这是@hpaulj提到的一个解决方案：

stack(tuple([np.array([1,2,3]), np.array([4,5,6])]))

[[1 2 3]
 [4 5 6]]

在numba中，vstack、hstack和concatenate只支持tuple作为输入，而不支持list。
他们说这是因为numba在编译Ref(https://github.com/numba/numba/issues/7476)时无法推断堆栈数组的维数。但我怀疑它实际上可以，因为你可以手动完成这一点，如下所示。
您可以通过以下间接方式堆叠list：

from numba import njit, prange
import numpy as np

@njit()
def test_list_stack(i, array_to_be_stacked):
    shape = (i,) + array_to_be_stacked.shape
    list_of_array = [array_to_be_stacked] * i
    stacked_array = np.empty(shape)
    for j in prange(i):
        stacked_array[j] = list_of_array[j]
    return stacked_array

if __name__ == "__main__":
    test_list_stack(10, np.ones((2, 3)))

或者，您可以定义一个自定义函数来执行堆栈工作：

from numba import njit, prange
from numba.typed import List
import numpy as np

@njit()
def stack(list_of_array):
    shape = (len(list_of_array),) + list_of_array[0].shape
    stacked_array = np.empty(shape)
    for j in prange(len(list_of_array)):
        stacked_array[j] = list_of_array[j]
    return stacked_array

if __name__ == "__main__":
    # Note that you have to use typed list provided by numba here.
    typed_list = List()
    [typed_list.append(np.ones((2, 3))) for _ in range(10)]
    stacked = stack(typed_list)
    print(stacked.shape)
    print(stacked)