Can't load Oxford Pets dataset due to NotImplementedError

[vfdev-5]

Short description

The following code is not working:

dataset, info = tfds.load('oxford_iiit_pet:4.*.*', with_info=True, download=True, data_dir="/tmp/data")

Error message:

File [/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/file_adapters.py:301](http://localhost:2211/lab/tree/jax-ai-stack/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/file_adapters.py#line=300), in ArrayRecordFileAdapter.make_tf_data(cls, filename, buffer_size)
    294 @classmethod
    295 def make_tf_data(
    296     cls,
    297     filename: epath.PathLike,
    298     buffer_size: int | None = None,
    299 ) -> tf.data.Dataset:
    300   """Returns TensorFlow Dataset comprising given array record file."""
--> 301   raise NotImplementedError(
    302       '`.as_dataset()` not implemented for ArrayRecord files. Please, use'
    303       ' `.as_data_source()`.'
    304   )

NotImplementedError: `.as_dataset()` not implemented for ArrayRecord files. Please, use `.as_data_source()`.

Environment information

• Operating System: ubuntu

• Python version: 3.11.9

• tensorflow-datasets/tfds-nightly version: tfds-nightly 4.9.7+nightly

• tensorflow/tf-nightly version: tf-nightly 2.19.0-dev20241105

• Does the issue still exists with the last tfds-nightly package (pip install --upgrade tfds-nightly) ?

yes

Reproduction instructions

import tensorflow_datasets as tfds

dataset, info = tfds.load('oxford_iiit_pet:4.*.*', with_info=True, download=True, data_dir="/tmp/data")

Link to logs

Error message

---------------------------------------------------------------------------
NotImplementedError                       Traceback (most recent call last)
Cell In[3], line 2
      1 get_ipython().system('mkdir -p [/tmp/data](http://localhost:2213/tmp/data)')
----> 2 dataset, info = tfds.load('oxford_iiit_pet:4.*.*', with_info=True, download=True, data_dir="[/tmp/data](http://localhost:2213/tmp/data)")

File [/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/logging/__init__.py:176](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/logging/__init__.py#line=175), in _FunctionDecorator.__call__(self, function, instance, args, kwargs)
    174 metadata = self._start_call()
    175 try:
--> 176   return function(*args, **kwargs)
    177 except Exception:
    178   metadata.mark_error()

File [/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/load.py:670](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/load.py#line=669), in load(name, split, data_dir, batch_size, shuffle_files, download, as_supervised, decoders, read_config, with_info, builder_kwargs, download_and_prepare_kwargs, as_dataset_kwargs, try_gcs)
    667 as_dataset_kwargs.setdefault('shuffle_files', shuffle_files)
    668 as_dataset_kwargs.setdefault('read_config', read_config)
--> 670 ds = dbuilder.as_dataset(**as_dataset_kwargs)
    671 if with_info:
    672   return ds, dbuilder.info

File [/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/logging/__init__.py:176](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/logging/__init__.py#line=175), in _FunctionDecorator.__call__(self, function, instance, args, kwargs)
    174 metadata = self._start_call()
    175 try:
--> 176   return function(*args, **kwargs)
    177 except Exception:
    178   metadata.mark_error()

File [/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/dataset_builder.py:1025](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/dataset_builder.py#line=1024), in DatasetBuilder.as_dataset(self, split, batch_size, shuffle_files, decoders, read_config, as_supervised)
   1016 # Create a dataset for each of the given splits
   1017 build_single_dataset = functools.partial(
   1018     self._build_single_dataset,
   1019     shuffle_files=shuffle_files,
   (...)
   1023     as_supervised=as_supervised,
   1024 )
-> 1025 all_ds = tree.map_structure(build_single_dataset, split)
   1026 return all_ds

File [/opt/conda/lib/python3.11/site-packages/tree/__init__.py:435](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tree/__init__.py#line=434), in map_structure(func, *structures, **kwargs)
    432 for other in structures[1:]:
    433   assert_same_structure(structures[0], other, check_types=check_types)
    434 return unflatten_as(structures[0],
--> 435                     [func(*args) for args in zip(*map(flatten, structures))])

