unable to train in docker

[wondergo2017]

I'm pretty interested in your work, but I met with some problems when running your code.

I pull docker haotang95/itergnn:main and run cmd in docker
python train.py lobster-unweighted DecIterGNN30-HomoPathSimConv-Max

but I get Runtime Error :

root@d21db24d6e17:/workspace# python train.py lobster-unweighted DecIterGNN30-HomoPathSimConv-Max
Saving logs to /workspace/record/ShortestPathLen/lobster_unweighted/4_30/100000/DecIterGNN_30/PathSimConv_homo/Max/1/Graph/logs
2020-12-12 14:13:29,919 root INFO: DatasetParam: {'dataset_name': 'ShortestPathLen', 'size': 100000, 'min_num_node': 4, 'num_num_node': 30, 'sparsity': 0.5, 'k': 8, 'dim': 2, 'lobster_prob': (0.2, 0.2), 'index_generator': 'lobster', 'min_edge_distance': 1.0, 'max_edge_distance': 1.0, 'device': device(type='cpu')}
2020-12-12 14:13:29,919 root INFO: ModelParam: {'embedding_layer_num': 2, 'architecture_name': 'DecIterGNN', 'layer_num': 30, 'module_num': 1, 'layer_name': 'PathSimConv', 'hidden_size': 64, 'input_feat_flag': True, 'homogeneous_flag': 1, 'readout_name': 'Max', 'confidence_layer_num': 1, 'head_layer_num': 1, 'model_type': 'Graph'}
2020-12-12 14:13:29,919 root INFO: GeneralParam: {'learning_rate': 0.001, 'epoch_num': 200, 'batch_size': 32, 'save_freq': 20, 'log_freq': 100, 'resume_flag': False, 'running_metric_name_list': ['relative_loss', 'mse_loss']}
2020-12-12 14:13:29,919 root INFO: cpu
Traceback (most recent call last):
File "/opt/conda/lib/python3.6/tarfile.py", line 2297, in next
tarinfo = self.tarinfo.fromtarfile(self)
File "/opt/conda/lib/python3.6/tarfile.py", line 1093, in fromtarfile
obj = cls.frombuf(buf, tarfile.encoding, tarfile.errors)
File "/opt/conda/lib/python3.6/tarfile.py", line 1031, in frombuf
raise TruncatedHeaderError("truncated header")
tarfile.TruncatedHeaderError: truncated header

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "/opt/conda/lib/python3.6/site-packages/torch/serialization.py", line 556, in _load
return legacy_load(f)
File "/opt/conda/lib/python3.6/site-packages/torch/serialization.py", line 467, in legacy_load
with closing(tarfile.open(fileobj=f, mode='r:', format=tarfile.PAX_FORMAT)) as tar,
File "/opt/conda/lib/python3.6/tarfile.py", line 1589, in open
return func(name, filemode, fileobj, **kwargs)
File "/opt/conda/lib/python3.6/tarfile.py", line 1619, in taropen
return cls(name, mode, fileobj, **kwargs)
File "/opt/conda/lib/python3.6/tarfile.py", line 1482, in init
self.firstmember = self.next()
File "/opt/conda/lib/python3.6/tarfile.py", line 2315, in next
raise ReadError(str(e))
tarfile.ReadError: truncated header

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "train.py", line 29, in
main(arg)
File "train.py", line 23, in main
train_model(dataset_param, model_param, general_param)
File "/workspace/main/train.py", line 62, in train_model
train_dataset = param2dataset(dataset_param, train_flag=True)
File "/workspace/main/parameters.py", line 68, in param2dataset
return globals()[param.dataset_name+'Dataset'](train_flag=train_flag, **other_param)
File "/workspace/dataset/base.py", line 29, in init
pre_transform=None, pre_filter=pre_filter)
File "/opt/conda/lib/python3.6/site-packages/torch_geometric/data/dataset.py", line 88, in init
if osp.exists(path) and torch.load(path) != repr(pre_transform):
File "/opt/conda/lib/python3.6/site-packages/torch/serialization.py", line 387, in load
return _load(f, map_location, pickle_module, **pickle_load_args)
File "/opt/conda/lib/python3.6/site-packages/torch/serialization.py", line 560, in _load
raise RuntimeError("{} is a zip archive (did you mean to use torch.jit.load()?)".format(f.name))
RuntimeError: /workspace/dataset/FAKEDataset/processed/pre_transform.pt is a zip archive (did you mean to use torch.jit.load()?)
root@d21db24d6e17:/workspace# ls

It must be something wrong, can you help me? Thank you for your attention!

[silent567]

Hi, thanks for your interest in our work! However, I will be quite busy with the application in the next few days and may not have enough time to check the details.

I have briefly gone through your error messages and it seems that there is something wrong with the datasets, which should be quite stable in my experiments. So, a quick check is to do rm -r dataset/FAKEDataset, then mkdir dataset/FAKEDataset and finally run the corresponding training commands. If the error still exists, could you please help provide a method to reproduce the error as well as your system's details? I will check them as soon as possible (maybe after Dec. 16th). Thanks!

[wondergo2017]

It works, thank you!