代码拉取完成,页面将自动刷新
Compared with the previous version, the added, deleted and supported platforms change information of mindspore.ops
operators in MindSpore, is shown in the following table.
API | Status | Description | Support Platform | Class |
---|---|---|---|---|
mindspore.ops.cummin | New | Computation of the cumulative minimum of elements of ‘x’ in the dimension axis, and the index location of each maximum value found in the dimension ‘axis’. | r1.6: Ascend | composite |
mindspore.ops.Map | New | Map will apply the set operation on input sequences. | r1.6: Ascend/GPU/CPU | composite |
mindspore.ops.clip_by_global_norm | Changed | Clips tensor values by the ratio of the sum of their norms. | r1.5: Ascend/GPU => r1.6: Ascend/GPU/CPU | composite |
mindspore.ops.clip_by_value | Changed | Clips tensor values to a specified min and max. | r1.5: Ascend/GPU => r1.6: Ascend/GPU/CPU | composite |
mindspore.ops.check_bprop | New | Refer to mindspore.ops.CheckBprop. | r1.6: Ascend/GPU/CPU | functional |
mindspore.ops.partial | New | Refer to mindspore.ops.Partial. | r1.6: Ascend/GPU/CPU | functional |
mindspore.ops.select | Deleted | Refer to mindspore.ops.Select. | functional | |
mindspore.ops.AiCPURegOp | New | Class for AiCPU operator information register. | op_info_register | |
mindspore.ops.custom_info_register | New | A decorator which is used to bind the registration information to the func parameter of mindspore.ops.Custom. | op_info_register | |
mindspore.ops.CustomRegOp | New | Class used for generating the registration information for the func parameter of mindspore.ops.Custom. | op_info_register | |
mindspore.ops.ExtractVolumePatches | New | Extract patches from input and put them in the “depth” output dimension. | r1.6: Ascend | operations--Array Operators |
mindspore.ops.SpaceToBatch | Changed | SpaceToBatch is deprecated. | r1.5: Ascend/GPU => r1.6: Deprecated | operations--Array Operators |
mindspore.ops.AlltoAll | New | AlltoAll is a collective operation. | r1.6: Ascend | operations--Communication Operators |
mindspore.ops.NeighborExchange | New | NeighborExchange is a collective operation. | r1.6: Ascend | operations--Communication Operators |
mindspore.ops.NeighborExchangeV2 | New | NeighborExchangeV2 is a collective operation. | r1.6: Ascend | operations--Communication Operators |
mindspore.ops.Custom | New | Custom primitive is used for user defined operators and is to enhance the expressive ability of built-in primitives. | r1.6: Ascend/GPU/CPU | operations--Custom Operators |
mindspore.ops.HookBackward | Changed | This operation is used as a tag to hook gradient in intermediate variables. | r1.5: To Be Developed => r1.6: Ascend/GPU/CPU | operations--Debug Operators |
mindspore.ops.Cdist | New | Computes batched the p-norm distance between each pair of the two collections of row vectors. | r1.6: Ascend | operations--Math Operators |
mindspore.ops.Ger | New | Ger product of x1 and x2 . | r1.6: Ascend | operations--Math Operators |
mindspore.ops.LpNorm | New | Returns the matrix norm or vector norm of a given tensor. | r1.6: Ascend | operations--Math Operators |
mindspore.ops.IsInf | Changed | Determines which elements are inf or -inf for each position | r1.5: GPU => r1.6: GPU/CPU | operations--Math Operators |
mindspore.ops.MatrixInverse | Changed | Returns the inverse of the input matrix. | r1.5: GPU => r1.6: GPU/CPU | operations--Math Operators |
mindspore.ops.Reciprocal | Changed | Returns reciprocal of a tensor element-wise. | r1.5: Ascend/GPU/CPU => r1.6: Ascend/GPU | operations--Math Operators |
mindspore.ops.Conv2DTranspose | New | Compute a 2D transposed convolution, which is also known as a deconvolution (although it is not an actual deconvolution). | r1.6: Ascend/GPU/CPU | operations--Neural Network Operators |
mindspore.ops.Conv2DBackpropInput | Changed | The Conv2DBackpropInput interface is deprecated, please refer to mindspore.ops.Conv2dTranspose if you want to do unsampling. | r1.5: Ascend/GPU/CPU => r1.6: Deprecated | operations--Neural Network Operators |
mindspore.ops.DepthwiseConv2dNative | Changed | DepthwiseConv2dNative will be deprecated in the future. | r1.5: Ascend => r1.6: Deprecated | operations--Neural Network Operators |
mindspore.ops.DropoutDoMask | Changed | The DropoutDoMask interface is deprecated, please use the mindspore.ops.Dropout instead. | r1.5: Ascend => r1.6: Deprecated | operations--Neural Network Operators |
mindspore.ops.DropoutGenMask | Changed | The DropoutGenMask interface is deprecated, please use the mindspore.ops.Dropout instead. | r1.5: Ascend => r1.6: Deprecated | operations--Neural Network Operators |
mindspore.ops.CheckBprop | New | Checks whether the data type and the shape of corresponding elements from tuples x and y are the same. | r1.6: Ascend/GPU/CPU | operations--Other Operators |
mindspore.ops.Partial | New | Makes a partial function instance. | r1.6: Ascend/GPU/CPU | operations--Other Operators |
mindspore.ops.Abs | Changed | Returns absolute value of a tensor element-wise. | r1.6: Ascend/GPU/CPU | operations--r1.5: functional => r1.6: Math Operators |
此处可能存在不合适展示的内容,页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容,可点击提交进行申诉,我们将尽快为您处理。