I have encountered this question below when I run the demo.py you provided.Best wishes!

[fhjfjhgj]

File "/2TB/download/FreeReg-FFreeReg/tools/YOHO/utils/network.py", line 82, in data_process
data=data[:,:,self.Nei_in_SO3]
IndexError: tensors used as indices must be long, byte or bool tensors

[fhjfjhgj]

raceback (most recent call last):
File "/2TB/download/FreeReg-FFreeReg/demo.py", line 119, in
pipe(img,pcd)
File "/2TB/download/FreeReg-FFreeReg/demo.py", line 92, in call
img_kpts, img_feats = self._extract_yoho(img_pc,self.nkpts)
File "/home/zzy/anaconda3/envs/zzy/lib/python3.8/site-packages/torch/autograd/grad_mode.py", line 27, in decorate_context
return func(*args, **kwargs)
File "/2TB/download/FreeReg-FFreeReg/demo.py", line 63, in _extract_yoho
kpts, yoho_desc, yoho_eqv = self.yoho.run(pc, self.vs, nkpts)
File "/2TB/download/FreeReg-FFreeReg/tools/YOHO/yoho_extract.py", line 148, in run
kpts, feat_inv, feat_eqv = self.extract_features(pc, voxel_size, nkpts=nkpts)
File "/2TB/download/FreeReg-FFreeReg/tools/YOHO/yoho_extract.py", line 134, in extract_features
output = self.network(batch)
File "/home/zzy/anaconda3/envs/zzy/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
return forward_call(*input, **kwargs)
File "/2TB/download/FreeReg-FFreeReg/tools/YOHO/utils/network.py", line 147, in forward
return self.PartI_net(group_feat)
File "/home/zzy/anaconda3/envs/zzy/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
return forward_call(*input, **kwargs)
File "/2TB/download/FreeReg-FFreeReg/tools/YOHO/utils/network.py", line 97, in forward
feats_eqv=self.SO3_Conv(feats)# bn,f,gn
File "/2TB/download/FreeReg-FFreeReg/tools/YOHO/utils/network.py", line 87, in SO3_Conv
data=self.data_process(data)
File "/2TB/download/FreeReg-FFreeReg/tools/YOHO/utils/network.py", line 82, in data_process
data=data[:,:,self.Nei_in_SO3]
IndexError: tensors used as indices must be long, byte or bool tensors

[HpWang-whu]

Hi @fhjfjhgj ,
It is waird, you might try to print self.Nei_in_SO3 to check whether it is correctly loaded.
It should load tools/YOHO/utils/group_related/Nei_Index_in_SO3_ordered_13.npy at here.

Yours,