File [/opt/conda/lib/python3.11/site-packages/tree/__init__.py:435](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tree/__init__.py#line=434), in <listcomp>(.0)
    432 for other in structures[1:]:
    433   assert_same_structure(structures[0], other, check_types=check_types)
    434 return unflatten_as(structures[0],
--> 435                     [func(*args) for args in zip(*map(flatten, structures))])

File [/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/dataset_builder.py:1043](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/dataset_builder.py#line=1042), in DatasetBuilder._build_single_dataset(self, split, batch_size, shuffle_files, decoders, read_config, as_supervised)
   1040   batch_size = self.info.splits.total_num_examples or sys.maxsize
   1042 # Build base dataset
-> 1043 ds = self._as_dataset(
   1044     split=split,
   1045     shuffle_files=shuffle_files,
   1046     decoders=decoders,
   1047     read_config=read_config,
   1048 )
   1049 # Auto-cache small datasets which are small enough to fit in memory.
   1050 if self._should_cache_ds(
   1051     split=split, shuffle_files=shuffle_files, read_config=read_config
   1052 ):

File [/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/dataset_builder.py:1497](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/dataset_builder.py#line=1496), in FileReaderBuilder._as_dataset(self, split, decoders, read_config, shuffle_files)
   1491 reader = reader_lib.Reader(
   1492     self.data_dir,
   1493     example_specs=example_specs,
   1494     file_format=self.info.file_format,
   1495 )
   1496 decode_fn = functools.partial(features.decode_example, decoders=decoders)
-> 1497 return reader.read(
   1498     instructions=split,
   1499     split_infos=self.info.splits.values(),
   1500     decode_fn=decode_fn,
   1501     read_config=read_config,
   1502     shuffle_files=shuffle_files,
   1503     disable_shuffling=self.info.disable_shuffling,
   1504 )

File [/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/reader.py:430](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/reader.py#line=429), in Reader.read(self, instructions, split_infos, read_config, shuffle_files, disable_shuffling, decode_fn)
    421   file_instructions = splits_dict[instruction].file_instructions
    422   return self.read_files(
    423       file_instructions,
    424       read_config=read_config,
   (...)
    427       decode_fn=decode_fn,
    428   )
--> 430 return tree.map_structure(_read_instruction_to_ds, instructions)

File [/opt/conda/lib/python3.11/site-packages/tree/__init__.py:435](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tree/__init__.py#line=434), in map_structure(func, *structures, **kwargs)
    432 for other in structures[1:]:
    433   assert_same_structure(structures[0], other, check_types=check_types)
    434 return unflatten_as(structures[0],
--> 435                     [func(*args) for args in zip(*map(flatten, structures))])

File [/opt/conda/lib/python3.11/site-packages/tree/__init__.py:435](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tree/__init__.py#line=434), in <listcomp>(.0)
    432 for other in structures[1:]:
    433   assert_same_structure(structures[0], other, check_types=check_types)
    434 return unflatten_as(structures[0],
--> 435                     [func(*args) for args in zip(*map(flatten, structures))])

File [/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/reader.py:422](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/reader.py#line=421), in Reader.read.<locals>._read_instruction_to_ds(instruction)
    420 def _read_instruction_to_ds(instruction):
    421   file_instructions = splits_dict[instruction].file_instructions
--> 422   return self.read_files(
    423       file_instructions,
    424       read_config=read_config,
    425       shuffle_files=shuffle_files,
    426       disable_shuffling=disable_shuffling,
    427       decode_fn=decode_fn,
    428   )

