyrefly: ignore[missing-attribute]
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.VALIDATION,
+            name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             gen_kwargs=dict(split='valid', paths=paths),
         ),
     ]
diff --git a/tensorflow_datasets/datasets/paws_wiki/paws_wiki_dataset_builder.py b/tensorflow_datasets/datasets/paws_wiki/paws_wiki_dataset_builder.py
index 10c1958f20f..6d34d6e7fae 100644
--- a/tensorflow_datasets/datasets/paws_wiki/paws_wiki_dataset_builder.py
+++ b/tensorflow_datasets/datasets/paws_wiki/paws_wiki_dataset_builder.py
@@ -44,9 +44,9 @@
     _UNLABELED_FINAL: ["train", "dev"],
 }
 _SPLIT_MAPPINGS = {
-    "train": tfds.Split.TRAIN,
-    "dev": tfds.Split.VALIDATION,
-    "test": tfds.Split.TEST,
+    "train": tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
+    "dev": tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
+    "test": tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
 }
 
 _CLASS_LABELS = ["different_meaning", "paraphrase"]
@@ -101,10 +101,10 @@ def _info(self):
   def _split_generators(self, dl_manager):
     """Returns SplitGenerators."""
     dl_paths = dl_manager.download_and_extract(_DOWNLOAD_URLS)
-    subset = self.builder_config.subset
+    subset = self.builder_config.subset  # pyrefly: ignore[missing-attribute]
     labels_path = os.path.join(dl_paths[subset], _EXTRACTED_FOLDERS[subset])
 
-    if self.builder_config.tokenized:
+    if self.builder_config.tokenized:  # pyrefly: ignore[missing-attribute]
       mappings_path = ""
       tags2texts = {}
     else:
@@ -193,7 +193,7 @@ def _generate_examples(
           key = row["id"]
           label_str = (
               "noisy_label"
-              if self.builder_config.subset == _UNLABELED_FINAL
+              if self.builder_config.subset == _UNLABELED_FINAL  # pyrefly: ignore[missing-attribute]
               else "label"
           )
           example = {
diff --git a/tensorflow_datasets/datasets/paws_x_wiki/paws_x_wiki_dataset_builder.py b/tensorflow_datasets/datasets/paws_x_wiki/paws_x_wiki_dataset_builder.py
index 7ed5aff4a1f..41fb2f5fac6 100644
--- a/tensorflow_datasets/datasets/paws_x_wiki/paws_x_wiki_dataset_builder.py
+++ b/tensorflow_datasets/datasets/paws_x_wiki/paws_x_wiki_dataset_builder.py
@@ -75,32 +75,32 @@ def _split_generators(self, dl_manager):
     base_path = os.path.join(dl_path["x-final"], "x-final")
 
     # Name of file for training for 'en' is different from other languages
-    if self.builder_config.language == "en":
+    if self.builder_config.language == "en":  # pyrefly: ignore[missing-attribute]
       training_path = os.path.join(
           base_path, self.builder_config.language, "train.tsv"
       )
     else:
-      training_path = os.path.join(
+      training_path = os.path.join(  # pyrefly: ignore[no-matching-overload]
           base_path, self.builder_config.language, "translated_train.tsv"
       )
 
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"path": training_path},
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
-                "path": os.path.join(
+                "path": os.path.join(  # pyrefly: ignore[no-matching-overload]
                     base_path, self.builder_config.language, "test_2k.tsv"
                 )
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.VALIDATION,
+            name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
-                "path": os.path.join(
+                "path": os.path.join(  # pyrefly: ignore[no-matching-overload]
                     base_path, self.builder_config.language, "dev_2k.tsv"
                 )
             },
diff --git a/tensorflow_datasets/datasets/penguins/penguins_dataset_builder.py b/tensorflow_datasets/datasets/penguins/penguins_dataset_builder.py
index 2254fcffff9..8bc4dd4683a 100644
--- a/tensorflow_datasets/datasets/penguins/penguins_dataset_builder.py
+++ b/tensorflow_datasets/datasets/penguins/penguins_dataset_builder.py
@@ -159,11 +159,11 @@ class Builder(tfds.core.GeneratorBasedBuilder):
   def _info(self) -> tfds.core.DatasetInfo:
     """Returns the dataset metadata."""
     supervised_keys = None
-    features = self.builder_config.features
+    features = self.builder_config.features  # pyrefly: ignore[missing-attribute]
     supervised_features = features.copy()
-    label_name = self.builder_config.label
+    label_name = self.builder_config.label  # pyrefly: ignore[missing-attribute]
 
-    if self.builder_config.name == 'processed':
+    if self.builder_config.name == 'processed':  # pyrefly: ignore[missing-attribute]
       label_feature = supervised_features.pop(label_name, None)
       supervised_keys = ('features', label_name)
       features = {
@@ -187,7 +187,7 @@ def _info(self) -> tfds.core.DatasetInfo:
 
   def _split_generators(self, dl_manager: tfds.download.DownloadManager):
     """Returns SplitGenerators."""
-    path = dl_manager.download(_PENGUINS_PATH + self.builder_config.file_name)
+    path = dl_manager.download(_PENGUINS_PATH + self.builder_config.file_name)  # pyrefly: ignore[missing-attribute]
     return {'train': self._generate_examples(path)}
 
   def _generate_examples(self, path):
@@ -197,8 +197,8 @@ def _generate_examples(self, path):
         row = {f: self._clean_up(f, v) for f, v in row.items()}
 
         # Pack features if requested.
-        if self.builder_config.name == 'processed':
-          label_name = self.builder_config.label
+        if self.builder_config.name == 'processed':  # pyrefly: ignore[missing-attribute]
+          label_name = self.builder_config.label  # pyrefly: ignore[missing-attribute]
           label = row.pop(label_name, None)
           row = list(row.values())
           yield i, {'features': row, label_name: label}
@@ -207,21 +207,21 @@ def _generate_examples(self, path):
 
   def _clean_up(self, field, value):
     """Applies field-level pre-processing, if needed."""
-    if not self.builder_config.cleanup:
+    if not self.builder_config.cleanup:  # pyrefly: ignore[missing-attribute]
       return value
     if field not in self.builder_config.cleanup:
       return value
 
-    feature_type = self.builder_config.features[field]
+    feature_type = self.builder_config.features[field]  # pyrefly: ignore[missing-attribute]
     if feature_type == np.float32:
       # Field is a float. If it won't parse, clean it up.
       try:
         return float(value)
       except ValueError:
-        return self.builder_config.cleanup[field]
+        return self.builder_config.cleanup[field]  # pyrefly: ignore[bad-index]
     elif isinstance(feature_type, tfds.features.ClassLabel):
       # Field is a class. If it's OOV, clean it up.
       if value not in feature_type.names:
-        return self.builder_config.cleanup[field]
+        return self.builder_config.cleanup[field]  # pyrefly: ignore[bad-index]
 
     return value
diff --git a/tensorflow_datasets/datasets/pet_finder/pet_finder_dataset_builder.py b/tensorflow_datasets/datasets/pet_finder/pet_finder_dataset_builder.py
index 18ddf3d1d0b..8a1d751d9ad 100644
--- a/tensorflow_datasets/datasets/pet_finder/pet_finder_dataset_builder.py
+++ b/tensorflow_datasets/datasets/pet_finder/pet_finder_dataset_builder.py
@@ -100,7 +100,7 @@ def _split_generators(self, dl_manager):
 
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "csv_name": "train.csv",
                 "csv_paths": dl_paths["train"],
@@ -108,7 +108,7 @@ def _split_generators(self, dl_manager):
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "csv_name": "test.csv",
                 "csv_paths": dl_paths["test"],
diff --git a/tensorflow_datasets/datasets/pg19/pg19_dataset_builder.py b/tensorflow_datasets/datasets/pg19/pg19_dataset_builder.py
index ee50fb17fd2..d8535977a2e 100644
--- a/tensorflow_datasets/datasets/pg19/pg19_dataset_builder.py
+++ b/tensorflow_datasets/datasets/pg19/pg19_dataset_builder.py
@@ -59,21 +59,21 @@ def _split_generators(self, dl_manager):
 
