[Feature]: Add image_only param (#1613)

* [Feature]: Add image_only param

* [Feature]: Use image_only

mmpretrain/datasets/scienceqa.py

@@ -23,6 +23,8 @@ class ScienceQA(BaseDataset):
         split_file (str): The split file of dataset, which contains the
             ids of data samples in the split.
         ann_file (str): Annotation file path.
+        image_only (bool): Whether only to load data with image. Defaults to
+            False.
         data_prefix (dict): Prefix for data field. Defaults to
             ``dict(img_path='')``.
         pipeline (Sequence): Processing pipeline. Defaults to an empty tuple.
@@ -34,15 +36,16 @@ class ScienceQA(BaseDataset):
                  split: str,
                  split_file: str,
                  ann_file: str,
+                 image_only: bool = False,
                  data_prefix: dict = dict(img_path=''),
                  pipeline: Sequence[Callable] = (),
                  **kwargs):
-
         assert split in [
             'train', 'val', 'test', 'trainval', 'minival', 'minitest'
         ], f'Invalid split {split}'
         self.split = split
         self.split_file = os.path.join(data_root, split_file)
+        self.image_only = image_only
 
         super().__init__(
             data_root=data_root,
@@ -62,6 +65,8 @@ class ScienceQA(BaseDataset):
         data_list = []
         for data_id in current_data_split:
             ann = annotations[data_id]
+            if self.image_only and ann['image'] is None:
+                continue
             data_info = {
                 'image_id':
                 data_id,