File [/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/reader.py:462](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/reader.py#line=461), in Reader.read_files(self, file_instructions, read_config, shuffle_files, disable_shuffling, decode_fn)
    459   raise ValueError(msg)
    461 # Read serialized example (eventually with `tfds_id`)
--> 462 ds = _read_files(
    463     file_instructions=file_instructions,
    464     read_config=read_config,
    465     shuffle_files=shuffle_files,
    466     disable_shuffling=disable_shuffling,
    467     file_format=self._file_format,
    468 )
    470 # Parse and decode
    471 def parse_and_decode(ex: Tensor) -> TreeDict[Tensor]:
    472   # TODO(pierrot): `parse_example` uses
    473   # `tf.io.parse_single_example`. It might be faster to use `parse_example`,
    474   # after batching.
    475   # https://www.tensorflow.org/api_docs/python/tf/io/parse_example

File /opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/reader.py:302, in _read_files(file_instructions, read_config, shuffle_files, disable_shuffling, file_format)
    295 if (
    296     shuffle_files
    297     and read_config.shuffle_seed is None
    298     and tf_compat.get_option_deterministic(read_config.options) is None
    299 ):
    300   deterministic = False
--> 302 ds = instruction_ds.interleave(
    303     functools.partial(
    304         _get_dataset_from_filename,
    305         do_skip=do_skip,
    306         do_take=do_take,
    307         file_format=file_format,
    308         add_tfds_id=read_config.add_tfds_id,
    309         override_buffer_size=read_config.override_buffer_size,
    310     ),
    311     cycle_length=cycle_length,
    312     block_length=block_length,
    313     num_parallel_calls=read_config.num_parallel_calls_for_interleave_files,
    314     deterministic=deterministic,
    315 )
    317 return assert_cardinality_and_apply_options(ds)

File [/opt/conda/lib/python3.11/site-packages/tensorflow/python/data/ops/dataset_ops.py:2534](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow/python/data/ops/dataset_ops.py#line=2533), in DatasetV2.interleave(self, map_func, cycle_length, block_length, num_parallel_calls, deterministic, name)
   2530 # Loaded lazily due to a circular dependency (
   2531 # dataset_ops -> interleave_op -> dataset_ops).
   2532 # pylint: disable=g-import-not-at-top,protected-access
   2533 from tensorflow.python.data.ops import interleave_op
-> 2534 return interleave_op._interleave(self, map_func, cycle_length, block_length,
   2535                                  num_parallel_calls, deterministic, name)

File [/opt/conda/lib/python3.11/site-packages/tensorflow/python/data/ops/interleave_op.py:49](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow/python/data/ops/interleave_op.py#line=48), in _interleave(input_dataset, map_func, cycle_length, block_length, num_parallel_calls, deterministic, name)
     46   return _InterleaveDataset(
     47       input_dataset, map_func, cycle_length, block_length, name=name)
     48 else:
---> 49   return _ParallelInterleaveDataset(
     50       input_dataset,
     51       map_func,
     52       cycle_length,
     53       block_length,
     54       num_parallel_calls,
     55       deterministic=deterministic,
     56       name=name)

File [/opt/conda/lib/python3.11/site-packages/tensorflow/python/data/ops/interleave_op.py:119](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow/python/data/ops/interleave_op.py#line=118), in _ParallelInterleaveDataset.__init__(self, input_dataset, map_func, cycle_length, block_length, num_parallel_calls, buffer_output_elements, prefetch_input_elements, deterministic, name)
    117 """See `Dataset.interleave()` for details."""
    118 self._input_dataset = input_dataset
--> 119 self._map_func = structured_function.StructuredFunctionWrapper(
    120     map_func, self._transformation_name(), dataset=input_dataset)
    121 if not isinstance(self._map_func.output_structure, dataset_ops.DatasetSpec):
    122   raise TypeError(
    123       "The `map_func` argument must return a `Dataset` object. Got "
    124       f"{dataset_ops.get_type(self._map_func.output_structure)!r}.")

