ValueError in demo_polyrnn

[Suvi-dha]

I am receiving following error while initializing polygon rnn m odel in cell 5 line 2.
graph_def is invalid at node 'GatherTree': Input types mismatch (expected 'int32, int32, int32, int32' but got 'int32, int32, int32').

Traceback of the error is:
`ValueError                                Traceback (most recent call last)
<ipython-input-5-680ad34db7a9> in <module>()
      1 #Initializing and restoring PolyRNN++
----> 2 model = PolygonModel(PolyRNN_metagraph, polyGraph)
      3 model.register_eval_fn(lambda input_: evaluator.do_test(evalSess, input_))
      4 polySess = tf.Session(config=tf.ConfigProto(
      5     allow_soft_placement=True

D:\SUVIDHA\polyrnn-pp-master\src\PolygonModel.py in __init__(self, meta_graph_path, graph)
     30         self.saver = None
     31         self.eval_pred_fn = None
---> 32         self._restore_graph(meta_graph_path)
     33 
     34     def _restore_graph(self, meta_graph_path):

D:\SUVIDHA\polyrnn-pp-master\src\PolygonModel.py in _restore_graph(self, meta_graph_path)
     34     def _restore_graph(self, meta_graph_path):
     35         with self.graph.as_default():
---> 36             self.saver = tf.train.import_meta_graph(meta_graph_path, clear_devices=True)
     37 
     38     def _prediction(self):

~\Anaconda3\lib\site-packages\tensorflow\python\training\saver.py in import_meta_graph(meta_graph_or_file, clear_devices, import_scope, **kwargs)
   1925                                       clear_devices=clear_devices,
   1926                                       import_scope=import_scope,
-> 1927                                       **kwargs)
   1928   if meta_graph_def.HasField("saver_def"):
   1929     return Saver(saver_def=meta_graph_def.saver_def, name=import_scope)

~\Anaconda3\lib\site-packages\tensorflow\python\framework\meta_graph.py in import_scoped_meta_graph(meta_graph_or_file, clear_devices, graph, import_scope, input_map, unbound_inputs_col_name, restore_collections_predicate)
    739     importer.import_graph_def(
    740         input_graph_def, name=(import_scope or ""), input_map=input_map,
--> 741         producer_op_list=producer_op_list)
    742 
    743     # Restores all the other collections.

~\Anaconda3\lib\site-packages\tensorflow\python\util\deprecation.py in new_func(*args, **kwargs)
    430                 'in a future version' if date is None else ('after %s' % date),
    431                 instructions)
--> 432       return func(*args, **kwargs)
    433     return tf_decorator.make_decorator(func, new_func, 'deprecated',
    434                                        _add_deprecated_arg_notice_to_docstring(

~\Anaconda3\lib\site-packages\tensorflow\python\framework\importer.py in import_graph_def(graph_def, input_map, return_elements, name, op_dict, producer_op_list)
    678                   'Input types mismatch (expected %r but got %r)'
    679                   % (', '.join(dtypes.as_dtype(x).name for x in input_types),
--> 680                      ', '.join(x.name for x in op._input_types))))
    681         # pylint: enable=protected-access
    682 

ValueError: graph_def is invalid at node 'GatherTree': Input types mismatch (expected 'int32, int32, int32, int32' but got 'int32, int32, int32').`

I have setup the environment in anaconda on windows machine.

[David Acuna]

Please check that you are using tf 1.3.
Also check issues 2 and 3

Thanks

[Suvi-dha]

ok, solved, thanks. Hoping the training code will be released soon. I wan to try on my custom dataset.