IndexError: index 64 is out of bounds for dimension 0 with size 64

[chenming1656161936]

File "/code/train.py", line 411, in
[2025-07-22 16:32:14] main()
[2025-07-22 16:32:14] File "/code/train.py", line 167, in main
[2025-07-22 16:32:14] l0_q, l0_t, l1_q, l1_t, l2_q, l2_t, l3_q, l3_t, pc1_ouput, q_gt, t_gt, w_x, w_q = model(pos2, pos1, T_gt, T_trans, T_trans_inv)
[2025-07-22 16:32:14] File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1553, in _wrapped_call_impl
[2025-07-22 16:32:14] return self._call_impl(*args, **kwargs)
[2025-07-22 16:32:14] File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1562, in _call_impl
[2025-07-22 16:32:14] return forward_call(*args, **kwargs)
[2025-07-22 16:32:14] File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/parallel/data_parallel.py", line 186, in forward
[2025-07-22 16:32:14] outputs = self.parallel_apply(replicas, inputs, module_kwargs)
[2025-07-22 16:32:14] File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/parallel/data_parallel.py", line 201, in parallel_apply
[2025-07-22 16:32:14] return parallel_apply(replicas, inputs, kwargs, self.device_ids[:len(replicas)])
[2025-07-22 16:32:14] File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/parallel/parallel_apply.py", line 109, in parallel_apply
[2025-07-22 16:32:14] output.reraise()
[2025-07-22 16:32:14] File "/root/miniconda3/lib/python3.10/site-packages/torch/_utils.py", line 706, in reraise
[2025-07-22 16:32:14] raise exception
[2025-07-22 16:32:14] IndexError: Caught IndexError in replica 0 on device 0.
[2025-07-22 16:32:14] Original Traceback (most recent call last):
[2025-07-22 16:32:14] File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/parallel/parallel_apply.py", line 84, in _worker
[2025-07-22 16:32:14] output = module(*input, **kwargs)
[2025-07-22 16:32:14] File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1553, in _wrapped_call_impl
[2025-07-22 16:32:14] return self._call_impl(*args, **kwargs)
[2025-07-22 16:32:14] File "/root/miniconda3/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1562, in _call_impl
[2025-07-22 16:32:14] return forward_call(*args, **kwargs)
[2025-07-22 16:32:14] File "/code/regformer_model.py", line 329, in forward
[2025-07-22 16:32:14] input_xyz_aug_f1, input_xyz_aug_f2, q_gt, t_gt = PreProcess(input_xyz_f1, input_xyz_f2, T_gt)
[2025-07-22 16:32:14] File "/code/regformer_model_utils.py", line 275, in PreProcess
[2025-07-22 16:32:14] cur_T_gt = T_gt[i, :, :].to(torch.float32)
[2025-07-22 16:32:14] IndexError: index 64 is out of bounds for dimension 0 with size 64
Dear experts, I've been encountering an error in the for loop inside the preprocess function when doing multi-GPU training. How can I solve this issue? Thank you.

[liujiuming123]

Hi thank you for following our work. Actually, we haven't met this error before, but I suggest you could print the size of T_gt to see what exact problems lie in. And also you could check the parameters used in multi-GPU settings.