File [/opt/conda/lib/python3.11/site-packages/tensorflow/python/data/ops/structured_function.py:265](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow/python/data/ops/structured_function.py#line=264), in StructuredFunctionWrapper.__init__(self, func, transformation_name, dataset, input_classes, input_shapes, input_types, input_structure, add_to_graph, use_legacy_function, defun_kwargs)
    258       warnings.warn(
    259           "Even though the `tf.config.experimental_run_functions_eagerly` "
    260           "option is set, this option does not apply to tf.data functions. "
    261           "To force eager execution of tf.data functions, please use "
    262           "`tf.data.experimental.enable_debug_mode()`.")
    263     fn_factory = trace_tf_function(defun_kwargs)
--> 265 self._function = fn_factory()
    266 # There is no graph to add in eager mode.
    267 add_to_graph &= not context.executing_eagerly()

File [/opt/conda/lib/python3.11/site-packages/tensorflow/python/eager/polymorphic_function/polymorphic_function.py:1256](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow/python/eager/polymorphic_function/polymorphic_function.py#line=1255), in Function.get_concrete_function(self, *args, **kwargs)
   1254 def get_concrete_function(self, *args, **kwargs):
   1255   # Implements PolymorphicFunction.get_concrete_function.
-> 1256   concrete = self._get_concrete_function_garbage_collected(*args, **kwargs)
   1257   concrete._garbage_collector.release()  # pylint: disable=protected-access
   1258   return concrete

File [/opt/conda/lib/python3.11/site-packages/tensorflow/python/eager/polymorphic_function/polymorphic_function.py:1226](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow/python/eager/polymorphic_function/polymorphic_function.py#line=1225), in Function._get_concrete_function_garbage_collected(self, *args, **kwargs)
   1224   if self._variable_creation_config is None:
   1225     initializers = []
-> 1226     self._initialize(args, kwargs, add_initializers_to=initializers)
   1227     self._initialize_uninitialized_variables(initializers)
   1229 if self._created_variables:
   1230   # In this case we have created variables on the first call, so we run the
   1231   # version which is guaranteed to never create variables.

File [/opt/conda/lib/python3.11/site-packages/tensorflow/python/eager/polymorphic_function/polymorphic_function.py:696](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow/python/eager/polymorphic_function/polymorphic_function.py#line=695), in Function._initialize(self, args, kwds, add_initializers_to)
    691 self._variable_creation_config = self._generate_scoped_tracing_options(
    692     variable_capturing_scope,
    693     tracing_compilation.ScopeType.VARIABLE_CREATION,
    694 )
    695 # Force the definition of the function for these arguments
--> 696 self._concrete_variable_creation_fn = tracing_compilation.trace_function(
    697     args, kwds, self._variable_creation_config
    698 )
    700 def invalid_creator_scope(*unused_args, **unused_kwds):
    701   """Disables variable creation."""

File [/opt/conda/lib/python3.11/site-packages/tensorflow/python/eager/polymorphic_function/tracing_compilation.py:178](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow/python/eager/polymorphic_function/tracing_compilation.py#line=177), in trace_function(args, kwargs, tracing_options)
    175     args = tracing_options.input_signature
    176     kwargs = {}
--> 178   concrete_function = _maybe_define_function(
    179       args, kwargs, tracing_options
    180   )
    182 if not tracing_options.bind_graph_to_function:
    183   concrete_function._garbage_collector.release()  # pylint: disable=protected-access

File [/opt/conda/lib/python3.11/site-packages/tensorflow/python/eager/polymorphic_function/tracing_compilation.py:283](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow/python/eager/polymorphic_function/tracing_compilation.py#line=282), in _maybe_define_function(args, kwargs, tracing_options)
    281 else:
    282   target_func_type = lookup_func_type
--> 283 concrete_function = _create_concrete_function(
    284     target_func_type, lookup_func_context, func_graph, tracing_options
    285 )
    287 if tracing_options.function_cache is not None:
    288   tracing_options.function_cache.add(
    289       concrete_function, current_func_context
    290   )

