Could you help me with error please？

[HuoQue]

Dear author, since I am not familiar with the training steps, after downloading the UBnorma data set locally and using your readme file to train, I found the following error: (simclr) C:\Users\34252\Desktop\MoCoDAD-main>python train_MoCoDAD.py --config config/UBnormal/mocodad_train.yaml Namespace(config='config/UBnormal/mocodad_train.yaml') config/UBnormal/mocodad_train.yaml
Experiment directories created in ./checkpoints\UBnormal\train_experiment 'cp' is not an internal or external command, nor a runnable program Or batch files.
Seed set to 999
Traceback (most recent call last):
File "C:\Users\34252\Desktop\MoCoDAD-main\trainMoCoDAD.py", line 65, in
, trainloader, , val_loader = get_dataset_and_loader(args, split=args.split, validation=args.validation) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\Users\34252\Desktop\MoCoDAD-main\utils\dataset.py", line 310, in get_dataset_and_loader
dataset = PoseDatasetRobust(path_to_data=args.data_dir,
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\34252\Desktop\MoCoDAD-main\utils\dataset.py", line 223, in init
arr = _read(fname, dtype=dtype, comment=comment, delimiter=delimiter, ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\Users\34252.conda\envs\simclr\Lib\site-packages\numpy\lib\npyio.py", line 999, in _read arr = _load_from_filelike( ^^^^^^^^^^^^^^^^^^^^ UnicodeDecodeError: 'gbk' codec can't decode byte 0xc0 in position 51: illegal multibyte sequence I would like to know how to set up the data set structure if I only train on the ubnor data set and use the validation set to identify abnormal behavior.Looking forward to your early reply！I would be appreciated it！

[aleflabo]

Hi,

We tested MoCoDAD only in Linux environments. Looking at the path in your code, it seems you're using Windows and the paths you're using are not formatted in the proper way.

[HuoQue]

I am using the UBnormal data set, I am not very clear about the path of the UBnormal data set structure in the data directory, so I want to consult you about it.

[aleflabo]

Once you've downloaded the UBnormal dataset, you will find a README.md file, which includes the following dataset structure:

UBnormal
|
|__________ hr_bool_masks
|           |
|           |__________ testing
|           |           |
|           |           |__________ test_frame_mask
|           |                       |_______________{scene_id}_{clip_id}.npy
|           |                       |_______________...
|           |                       |_______________{scene_id}_{clip_id}.npy
|           |
|           |__________ validating
|                       |
|                       |__________ test_frame_mask
|                                   |_______________{scene_id}_{clip_id}.npy
|                                   |_______________...
|                                   |_______________{scene_id}_{clip_id}.npy
|
|__________ training
|           |
|           |__________ trajectories
|                       |
|                       |_________{scene_id}_{clip_id}
|                                 |
|                                 |_________00001.csv
|                                 |_________...
|                                 |_________0000{n}.csv
|
|__________ testing
|           |
|           |__________ trajectories
|           |           |
|           |           |_________{scene_id}_{clip_id}
|           |                     |
|           |                     |_________00001.csv
|           |                     |_________...
|           |                     |_________0000{n}.csv
|           |
|           |__________ test_frame_mask
|                       |
|                       |_______________{scene_id}_{clip_id}.npy
|                       |_______________...
|                       |_______________{scene_id}_{clip_id}.npy
|
|__________ validating
            |
            |__________ trajectories
            |           |
            |           |_________{scene_id}_{clip_id}
            |                     |
            |                     |_________00001.csv
            |                     |_________...
            |                     |_________0000{n}.csv
            |
            |__________ test_frame_mask
                        |
                        |_______________{scene_id}_{clip_id}.npy
                        |_______________...
                        |_______________{scene_id}_{clip_id}.npy

Using the config/UBnormal/mocodad_train.yaml to train MoCoDAD on UBnormal, the validation set is used by default. Set the validation parameter to false to disable it.

[HuoQue]

My computer had only one GPU and it had the following problem: File "C:\Users\34252.conda\envs\simclr\Lib\site-packages\torch\distributed\distributed_c10d.py", line 1302, in _new_process_group_helper raise RuntimeError("Distributed package doesn't have NCCL built in") RuntimeError: Distributed package doesn't have NCCL built in If you know how to solve this problem, I would really appreciate it! Thank you very much for your guidance these days.

[aleflabo]

This is something related to the environment you have and not related to the code. You can try to work in a Linux environment and see if the problem is solved.

You can check for the error online to solve it: https://discuss.pytorch.org/t/runtimeerror-distributed-package-doesnt-have-nccl-built-in/176744/3