Tensor

  • paddle.fluid.Tensor()

Tensor用于表示多维张量,可以通过 np.array(tensor) 方法转换为numpy.ndarray。

示例代码

  1. import paddle.fluid as fluid
  2.  
  3. t = fluid.Tensor()
  • set(args, *kwargs)

该接口根据输入的numpy array和设备place,设置Tensor的数据。

重载函数:

  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[float32], place: paddle::platform::CPUPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[int32], place: paddle::platform::CPUPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[float64], place: paddle::platform::CPUPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[int64], place: paddle::platform::CPUPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[bool], place: paddle::platform::CPUPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[uint16], place: paddle::platform::CPUPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[uint8], place: paddle::platform::CPUPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[int8], place: paddle::platform::CPUPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[float32], place: paddle::platform::CUDAPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[int32], place: paddle::platform::CUDAPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[float64], place: paddle::platform::CUDAPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[int64], place: paddle::platform::CUDAPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[bool], place: paddle::platform::CUDAPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[uint16], place: paddle::platform::CUDAPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[uint8], place: paddle::platform::CUDAPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[int8], place: paddle::platform::CUDAPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[float32], place: paddle::platform::CUDAPinnedPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[int32], place: paddle::platform::CUDAPinnedPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[float64], place: paddle::platform::CUDAPinnedPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[int64], place: paddle::platform::CUDAPinnedPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[bool], place: paddle::platform::CUDAPinnedPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[uint16], place: paddle::platform::CUDAPinnedPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[uint8], place: paddle::platform::CUDAPinnedPlace) -> None
  • set(self: paddle.fluid.core_avx.Tensor, array: numpy.ndarray[int8], place: paddle::platform::CUDAPinnedPlace) -> None
  • 参数:
    • array (numpy.ndarray) - 要设置的numpy array,支持的数据类型为bool, float32, float64, int8, int32, int64, uint8, uint16。
    • place (CPUPlace|CUDAPlace|CUDAPinnedPlace) - 要设置的Tensor所在的设备。

返回:无。

示例代码

  1. import paddle.fluid as fluid
  2. import numpy as np
  3.  
  4. t = fluid.Tensor()
  5. t.set(np.ndarray([5, 30]), fluid.CPUPlace())
  • shape(self: paddle.fluid.core_avx.Tensor) → List[int]

该接口返回Tensor的shape。

返回:Tensor的shape。

返回类型:List[int] 。

示例代码

  1. import paddle.fluid as fluid
  2. import numpy as np
  3.  
  4. t = fluid.Tensor()
  5. t.set(np.ndarray([5, 30]), fluid.CPUPlace())
  6. print(t.shape()) # [5, 30]