File [/opt/conda/lib/python3.11/site-packages/tensorflow/python/eager/polymorphic_function/tracing_compilation.py:310](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow/python/eager/polymorphic_function/tracing_compilation.py#line=309), in _create_concrete_function(function_type, type_context, func_graph, tracing_options)
    303   placeholder_bound_args = function_type.placeholder_arguments(
    304       placeholder_context
    305   )
    307 disable_acd = tracing_options.attributes and tracing_options.attributes.get(
    308     attributes_lib.DISABLE_ACD, False
    309 )
--> 310 traced_func_graph = func_graph_module.func_graph_from_py_func(
    311     tracing_options.name,
    312     tracing_options.python_function,
    313     placeholder_bound_args.args,
    314     placeholder_bound_args.kwargs,
    315     None,
    316     func_graph=func_graph,
    317     add_control_dependencies=not disable_acd,
    318     arg_names=function_type_utils.to_arg_names(function_type),
    319     create_placeholders=False,
    320 )
    322 transform.apply_func_graph_transforms(traced_func_graph)
    324 graph_capture_container = traced_func_graph.function_captures

File [/opt/conda/lib/python3.11/site-packages/tensorflow/python/framework/func_graph.py:1059](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow/python/framework/func_graph.py#line=1058), in func_graph_from_py_func(name, python_func, args, kwargs, signature, func_graph, add_control_dependencies, arg_names, op_return_value, collections, capture_by_value, create_placeholders)
   1056   return x
   1058 _, original_func = tf_decorator.unwrap(python_func)
-> 1059 func_outputs = python_func(*func_args, **func_kwargs)
   1061 # invariant: `func_outputs` contains only Tensors, CompositeTensors,
   1062 # TensorArrays and `None`s.
   1063 func_outputs = variable_utils.convert_variables_to_tensors(func_outputs)

File [/opt/conda/lib/python3.11/site-packages/tensorflow/python/eager/polymorphic_function/polymorphic_function.py:599](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow/python/eager/polymorphic_function/polymorphic_function.py#line=598), in Function._generate_scoped_tracing_options.<locals>.wrapped_fn(*args, **kwds)
    595 with default_graph._variable_creator_scope(scope, priority=50):  # pylint: disable=protected-access
    596   # __wrapped__ allows AutoGraph to swap in a converted function. We give
    597   # the function a weak reference to itself to avoid a reference cycle.
    598   with OptionalXlaContext(compile_with_xla):
--> 599     out = weak_wrapped_fn().__wrapped__(*args, **kwds)
    600   return out

File [/opt/conda/lib/python3.11/site-packages/tensorflow/python/data/ops/structured_function.py:231](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow/python/data/ops/structured_function.py#line=230), in StructuredFunctionWrapper.__init__.<locals>.trace_tf_function.<locals>.wrapped_fn(*args)
    230 def wrapped_fn(*args):  # pylint: disable=missing-docstring
--> 231   ret = wrapper_helper(*args)
    232   ret = structure.to_tensor_list(self._output_structure, ret)
    233   return [ops.convert_to_tensor(t) for t in ret]

File [/opt/conda/lib/python3.11/site-packages/tensorflow/python/data/ops/structured_function.py:161](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow/python/data/ops/structured_function.py#line=160), in StructuredFunctionWrapper.__init__.<locals>.wrapper_helper(*args)
    159 if not _should_unpack(nested_args):
    160   nested_args = (nested_args,)
--> 161 ret = autograph.tf_convert(self._func, ag_ctx)(*nested_args)
    162 ret = variable_utils.convert_variables_to_tensors(ret)
    163 if _should_pack(ret):

File [/opt/conda/lib/python3.11/site-packages/tensorflow/python/autograph/impl/api.py:690](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow/python/autograph/impl/api.py#line=689), in convert.<locals>.decorator.<locals>.wrapper(*args, **kwargs)
    688 try:
    689   with conversion_ctx:
--> 690     return converted_call(f, args, kwargs, options=options)
    691 except Exception as e:  # pylint:disable=broad-except
    692   if hasattr(e, 'ag_error_metadata'):

