Python 实现 Tensor-Train 分解

第一步：安装Tensorly?

可移步到该篇博客中：Anaconda 安装 tensorly_小木王子、的博客-CSDN博客

第二步：借助tensor.decomposition模块进行分解

import numpy as np

tensor = np.random.random([10,10,10])

from tensorly.decomposition import matrix_product_state

factors = matrix_product_state(state,[1,2,3,1])

len(factors)→2

[f.shape for f in factors] → [(1,10,2),(2,10,3),(3,10,1)]