Transform state error

[ardila]

@yamins81 Here is the error. It can be reproduced by trying to render image
1105108

of the rosch dataset. i.e.

import dldata.stimulus_sets.synthetic.synthetic_datasets as sd
import numpy as np
preproc = {'crop': None,
            'crop_rand': None,
            'dtype': 'float32',
            'mask': None,
            'mode': 'RGB',
            'normalize': False,
            'resize_to': (256, 256, 3),
            'seed': 0}
dataset = sd.RoschDataset()   
I = dataset.get_images(preproc)
I[1105108]

Assertion failed: !pos.is_nan() at line 381 of panda/src/pgraph/transformState.cxx
Traceback (most recent call last):
  File "render_rosch.py", line 25, in <module>
    render_results = [render_index(i, I) for i in inds[-1000:]]
  File "render_rosch.py", line 14, in render_index
    I[ind]
  File "/mindhive/dicarlolab/u/ardila/virtual_environments/default/src/skdata/skdata/larray.py", line 450, in __getitem__
    obj_item = self.obj[item]
  File "/mindhive/dicarlolab/u/ardila/virtual_environments/default/src/skdata/skdata/larray.py", line 146, in __getitem__
    return self.fn(*[o[idx] for o in self.objs])
  File "/mindhive/dicarlolab/u/ardila/virtual_environments/default/src/genthor/genthor/renderer/imager.py", line 136, in __call__
    raise e
AssertionError: !pos.is_nan() at line 381 of panda/src/pgraph/transformState.cxx

[yamins81]

I haven't been able to reproduce this problem. I'm generating the stimuli (hopefully with all the images this time) and hope to start training shortly.

In terms of figuring out what the problem is, I'm going to look into what the differences are between your setup and mine.