I encountered a problem in the preprocessing stage, how should I solve it?

[Yuanyangkai]

(MeNeXt) PS E:\PythonProject\MedNeXt> mednextv1_plan_and_preprocess -t 521 -pl3d ExperimentPlanner3D_v21_customTargetSpacing_1x1x1
E:\PythonProject\MedNeXt\mednext\DATASET\nnUNet_raw_data_base\nnUNet_raw_data\Task521_Lung\imagesTr\CTA_007
E:\PythonProject\MedNeXt\mednext\DATASET\nnUNet_raw_data_base\nnUNet_raw_data\Task521_Lung\imagesTr\CTA_001
E:\PythonProject\MedNeXt\mednext\DATASET\nnUNet_raw_data_base\nnUNet_raw_data\Task521_Lung\imagesTr\CTA_002
E:\PythonProject\MedNeXt\mednext\DATASET\nnUNet_raw_data_base\nnUNet_raw_data\Task521_Lung\imagesTr\CTA_006
E:\PythonProject\MedNeXt\mednext\DATASET\nnUNet_raw_data_base\nnUNet_raw_data\Task521_Lung\imagesTr\CTA_005
E:\PythonProject\MedNeXt\mednext\DATASET\nnUNet_raw_data_base\nnUNet_raw_data\Task521_Lung\imagesTr\CTA_004
E:\PythonProject\MedNeXt\mednext\DATASET\nnUNet_raw_data_base\nnUNet_raw_data\Task521_Lung\imagesTr\CTA_003

Task521_Lung
number of threads: (8, 8)

D:\Anaconda3\envs\MeNeXT\lib\site-packages\numpy\core\fromnumeric.py:3504: RuntimeWarning: Mean of empty slice.
return _methods._mean(a, axis=axis, dtype=dtype,
D:\Anaconda3\envs\MeNeXT\lib\site-packages\numpy\core_methods.py:129: RuntimeWarning: invalid value encountered in scalar divide
ret = ret.dtype.type(ret / rcount)
not using nonzero mask for normalization
Are we using the nonzero mask for normalization? OrderedDict([(0, False)])
Traceback (most recent call last):
File "\?\D:\Anaconda3\envs\MeNeXt\Scripts\mednextv1_plan_and_preprocess-script.py", line 33, in
sys.exit(load_entry_point('mednextv1', 'console_scripts', 'mednextv1_plan_and_preprocess')())
File "e:\pythonproject\mednext\mednext\nnunet_mednext\experiment_planning\nnUNet_plan_and_preprocess.py", line 159, in main
exp_planner.plan_experiment()
File "e:\pythonproject\mednext\mednext\nnunet_mednext\experiment_planning\experiment_planner_baseline_3DUNet.py", line 266, in plan_experiment
median_shape = np.median(np.vstack(new_shapes), 0)
File "D:\Anaconda3\envs\MeNeXT\lib\site-packages\numpy\core\shape_base.py", line 289, in vstack
return _nx.concatenate(arrs, 0, dtype=dtype, casting=casting)
ValueError: need at least one array to concatenate

[saikat-roy]

Hey @Yuanyangkai, have you managed to solve this issue?

[Yuanyangkai]

嘿，你设法解决了这个问题吗？

I didn't solve it

[Yuanyangkai]

Hey @Yuanyangkai, have you managed to solve this issue?

This is the command I entered.
"mednextv1_plan_and_preprocess -t 521 -pl3d ExperimentPlanner3D_v21_customTargetSpacing_1x1x1"
This is my json file.
{
"name": "Lung",
"description": "artery and vein segmentation",
"reference": "https://example.com/my_dataset",
"licence": "CC BY-NC-SA 4.0",
"tensorImageSize": "3d",
"modality": {
"0": "C"
},
"labels": {
"0": "background",
"1": "vein",
"2": "artery"
},

