questions

[1squarewave]

配置环境一堆问题，跑脚本跑不通，目前完全无法复现，让人怀疑真实性
Setting up the environment has a bunch of issues, scripts won't run, currently can't reproduce it at all, makes me doubt if it's real

[1squarewave]

(forcevla) root@DESKTOP-K1VG2CP:~/ForceVLA# /root/miniconda3/envs/forcevla/bin/python /root/ForceVLA/scripts/compute_norm_stats.py
2026-07-17 14:58:40.864615: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNNwhen one has already been registered
2026-07-17 14:58:40.864720: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2026-07-17 14:58:41.051332: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2026-07-17 14:58:50.191417: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Traceback (most recent call last):
File "/root/ForceVLA/scripts/compute_norm_stats.py", line 119, in
tyro.cli(main(config_name="forcevla_lora "))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/root/ForceVLA/scripts/compute_norm_stats.py", line 90, in main
config = _config.get_config(config_name)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/root/ForceVLA/openpi/src/openpi/training/config.py", line 987, in get_config
raise ValueError(f"Config '{config_name}' not found.{closest_str}")
ValueError: Config 'forcevla_lora ' not found. Did you mean 'pi0_aloha'?

please tell me why？

[1squarewave]

Exception has occurred: TypeError
stack(): argument 'tensors' (position 1) must be tuple of Tensors, not Column
File "/root/ForceVLA/lerobot/lerobot/common/datasets/lerobot_dataset.py", line 508, in init
timestamps = torch.stack(self.hf_dataset["timestamp"]).numpy()
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/root/ForceVLA/openpi/src/openpi/training/data_loader.py", line 141, in create_torch_dataset
data_config.repo_id,

    delta_timestamps={

        key: [t / dataset_meta.fps for t in range(action_horizon)] for key in data_config.action_sequence_keys

    },

)
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

File "/root/ForceVLA/scripts/compute_norm_stats.py", line 34, in create_torch_dataloader
dataset = _data_loader.create_torch_dataset(data_config, action_horizon, model_config)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/root/ForceVLA/scripts/compute_norm_stats.py", line 98, in main
data_config, config.model.action_horizon, config.batch_size, config.model, config.num_workers, max_frames

    )
                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

File "/root/ForceVLA/scripts/compute_norm_stats.py", line 117, in
tyro.cli(main(config_name="pi0_fast_libero"))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
TypeError: stack(): argument 'tensors' (position 1) must be tuple of Tensors, not Column