KeyError: "Unable to open object (object 'data' doesn't exist)"

[qixuxiang]

Thanks for your awesome code share!

I run the sem_seg code following readme step by step, but when I runpython train.py --log_dir log6 --test_area 6, there is an error:KeyError: "Unable to open object (object 'data' doesn't exist)", here is details:

Traceback (most recent call last):
  File "train.py", line 70, in <module>
    data_batch, label_batch = provider.loadDataFile(h5_filename)
  File "/usr/Downloads/pointnet/provider.py", line 97, in loadDataFile
    return load_h5(filename)
  File "/usr/Downloads/pointnet/provider.py", line 92, in load_h5
    data = f['data'][:]
  File "h5py/_objects.pyx", line 54, in h5py._objects.with_phil.wrapper
  File "h5py/_objects.pyx", line 55, in h5py._objects.with_phil.wrapper
  File "/usr/local/lib/python2.7/dist-packages/h5py/_hl/group.py", line 177, in __getitem__
    oid = h5o.open(self.id, self._e(name), lapl=self._lapl)
  File "h5py/_objects.pyx", line 54, in h5py._objects.with_phil.wrapper
  File "h5py/_objects.pyx", line 55, in h5py._objects.with_phil.wrapper
  File "h5py/h5o.pyx", line 190, in h5py.h5o.open
KeyError: "Unable to open object (object 'data' doesn't exist)"

I run the code in tensorflow docker 1.7.1-devel-gpu of Python 2.7, and I solve all the problem of dependency.

Looking forward to your response, thanks a lot!

[L-YeZhu]

Have you downloaded or prepared your hdf5 files properly? Usually there will be a folder named "indoor3d_sem_seg_hdf5_data" in folder sem_seg. And there should be two separate parts "data" and "label" in each .h5 file. I would recommend you to check the contents of your hdf5 files.

[dvirginz]

When you run python train.py for the first time it downloads the data for you I have the same error, when running it locally (with sudo privileges) it works, but on a remote server with a virtualenv I have the same error like qixuxiang

[zkm98]

I have the same question while using data from indoor3d_sem_seg_hdf5_data

[vajrag]

I faced this issue too. This can happen if the xxx.h5 files are broken. Try to download the dataset again and train. This worked for me.