mindspore.mint

mindspore.mint provides a large number of functional, nn, optimizer interfaces. The API usages and functions are consistent with the mainstream usage in the industry for easy reference. The mint interface is currently an experimental interface and performs better than ops in graph mode of O0 and PyNative mode. Currently, the O2 (graph sinking mode) and CPU/GPU backend are not supported, and it will be gradually improved in the future.

The module import method is as follows:

from mindspore import mint

Compared with the previous version, the added, deleted and supported platforms change information of mindspore.mint operators in MindSpore, please refer to the link mindspore.mint API Interface Change .

Tensor

Creation Operations

API Name	Description	Supported Platforms
mindspore.mint.arange	Creates a sequence of numbers that begins at start and extends by increments of step up to but not including end.	Ascend
mindspore.mint.bernoulli	Sample from the Bernoulli distribution and randomly set the i^{th} element of the output to (0 or 1) according to the i^{th} probability value given in the input.	Ascend
mindspore.mint.bincount	Count the occurrences of each value in the input.	Ascend
mindspore.mint.clone	Returns a copy of the input tensor.	Ascend
mindspore.mint.eye	Returns a tensor with ones on the diagonal and zeros in the rest.	Ascend
mindspore.mint.einsum	According to the Einstein summation Convention (Einsum), the product of the input tensor elements is summed along the specified dimension.	Ascend
mindspore.mint.empty	Creates a tensor with uninitialized data, whose shape, dtype and device are described by the argument size, dtype and device respectively.	Ascend
mindspore.mint.empty_like	Returns an uninitialized Tensor with the same shape as the input.	Ascend
mindspore.mint.full	Create a Tensor of the specified shape and fill it with the specified value.	Ascend
mindspore.mint.full_like	Return a Tensor of the same shape as input and filled with fill_value.	Ascend
mindspore.mint.linspace	Generate a one-dimensional tensor with steps elements, evenly distributed in the interval [start, end].	Ascend
mindspore.mint.ones	Creates a tensor filled with value ones.	Ascend
mindspore.mint.ones_like	Creates a tensor filled with 1, with the same shape as input, and its data type is determined by the given dtype.	Ascend
mindspore.mint.randint	Returns a new tensor filled with integer numbers from the uniform distribution over an interval $[low,high)$ based on the given shape and dtype.	Ascend
mindspore.mint.randint_like	Returns a new tensor filled with integer numbers from the uniform distribution over an interval $[low,high)$ based on the given dtype and shape of the input tensor.	Ascend
mindspore.mint.randn	Returns a new tensor filled with numbers from the normal distribution over an interval $[0,1)$ based on the given shape and dtype.	Ascend
mindspore.mint.randn_like	Returns a new tensor filled with numbers from the normal distribution over an interval $[0,1)$ based on the given dtype and shape of the input tensor.	Ascend
mindspore.mint.randperm	Generates random permutation of integers from 0 to n-1.	Ascend
mindspore.mint.zeros	Creates a tensor filled with 0 with shape described by size and fills it with value 0 in type of dtype.	Ascend
mindspore.mint.zeros_like	Creates a tensor filled with 0, with the same size as input.	Ascend