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 'metadata': metadata_dict,
                 'filepath': os.path.join(_DATA_DIR, 'train'),
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.VALIDATION,
+            name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 'metadata': metadata_dict,
                 'filepath': os.path.join(_DATA_DIR, 'validation'),
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 'metadata': metadata_dict,
                 'filepath': os.path.join(_DATA_DIR, 'test'),
diff --git a/tensorflow_datasets/datasets/piqa/piqa_dataset_builder.py b/tensorflow_datasets/datasets/piqa/piqa_dataset_builder.py
index ecda74f02fc..c72237fd8b8 100644
--- a/tensorflow_datasets/datasets/piqa/piqa_dataset_builder.py
+++ b/tensorflow_datasets/datasets/piqa/piqa_dataset_builder.py
@@ -58,7 +58,7 @@ def _split_generators(self, dl_manager):
     # Specify the splits
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "data_path": os.path.join(
                     extracted_path, "physicaliqa-train-dev/train.jsonl"
@@ -69,7 +69,7 @@ def _split_generators(self, dl_manager):
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.VALIDATION,
+            name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "data_path": os.path.join(
                     extracted_path, "physicaliqa-train-dev/dev.jsonl"
diff --git a/tensorflow_datasets/datasets/places365_small/places365_small_dataset_builder.py b/tensorflow_datasets/datasets/places365_small/places365_small_dataset_builder.py
index 04d4326998c..c29e488b475 100644
--- a/tensorflow_datasets/datasets/places365_small/places365_small_dataset_builder.py
+++ b/tensorflow_datasets/datasets/places365_small/places365_small_dataset_builder.py
@@ -18,7 +18,7 @@
 import os
 
 from etils import epath
-import six.moves.urllib as urllib
+import six.moves.urllib as urllib  # pyrefly: ignore[missing-source-for-stubs]
 import tensorflow_datasets.public_api as tfds
 
 _BASE_URL = "http://data.csail.mit.edu/places/places365/"
diff --git a/tensorflow_datasets/datasets/plant_leaves/plant_leaves_dataset_builder.py b/tensorflow_datasets/datasets/plant_leaves/plant_leaves_dataset_builder.py
index 1973f03f86c..4bc493d3bbd 100644
--- a/tensorflow_datasets/datasets/plant_leaves/plant_leaves_dataset_builder.py
+++ b/tensorflow_datasets/datasets/plant_leaves/plant_leaves_dataset_builder.py
@@ -80,7 +80,7 @@ def _split_generators(self, dl_manager):
     archive_path = dl_manager.download(_DOWNLOAD_URL)
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"archive": dl_manager.iter_archive(archive_path)},
         )
     ]
diff --git a/tensorflow_datasets/datasets/plant_village/plant_village_dataset_builder.py b/tensorflow_datasets/datasets/plant_village/plant_village_dataset_builder.py
index e412103fe15..a7daaa106bd 100644
--- a/tensorflow_datasets/datasets/plant_village/plant_village_dataset_builder.py
+++ b/tensorflow_datasets/datasets/plant_village/plant_village_dataset_builder.py
@@ -84,7 +84,7 @@ def _split_generators(self, dl_manager):
     path = dl_manager.download_and_extract(_URL)
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN, gen_kwargs={"datapath": path}
+            name=tfds.Split.TRAIN, gen_kwargs={"datapath": path}  # pyrefly: ignore[missing-attribute]
         )
     ]
 
diff --git a/tensorflow_datasets/datasets/plantae_k/plantae_k_dataset_builder.py b/tensorflow_datasets/datasets/plantae_k/plantae_k_dataset_builder.py
index 8d3d438132a..f4420b53f3e 100644
--- a/tensorflow_datasets/datasets/plantae_k/plantae_k_dataset_builder.py
+++ b/tensorflow_datasets/datasets/plantae_k/plantae_k_dataset_builder.py
@@ -91,7 +91,7 @@ def _split_generators(self, dl_manager):
             )
       return [
           tfds.core.SplitGenerator(
-              name=tfds.Split.TRAIN, gen_kwargs={"image_files": image_files}
+              name=tfds.Split.TRAIN, gen_kwargs={"image_files": image_files}  # pyrefly: ignore[missing-attribute]
           )
       ]
 
diff --git a/tensorflow_datasets/datasets/protein_net/protein_net_dataset_builder.py b/tensorflow_datasets/datasets/protein_net/protein_net_dataset_builder.py
index 50bf3d1a1a7..177546fb1e7 100644
--- a/tensorflow_datasets/datasets/protein_net/protein_net_dataset_builder.py
+++ b/tensorflow_datasets/datasets/protein_net/protein_net_dataset_builder.py
@@ -34,7 +34,7 @@
 
 def _parse_array(lines: Sequence[str]) -> np.ndarray:
   """Parse lines of tab-separated numbers into an array."""
-  lines = [x.split('\t') for x in lines]
+  lines = [x.split('\t') for x in lines]  # pyrefly: ignore[bad-assignment]
   return np.array(lines, dtype=np.float32)
 
 
@@ -165,16 +165,16 @@ def _split_generators(
       self, dl_manager: tfds.download.DownloadManager
   ) -> Dict[Union[str, tfds.Split], _ExampleIterator]:
     """Returns SplitGenerators."""
-    name = self.builder_config.name  # Configurable dataset (config) name.
+    name = self.builder_config.name  # Configurable dataset (config) name.  # pyrefly: ignore[missing-attribute]
     path = dl_manager.download_and_extract(
         urllib.parse.urljoin(self.URL, self.FILES[name])
     )
 
     splits = {
-        tfds.Split.VALIDATION: self._generate_examples(
+        tfds.Split.VALIDATION: self._generate_examples(  # pyrefly: ignore[missing-attribute]
             os.path.join(path, name, 'validation')
         ),
-        tfds.Split.TEST: self._generate_examples(
+        tfds.Split.TEST: self._generate_examples(  # pyrefly: ignore[missing-attribute]
             os.path.join(path, name, 'testing')
         ),
     }
diff --git a/tensorflow_datasets/datasets/qa4mre/qa4mre_dataset_builder.py b/tensorflow_datasets/datasets/qa4mre/qa4mre_dataset_builder.py
index 45ad4406f19..8c87d9c80ff 100644
--- a/tensorflow_datasets/datasets/qa4mre/qa4mre_dataset_builder.py
+++ b/tensorflow_datasets/datasets/qa4mre/qa4mre_dataset_builder.py
@@ -92,10 +92,10 @@ def _get_question(
       'document_id': document_id,
       'document_str': document_str,
       'question_id': question_id,
-      'question_str': question_str,
+      'question_str': question_str,  # pyrefly: ignore[unbound-name]
       'answer_options': possible_answers,
-      'correct_answer_id': correct_answer_id,
-      'correct_answer_str': correct_answer_str,
+      'correct_answer_id': correct_answer_id,  # pyrefly: ignore[unbound-name]
+      'correct_answer_str': correct_answer_str,  # pyrefly: ignore[unbound-name]
   }
 
   return id_, feats
@@ -247,17 +247,17 @@ def _split_generators(self, dl_manager):
     cfg = self.builder_config
     download_urls = dict()
 
-    if cfg.track == 'main':
-      download_urls['{}.main.{}'.format(cfg.year, cfg.lang)] = os.path.join(
-          _BASE_URL, PATHS[cfg.year]['_PATH_TMPL_MAIN_GS'].format(cfg.lang)
+    if cfg.track == 'main':  # pyrefly: ignore[missing-attribute]
+      download_urls['{}.main.{}'.format(cfg.year, cfg.lang)] = os.path.join(  # pyrefly: ignore[missing-attribute]
+          _BASE_URL, PATHS[cfg.year]['_PATH_TMPL_MAIN_GS'].format(cfg.lang)  # pyrefly: ignore[missing-attribute]
       )  # pytype: disable=attribute-error
 