"numTest": 3,
"numTraining": 7,

"test": [
    "E:/PythonProject/MedNeXt/DATASET/nnUNet_raw_data_base/nnUNet_raw_data/Task521_Lung/imagesTs/CTA_008.nii.gz",
    "E:/PythonProject/MedNeXt/DATASET/nnUNet_raw_data_base/nnUNet_raw_data/Task521_Lung/imagesTs/CTA_009.nii.gz",
    "E:/PythonProject/MedNeXt/DATASET/nnUNet_raw_data_base/nnUNet_raw_data/Task521_Lung/imagesTs/CTA_010.nii.gz"
],
"training": [
    {
        "image": "E:/PythonProject/MedNeXt/DATASET/nnUNet_raw_data_base/nnUNet_raw_data/Task521_Lung/imagesTr/CTA_001.nii.gz",
        "label": "E:/PythonProject/MedNeXt/DATASET/nnUNet_raw_data_base/nnUNet_raw_data/Task521_Lung/labelsTr/CTA_001.nii.gz"
    },
    {
        "image": "E:/PythonProject/MedNeXt/DATASET/nnUNet_raw_data_base/nnUNet_raw_data/Task521_Lung/imagesTr/CTA_002.nii.gz",
        "label": "E:/PythonProject/MedNeXt/DATASET/nnUNet_raw_data_base/nnUNet_raw_data/Task521_Lung/labelsTr/CTA_002.nii.gz"
    },
    {
        "image": "E:/PythonProject/MedNeXt/DATASET/nnUNet_raw_data_base/nnUNet_raw_data/Task521_Lung/imagesTr/CTA_003.nii.gz",
        "label": "E:/PythonProject/MedNeXt/DATASET/nnUNet_raw_data_base/nnUNet_raw_data/Task521_Lung/labelsTr/CTA_003.nii.gz"
    },
    {
        "image": "E:/PythonProject/MedNeXt/DATASET/nnUNet_raw_data_base/nnUNet_raw_data/Task521_Lung/imagesTr/CTA_004.nii.gz",
        "label": "E:/PythonProject/MedNeXt/DATASET/nnUNet_raw_data_base/nnUNet_raw_data/Task521_Lung/labelsTr/CTA_004.nii.gz"
    },
    {
        "image": "E:/PythonProject/MedNeXt/DATASET/nnUNet_raw_data_base/nnUNet_raw_data/Task521_Lung/imagesTr/CTA_005.nii.gz",
        "label": "E:/PythonProject/MedNeXt/DATASET/nnUNet_raw_data_base/nnUNet_raw_data/Task521_Lung/labelsTr/CTA_005.nii.gz"
    },
    {
        "image": "E:/PythonProject/MedNeXt/DATASET/nnUNet_raw_data_base/nnUNet_raw_data/Task521_Lung/imagesTr/CTA_006.nii.gz",
        "label": "E:/PythonProject/MedNeXt/DATASET/nnUNet_raw_data_base/nnUNet_raw_data/Task521_Lung/labelsTr/CTA_006.nii.gz"
    },
    {
        "image": "E:/PythonProject/MedNeXt/DATASET/nnUNet_raw_data_base/nnUNet_raw_data/Task521_Lung/imagesTr/CTA_007.nii.gz",
        "label": "E:/PythonProject/MedNeXt/DATASET/nnUNet_raw_data_base/nnUNet_raw_data/Task521_Lung/labelsTr/CTA_007.nii.gz"
    }
]

}

[Yuanyangkai]

Hey @Yuanyangkai, have you managed to solve this issue?

When I run my tests_mednext file, I also get an error.

D:\Anaconda3\envs\MeNeXt\python.exe "D:/Pycharm/PyCharm Community Edition 2024.1/plugins/python-ce/helpers/pycharm/_jb_pytest_runner.py" --target tests_mednext_miccai_architectures.py::Test_MedNeXt_archs
Testing started at 下午7:30 ...
Launching pytest with arguments tests_mednext_miccai_architectures.py::Test_MedNeXt_archs --no-header --no-summary -q in E:\PythonProject\MedNeXt\mednext\tests

============================= test session starts =============================
collecting ... collected 8 items

tests_mednext_miccai_architectures.py::Test_MedNeXt_archs::test_init_and_forward[S-3]
tests_mednext_miccai_architectures.py::Test_MedNeXt_archs::test_init_and_forward[B-3]
tests_mednext_miccai_architectures.py::Test_MedNeXt_archs::test_init_and_forward[M-3]
tests_mednext_miccai_architectures.py::Test_MedNeXt_archs::test_init_and_forward[L-3]
tests_mednext_miccai_architectures.py::Test_MedNeXt_archs::test_init_and_forward[S-5]
tests_mednext_miccai_architectures.py::Test_MedNeXt_archs::test_init_and_forward[B-5]
tests_mednext_miccai_architectures.py::Test_MedNeXt_archs::test_init_and_forward[M-5]
tests_mednext_miccai_architectures.py::Test_MedNeXt_archs::test_init_and_forward[L-5]

