【mindspore产品】【model.功能】预测值维度为什么要比实际值维度多一维

问题描述：

【功能模块】

nn.metric

【操作步骤&问题现象】

1、对训练过的模型进行验证时，报错：ValueError: Classification case, dims of y_pred equal dims of y add 1, but got y_pred: 2 dims and y: 2 dims

2、

【截图信息】

def _check_shape(self, y_pred, y):

    """

    Checks the shapes of y_pred and y.

    Args:

        y_pred (Tensor): Predict array.

        y (Tensor): Target array.

    """

    if self._type == 'classification':

        if y_pred.ndim != y.ndim + 1:

            raise ValueError('Classification case, dims of y_pred equal dims of y add 1, '

                             'but got y_pred: {} dims and y: {} dims'.format(y_pred.ndim, y.ndim))

        if y.shape != (y_pred.shape[0],) + y_pred.shape[2:]:

            raise ValueError('Classification case, y_pred shape and y shape can not match. '

                             'got y_pred shape is {} and y shape is {}'.format(y_pred.shape, y.shape))

解答：

`y_pred` stands for logits, `y` stands for labels. `y_pred` and `y` must be a `Tensor`, a list or an array.

For the 'classification' evaluation type:

`y_pred` is a list of floating numbers in range :math:`[0, 1]` and the shape is :math:`(N, C)` in most cases (not strictly), where :math:`N` is the number of cases and :math:`C` is the number of categories.

`y` must be in one-hot format that shape is :math:`(N, C)`, or can be transformed to one-hot format that shape is :math:`(N,)`.

For 'multilabel' evaluation type, the value of `y_pred` and `y` can only be 0 or 1, indices with 1 indicate the positive category. The shape of `y_pred` and `y` are both :math:`(N, C)`.

可以参考链接：

https://gitee.com/sun_zhongqian/mindspore/blob/r1.6/mindspore/python/mindspore/nn/metrics/accuracy.py

https://www.mindspore.cn/docs/api/zh-CN/r1.6/api_python/nn/mindspore.nn.Accuracy.html

自定义调试信息 — MindSpore master documentation

可以参考这部分代码去实现model.eval()

from mindspore import Tensor

from mindspore.nn import Accuracy

import numpy as np

x = Tensor(np.array([[0.2, 0.5], [0.3, 0.1], [0.9, 0.6]]))

y = Tensor(np.array([1, 0, 1]))

metric = Accuracy()

metric.clear()

metric.update(x, y)

accuracy = metric.eval()

print('Accuracy is ', accuracy)

最后

以上就是勤恳斑马为你收集整理的【mindspore产品】【model.功能】预测值维度为什么要比实际值维度多一维的全部内容，希望文章能够帮你解决【mindspore产品】【model.功能】预测值维度为什么要比实际值维度多一维所遇到的程序开发问题。

如果觉得靠谱客网站的内容还不错，欢迎将靠谱客网站推荐给程序员好友。