-    if cfg.year in ['2012', '2013'] and cfg.track == 'alzheimers':
-      download_urls['{}.alzheimers.EN'.format(cfg.year)] = os.path.join(
+    if cfg.year in ['2012', '2013'] and cfg.track == 'alzheimers':  # pyrefly: ignore[missing-attribute]
+      download_urls['{}.alzheimers.EN'.format(cfg.year)] = os.path.join(  # pyrefly: ignore[no-matching-overload]
           _BASE_URL, PATHS[cfg.year]['_PATH_ALZHEIMER']
       )
 
-    if cfg.year == '2013' and cfg.track == 'entrance_exam':
+    if cfg.year == '2013' and cfg.track == 'entrance_exam':  # pyrefly: ignore[missing-attribute]
       download_urls['2013.entrance_exam.EN'] = os.path.join(
           _BASE_URL, PATHS[cfg.year]['_PATH_ENTRANCE_EXAM']
       )
@@ -266,7 +266,7 @@ def _split_generators(self, dl_manager):
 
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 'filepath': downloaded_files[
                     '{}.{}.{}'.format(cfg.year, cfg.track, cfg.lang)
@@ -293,7 +293,7 @@ def _generate_examples(self, filepath):
                 topic_id,
                 topic_name,
                 test_id,
-                document_id,
-                document_str,
+                document_id,  # pyrefly: ignore[unbound-name]
+                document_str,  # pyrefly: ignore[unbound-name]
                 question,
             )
diff --git a/tensorflow_datasets/datasets/qasc/qasc_dataset_builder.py b/tensorflow_datasets/datasets/qasc/qasc_dataset_builder.py
index 7520bb4fbce..9279f0d1049 100644
--- a/tensorflow_datasets/datasets/qasc/qasc_dataset_builder.py
+++ b/tensorflow_datasets/datasets/qasc/qasc_dataset_builder.py
@@ -55,15 +55,15 @@ def _split_generators(self, dl_manager):
     data_dir = os.path.join(dl_dir["QASC_Dataset"], "QASC_Dataset")
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"filepath": os.path.join(data_dir, "train.jsonl")},
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"filepath": os.path.join(data_dir, "test.jsonl")},
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.VALIDATION,
+            name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"filepath": os.path.join(data_dir, "dev.jsonl")},
         ),
     ]
diff --git a/tensorflow_datasets/datasets/qm9/qm9_dataset_builder.py b/tensorflow_datasets/datasets/qm9/qm9_dataset_builder.py
index bdf5c5be89d..722d7dec19a 100644
--- a/tensorflow_datasets/datasets/qm9/qm9_dataset_builder.py
+++ b/tensorflow_datasets/datasets/qm9/qm9_dataset_builder.py
@@ -131,7 +131,7 @@ def _process_molecule(
     thermo = 0
     for z in atoms['Z'].values:
       thermo += v[z]
-    example[f'{k}_atomization'] = example[k] - thermo
+    example[f'{k}_atomization'] = example[k] - thermo  # pyrefly: ignore[unsupported-operation]
 
   return example
 
@@ -234,7 +234,7 @@ class Builder(tfds.core.GeneratorBasedBuilder):
   def _info(self) -> tfds.core.DatasetInfo:
     """Returns the dataset metadata."""
     return self.dataset_info_from_configs(
-        disable_shuffling=self.builder_config.permutation_seed is None,
+        disable_shuffling=self.builder_config.permutation_seed is None,  # pyrefly: ignore[missing-attribute]
         features=tfds.features.FeaturesDict({
             'num_atoms': tfds.features.Tensor(shape=(), dtype=np.int64),
             'charges': tfds.features.Tensor(shape=(29,), dtype=np.int64),
@@ -306,9 +306,9 @@ def _split_generators(
 
     split_ids = _get_split_ids(
         uncharacterized,
-        permutation_seed=self.builder_config.permutation_seed,
-        train_size=self.builder_config.train_size,
-        validation_size=self.builder_config.validation_size,
+        permutation_seed=self.builder_config.permutation_seed,  # pyrefly: ignore[missing-attribute]
+        train_size=self.builder_config.train_size,  # pyrefly: ignore[missing-attribute]
+        validation_size=self.builder_config.validation_size,  # pyrefly: ignore[missing-attribute]
     )
 
     return {
diff --git a/tensorflow_datasets/datasets/quality/quality_dataset_builder.py b/tensorflow_datasets/datasets/quality/quality_dataset_builder.py
index c5c60088c08..bb01dfa2cb1 100644
--- a/tensorflow_datasets/datasets/quality/quality_dataset_builder.py
+++ b/tensorflow_datasets/datasets/quality/quality_dataset_builder.py
@@ -88,7 +88,7 @@ def _split_generators(self, dl_manager: tfds.download.DownloadManager):
     """Returns SplitGenerators."""
     path = dl_manager.download_and_extract(_DOWNLOAD_URL)
 
-    if self.builder_config.stripped:
+    if self.builder_config.stripped:  # pyrefly: ignore[missing-attribute]
       return {
           'train': self._generate_examples(
               path / 'QuALITY.v0.9.htmlstripped.train', 'train'
diff --git a/tensorflow_datasets/datasets/quickdraw_bitmap/quickdraw_bitmap_dataset_builder.py b/tensorflow_datasets/datasets/quickdraw_bitmap/quickdraw_bitmap_dataset_builder.py
index 24240df11d7..82390ceb69e 100644
--- a/tensorflow_datasets/datasets/quickdraw_bitmap/quickdraw_bitmap_dataset_builder.py
+++ b/tensorflow_datasets/datasets/quickdraw_bitmap/quickdraw_bitmap_dataset_builder.py
@@ -67,7 +67,7 @@ def _split_generators(self, dl_manager):
     # There is no predefined train/test split for this dataset.
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "file_paths": file_paths,
             },
diff --git a/tensorflow_datasets/datasets/race/race_dataset_builder.py b/tensorflow_datasets/datasets/race/race_dataset_builder.py
index 26b0b1acf07..bfc9eebd747 100644
--- a/tensorflow_datasets/datasets/race/race_dataset_builder.py
+++ b/tensorflow_datasets/datasets/race/race_dataset_builder.py
@@ -64,7 +64,7 @@ def _split_generators(self, dl_manager: tfds.download.DownloadManager):
     )
 
     path = path / "RACE"
-    task = self.builder_config.name
+    task = self.builder_config.name  # pyrefly: ignore[missing-attribute]
 
     return {
         "train": self._generate_examples(path / "train" / task),
diff --git a/tensorflow_datasets/datasets/reddit/reddit_dataset_builder.py b/tensorflow_datasets/datasets/reddit/reddit_dataset_builder.py
index 207a16df27a..6f4ca8987f6 100644
--- a/tensorflow_datasets/datasets/reddit/reddit_dataset_builder.py
+++ b/tensorflow_datasets/datasets/reddit/reddit_dataset_builder.py
@@ -57,7 +57,7 @@ def _split_generators(self, dl_manager):
     dl_path = dl_manager.download_and_extract(_URL)
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "path": os.path.join(dl_path, "corpus-webis-tldr-17.json")
             },
diff --git a/tensorflow_datasets/datasets/reddit_disentanglement/reddit_disentanglement_dataset_builder.py b/tensorflow_datasets/datasets/reddit_disentanglement/reddit_disentanglement_dataset_builder.py
index 144afeedd81..68ffac3981d 100644
--- a/tensorflow_datasets/datasets/reddit_disentanglement/reddit_disentanglement_dataset_builder.py
+++ b/tensorflow_datasets/datasets/reddit_disentanglement/reddit_disentanglement_dataset_builder.py
@@ -104,17 +104,17 @@ def _split_generators(self, dl_manager):
     """Returns SplitGenerators."""
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "path": os.path.join(dl_manager.manual_dir, "train.csv")
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.VALIDATION,
+            name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"path": os.path.join(dl_manager.manual_dir, "val.csv")},
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "path": os.path.join(dl_manager.manual_dir, "test.csv")
             },
diff --git a/tensorflow_datasets/datasets/reddit_tifu/reddit_tifu_dataset_builder.py b/tensorflow_datasets/datasets/reddit_tifu/reddit_tifu_dataset_builder.py
index 1ae65eecb5e..32eb0e784d1 100644
--- a/tensorflow_datasets/datasets/reddit_tifu/reddit_tifu_dataset_builder.py
+++ b/tensorflow_datasets/datasets/reddit_tifu/reddit_tifu_dataset_builder.py
@@ -90,13 +90,13 @@ def _info(self):
     )
     return self.dataset_info_from_configs(
         features=tfds.features.FeaturesDict(features),
-        supervised_keys=(_DOCUMENT, self.builder_config.summary_key),
+        supervised_keys=(_DOCUMENT, self.builder_config.summary_key),  # pyrefly: ignore[missing-attribute]
         homepage="https://github.com/ctr4si/MMN",
     )
 