============================== 8 failed in 8.25s ==============================
FAILED [ 12%]
tests\tests_mednext_miccai_architectures.py:6 (Test_MedNeXt_archs.test_init_and_forward[S-3])
torch.Size([4, 128, 128, 128]) != (1, 4, 128, 128, 128)

预期:(1, 4, 128, 128, 128)
实际:torch.Size([4, 128, 128, 128])
<点击以查看差异>

self = <tests_mednext_miccai_architectures.Test_MedNeXt_archs object at 0x0000020E4982E100>
model_size = 'S', kernel_size = 3

@pytest.mark.parametrize("model_size, kernel_size", [
('S', 3),
('B', 3),
('M', 3),
('L', 3),
('S', 5),
('B', 5),
('M', 5),
('L', 5),
])
def test_init_and_forward(self, model_size, kernel_size):
    m = create_mednext_v1(2, 4, model_size, kernel_size).cuda()
    input = torch.zeros((1,2,128,128,128), requires_grad=False).cuda()
    with torch.no_grad():
        output = m(input)
    del m
    inp_shape = input.shape

  assert output[0].shape == (inp_shape[0], 4, *inp_shape[2:])

E assert torch.Size([4, 128, 128, 128]) == (1, 4, 128, 128, 128)
E
E At index 0 diff: 4 != 1
E Right contains one more item: 128
E
E Full diff:
E + torch.Size([4, 128, 128, 128])
E - (
E - 1,
E - 4,
E - 128,
E - 128,
E - 128,
E - )

tests_mednext_miccai_architectures.py:24: AssertionError
FAILED [ 25%]
tests\tests_mednext_miccai_architectures.py:6 (Test_MedNeXt_archs.test_init_and_forward[B-3])
torch.Size([4, 128, 128, 128]) != (1, 4, 128, 128, 128)

预期:(1, 4, 128, 128, 128)
实际:torch.Size([4, 128, 128, 128])
<点击以查看差异>

self = <tests_mednext_miccai_architectures.Test_MedNeXt_archs object at 0x0000020E53DFAA30>
model_size = 'B', kernel_size = 3

@pytest.mark.parametrize("model_size, kernel_size", [
('S', 3),
('B', 3),
('M', 3),
('L', 3),
('S', 5),
('B', 5),
('M', 5),
('L', 5),
])
def test_init_and_forward(self, model_size, kernel_size):
    m = create_mednext_v1(2, 4, model_size, kernel_size).cuda()
    input = torch.zeros((1,2,128,128,128), requires_grad=False).cuda()
    with torch.no_grad():
        output = m(input)
    del m
    inp_shape = input.shape

  assert output[0].shape == (inp_shape[0], 4, *inp_shape[2:])

E assert torch.Size([4, 128, 128, 128]) == (1, 4, 128, 128, 128)
E
E At index 0 diff: 4 != 1
E Right contains one more item: 128
E
E Full diff:
E + torch.Size([4, 128, 128, 128])
E - (
E - 1,
E - 4,
E - 128,
E - 128,
E - 128,
E - )

tests_mednext_miccai_architectures.py:24: AssertionError
FAILED [ 37%]
tests\tests_mednext_miccai_architectures.py:6 (Test_MedNeXt_archs.test_init_and_forward[M-3])
torch.Size([4, 128, 128, 128]) != (1, 4, 128, 128, 128)

预期:(1, 4, 128, 128, 128)
实际:torch.Size([4, 128, 128, 128])
<点击以查看差异>

self = <tests_mednext_miccai_architectures.Test_MedNeXt_archs object at 0x0000020E53DFAF70>
model_size = 'M', kernel_size = 3

@pytest.mark.parametrize("model_size, kernel_size", [
('S', 3),
('B', 3),
('M', 3),
('L', 3),
('S', 5),
('B', 5),
('M', 5),
('L', 5),
])
def test_init_and_forward(self, model_size, kernel_size):
    m = create_mednext_v1(2, 4, model_size, kernel_size).cuda()
    input = torch.zeros((1,2,128,128,128), requires_grad=False).cuda()
    with torch.no_grad():
        output = m(input)
    del m
    inp_shape = input.shape

  assert output[0].shape == (inp_shape[0], 4, *inp_shape[2:])

