search

WHU-USI3DV / CoFiI2P

[IEEE RA-L 2024] CoFiI2P: Coarse-to-Fine Correspondences-Based Image-to-Point Cloud Registration
https://whu-usi3dv.github.io/CoFiI2P/
56 stars 3 forks source link

Training report error #15

Open FDZ-XL opened 2 days ago

FDZ-XL commented 2 days ago

Hello, sorry to bother you again. I encountered the following error during training. After checking, it seems to be caused by patc_list being empty. Have you encountered this error before and how can I correct it? I hope you can give me some guidance.

(dlpeizhun) zxl@zxl-Legion-R9000P-ARX8:~/桌面/peizhun/CoFiI2P-main$ python -m train kitti train set: 40818 frames load train data complete val set: 5584 frames load val data complete 2024-09-18 20:06:43,771 - main - INFO - <data.options.Options_KITTI object at 0x7f89afe5d100> 2024-09-18 20:06:52,445 - main - INFO - train-0-10, loss: 8.878502, loss_desc: 3.459542, loss_coarse: 0.596121, loss_fine: 4.822839 2024-09-18 20:06:58,247 - main - INFO - train-0-20, loss: 9.007816, loss_desc: 3.502417, loss_coarse: 0.678173, loss_fine: 4.827226 2024-09-18 20:07:04,232 - main - INFO - train-0-30, loss: 9.128189, loss_desc: 3.467377, loss_coarse: 0.819809, loss_fine: 4.841004 2024-09-18 20:07:10,057 - main - INFO - train-0-40, loss: 8.907209, loss_desc: 3.447842, loss_coarse: 0.620799, loss_fine: 4.838569 2024-09-18 20:07:15,962 - main - INFO - train-0-50, loss: 8.790735, loss_desc: 3.445036, loss_coarse: 0.526920, loss_fine: 4.818779 2024-09-18 20:07:22,159 - main - INFO - train-0-60, loss: 8.886412, loss_desc: 3.436871, loss_coarse: 0.625806, loss_fine: 4.823734 2024-09-18 20:07:27,962 - main - INFO - train-0-70, loss: 9.152025, loss_desc: 3.467515, loss_coarse: 0.860146, loss_fine: 4.824364 2024-09-18 20:07:34,144 - main - INFO - train-0-80, loss: 8.793801, loss_desc: 3.452446, loss_coarse: 0.524749, loss_fine: 4.816607 2024-09-18 20:07:40,282 - main - INFO - train-0-90, loss: 8.711649, loss_desc: 3.439976, loss_coarse: 0.447469, loss_fine: 4.824203 2024-09-18 20:07:46,364 - main - INFO - train-0-100, loss: 8.819649, loss_desc: 3.444129, loss_coarse: 0.556634, loss_fine: 4.818886 Traceback (most recent call last): File "/home/zxl/anaconda3/envs/dlpeizhun/lib/python3.8/runpy.py", line 194, in _run_module_as_main return _run_code(code, main_globals, None, File "/home/zxl/anaconda3/envs/dlpeizhun/lib/python3.8/runpy.py", line 87, in _run_code exec(code, run_globals) File "/home/zxl/桌面/peizhun/CoFiI2P-main/train.py", line 307, in acc = test_acc(model,testloader,opt) File "/home/zxl/桌面/peizhun/CoFiI2P-main/train.py", line 70, in test_acc , fine_center_xy, coarse_pc_points=model(pc_data_dict,img, fine_center_kpt_coors,fine_xy, fine_pc_inline_index, mode) # [128, 20, 64] ,[128, 2560] File "/home/zxl/anaconda3/envs/dlpeizhun/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1194, in _call_impl return forward_call(*input, **kwargs) File "/home/zxl/桌面/peizhun/CoFiI2P-main/model/network.py", line 152, in forward fine_img_feature_patch = torch.squeeze(extract_patch(img_upsample_s2, fine_center_xy)) File "/home/zxl/桌面/peizhun/CoFiI2P-main/model/network.py", line 220, in extract_patch patch_tensor = torch.stack(patch_list) RuntimeError: stack expects a non-empty TensorList

martin-liao commented 1 day ago

Sorry for not replying in time. I will check the error.

martin-liao commented 1 day ago

We have fixed the validation error here.

FDZ-XL commented 1 day ago

Thank you very much. I wish you smooth scientific research and good health. Additionally, I have a question I would like to ask you and hope you can provide me with an answer. On line 112 of train.py: parser.add.argument ("-- ft_from", type=str, required=False, help="fine tune from existing checkpoint") Does this non essential parameter '-- ft_from' refer to the pre training weight you provided (cofi2pukitti. T7)? How was this pre training weight obtained?