   def _split_generators(self, dl_manager):
     """Returns SplitGenerators."""
-    if self.builder_config.name == "long_split":
+    if self.builder_config.name == "long_split":  # pyrefly: ignore[missing-attribute]
       paths = dl_manager.download_and_extract(
           {"data": _URL, "split": _LONG_SPLIT}
       )
@@ -104,7 +104,7 @@ def _split_generators(self, dl_manager):
       split_path = paths["split"]
       return [
           tfds.core.SplitGenerator(
-              name=tfds.Split.TRAIN,
+              name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
               gen_kwargs={
                   "path": dl_path,
                   "split_path": split_path,
@@ -112,7 +112,7 @@ def _split_generators(self, dl_manager):
               },
           ),
           tfds.core.SplitGenerator(
-              name=tfds.Split.TEST,
+              name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
               gen_kwargs={
                   "path": dl_path,
                   "split_path": split_path,
@@ -120,7 +120,7 @@ def _split_generators(self, dl_manager):
               },
           ),
           tfds.core.SplitGenerator(
-              name=tfds.Split.VALIDATION,
+              name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
               gen_kwargs={
                   "path": dl_path,
                   "split_path": split_path,
@@ -132,7 +132,7 @@ def _split_generators(self, dl_manager):
       dl_path = dl_manager.download_and_extract(_URL)
       return [
           tfds.core.SplitGenerator(
-              name=tfds.Split.TRAIN,
+              name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
               gen_kwargs={"path": dl_path},
           )
       ]
@@ -149,7 +149,7 @@ def _generate_examples(self, path=None, split_path=None, split=None):
         #   'selftext','trimmed_title','selftext_without_tldr_tokenized',
         #   'id','selftext_without_tldr'
         d = json.loads(line)
-        if not split or (split and d["id"] in split_data[split]):
+        if not split or (split and d["id"] in split_data[split]):  # pyrefly: ignore[unbound-name]
           r = {
               _DOCUMENT: d["selftext_without_tldr"].strip(),
               _TITLE: d["trimmed_title"].strip(),
@@ -158,5 +158,5 @@ def _generate_examples(self, path=None, split_path=None, split=None):
           }
           r.update({k: d[k] for k in _ADDITIONAL_FEATURES})
           # skip if document or summary is empty
-          if r[_DOCUMENT] and r[self.builder_config.summary_key]:
+          if r[_DOCUMENT] and r[self.builder_config.summary_key]:  # pyrefly: ignore[missing-attribute]
             yield i, r
diff --git a/tensorflow_datasets/datasets/ref_coco/ref_coco_dataset_builder.py b/tensorflow_datasets/datasets/ref_coco/ref_coco_dataset_builder.py
index 00be8eac326..3f514ee9ac2 100644
--- a/tensorflow_datasets/datasets/ref_coco/ref_coco_dataset_builder.py
+++ b/tensorflow_datasets/datasets/ref_coco/ref_coco_dataset_builder.py
@@ -218,35 +218,35 @@ def _info(self):
   def _split_generators(self, dl_manager):
     allowed_splits = {
         ('refcoco', 'google'): [
-            tfds.Split.TRAIN,
-            tfds.Split.VALIDATION,
-            tfds.Split.TEST,
+            tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
+            tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
+            tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
         ],
         ('refcoco', 'unc'): [
-            tfds.Split.TRAIN,
-            tfds.Split.VALIDATION,
+            tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
+            tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             'testA',
             'testB',
         ],
         ('refcocoplus', 'unc'): [
-            tfds.Split.TRAIN,
-            tfds.Split.VALIDATION,
+            tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
+            tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             'testA',
             'testB',
         ],
-        ('refcocog', 'google'): [tfds.Split.TRAIN, tfds.Split.VALIDATION],
+        ('refcocog', 'google'): [tfds.Split.TRAIN, tfds.Split.VALIDATION],  # pyrefly: ignore[missing-attribute]
         ('refcocog', 'umd'): [
-            tfds.Split.TRAIN,
-            tfds.Split.VALIDATION,
-            tfds.Split.TEST,
+            tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
+            tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
+            tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
         ],
     }
     bc = self.builder_config
-    splits = allowed_splits[(bc.dataset, bc.dataset_partition)]
+    splits = allowed_splits[(bc.dataset, bc.dataset_partition)]  # pyrefly: ignore[missing-attribute]
 
     return {
         split: self._generate_examples(
-            bc.dataset, bc.dataset_partition, split, dl_manager
+            bc.dataset, bc.dataset_partition, split, dl_manager  # pyrefly: ignore[missing-attribute]
         )
         for split in splits
     }
@@ -259,7 +259,7 @@ def _generate_examples(self, dataset, dataset_partition, split, dl_manager):
 
     if dataset == 'refcocoplus':
       dataset = 'refcoco+'
-    if split == tfds.Split.VALIDATION:
+    if split == tfds.Split.VALIDATION:  # pyrefly: ignore[missing-attribute]
       split = 'val'
 
     for image_id, example in _generate_examples(
diff --git a/tensorflow_datasets/datasets/resisc45/resisc45_dataset_builder.py b/tensorflow_datasets/datasets/resisc45/resisc45_dataset_builder.py
index 3df92d1f80c..7eb4fbef7ce 100644
--- a/tensorflow_datasets/datasets/resisc45/resisc45_dataset_builder.py
+++ b/tensorflow_datasets/datasets/resisc45/resisc45_dataset_builder.py
@@ -103,7 +103,7 @@ def _split_generators(self, dl_manager):
       )
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={'path': path},
         ),
     ]
diff --git a/tensorflow_datasets/datasets/robomimic_mh/robomimic_mh_dataset_builder.py b/tensorflow_datasets/datasets/robomimic_mh/robomimic_mh_dataset_builder.py
index 1ce49b41099..c17c950cb8f 100644
--- a/tensorflow_datasets/datasets/robomimic_mh/robomimic_mh_dataset_builder.py
+++ b/tensorflow_datasets/datasets/robomimic_mh/robomimic_mh_dataset_builder.py
@@ -35,7 +35,7 @@ class Builder(utils.RobomimicBuilder):
   DATASET_NAME = 'robomimic_mh'
 
   def _get_metadata(self) -> Dict[Any, Any]:
-    if self.builder_config.task == utils.Task.TRANSPORT:
+    if self.builder_config.task == utils.Task.TRANSPORT:  # pyrefly: ignore[missing-attribute]
       episode_metadata = {
           '20_percent': np.bool_,
           '20_percent_train': np.bool_,
diff --git a/tensorflow_datasets/datasets/robomimic_ph/robomimic_ph_dataset_builder.py b/tensorflow_datasets/datasets/robomimic_ph/robomimic_ph_dataset_builder.py
index baabb7053c4..2bc223cb7e8 100644
--- a/tensorflow_datasets/datasets/robomimic_ph/robomimic_ph_dataset_builder.py
+++ b/tensorflow_datasets/datasets/robomimic_ph/robomimic_ph_dataset_builder.py
@@ -36,7 +36,7 @@ class Builder(utils.RobomimicBuilder):
   DATASET_NAME = 'robomimic_ph'
 
