softmax¶
- paddle.sparse.nn.functional. softmax ( x, axis=- 1, name=None ) [source]
-
sparse softmax activation, requiring x to be a SparseCooTensor or SparseCsrTensor.
- Note:
-
Only support axis=-1 for SparseCsrTensor, which is faster when read data by row (axis=-1).
From the point of view of dense matrix, for each row \(i\) and each column \(j\) in the matrix, we have:
\[softmax_ij =\]rac{exp(x_ij - max_j(x_ij))}{sum_j(exp(x_ij - max_j(x_ij))}
- Parameters:
-
x (Tensor): The input tensor. It can be SparseCooTensor/SparseCsrTensor. The data type can be float32 or float64. axis (int, optional): The axis along which to perform softmax calculations. Only support -1 for SparseCsrTensor. name (str, optional): Name for the operation (optional, default is None).
For more information, please refer to Name.
- Returns:
-
Tensor: SparseCoo or SparseCsr, whose layout is the same with x .
- Examples:
-
import paddle import numpy as np paddle.seed(100) mask = np.random.rand(3, 4) < 0.5 np_x = np.random.rand(3, 4) * mask # [[0. 0. 0.96823406 0.19722934] # [0.94373937 0. 0.02060066 0.71456372] # [0. 0. 0. 0.98275049]] csr = paddle.to_tensor(np_x).to_sparse_csr() # Tensor(shape=[3, 4], dtype=paddle.float64, place=Place(gpu:0), stop_gradient=True, # crows=[0, 2, 5, 6], # cols=[2, 3, 0, 2, 3, 3], # values=[0.96823406, 0.19722934, 0.94373937, 0.02060066, 0.71456372, # 0.98275049]) out = paddle.sparse.nn.functional.softmax(csr) # Tensor(shape=[3, 4], dtype=paddle.float64, place=Place(gpu:0), stop_gradient=True, # crows=[0, 2, 5, 6], # cols=[2, 3, 0, 2, 3, 3], # values=[0.68373820, 0.31626180, 0.45610887, 0.18119845, 0.36269269, # 1. ])