E assert torch.Size([4, 128, 128, 128]) == (1, 4, 128, 128, 128)
E
E At index 0 diff: 4 != 1
E Right contains one more item: 128
E
E Full diff:
E + torch.Size([4, 128, 128, 128])
E - (
E - 1,
E - 4,
E - 128,
E - 128,
E - 128,
E - )

tests_mednext_miccai_architectures.py:24: AssertionError
FAILED [ 50%]
tests\tests_mednext_miccai_architectures.py:6 (Test_MedNeXt_archs.test_init_and_forward[L-3])
torch.Size([4, 128, 128, 128]) != (1, 4, 128, 128, 128)

预期:(1, 4, 128, 128, 128)
实际:torch.Size([4, 128, 128, 128])
<点击以查看差异>

self = <tests_mednext_miccai_architectures.Test_MedNeXt_archs object at 0x0000020E53E1C070>
model_size = 'L', kernel_size = 3

@pytest.mark.parametrize("model_size, kernel_size", [
('S', 3),
('B', 3),
('M', 3),
('L', 3),
('S', 5),
('B', 5),
('M', 5),
('L', 5),
])
def test_init_and_forward(self, model_size, kernel_size):
    m = create_mednext_v1(2, 4, model_size, kernel_size).cuda()
    input = torch.zeros((1,2,128,128,128), requires_grad=False).cuda()
    with torch.no_grad():
        output = m(input)
    del m
    inp_shape = input.shape

  assert output[0].shape == (inp_shape[0], 4, *inp_shape[2:])

E assert torch.Size([4, 128, 128, 128]) == (1, 4, 128, 128, 128)
E
E At index 0 diff: 4 != 1
E Right contains one more item: 128
E
E Full diff:
E + torch.Size([4, 128, 128, 128])
E - (
E - 1,
E - 4,
E - 128,
E - 128,
E - 128,
E - )

tests_mednext_miccai_architectures.py:24: AssertionError
FAILED [ 62%]
tests\tests_mednext_miccai_architectures.py:6 (Test_MedNeXt_archs.test_init_and_forward[S-5])
torch.Size([4, 128, 128, 128]) != (1, 4, 128, 128, 128)

预期:(1, 4, 128, 128, 128)
实际:torch.Size([4, 128, 128, 128])
<点击以查看差异>

self = <tests_mednext_miccai_architectures.Test_MedNeXt_archs object at 0x0000020E53E1C130>
model_size = 'S', kernel_size = 5

@pytest.mark.parametrize("model_size, kernel_size", [
('S', 3),
('B', 3),
('M', 3),
('L', 3),
('S', 5),
('B', 5),
('M', 5),
('L', 5),
])
def test_init_and_forward(self, model_size, kernel_size):
    m = create_mednext_v1(2, 4, model_size, kernel_size).cuda()
    input = torch.zeros((1,2,128,128,128), requires_grad=False).cuda()
    with torch.no_grad():
        output = m(input)
    del m
    inp_shape = input.shape

  assert output[0].shape == (inp_shape[0], 4, *inp_shape[2:])

E assert torch.Size([4, 128, 128, 128]) == (1, 4, 128, 128, 128)
E
E At index 0 diff: 4 != 1
E Right contains one more item: 128
E
E Full diff:
E + torch.Size([4, 128, 128, 128])
E - (
E - 1,
E - 4,
E - 128,
E - 128,
E - 128,
E - )

tests_mednext_miccai_architectures.py:24: AssertionError
FAILED [ 75%]
tests\tests_mednext_miccai_architectures.py:6 (Test_MedNeXt_archs.test_init_and_forward[B-5])
torch.Size([4, 128, 128, 128]) != (1, 4, 128, 128, 128)

预期:(1, 4, 128, 128, 128)
实际:torch.Size([4, 128, 128, 128])
<点击以查看差异>

self = <tests_mednext_miccai_architectures.Test_MedNeXt_archs object at 0x0000020E53E1C1F0>
model_size = 'B', kernel_size = 5