   def _get_metadata(self) -> Dict[Any, Any]:
-    if self.builder_config.task != utils.Task.TOOL_HANG:
+    if self.builder_config.task != utils.Task.TOOL_HANG:  # pyrefly: ignore[missing-attribute]
       episode_metadata = {
           '20_percent': np.bool_,
           '20_percent_train': np.bool_,
diff --git a/tensorflow_datasets/datasets/robonet/robonet_dataset_builder.py b/tensorflow_datasets/datasets/robonet/robonet_dataset_builder.py
index 067a1dc5fe9..406ac1e7de9 100644
--- a/tensorflow_datasets/datasets/robonet/robonet_dataset_builder.py
+++ b/tensorflow_datasets/datasets/robonet/robonet_dataset_builder.py
@@ -137,8 +137,8 @@ class Builder(tfds.core.BeamBasedBuilder):
   ]
 
   def _info(self):
-    if self.builder_config.width is not None:
-      if self.builder_config.height is None:
+    if self.builder_config.width is not None:  # pyrefly: ignore[missing-attribute]
+      if self.builder_config.height is None:  # pyrefly: ignore[missing-attribute]
         raise ValueError('Provide either both height and width or none.')
       ffmpeg_extra_args = (
           '-vf',
@@ -151,7 +151,7 @@ def _info(self):
 
     video_shape = (
         None,
-        self.builder_config.height,
+        self.builder_config.height,  # pyrefly: ignore[missing-attribute]
         self.builder_config.width,
         3,
     )
@@ -159,7 +159,7 @@ def _info(self):
     features = tfds.features.FeaturesDict({
         # Video frames: uint8 [None, Time, Width, Height, Channels]
         'video': tfds.features.Video(
-            video_shape,
+            video_shape,  # pyrefly: ignore[bad-argument-type]
             ffmpeg_extra_args=ffmpeg_extra_args,
             encoding_format='png',
         ),  # pytype: disable=wrong-arg-types  # gen-stub-imports
@@ -182,11 +182,11 @@ def _info(self):
 
   def _split_generators(self, dl_manager):
     files = dl_manager.download_and_extract(
-        DATA_URL_SAMPLE if self.builder_config.sample_dataset else DATA_URL
+        DATA_URL_SAMPLE if self.builder_config.sample_dataset else DATA_URL  # pyrefly: ignore[missing-attribute]
     )
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 'filedir': os.path.join(files, 'hdf5'),
             },
diff --git a/tensorflow_datasets/datasets/rock_paper_scissors/rock_paper_scissors_dataset_builder.py b/tensorflow_datasets/datasets/rock_paper_scissors/rock_paper_scissors_dataset_builder.py
index 0dce7245acd..08ffdb64fa6 100644
--- a/tensorflow_datasets/datasets/rock_paper_scissors/rock_paper_scissors_dataset_builder.py
+++ b/tensorflow_datasets/datasets/rock_paper_scissors/rock_paper_scissors_dataset_builder.py
@@ -56,13 +56,13 @@ def _split_generators(self, dl_manager):
 
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "archive": dl_manager.iter_archive(train_path),
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "archive": dl_manager.iter_archive(test_path),
             },
diff --git a/tensorflow_datasets/datasets/s3o4d/s3o4d_dataset_builder.py b/tensorflow_datasets/datasets/s3o4d/s3o4d_dataset_builder.py
index 18e6cef9a13..380ccb0102a 100644
--- a/tensorflow_datasets/datasets/s3o4d/s3o4d_dataset_builder.py
+++ b/tensorflow_datasets/datasets/s3o4d/s3o4d_dataset_builder.py
@@ -68,8 +68,8 @@ def _split_generators(self, dl_manager: tfds.download.DownloadManager):
                 '_'.join([a, b]),
                 self._generate_examples(
                     dl_manager,
-                    paths['_'.join([a, b, 'img'])],
-                    paths['_'.join([a, b, 'latent'])],
+                    paths['_'.join([a, b, 'img'])],  # pyrefly: ignore[bad-index]
+                    paths['_'.join([a, b, 'latent'])],  # pyrefly: ignore[bad-index]
                     a,
                 ),
             )
diff --git a/tensorflow_datasets/datasets/salient_span_wikipedia/salient_span_wikipedia_dataset_builder.py b/tensorflow_datasets/datasets/salient_span_wikipedia/salient_span_wikipedia_dataset_builder.py
index e94a557ab1f..fc11a5f26bd 100644
--- a/tensorflow_datasets/datasets/salient_span_wikipedia/salient_span_wikipedia_dataset_builder.py
+++ b/tensorflow_datasets/datasets/salient_span_wikipedia/salient_span_wikipedia_dataset_builder.py
@@ -66,7 +66,7 @@ def _info(self):
             "type": np.str_,
         }),
     }
-    if not self.builder_config.split_sentences:
+    if not self.builder_config.split_sentences:  # pyrefly: ignore[missing-attribute]
       feature_dict["sentences"] = tfds.features.Sequence({
           "start": np.int32,
           "limit": np.int32,
@@ -79,8 +79,8 @@ def _info(self):
   def _split_generators(self, dl_manager):
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
-            gen_kwargs={"split_sentences": self.builder_config.split_sentences},
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
+            gen_kwargs={"split_sentences": self.builder_config.split_sentences},  # pyrefly: ignore[missing-attribute]
         ),
     ]
 
diff --git a/tensorflow_datasets/datasets/samsum/samsum_dataset_builder.py b/tensorflow_datasets/datasets/samsum/samsum_dataset_builder.py
index a53739cd988..7b2b1eb45d7 100644
--- a/tensorflow_datasets/datasets/samsum/samsum_dataset_builder.py
+++ b/tensorflow_datasets/datasets/samsum/samsum_dataset_builder.py
@@ -53,19 +53,19 @@ def _split_generators(
     """Returns SplitGenerators."""
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "path": os.path.join(dl_manager.manual_dir, "train.json")
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.VALIDATION,
+            name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "path": os.path.join(dl_manager.manual_dir, "val.json")
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "path": os.path.join(dl_manager.manual_dir, "test.json")
             },
diff --git a/tensorflow_datasets/datasets/savee/savee_dataset_builder.py b/tensorflow_datasets/datasets/savee/savee_dataset_builder.py
index f362192010d..96cae330253 100644
--- a/tensorflow_datasets/datasets/savee/savee_dataset_builder.py
+++ b/tensorflow_datasets/datasets/savee/savee_dataset_builder.py
@@ -158,15 +158,15 @@ def _split_generators(self, dl_manager):
 
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={'file_names': splits['train']},
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.VALIDATION,
+            name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={'file_names': splits['validation']},
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={'file_names': splits['test']},
         ),
     ]
diff --git a/tensorflow_datasets/datasets/scan/scan_dataset_builder.py b/tensorflow_datasets/datasets/scan/scan_dataset_builder.py
index bfcd7d9656b..ad1ac8fb125 100644
--- a/tensorflow_datasets/datasets/scan/scan_dataset_builder.py
+++ b/tensorflow_datasets/datasets/scan/scan_dataset_builder.py
@@ -102,10 +102,10 @@ def _split_generators(self, dl_manager):
     """Returns SplitGenerators."""
     data_dir = dl_manager.download_and_extract(_DATA_URL)
     data_dir = os.path.join(
-        data_dir, 'SCAN-master', self.builder_config.directory
+        data_dir, 'SCAN-master', self.builder_config.directory  # pyrefly: ignore[missing-attribute]
     )
-    split = self.builder_config.name
-    splitfile = self.builder_config.splitfile
+    split = self.builder_config.name  # pyrefly: ignore[missing-attribute]
+    splitfile = self.builder_config.splitfile  # pyrefly: ignore[missing-attribute]
     if 'mcd' in split:
       split_dir = dl_manager.download_and_extract(_MCD_SPLITS_URL)
       split_dir = os.path.join(split_dir, 'scan-splits')
@@ -128,9 +128,9 @@ def _split_generators(self, dl_manager):
       }
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN, gen_kwargs=train_kwargs
+            name=tfds.Split.TRAIN, gen_kwargs=train_kwargs  # pyrefly: ignore[missing-attribute]
         ),
-        tfds.core.SplitGenerator(name=tfds.Split.TEST, gen_kwargs=test_kwargs),
+        tfds.core.SplitGenerator(name=tfds.Split.TEST, gen_kwargs=test_kwargs),  # pyrefly: ignore[missing-attribute]
     ]
 
   def _read_examples(self, datapath):
diff --git a/tensorflow_datasets/datasets/scene_parse150/scene_parse150_dataset_builder.py b/tensorflow_datasets/datasets/scene_parse150/scene_parse150_dataset_builder.py
index dfe90633df2..8b43291e0ba 100644
--- a/tensorflow_datasets/datasets/scene_parse150/scene_parse150_dataset_builder.py
+++ b/tensorflow_datasets/datasets/scene_parse150/scene_parse150_dataset_builder.py
@@ -51,7 +51,7 @@ def _split_generators(self, dl_manager):
 
