error of "stack(): argument 'tensors' (position 1) must be tuple of Tensors, not Column"

[Jenny199099]

Hi author, thank you for sharing your great work. I have encountered an error during training. The error info are shown as following:

Traceback (most recent call last):
File "/home/wangjingyi/code/OneTwoVLA-main/scripts/train.py", line 428, in
main(_config.cli())
File "/home/wangjingyi/code/OneTwoVLA-main/scripts/train.py", line 337, in main
data_loader, val_data_loader = _data_loader.create_data_loader(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/wangjingyi/code/OneTwoVLA-main/src/openpi/training/data_loader.py", line 166, in create_data_loader
dataset, val_dataset = create_dataset(data_config, config.model)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/wangjingyi/code/OneTwoVLA-main/src/openpi/training/data_loader.py", line 97, in create_dataset
dataset = umi_dataset.UMIDataset(data_config, model_config.action_horizon)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/wangjingyi/code/OneTwoVLA-main/src/openpi/policies/umi_dataset.py", line 27, in init
super().init(data_config.repo_id, local_files_only=data_config.local_files_only)
File "/home/wangjingyi/code/OneTwoVLA-main/.venv/lib/python3.11/site-packages/lerobot/common/datasets/lerobot_dataset.py", line 464, in init
check_timestamps_sync(self.hf_dataset, self.episode_data_index, self.fps, self.tolerance_s)
File "/home/wangjingyi/code/OneTwoVLA-main/.venv/lib/python3.11/site-packages/lerobot/common/datasets/utils.py", line 410, in check_timestamps_sync
timestamps = torch.stack(hf_dataset["timestamp"])
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
TypeError: stack(): argument 'tensors' (position 1) must be tuple of Tensors, not Column

I found that someone said it is because the unmatched versions of lerobot and datasets. But I did not solve this error by changing the version of datasets. Could you please help me with this? Thank you.