@pytest.mark.parametrize("model_size, kernel_size", [
('S', 3),
('B', 3),
('M', 3),
('L', 3),
('S', 5),
('B', 5),
('M', 5),
('L', 5),
])
def test_init_and_forward(self, model_size, kernel_size):
    m = create_mednext_v1(2, 4, model_size, kernel_size).cuda()
    input = torch.zeros((1,2,128,128,128), requires_grad=False).cuda()
    with torch.no_grad():
        output = m(input)
    del m
    inp_shape = input.shape

  assert output[0].shape == (inp_shape[0], 4, *inp_shape[2:])

E assert torch.Size([4, 128, 128, 128]) == (1, 4, 128, 128, 128)
E
E At index 0 diff: 4 != 1
E Right contains one more item: 128
E
E Full diff:
E + torch.Size([4, 128, 128, 128])
E - (
E - 1,
E - 4,
E - 128,
E - 128,
E - 128,
E - )

tests_mednext_miccai_architectures.py:24: AssertionError
FAILED [ 87%]
tests\tests_mednext_miccai_architectures.py:6 (Test_MedNeXt_archs.test_init_and_forward[M-5])
torch.Size([4, 128, 128, 128]) != (1, 4, 128, 128, 128)

预期:(1, 4, 128, 128, 128)
实际:torch.Size([4, 128, 128, 128])
<点击以查看差异>

self = <tests_mednext_miccai_architectures.Test_MedNeXt_archs object at 0x0000020E53E1C2B0>
model_size = 'M', kernel_size = 5

@pytest.mark.parametrize("model_size, kernel_size", [
('S', 3),
('B', 3),
('M', 3),
('L', 3),
('S', 5),
('B', 5),
('M', 5),
('L', 5),
])
def test_init_and_forward(self, model_size, kernel_size):
    m = create_mednext_v1(2, 4, model_size, kernel_size).cuda()
    input = torch.zeros((1,2,128,128,128), requires_grad=False).cuda()
    with torch.no_grad():
        output = m(input)
    del m
    inp_shape = input.shape

  assert output[0].shape == (inp_shape[0], 4, *inp_shape[2:])

E assert torch.Size([4, 128, 128, 128]) == (1, 4, 128, 128, 128)
E
E At index 0 diff: 4 != 1
E Right contains one more item: 128
E
E Full diff:
E + torch.Size([4, 128, 128, 128])
E - (
E - 1,
E - 4,
E - 128,
E - 128,
E - 128,
E - )

tests_mednext_miccai_architectures.py:24: AssertionError
FAILED [100%]
tests\tests_mednext_miccai_architectures.py:6 (Test_MedNeXt_archs.test_init_and_forward[L-5])
torch.Size([4, 128, 128, 128]) != (1, 4, 128, 128, 128)

预期:(1, 4, 128, 128, 128)
实际:torch.Size([4, 128, 128, 128])
<点击以查看差异>

self = <tests_mednext_miccai_architectures.Test_MedNeXt_archs object at 0x0000020E53E1C370>
model_size = 'L', kernel_size = 5

@pytest.mark.parametrize("model_size, kernel_size", [
('S', 3),
('B', 3),
('M', 3),
('L', 3),
('S', 5),
('B', 5),
('M', 5),
('L', 5),
])
def test_init_and_forward(self, model_size, kernel_size):
    m = create_mednext_v1(2, 4, model_size, kernel_size).cuda()
    input = torch.zeros((1,2,128,128,128), requires_grad=False).cuda()
    with torch.no_grad():
        output = m(input)
    del m
    inp_shape = input.shape

  assert output[0].shape == (inp_shape[0], 4, *inp_shape[2:])

E assert torch.Size([4, 128, 128, 128]) == (1, 4, 128, 128, 128)
E
E At index 0 diff: 4 != 1
E Right contains one more item: 128
E
E Full diff:
E + torch.Size([4, 128, 128, 128])
E - (
E - 1,
E - 4,
E - 128,
E - 128,
E - 128,
E - )

tests_mednext_miccai_architectures.py:24: AssertionError

进程已结束，退出代码为 1

[Yuanyangkai]

Hey @Yuanyangkai, have you managed to solve this issue?

My data set .nii.gz file successfully ran in the nnunet_v2 official file, but after I modified the json file, it failed to run in the mednext file. I did not find the problem.