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "images_dir_path": os.path.join(
                     dl_paths["images"], "images/training"
@@ -62,7 +62,7 @@ def _split_generators(self, dl_manager):
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "images_dir_path": os.path.join(
                     dl_paths["images"], "images/validation"
diff --git a/tensorflow_datasets/datasets/sci_tail/sci_tail_dataset_builder.py b/tensorflow_datasets/datasets/sci_tail/sci_tail_dataset_builder.py
index 8f777cab1d1..6fce065cf4a 100644
--- a/tensorflow_datasets/datasets/sci_tail/sci_tail_dataset_builder.py
+++ b/tensorflow_datasets/datasets/sci_tail/sci_tail_dataset_builder.py
@@ -58,13 +58,13 @@ def _split_generators(self, dl_manager: tfds.download.DownloadManager):
 
     # Returns the Dict[split names, Iterator[Key, Example]]
     return {
-        tfds.Split.TRAIN: self._generate_examples(
+        tfds.Split.TRAIN: self._generate_examples(  # pyrefly: ignore[missing-attribute]
             path=data_dir / _SCITAIL_DIR / _TSV_DIR / 'scitail_1.0_train.tsv'
         ),
-        tfds.Split.VALIDATION: self._generate_examples(
+        tfds.Split.VALIDATION: self._generate_examples(  # pyrefly: ignore[missing-attribute]
             path=data_dir / _SCITAIL_DIR / _TSV_DIR / 'scitail_1.0_dev.tsv'
         ),
-        tfds.Split.TEST: self._generate_examples(
+        tfds.Split.TEST: self._generate_examples(  # pyrefly: ignore[missing-attribute]
             path=data_dir / _SCITAIL_DIR / _TSV_DIR / 'scitail_1.0_test.tsv'
         ),
     }
diff --git a/tensorflow_datasets/datasets/scicite/scicite_dataset_builder.py b/tensorflow_datasets/datasets/scicite/scicite_dataset_builder.py
index 5a747844bbe..0383eafcd2c 100644
--- a/tensorflow_datasets/datasets/scicite/scicite_dataset_builder.py
+++ b/tensorflow_datasets/datasets/scicite/scicite_dataset_builder.py
@@ -78,22 +78,22 @@ def _split_generators(self, dl_manager):
     path = os.path.join(dl_paths["scicite"], "scicite")
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"path": os.path.join(path, "train.jsonl")},
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.VALIDATION,
+            name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"path": os.path.join(path, "dev.jsonl")},
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"path": os.path.join(path, "test.jsonl")},
         ),
     ]
 
   def _generate_examples(self, path=None):
     """Yields examples."""
-    with epath.Path(path).open() as f:
+    with epath.Path(path).open() as f:  # pyrefly: ignore[bad-argument-type]
       unique_ids = {}
       for line in f:
         d = json.loads(line)
diff --git a/tensorflow_datasets/datasets/scientific_papers/scientific_papers_dataset_builder.py b/tensorflow_datasets/datasets/scientific_papers/scientific_papers_dataset_builder.py
index 4c43403c97d..f0c755b5685 100644
--- a/tensorflow_datasets/datasets/scientific_papers/scientific_papers_dataset_builder.py
+++ b/tensorflow_datasets/datasets/scientific_papers/scientific_papers_dataset_builder.py
@@ -77,26 +77,26 @@ def _split_generators(self, dl_manager):
     dl_paths = dl_manager.download_and_extract(_URLS)
     path = os.path.join(
         dl_paths[self.builder_config.name],
-        self.builder_config.name + "-dataset",
+        self.builder_config.name + "-dataset",  # pyrefly: ignore[missing-attribute]
     )
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"path": os.path.join(path, "train.txt")},
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.VALIDATION,
+            name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"path": os.path.join(path, "val.txt")},
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"path": os.path.join(path, "test.txt")},
         ),
     ]
 
   def _generate_examples(self, path=None):
     """Yields examples."""
-    with epath.Path(path).open() as f:
+    with epath.Path(path).open() as f:  # pyrefly: ignore[bad-argument-type]
       for line in f:
         # Possible keys are:
         # "article_id": str
diff --git a/tensorflow_datasets/datasets/segment_anything/segment_anything_dataset_builder.py b/tensorflow_datasets/datasets/segment_anything/segment_anything_dataset_builder.py
index f4c6030da09..a3767225c83 100644
--- a/tensorflow_datasets/datasets/segment_anything/segment_anything_dataset_builder.py
+++ b/tensorflow_datasets/datasets/segment_anything/segment_anything_dataset_builder.py
@@ -151,7 +151,7 @@ def _split_generators(self, dl_manager: tfds.download.DownloadManager):
         dict_of_urls[name] = url
     paths = dl_manager.download_and_extract(dict_of_urls)
     return {
-        'train': self._generate_examples(paths),
+        'train': self._generate_examples(paths),  # pyrefly: ignore[bad-argument-type]
     }
 
   def _generate_examples(self, paths: Dict[str, epath.Path]):
diff --git a/tensorflow_datasets/datasets/sentiment140/sentiment140_dataset_builder.py b/tensorflow_datasets/datasets/sentiment140/sentiment140_dataset_builder.py
index b6592e2334d..1386b0ad34a 100644
--- a/tensorflow_datasets/datasets/sentiment140/sentiment140_dataset_builder.py
+++ b/tensorflow_datasets/datasets/sentiment140/sentiment140_dataset_builder.py
@@ -53,7 +53,7 @@ def _split_generators(self, dl_manager):
 
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "path": os.path.join(
                     dl_paths, "training.1600000.processed.noemoticon.csv"
@@ -61,7 +61,7 @@ def _split_generators(self, dl_manager):
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "path": os.path.join(dl_paths, "testdata.manual.2009.06.14.csv")
             },
diff --git a/tensorflow_datasets/datasets/shapes3d/shapes3d_dataset_builder.py b/tensorflow_datasets/datasets/shapes3d/shapes3d_dataset_builder.py
index a2dc9729fec..3afc272a741 100644
--- a/tensorflow_datasets/datasets/shapes3d/shapes3d_dataset_builder.py
+++ b/tensorflow_datasets/datasets/shapes3d/shapes3d_dataset_builder.py
@@ -59,7 +59,7 @@ def _split_generators(self, dl_manager):
     # There is no predefined train/val/test split for this dataset.
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN, gen_kwargs=dict(filepath=filepath)
+            name=tfds.Split.TRAIN, gen_kwargs=dict(filepath=filepath)  # pyrefly: ignore[missing-attribute]
         ),
     ]
 
