NotFoundError: No registered 'ResourceStridedSliceAssign' OpKernel for 'GPU' devices compatible with node {{node ResourceStridedSliceAssign}}

Hi, I 'm using a 2016 Intel based MBP with an AMD Radeon Pro 460 running OS X Monterey 12.1.

When I ran the code here:

https://d2l.ai/chapter_preliminaries/ndarray.html#indexing-and-slicing

X_var = tf.Variable(X)
X_var[1, 2].assign(9)

I got the error output:

 ---------------------------------------------------------------------------
NotFoundError                             Traceback (most recent call last)
/Volumes/WD 1T/Repos/Dive into Deep Learning/tensorflow/chapter_4_preliminaries/ndarray.ipynb Cell 38' in <module>
      3     X_var[1, 2].assign(9)
      4 X_var = tf.Variable(X)
----> 5 X_var[1, 2].assign(9)
      6 X_var

File ~/.conda/envs/tensorflow-metal/lib/python3.8/site-packages/tensorflow/python/ops/array_ops.py:1236, in strided_slice.<locals>.assign(val, name)
   1233 if name is None:
   1234   name = parent_name + "_assign"
-> 1236 return var._strided_slice_assign(
   1237     begin=begin,
   1238     end=end,
   1239     strides=strides,
   1240     value=val,
   1241     name=name,
   1242     begin_mask=begin_mask,
   1243     end_mask=end_mask,
   1244     ellipsis_mask=ellipsis_mask,
   1245     new_axis_mask=new_axis_mask,
   1246     shrink_axis_mask=shrink_axis_mask)

File ~/.conda/envs/tensorflow-metal/lib/python3.8/site-packages/tensorflow/python/ops/resource_variable_ops.py:1407, in BaseResourceVariable._strided_slice_assign(self, begin, end, strides, value, name, begin_mask, end_mask, ellipsis_mask, new_axis_mask, shrink_axis_mask)
   1402 def _strided_slice_assign(self, begin, end, strides, value, name, begin_mask,
   1403                           end_mask, ellipsis_mask, new_axis_mask,
   1404                           shrink_axis_mask):
   1405   with _handle_graph(self.handle), self._assign_dependencies():
   1406     return self._lazy_read(
-> 1407         gen_array_ops.resource_strided_slice_assign(
   1408             ref=self.handle,
   1409             begin=begin,
   1410             end=end,
   1411             strides=strides,
   1412             value=ops.convert_to_tensor(value, dtype=self.dtype),
   1413             name=name,
   1414             begin_mask=begin_mask,
   1415             end_mask=end_mask,
   1416             ellipsis_mask=ellipsis_mask,
   1417             new_axis_mask=new_axis_mask,
   1418             shrink_axis_mask=shrink_axis_mask))

File ~/.conda/envs/tensorflow-metal/lib/python3.8/site-packages/tensorflow/python/ops/gen_array_ops.py:8612, in resource_strided_slice_assign(ref, begin, end, strides, value, begin_mask, end_mask, ellipsis_mask, new_axis_mask, shrink_axis_mask, name)
   8610   return _result
   8611 except _core._NotOkStatusException as e:
-> 8612   _ops.raise_from_not_ok_status(e, name)
   8613 except _core._FallbackException:
   8614   pass

File ~/.conda/envs/tensorflow-metal/lib/python3.8/site-packages/tensorflow/python/framework/ops.py:7107, in raise_from_not_ok_status(e, name)
   7105 def raise_from_not_ok_status(e, name):
   7106   e.message += (" name: " + name if name is not None else "")
-> 7107   raise core._status_to_exception(e) from None

NotFoundError: No registered 'ResourceStridedSliceAssign' OpKernel for 'GPU' devices compatible with node {{node ResourceStridedSliceAssign}}
	 (OpKernel was found, but attributes didn't match) Requested Attributes: Index=DT_INT32, T=DT_FLOAT, begin_mask=0, ellipsis_mask=0, end_mask=0, new_axis_mask=0, shrink_axis_mask=3
	.  Registered:  device='XLA_CPU_JIT'; Index in [DT_INT32, DT_INT64]; T in [DT_FLOAT, DT_DOUBLE, DT_INT32, DT_UINT8, DT_INT16, DT_INT8, DT_COMPLEX64, DT_INT64, DT_BOOL, DT_QINT8, DT_QUINT8, DT_QINT32, DT_BFLOAT16, DT_UINT16, DT_COMPLEX128, DT_HALF, DT_UINT32, DT_UINT64]
  device='DEFAULT'; T in [DT_INT32]
  device='CPU'; T in [DT_UINT64]
  device='CPU'; T in [DT_INT64]
  device='CPU'; T in [DT_UINT32]
  device='CPU'; T in [DT_UINT16]
  device='CPU'; T in [DT_INT16]
  device='CPU'; T in [DT_UINT8]
  device='CPU'; T in [DT_INT8]
  device='CPU'; T in [DT_INT32]
  device='CPU'; T in [DT_HALF]
  device='CPU'; T in [DT_BFLOAT16]
  device='CPU'; T in [DT_FLOAT]
  device='CPU'; T in [DT_DOUBLE]
  device='CPU'; T in [DT_COMPLEX64]
  device='CPU'; T in [DT_COMPLEX128]
  device='CPU'; T in [DT_BOOL]
  device='CPU'; T in [DT_STRING]
  device='CPU'; T in [DT_RESOURCE]
  device='CPU'; T in [DT_VARIANT]
 [Op:ResourceStridedSliceAssign] name: strided_slice/_assign

It seems that the operation is somehow not supported on GPU, I have to specify CPU as the device of the operation and it works.

with tf.device('/CPU:0'):
    X_var = tf.Variable(X)
    X_var[1, 2].assign(9)