當前位置: 首頁>>代碼示例>>Python>>正文


Python NervanaGPU.power方法代碼示例

本文整理匯總了Python中nervanagpu.NervanaGPU.power方法的典型用法代碼示例。如果您正苦於以下問題:Python NervanaGPU.power方法的具體用法?Python NervanaGPU.power怎麽用?Python NervanaGPU.power使用的例子?那麽, 這裏精選的方法代碼示例或許可以為您提供幫助。您也可以進一步了解該方法所在nervanagpu.NervanaGPU的用法示例。


在下文中一共展示了NervanaGPU.power方法的1個代碼示例,這些例子默認根據受歡迎程度排序。您可以為喜歡或者感覺有用的代碼點讚,您的評價將有助於係統推薦出更棒的Python代碼示例。

示例1: GPU

# 需要導入模塊: from nervanagpu import NervanaGPU [as 別名]
# 或者: from nervanagpu.NervanaGPU import power [as 別名]

#.........這裏部分代碼省略.........
            order (int): The order or p upon which the norm is calculated.
                         Valid values include:
                         None, inf, -inf, 0, 1, -1, 2, -2, ...
            axis (int): The axis along which to compute vector norms.
            out (GPUTensor): where to write the results to.  Must be
                             of the expected result shape.

        Returns:
            GPUTensor: p-norm of tsr along the specified axis.

        Raises:
            IndexError if invalid axis specified
            AttributeError if invalid order specified

        See Also:
            `numpy.linalg.norm`
        """
        if not isinstance(axis, int) or axis < 0 or axis >= len(tsr.shape):
            raise IndexError("invalid axis value: %s", axis)
        if not isinstance(order, (int, float)):
            raise AttributeError("invalid order value: %s", order)
        if out is None:
            raise AttributeError("No output tensor speficied", order)
        if order == float('Inf'):
            self.ng.max(self.fabs(tsr), axis, out)
        elif order == float('-Inf'):
            self.ng.min(self.fabs(tsr), axis, out)
        elif order == 0:
            tmp = self.zeros(tsr.shape)
            self.ng.not_equal(tsr, tmp, tmp)
            self.ng.sum(tmp, axis, out)
        else:
            tmp = self.empty(tsr.shape)
            self.ng.power(self.fabs(tsr), order, tmp)
            self.ng.sum(tmp, axis, out)
            self.ng.power(out, (1.0 / order), out)
        return out

    def mean(self, tsr, axes, out):
        """
        Calculates the arithmetic mean of the elements along the specified
        axes.

        Arguments:
            tsr (GPUTensor): Input tensor
            axes (int): Axis along which the reduction is performed. If axes
                        is None,  the tensor is flattened and reduced over
                        both dimensions.
            out (GPUTensor): Output tensor
        """
        if axes is None:
            sze = tsr.shape[0]*tsr.shape[1]
            self.ng.mean(tsr.reshape(sze, 1), axis=0, out=out)
        else:
            self.ng.mean(tsr, axis=axes, out=out)
        return out

    def min(self, tsr, axes, out):
        """
        Calculates the minimum of the elements along the specified
        axes.

        Arguments:
            tsr (GPUTensor): Input tensor
            axes (int): Axis along which the reduction is performed. If axes
                        is None,  the tensor is flattened and reduced over
開發者ID:neuroidss,項目名稱:neon,代碼行數:70,代碼來源:gpu.py


注:本文中的nervanagpu.NervanaGPU.power方法示例由純淨天空整理自Github/MSDocs等開源代碼及文檔管理平台,相關代碼片段篩選自各路編程大神貢獻的開源項目,源碼版權歸原作者所有,傳播和使用請參考對應項目的License;未經允許,請勿轉載。