diff --git a/tensorflow_datasets/datasets/siscore/siscore_dataset_builder.py b/tensorflow_datasets/datasets/siscore/siscore_dataset_builder.py
index 477de68328f..86c54037974 100644
--- a/tensorflow_datasets/datasets/siscore/siscore_dataset_builder.py
+++ b/tensorflow_datasets/datasets/siscore/siscore_dataset_builder.py
@@ -81,7 +81,7 @@ def _info(self):
   def _split_generators(self, dl_manager):
     """Returns SplitGenerator."""
     # using rotation link only for now
-    variant = self.builder_config.variant
+    variant = self.builder_config.variant  # pyrefly: ignore[missing-attribute]
     dataset_url = "/".join((_BASE_URL, f"{variant}.zip"))
     path = dl_manager.download_and_extract(dataset_url)
     path = os.path.join(path, _VARIANT_EXPANDED_DIR_NAMES[variant])
diff --git a/tensorflow_datasets/datasets/smallnorb/smallnorb_dataset_builder.py b/tensorflow_datasets/datasets/smallnorb/smallnorb_dataset_builder.py
index 8a861df15f6..22006062e21 100644
--- a/tensorflow_datasets/datasets/smallnorb/smallnorb_dataset_builder.py
+++ b/tensorflow_datasets/datasets/smallnorb/smallnorb_dataset_builder.py
@@ -59,7 +59,7 @@ def _info(self):
         "label_lighting": tfds.features.ClassLabel(num_classes=6),
     }
     if self.version > "2.0.0":
-      features_dict["id"] = tfds.features.Text()
+      features_dict["id"] = tfds.features.Text()  # pyrefly: ignore[bad-assignment]
     return self.dataset_info_from_configs(
         features=tfds.features.FeaturesDict(features_dict),
         homepage="https://cs.nyu.edu/~ylclab/data/norb-v1.0-small/",
@@ -81,7 +81,7 @@ def _split_generators(self, dl_manager):
 
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs=dict(
                 split_prefix="train_",
                 dat_path=files["training_dat"],
@@ -90,7 +90,7 @@ def _split_generators(self, dl_manager):
             ),
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs=dict(
                 split_prefix="test_",
                 dat_path=files["testing_dat"],
diff --git a/tensorflow_datasets/datasets/snli/snli_dataset_builder.py b/tensorflow_datasets/datasets/snli/snli_dataset_builder.py
index fe8b0035398..2e20c615d93 100644
--- a/tensorflow_datasets/datasets/snli/snli_dataset_builder.py
+++ b/tensorflow_datasets/datasets/snli/snli_dataset_builder.py
@@ -49,17 +49,17 @@ def _split_generators(self, dl_manager):
     data_dir = os.path.join(dl_dir, 'snli_1.0')
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 'filepath': os.path.join(data_dir, 'snli_1.0_test.txt')
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.VALIDATION,
+            name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={'filepath': os.path.join(data_dir, 'snli_1.0_dev.txt')},
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 'filepath': os.path.join(data_dir, 'snli_1.0_train.txt')
             },
diff --git a/tensorflow_datasets/datasets/so2sat/so2sat_dataset_builder.py b/tensorflow_datasets/datasets/so2sat/so2sat_dataset_builder.py
index 2198dce2a5b..2a180981fae 100644
--- a/tensorflow_datasets/datasets/so2sat/so2sat_dataset_builder.py
+++ b/tensorflow_datasets/datasets/so2sat/so2sat_dataset_builder.py
@@ -87,7 +87,7 @@ class Builder(tfds.core.GeneratorBasedBuilder):
   ]
 
   def _info(self):
-    if self.builder_config.selection == 'rgb':
+    if self.builder_config.selection == 'rgb':  # pyrefly: ignore[missing-attribute]
       features = tfds.features.FeaturesDict({
           'image': tfds.features.Image(shape=[32, 32, 3]),
           'label': tfds.features.ClassLabel(names=_LABELS),
@@ -107,8 +107,8 @@ def _info(self):
       })
       supervised_keys = None
     return self.dataset_info_from_configs(
-        features=features,
-        supervised_keys=supervised_keys,
+        features=features,  # pyrefly: ignore[unbound-name]
+        supervised_keys=supervised_keys,  # pyrefly: ignore[unbound-name]
         homepage='http://doi.org/10.14459/2018MP1454690',
     )
 
@@ -120,17 +120,17 @@ def _split_generators(self, dl_manager):
     })
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 'path': paths['train'],
-                'selection': self.builder_config.selection,
+                'selection': self.builder_config.selection,  # pyrefly: ignore[missing-attribute]
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.VALIDATION,
+            name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 'path': paths['val'],
-                'selection': self.builder_config.selection,
+                'selection': self.builder_config.selection,  # pyrefly: ignore[missing-attribute]
             },
         ),
     ]
@@ -155,7 +155,7 @@ def _generate_examples(self, path, selection):
               'label': np.argmax(label[i]).astype(int),
               'sample_id': i,
           }
-        yield i, record
+        yield i, record  # pyrefly: ignore[unbound-name]
 
 
 def _create_rgb(sen2_bands):
diff --git a/tensorflow_datasets/datasets/speech_commands/speech_commands_dataset_builder.py b/tensorflow_datasets/datasets/speech_commands/speech_commands_dataset_builder.py
index 4b1989b6b24..ec3f31a29f4 100644
--- a/tensorflow_datasets/datasets/speech_commands/speech_commands_dataset_builder.py
+++ b/tensorflow_datasets/datasets/speech_commands/speech_commands_dataset_builder.py
@@ -71,21 +71,21 @@ def _split_generators(self, dl_manager):
 
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 'archive': dl_manager.iter_archive(dl_path),
                 'file_list': train_paths,
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.VALIDATION,
+            name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 'archive': dl_manager.iter_archive(dl_path),
                 'file_list': validation_paths,
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 'archive': dl_manager.iter_archive(dl_test_path),
                 'file_list': None,
@@ -160,12 +160,12 @@ def _split_archive(self, train_archive):
         train_paths.append(path)
 
     # Original validation files did include silence - we add them manually here
-    validation_paths.append(os.path.join(BACKGROUND_NOISE, 'running_tap.wav'))
+    validation_paths.append(os.path.join(BACKGROUND_NOISE, 'running_tap.wav'))  # pyrefly: ignore[unbound-name]
 
     # The paths for the train set is just whichever paths that do not exist in
     # either the test or validation splits.
     train_paths = (
-        set(train_paths) - set(validation_paths) - set(train_test_paths)
+        set(train_paths) - set(validation_paths) - set(train_test_paths)  # pyrefly: ignore[unbound-name]
     )
 
     return train_paths, validation_paths
diff --git a/tensorflow_datasets/datasets/spoken_digit/spoken_digit_dataset_builder.py b/tensorflow_datasets/datasets/spoken_digit/spoken_digit_dataset_builder.py
index 33fbc7c6151..a0973a5b54a 100644
--- a/tensorflow_datasets/datasets/spoken_digit/spoken_digit_dataset_builder.py
+++ b/tensorflow_datasets/datasets/spoken_digit/spoken_digit_dataset_builder.py
@@ -50,7 +50,7 @@ def _split_generators(self, dl_manager):
     # There is no predefined train/val/test split for this dataset.
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN, gen_kwargs={"path": path}
+            name=tfds.Split.TRAIN, gen_kwargs={"path": path}  # pyrefly: ignore[missing-attribute]
         )
     ]
 
diff --git a/tensorflow_datasets/datasets/squad/squad_dataset_builder.py b/tensorflow_datasets/datasets/squad/squad_dataset_builder.py
index 7cffc5bfb6b..cebdeb301f2 100644
--- a/tensorflow_datasets/datasets/squad/squad_dataset_builder.py
+++ b/tensorflow_datasets/datasets/squad/squad_dataset_builder.py
@@ -130,7 +130,7 @@ class Builder(tfds.core.GeneratorBasedBuilder):
   }
 
   def _info(self):
-    if self.builder_config.name == "v1.1":
+    if self.builder_config.name == "v1.1":  # pyrefly: ignore[missing-attribute]
       features_dict = qa_utils.squadlike_features()
     elif self.builder_config.name == "v2.0":
       features_dict = _v2_features()
@@ -148,23 +148,23 @@ def _info(self):
   def _split_generators(self, dl_manager):
     """Returns SplitGenerators."""
     urls_to_download = {
-        "train": os.path.join(_URL, self.builder_config.train_file),
-        "dev": os.path.join(_URL, self.builder_config.dev_file),
+        "train": os.path.join(_URL, self.builder_config.train_file),  # pyrefly: ignore[missing-attribute]
+        "dev": os.path.join(_URL, self.builder_config.dev_file),  # pyrefly: ignore[missing-attribute]
     }
     downloaded_files = dl_manager.download_and_extract(urls_to_download)
 
