cupy.linalg.eigh不支持多个矩阵输入

如何解决cupy.linalg.eigh不支持多个矩阵输入

根据CuPy 8.0.0 https://docs.cupy.dev/en/stable/reference/generated/cupy.linalg.eigh.html上的文档，功能cupy.linalg.eigh应该能够将多个M * M矩阵作为输入。

Y = X[np.newaxis,:,:]
Y = np.repeat(Y,5,axis=0)
Y = cp.array(Y)
cp.linalg.eigh(Y)

我得到的错误是：

<ipython-input-6-6de0531e64a1> in <module>
      4 Y = cp.array(Y)
      5 X = cp.array(X)
----> 6 cp.linalg.eigh(Y)

~\Anaconda3\lib\site-packages\cupy\linalg\eigenvalue.py in eigh(a,UPLO)
    120     if not cuda.cusolver_enabled:
    121         raise RuntimeError('Current cupy only supports cusolver in CUDA 8.0')
--> 122     return _syevd(a,UPLO,True)
    123 
    124 

~\Anaconda3\lib\site-packages\cupy\linalg\eigenvalue.py in _syevd(a,with_eigen_vector)
     47     v = a.astype(inp_v_dtype,order='F',copy=True)
     48 
---> 49     m,lda = a.shape
     50     w = cupy.empty(m,inp_w_dtype)
     51     dev_info = cupy.empty((),'i')

ValueError: too many values to unpack (expected 2)

任何帮助将不胜感激。我在其他主题或论坛上没有发现任何帮助。