File "/LAV/lav/utils/datasets/lidar_painted_dataset.py", line 27, in __getitem__
lidar_painted = self.__class__.access('lidar_sem', lmdb_txn, index, 1).reshape(-1,len(self.seg_channels))
File "/LAV/lav/utils/datasets/basic_dataset.py", line 83, in <listcomp>
return np.stack([np.frombuffer(lmdb_txn.get((f'{tag}_{t:05d}{suffix}').encode()), dtype) for t in range(index,index+T)])
TypeError: a bytes-like object is required, not 'NoneType'
I also meet this issue, should I train the model depend your all the step ?
how about the dataset requirement ,I just download 5134 frames,is this too little to run a robust model?
continue the training step without a break
otherwise, there may be an error like here:
and these issues related: https://github.com/dotchen/LAV/issues/2#issuecomment-1148136434 https://github.com/dotchen/LAV/issues/9#issue-1194395147