No need for the new dataset_validate function.

deeplake/core/dataset/dataset.py

@@ -107,7 +107,6 @@
 )
 from deeplake.util.keys import (
     dataset_exists,
-    dataset_validate,
     get_dataset_info_key,
     get_dataset_meta_key,
     tensor_exists,
@@ -1761,7 +1760,6 @@ def diff(
     def _populate_meta(self, verbose=True):
         """Populates the meta information for the dataset."""
         if dataset_exists(self.storage):
-            dataset_validate(self.storage)
             load_meta(self)
 
         elif not self.storage.empty():

deeplake/util/keys.py

@@ -196,24 +196,10 @@ def dataset_exists(storage) -> bool:
         )
     except S3GetAccessError as err:
         raise AuthorizationException("The dataset storage cannot be accessed") from err
-    except S3GetError:
+    except (KeyError, S3GetError) as err:
         return False
 
 
-def dataset_validate(storage) -> None:
-    """
-    Checks the structure of the dataset and throws a descriptive DatasetCorruptError for any problems.
-    """
-    try:
-        for file in [get_dataset_meta_key(FIRST_COMMIT_ID), get_version_control_info_key()]:
-            if file not in storage:
-                raise DatasetCorruptError(f"Invalid dataset: missing file {file}")
-    except S3GetAccessError as err:
-        raise AuthorizationException("The dataset storage cannot be accessed") from err
-    except S3GetError:
-        raise DatasetCorruptError("Error reading from S3")
-
-
 def tensor_exists(key: str, storage, commit_id: str) -> bool:
     try:
         storage[get_tensor_meta_key(key, commit_id)]