[pymvpa] ValueError: FlattenMapper has not been trained for data shape '(47, 84, 80)' (known only '(50, 85, 80)').

[Stephanie A. Gagnon]

When I try to load an ROI mask made by 3DAFNItoNIFTI, I get this error:

loading dataset
Traceback (most recent call last):
  File "ROI_mask_wfeaturesel.py", line 8, in <module>
    ds = fmri_dataset(samples='tcat_bold_detrended.nii', targets=attr.targets,  
chunks=attr.chunks, mask='14mar10tc_divided_ROI_mask.nii')
  File "/usr/local/lib/python2.6/dist-packages/mvpa/datasets/mri.py", line 172,  
in fmri_dataset
    flatmask = ds.a.mapper.forward1(mask)
  File "/usr/local/lib/python2.6/dist-packages/mvpa/mappers/base.py", line 232,  
in forward1
    return self.forward(data[np.newaxis])[0]
  File "/usr/local/lib/python2.6/dist-packages/mvpa/mappers/base.py", line 221,  
in forward
    return self._forward_data(data)
  File "/usr/local/lib/python2.6/dist-packages/mvpa/mappers/flatten.py", line 107,  
in _forward_data
    % (str(sshape), str(oshape)))
ValueError: FlattenMapper has not been trained for data shape '(47, 84, 80)' (known  
only '(50, 85, 80)').


.py SCRIPT:
print "loading dataset"
#load nifti dataset
attr = SampleAttributes('31mar10js_contextvspace_onsets_allnumbers.txt')
ds = fmri_dataset(samples='tcat_bold_detrended.nii', targets=attr.targets, chunks=attr.chunks,  
mask='ROI_mask.nii')
print "Dataset"
print ds


Thanks,
Steph