[saikat-roy]

line 289, in vstack return _nx.concatenate(arrs, 0, dtype=dtype, casting=casting) ValueError: need at least one array to concatenate

So based on this, the nnunetv1 experiment planner cannot find any files to preprocess. I would start with making sure that the nnunetv1 code underneath can find the your files. The first thing I can see, which deviates from standard nnunetv1 recommendations are the paths in the dataset.json.

Could you try with relative paths instead of absolute paths in the image and label. Something like : {"image":"./imagesTr/CTA_001.nii.gz","label":"./labelsTr/CTA_001.nii.gz"}

[Yuanyangkai]

line 289, in vstack return _nx.concatenate(arrs, 0, dtype=dtype, casting=casting) ValueError: need at least one array to concatenate

So based on this, the nnunetv1 experiment planner cannot find any files to preprocess. I would start with making sure that the nnunetv1 code underneath can find the your files. The first thing I can see, which deviates from standard nnunetv1 recommendations are the paths in the dataset.json.

Could you try with relative paths instead of absolute paths in the image and label. Something like : {"image":"./imagesTr/CTA_001.nii.gz","label":"./labelsTr/CTA_001.nii.gz"}

Still reporting an error.

(MeNeXt) PS E:\PythonProject\MedNeXt> mednextv1_plan_and_preprocess -t 521 -pl3d ExperimentPlanner3D_v21_customTargetSpacing_1x1x1
E:\PythonProject\MedNeXt\mednext\DATASET\nnUNet_raw_data_base\nnUNet_raw_data\Task521_Lung\imagesTr\CTA_003
E:\PythonProject\MedNeXt\mednext\DATASET\nnUNet_raw_data_base\nnUNet_raw_data\Task521_Lung\imagesTr\CTA_001
E:\PythonProject\MedNeXt\mednext\DATASET\nnUNet_raw_data_base\nnUNet_raw_data\Task521_Lung\imagesTr\CTA_004
E:\PythonProject\MedNeXt\mednext\DATASET\nnUNet_raw_data_base\nnUNet_raw_data\Task521_Lung\imagesTr\CTA_002
E:\PythonProject\MedNeXt\mednext\DATASET\nnUNet_raw_data_base\nnUNet_raw_data\Task521_Lung\imagesTr\CTA_006
E:\PythonProject\MedNeXt\mednext\DATASET\nnUNet_raw_data_base\nnUNet_raw_data\Task521_Lung\imagesTr\CTA_007
E:\PythonProject\MedNeXt\mednext\DATASET\nnUNet_raw_data_base\nnUNet_raw_data\Task521_Lung\imagesTr\CTA_005

Task521_Lung
number of threads: (8, 8)

qqq <nnunet_mednext.experiment_planning.alternative_experiment_planning.target_spacing.experiment_planner_v21_isotropic1mm.ExperimentPlanner3D_v21_customTargetSpacing_1x1x1 object at 0x00000184087D9340>
D:\Anaconda3\envs\MeNeXt\lib\site-packages\numpy\core\fromnumeric.py:3504: RuntimeWarning: Mean of empty slice.
return _methods._mean(a, axis=axis, dtype=dtype,
D:\Anaconda3\envs\MeNeXt\lib\site-packages\numpy\core_methods.py:129: RuntimeWarning: invalid value encountered in scalar divide
ret = ret.dtype.type(ret / rcount)
not using nonzero mask for normalization
Are we using the nonzero mask for normalization? OrderedDict([(0, False)])
spacings: []
sizes: []
Traceback (most recent call last):
File "\?\D:\Anaconda3\envs\MeNeXt\Scripts\mednextv1_plan_and_preprocess-script.py", line 33, in
sys.exit(load_entry_point('mednextv1', 'console_scripts', 'mednextv1_plan_and_preprocess')())
File "e:\pythonproject\mednext\mednext\nnunet_mednext\experiment_planning\nnUNet_plan_and_preprocess.py", line 162, in main
exp_planner.plan_experiment()
File "e:\pythonproject\mednext\mednext\nnunet_mednext\experiment_planning\experiment_planner_baseline_3DUNet.py", line 268, in plan_experiment
median_shape = np.median(np.vstack(new_shapes), 0)
File "D:\Anaconda3\envs\MeNeXt\lib\site-packages\numpy\core\shape_base.py", line 289, in vstack
return _nx.concatenate(arrs, 0, dtype=dtype, casting=casting)
ValueError: need at least one array to concatenate