Provided Pre-trained ShapeNet Segmentation Model FAILS to Restore

[iskandergaba]

Hello,

I tried to run the testing on the ShapeNet pre-trained segmentation model that you provided. I followed all the instructions and ran the scripts to download and prepare the data. I also compiled the custom FarthestPointSampling operator successfully.

Then I used this command below to run the testing from pointcnn_seg directory:

./test_shapenet.sh -g 0 -x shapenet_x8_2048_fps -l ../../models/seg/pointcnn_seg_shapenet_x8_2048_fps_xxxx/ckpts/iter-544500 -r 10

Needless to say, I put the checkpoint files in this directory (relative to pointcnn_seg):

../../models/seg/pointcnn_seg_shapenet_x8_2048_fps_xxxx/ckpts/

The model apparently fails to load, giving errors regarding FPS and stating that:

Restoring from checkpoint failed. This is most likely due to a mismatch between the current graph and the graph from the checkpoint. Please ensure that you have not altered the graph expected based on the checkpoint.

I do not understand. Could it be a problem with the provided model? Or am I doing anything wrong?
I would appreciate any insight on this.

Thank you.

Full stack trace:

Traceback (most recent call last):
  File "/home/igaba/miniconda2/envs/pointcnn/lib/python3.7/site-packages/tensorflow/python/client/session.py", line 1334, in _do_call
    return fn(*args)
  File "/home/igaba/miniconda2/envs/pointcnn/lib/python3.7/site-packages/tensorflow/python/client/session.py", line 1317, in _run_fn
    self._extend_graph()
  File "/home/igaba/miniconda2/envs/pointcnn/lib/python3.7/site-packages/tensorflow/python/client/session.py", line 1352, in _extend_graph
    tf_session.ExtendSession(self._session)
tensorflow.python.framework.errors_impl.InvalidArgumentError: No OpKernel was registered to support Op 'FarthestPointSample' used by {{node FarthestPointSample}}with these attrs: [npoint=768]
Registered devices: [CPU, XLA_CPU]
Registered kernels:
  device='GPU'

	 [[{{node FarthestPointSample}}]]

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/home/igaba/miniconda2/envs/pointcnn/lib/python3.7/site-packages/tensorflow/python/training/saver.py", line 1276, in restore
    {self.saver_def.filename_tensor_name: save_path})
  File "/home/igaba/miniconda2/envs/pointcnn/lib/python3.7/site-packages/tensorflow/python/client/session.py", line 929, in run
    run_metadata_ptr)
  File "/home/igaba/miniconda2/envs/pointcnn/lib/python3.7/site-packages/tensorflow/python/client/session.py", line 1152, in _run
    feed_dict_tensor, options, run_metadata)
  File "/home/igaba/miniconda2/envs/pointcnn/lib/python3.7/site-packages/tensorflow/python/client/session.py", line 1328, in _do_run
    run_metadata)
  File "/home/igaba/miniconda2/envs/pointcnn/lib/python3.7/site-packages/tensorflow/python/client/session.py", line 1348, in _do_call
    raise type(e)(node_def, op, message)
tensorflow.python.framework.errors_impl.InvalidArgumentError: No OpKernel was registered to support Op 'FarthestPointSample' used by node FarthestPointSample (defined at <string>:148) with these attrs: [npoint=768]
Registered devices: [CPU, XLA_CPU]
Registered kernels:
  device='GPU'

	 [[node FarthestPointSample (defined at <string>:148) ]]

Caused by op 'FarthestPointSample', defined at:
  File "../test_shapenet_seg.py", line 156, in <module>
    main()
  File "../test_shapenet_seg.py", line 92, in main
    net = model.Net(points_sampled, features_sampled, is_training, setting)
  File "/home/igaba/miniconda2/envs/pointcnn/src/pointcnn_seg.py", line 11, in __init__
    PointCNN.__init__(self, points, features, is_training, setting)
  File "/home/igaba/miniconda2/envs/pointcnn/src/pointcnn.py", line 93, in __init__
    fps_indices = tf_sampling.farthest_point_sample(P, pts)
  File "/home/igaba/miniconda2/envs/pointcnn/src/sampling/tf_sampling.py", line 59, in farthest_point_sample
    return sampling_module.farthest_point_sample(inp, npoint)
  File "<string>", line 148, in farthest_point_sample
  File "/home/igaba/miniconda2/envs/pointcnn/lib/python3.7/site-packages/tensorflow/python/framework/op_def_library.py", line 788, in _apply_op_helper
    op_def=op_def)
  File "/home/igaba/miniconda2/envs/pointcnn/lib/python3.7/site-packages/tensorflow/python/util/deprecation.py", line 507, in new_func
    return func(*args, **kwargs)
  File "/home/igaba/miniconda2/envs/pointcnn/lib/python3.7/site-packages/tensorflow/python/framework/ops.py", line 3300, in create_op
    op_def=op_def)
  File "/home/igaba/miniconda2/envs/pointcnn/lib/python3.7/site-packages/tensorflow/python/framework/ops.py", line 1801, in __init__
    self._traceback = tf_stack.extract_stack()

