Add neuron to the ouput layer

[Niklas Dormagen]

Hi erveryone,
Is it possible to add a default neuron to the output layer? That if the analyzed structure does not correspond to any of the trained ones, I get an output like None

[Chengyuelan]

When I made the H5 data set, I encountered a problem. How could I solve it? I will be very grateful if I receive your reply.
@Alfajesto
/content/drive/Shareddrives/1000/pointnet/pointnet-master/data/stanford_indoor3d/pointcloudscene_0001.npy
Traceback (most recent call last):
File "gen_indoor3d_h5.py", line 77, in
random_sample=False, sample_num=None)
File "/content/drive/Shareddrives/1000/pointnet/pointnet-master/sem_seg/indoor3d_util.py", line 254, in room2blocks_wrapper_normalized
random_sample, sample_num, sample_aug)
File "/content/drive/Shareddrives/1000/pointnet/pointnet-master/sem_seg/indoor3d_util.py", line 230, in room2blocks_plus_normalized
random_sample, sample_num, sample_aug)
File "/content/drive/Shareddrives/1000/pointnet/pointnet-master/sem_seg/indoor3d_util.py", line 190, in room2blocks
return np.concatenate(block_data_list, 0),
File "<array_function internals>", line 6, in concatenate
ValueError: need at least one array to concatenate