File [/opt/conda/lib/python3.11/site-packages/tensorflow/python/autograph/impl/api.py:352](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow/python/autograph/impl/api.py#line=351), in converted_call(f, args, kwargs, caller_fn_scope, options)
    349   new_args = f.args + args
    350   logging.log(3, 'Forwarding call of partial %s with\n%s\n%s\n', f, new_args,
    351               new_kwargs)
--> 352   return converted_call(
    353       f.func,
    354       new_args,
    355       new_kwargs,
    356       caller_fn_scope=caller_fn_scope,
    357       options=options)
    359 if inspect_utils.isbuiltin(f):
    360   if f is eval:

File [/opt/conda/lib/python3.11/site-packages/tensorflow/python/autograph/impl/api.py:377](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow/python/autograph/impl/api.py#line=376), in converted_call(f, args, kwargs, caller_fn_scope, options)
    374   return _call_unconverted(f, args, kwargs, options)
    376 if not options.user_requested and conversion.is_allowlisted(f):
--> 377   return _call_unconverted(f, args, kwargs, options)
    379 # internal_convert_user_code is for example turned off when issuing a dynamic
    380 # call conversion from generated code while in nonrecursive mode. In that
    381 # case we evidently don't want to recurse, but we still have to convert
    382 # things like builtins.
    383 if not options.internal_convert_user_code:

File [/opt/conda/lib/python3.11/site-packages/tensorflow/python/autograph/impl/api.py:459](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow/python/autograph/impl/api.py#line=458), in _call_unconverted(f, args, kwargs, options, update_cache)
    456   return f.__self__.call(args, kwargs)
    458 if kwargs is not None:
--> 459   return f(*args, **kwargs)
    460 return f(*args)

File [/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/reader.py:69](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/reader.py#line=68), in _get_dataset_from_filename(instruction, do_skip, do_take, file_format, add_tfds_id, override_buffer_size)
     60 def _get_dataset_from_filename(
     61     instruction: _Instruction,
     62     do_skip: bool,
   (...)
     66     override_buffer_size: Optional[int] = None,
     67 ) -> tf.data.Dataset:
     68   """Returns a tf.data.Dataset instance from given instructions."""
---> 69   ds = file_adapters.ADAPTER_FOR_FORMAT[file_format].make_tf_data(
     70       instruction.filepath, buffer_size=override_buffer_size
     71   )
     72   if do_skip:
     73     ds = ds.skip(instruction.skip)

File [/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/file_adapters.py:301](http://localhost:2213/opt/conda/lib/python3.11/site-packages/tensorflow_datasets/core/file_adapters.py#line=300), in ArrayRecordFileAdapter.make_tf_data(cls, filename, buffer_size)
    294 @classmethod
    295 def make_tf_data(
    296     cls,
    297     filename: epath.PathLike,
    298     buffer_size: int | None = None,
    299 ) -> tf.data.Dataset:
    300   """Returns TensorFlow Dataset comprising given array record file."""
--> 301   raise NotImplementedError(
    302       '`.as_dataset()` not implemented for ArrayRecord files. Please, use'
    303       ' `.as_data_source()`.'
    304   )

NotImplementedError: `.as_dataset()` not implemented for ArrayRecord files. Please, use `.as_data_source()`.

[dsaha21]

Hi Team,

Just like @vfdev-5 mentioned, I am also facing the same issue. Also adding the issue thread #5416

[fineguy]

@vfdev-5 @dsaha21 sorry for the late reply, but could you try updating to the most recent version with pip install --upgrade tfds-nightly? I think this issues has been fixed.

[dsaha21]

Hi @fineguy,

I tried to perform just like you mentioned above.

However, when I am splitting the dataset into train and test, the train shape is coming like the following. I dont know might be I am making a mistake. Can you please verify once if possible. Thanks

[fineguy]

@dsaha21 did you expect a different result? It seems to be correct according to https://www.tensorflow.org/datasets/api_docs/python/tfds/load#args

[dsaha21]

Okay @fineguy , Thanks a lot for the help 👍