InvalidArgumentError (see above for traceback): No OpKernel was registered to support Op 'FarthestPointSample' used by node FarthestPointSample (defined at <string>:148) with these attrs: [npoint=768]
Registered devices: [CPU, XLA_CPU]
Registered kernels:
  device='GPU'

	 [[node FarthestPointSample (defined at <string>:148) ]]


During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "../test_shapenet_seg.py", line 156, in <module>
    main()
  File "../test_shapenet_seg.py", line 103, in main
    saver.restore(sess, args.load_ckpt)
  File "/home/igaba/miniconda2/envs/pointcnn/lib/python3.7/site-packages/tensorflow/python/training/saver.py", line 1312, in restore
    err, "a mismatch between the current graph and the graph")
tensorflow.python.framework.errors_impl.InvalidArgumentError: Restoring from checkpoint failed. This is most likely due to a mismatch between the current graph and the graph from the checkpoint. Please ensure that you have not altered the graph expected based on the checkpoint. Original error:

No OpKernel was registered to support Op 'FarthestPointSample' used by node FarthestPointSample (defined at <string>:148) with these attrs: [npoint=768]
Registered devices: [CPU, XLA_CPU]
Registered kernels:
  device='GPU'

	 [[node FarthestPointSample (defined at <string>:148) ]]

Caused by op 'FarthestPointSample', defined at:
  File "../test_shapenet_seg.py", line 156, in <module>
    main()
  File "../test_shapenet_seg.py", line 92, in main
    net = model.Net(points_sampled, features_sampled, is_training, setting)
  File "/home/igaba/miniconda2/envs/pointcnn/src/pointcnn_seg.py", line 11, in __init__
    PointCNN.__init__(self, points, features, is_training, setting)
  File "/home/igaba/miniconda2/envs/pointcnn/src/pointcnn.py", line 93, in __init__
    fps_indices = tf_sampling.farthest_point_sample(P, pts)
  File "/home/igaba/miniconda2/envs/pointcnn/src/sampling/tf_sampling.py", line 59, in farthest_point_sample
    return sampling_module.farthest_point_sample(inp, npoint)
  File "<string>", line 148, in farthest_point_sample
  File "/home/igaba/miniconda2/envs/pointcnn/lib/python3.7/site-packages/tensorflow/python/framework/op_def_library.py", line 788, in _apply_op_helper
    op_def=op_def)
  File "/home/igaba/miniconda2/envs/pointcnn/lib/python3.7/site-packages/tensorflow/python/util/deprecation.py", line 507, in new_func
    return func(*args, **kwargs)
  File "/home/igaba/miniconda2/envs/pointcnn/lib/python3.7/site-packages/tensorflow/python/framework/ops.py", line 3300, in create_op
    op_def=op_def)
  File "/home/igaba/miniconda2/envs/pointcnn/lib/python3.7/site-packages/tensorflow/python/framework/ops.py", line 1801, in __init__
    self._traceback = tf_stack.extract_stack()

InvalidArgumentError (see above for traceback): Restoring from checkpoint failed. This is most likely due to a mismatch between the current graph and the graph from the checkpoint. Please ensure that you have not altered the graph expected based on the checkpoint. Original error:

No OpKernel was registered to support Op 'FarthestPointSample' used by node FarthestPointSample (defined at <string>:148) with these attrs: [npoint=768]
Registered devices: [CPU, XLA_CPU]
Registered kernels:
  device='GPU'

	 [[node FarthestPointSample (defined at <string>:148) ]]

[aniketpant83]

Hey! did you get an answer for the above?

[iskandergaba]

Hey! did you get an answer for the above?

Hi, nope.