_pickle.UnpicklingError: invalid load key, 'v'.

[Leven-Cheung]

When I try to run nice-slam and point-slam, this situation will happen.The detail of this error is :

Traceback (most recent call last): File "/mnt/15disk2/zgz/anaconda/envs/xrdslam1/bin/ds-run", line 8, in sys.exit(entrypoint()) File "/mnt/15disk2/zgz/slam/xrdslam/scripts/run.py", line 28, in entrypoint main( File "/mnt/15disk2/zgz/slam/xrdslam/scripts/run.py", line 22, in main slam.setup() File "/mnt/15disk2/zgz/slam/xrdslam/slam/engine/xrdslamer.py", line 55, in setup self.slam = self.config.xrdslam.setup() File "/mnt/15disk2/zgz/slam/xrdslam/slam/configs/base_config.py", line 37, in setup return self._target(self, *kwargs) File "/mnt/15disk2/zgz/slam/xrdslam/slam/pipeline/xrdslam.py", line 52, in init self.algorithm = manager.ShareAlgorithm(config=self.config.algorithm, File "/mnt/15disk2/zgz/anaconda/envs/xrdslam1/lib/python3.10/multiprocessing/managers.py", line 723, in temp token, exp = self._create(typeid, args, **kwds) File "/mnt/15disk2/zgz/anaconda/envs/xrdslam1/lib/python3.10/multiprocessing/managers.py", line 608, in _create id, exposed = dispatch(conn, None, 'create', (typeid,)+args, kwds) File "/mnt/15disk2/zgz/anaconda/envs/xrdslam1/lib/python3.10/multiprocessing/managers.py", line 93, in dispatch raise convert_to_error(kind, result) multiprocessing.managers.RemoteError:

Traceback (most recent call last): File "/mnt/15disk2/zgz/anaconda/envs/xrdslam1/lib/python3.10/multiprocessing/managers.py", line 209, in _handle_request result = func(c, *args, *kwds) File "/mnt/15disk2/zgz/anaconda/envs/xrdslam1/lib/python3.10/multiprocessing/managers.py", line 387, in create obj = callable(args, kwds) File "/mnt/15disk2/zgz/slam/xrdslam/slam/algorithms/point_slam.py", line 77, in init self.model = model_config.setup(camera=camera) File "/mnt/15disk2/zgz/slam/xrdslam/slam/configs/base_config.py", line 37, in setup return self._target(self, kwargs) File "/mnt/15disk2/zgz/slam/xrdslam/slam/models/conv_onet_pointslam.py", line 85, in init super().init(config=config, File "/mnt/15disk2/zgz/slam/xrdslam/slam/models/base_model.py", line 40, in init self.populate_modules() # populate the modules File "/mnt/15disk2/zgz/slam/xrdslam/slam/models/conv_onet_pointslam.py", line 94, in populate_modules self.load_pretrain() File "/mnt/15disk2/zgz/slam/xrdslam/slam/models/conv_onet_pointslam.py", line 234, in load_pretrain ckpt = torch.load(self.config.pretrained_decoders_middle_fine, File "/mnt/15disk2/zgz/anaconda/envs/xrdslam1/lib/python3.10/site-packages/torch/serialization.py", line 1028, in load return _legacy_load(opened_file, map_location, pickle_module, pickle_load_args) File "/mnt/15disk2/zgz/anaconda/envs/xrdslam1/lib/python3.10/site-packages/torch/serialization.py", line 1246, in _legacy_load magic_number = pickle_module.load(f, pickle_load_args) _pickle.UnpicklingError: invalid load key, 'v'.

but it is good running when using co-slam and vox-fusion.I searched that and Chatgpt said it might be something wrong in pre-trained files. Am I doing something wrong?

[wangxiaomeng030]

Could you please check if the downloaded pre-trained model is complete? You can download it from the zip file and replace it.

[wangxiaomeng030]

Here is a download link for the pre-trained model of xrdslam： https://drive.google.com/file/d/14aro7iugO6ey62iVpXrVAXLTzP6Ned3L/view?usp=drive_link

[Leven-Cheung]

This problem is solved, thank you very much, I guess it is probably caused by the network wall that can not use GLFS.Thank you for your help