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"filepath": downloaded_files["train"]},
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.VALIDATION,
+            name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"filepath": downloaded_files["dev"]},
         ),
     ]
 
   def _generate_examples(self, filepath):
-    if self.builder_config.name == "v1.1":
+    if self.builder_config.name == "v1.1":  # pyrefly: ignore[missing-attribute]
       return qa_utils.generate_squadlike_examples(filepath)
     return _generate_v2_examples(filepath)
diff --git a/tensorflow_datasets/datasets/stanford_dogs/stanford_dogs_dataset_builder.py b/tensorflow_datasets/datasets/stanford_dogs/stanford_dogs_dataset_builder.py
index 1dcd2a1e6d0..c271d5c5945 100644
--- a/tensorflow_datasets/datasets/stanford_dogs/stanford_dogs_dataset_builder.py
+++ b/tensorflow_datasets/datasets/stanford_dogs/stanford_dogs_dataset_builder.py
@@ -91,29 +91,29 @@ def parse_mat_file(file_name):
       elif "test" in fname:
         test_list, _ = parse_mat_file(full_file_name)
 
-    self.info.features["label"].names = sorted(label_names)
+    self.info.features["label"].names = sorted(label_names)  # pyrefly: ignore[unbound-name]
 
     for root, _, files in tf.io.gfile.walk(annotation_path):
       # Parsing the XML file which have the image annotations
       for fname in files:
         annotation_file_name = os.path.join(root, fname)
         with tf.io.gfile.GFile(annotation_file_name, "rb") as f:
-          xml_file_list[fname] = ET.parse(f)
+          xml_file_list[fname] = ET.parse(f)  # pyrefly: ignore[unsupported-operation]
 
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "archive": dl_manager.iter_archive(images_path),
-                "file_names": train_list,
+                "file_names": train_list,  # pyrefly: ignore[unbound-name]
                 "annotation_files": xml_file_list,
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "archive": dl_manager.iter_archive(images_path),
-                "file_names": test_list,
+                "file_names": test_list,  # pyrefly: ignore[unbound-name]
                 "annotation_files": xml_file_list,
             },
         ),
diff --git a/tensorflow_datasets/datasets/stanford_online_products/stanford_online_products_dataset_builder.py b/tensorflow_datasets/datasets/stanford_online_products/stanford_online_products_dataset_builder.py
index 8d3f53767dd..1c99b1f36af 100644
--- a/tensorflow_datasets/datasets/stanford_online_products/stanford_online_products_dataset_builder.py
+++ b/tensorflow_datasets/datasets/stanford_online_products/stanford_online_products_dataset_builder.py
@@ -22,7 +22,7 @@
 import tensorflow_datasets.public_api as tfds
 
 _DOWNLOAD_LINK = "ftp://cs.stanford.edu/cs/cvgl/Stanford_Online_Products.zip"
-_SPLITS = {tfds.Split.TRAIN: "Ebay_train", tfds.Split.TEST: "Ebay_test"}
+_SPLITS = {tfds.Split.TRAIN: "Ebay_train", tfds.Split.TEST: "Ebay_test"}  # pyrefly: ignore[missing-attribute]
 
 _SUPER_CLASSES = [
     "bicycle",
diff --git a/tensorflow_datasets/datasets/star_cfq/star_cfq_dataset_builder.py b/tensorflow_datasets/datasets/star_cfq/star_cfq_dataset_builder.py
index 249dedd09fa..0b007ccdf6b 100644
--- a/tensorflow_datasets/datasets/star_cfq/star_cfq_dataset_builder.py
+++ b/tensorflow_datasets/datasets/star_cfq/star_cfq_dataset_builder.py
@@ -314,12 +314,12 @@ def _info(self) -> tfds.core.DatasetInfo:
   def _split_generators(self, dl_manager):
     """Returns SplitGenerators."""
     split_dir = dl_manager.download_and_extract(
-        '%s/%s' % (_DATA_URL, self.builder_config.split_archive_path)
+        '%s/%s' % (_DATA_URL, self.builder_config.split_archive_path)  # pyrefly: ignore[missing-attribute]
     )
-    split_path = os.path.join(split_dir, self.builder_config.split_path)
+    split_path = os.path.join(split_dir, self.builder_config.split_path)  # pyrefly: ignore[missing-attribute]
 
     dataset_paths = {}
-    if self.builder_config.compound_divergence:
+    if self.builder_config.compound_divergence:  # pyrefly: ignore[missing-attribute]
       extracted_dataset_path = dl_manager.download_and_extract(
           _DATA_URL
           + '/datasets/u-cfq-for-divergence-splits-1.0-compact-combined.tar.gz'
@@ -364,7 +364,7 @@ def _split_generators(self, dl_manager):
 
   def _generate_examples(self, dataset_paths, split_path, split_id):
     """Yields examples."""
-    if self.builder_config.compound_divergence:
+    if self.builder_config.compound_divergence:  # pyrefly: ignore[missing-attribute]
       samples_path = os.path.join(dataset_paths[_UCFQ_POOL], 'dataset.json')
       with epath.Path(samples_path).open() as samples_file:
         logging.info('Reading json from %s into memory...', samples_path)
@@ -404,7 +404,7 @@ def _generate_examples(self, dataset_paths, split_path, split_id):
               # be unique for each instance.
               key = f'{dataset}-{absolute_index}-{slice_index}'
               yield key, {
-                  _QUESTION: pool[idx][_QUESTION_FIELD],
-                  _QUERY: pool[idx][_QUERY_FIELD],
+                  _QUESTION: pool[idx][_QUESTION_FIELD],  # pyrefly: ignore[unsupported-operation]
+                  _QUERY: pool[idx][_QUERY_FIELD],  # pyrefly: ignore[unsupported-operation]
               }
             slice_index += 1
diff --git a/tensorflow_datasets/datasets/starcraft_video/starcraft_video_dataset_builder.py b/tensorflow_datasets/datasets/starcraft_video/starcraft_video_dataset_builder.py
index 8341f269fcd..5c376f4d62a 100644
--- a/tensorflow_datasets/datasets/starcraft_video/starcraft_video_dataset_builder.py
+++ b/tensorflow_datasets/datasets/starcraft_video/starcraft_video_dataset_builder.py
@@ -109,8 +109,8 @@ def _info(self):
             "rgb_screen": tfds.features.Video(
                 shape=(
                     None,
-                    self.builder_config.resolution,
-                    self.builder_config.resolution,
+                    self.builder_config.resolution,  # pyrefly: ignore[missing-attribute]
+                    self.builder_config.resolution,  # pyrefly: ignore[missing-attribute]
                     3,
                 )
             ),  # pytype: disable=wrong-arg-types  # gen-stub-imports
@@ -123,9 +123,9 @@ def _info(self):
 
   def _split_generators(self, dl_manager):
     url = DATA_URL_DIR + "%s_%dx%d_png/" % (
-        self.builder_config.map_name,
-        self.builder_config.resolution,
-        self.builder_config.resolution,
+        self.builder_config.map_name,  # pyrefly: ignore[missing-attribute]
+        self.builder_config.resolution,  # pyrefly: ignore[missing-attribute]
+        self.builder_config.resolution,  # pyrefly: ignore[missing-attribute]
     )
 
     urls_to_download = {
@@ -139,7 +139,7 @@ def _split_generators(self, dl_manager):
 
     return [
         tfds.core.SplitGenerator(
-            name=tfds.Split.TRAIN,
+            name=tfds.Split.TRAIN,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={
                 "files": [
                     download
@@ -149,11 +149,11 @@ def _split_generators(self, dl_manager):
             },
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.TEST,
+            name=tfds.Split.TEST,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"files": [downloaded_urls["test"]]},
         ),
         tfds.core.SplitGenerator(
-            name=tfds.Split.VALIDATION,
+            name=tfds.Split.VALIDATION,  # pyrefly: ignore[missing-attribute]
             gen_kwargs={"files": [downloaded_urls["valid"]]},
         ),
     ]