Indexing, Slicing, Joining, Mutating Operations

API Name	Description	Supported Platforms
mindspore.mint.cat	Connect input tensors along with the given dimension.	Ascend
mindspore.mint.chunk	Cut the input Tensor into chunks sub-tensors along the specified axis.	Ascend
mindspore.mint.concat	Alias for mindspore.mint.cat().	Ascend
mindspore.mint.count_nonzero	Count the number of non-zero elements in the Tensor input on a given dimension dim.	Ascend
mindspore.mint.gather	Gather data from a tensor by indices.	Ascend
mindspore.mint.index_add	Accumulate the elements of alpha times source into the input by adding to the index in the order given in index.	Ascend
mindspore.mint.index_select	Generates a new Tensor that accesses the values of input along the specified dim dimension using the indices specified in index.	Ascend
mindspore.mint.masked_select	Return a new 1-D tensor which indexes the input tensor according to the boolean mask.	Ascend
mindspore.mint.permute	Permutes the dimensions of the input tensor according to input dims .	Ascend
mindspore.mint.reshape	Reshape the input tensor based on the given shape.	Ascend
mindspore.mint.scatter	Update the value in src to input according to the specified index.	Ascend
mindspore.mint.scatter_add	Add all elements in src to the index specified by index to input along dimension specified by dim.	Ascend
mindspore.mint.split	Splits the Tensor into chunks along the given dim.	Ascend
mindspore.mint.narrow	Obtains a tensor of a specified length at a specified start position along a specified axis.	Ascend
mindspore.mint.nonzero	Return the positions of all non-zero values.	Ascend
mindspore.mint.tile	Creates a new tensor by repeating the elements in the input tensor dims times.	Ascend
mindspore.mint.tril	Zero the input tensor above the diagonal specified.	Ascend
mindspore.mint.select	Slices the input tensor along the selected dimension at the given index.	Ascend
mindspore.mint.squeeze	Return the Tensor after deleting the dimension of size 1 in the specified dim.	Ascend
mindspore.mint.stack	Stacks a list of tensors in specified dim.	Ascend
mindspore.mint.swapaxes	Alias for mindspore.mint.transpose() .	Ascend
mindspore.mint.transpose	Interchange two axes of a tensor.	Ascend
mindspore.mint.triu	Zero the input tensor below the diagonal specified.	Ascend
mindspore.mint.unbind	Unbind a tensor dimension in specified axis.	Ascend
mindspore.mint.unique_consecutive	Returns the elements that are unique in each consecutive group of equivalent elements in the input tensor.	Ascend
mindspore.mint.unsqueeze	Adds an additional dimension to the input tensor at the given dimension.	Ascend
mindspore.mint.where	Selects elements from input or other based on condition and returns a tensor.	Ascend

Random Sampling

API Name	Description	Supported Platforms
mindspore.mint.multinomial	Returns a tensor sampled from the multinomial probability distribution located in the corresponding row of the input tensor.	Ascend
mindspore.mint.normal	Generates random numbers according to the standard Normal (or Gaussian) random number distribution.	Ascend
mindspore.mint.rand_like	Returns a new tensor that fills numbers from the uniform distribution over an interval $[0,1)$ based on the given dtype and shape of the input tensor.	Ascend
mindspore.mint.rand	Returns a new tensor that fills numbers from the uniform distribution over an interval $[0,1)$ based on the given shape and dtype.	Ascend

Math Operations

Pointwise Operations

API Name	Description	Supported Platforms
mindspore.mint.abs	Compute the absolute value of a tensor element-wise.	Ascend
mindspore.mint.add	Adds scaled other value to self.	Ascend
mindspore.mint.addmv	Performs a matrix-vector product of mat and vec, and add the input vector input to the final result.	Ascend
mindspore.mint.acos	Computes arccosine of input tensors element-wise.	Ascend
mindspore.mint.acosh	Computes inverse hyperbolic cosine of the inputs element-wise.	Ascend
mindspore.mint.arccos	Alias for mindspore.mint.acos() .	Ascend
mindspore.mint.arccosh	Alias for mindspore.mint.acosh().	Ascend
mindspore.mint.arcsin	Alias for mindspore.mint.asin().	Ascend
mindspore.mint.arcsinh	Alias for mindspore.mint.asinh().	Ascend
mindspore.mint.arctan	Alias for mindspore.mint.atan().	Ascend
mindspore.mint.arctan2	Alias for mindspore.mint.atan2().	Ascend
mindspore.mint.arctanh	Alias for mindspore.mint.atanh().	Ascend
mindspore.mint.asin	Computes arcsine of input tensors element-wise.	Ascend
mindspore.mint.asinh	Computes inverse hyperbolic sine of the input element-wise.	Ascend
mindspore.mint.atan	Computes the trigonometric inverse tangent of the input element-wise.	Ascend
mindspore.mint.atan2	Returns arctangent of input/other element-wise.	Ascend
mindspore.mint.atanh	Computes inverse hyperbolic tangent of the input element-wise.	Ascend
mindspore.mint.bitwise_and	Returns bitwise and of two tensors element-wise.	Ascend
mindspore.mint.bitwise_or	Returns bitwise or of two tensors element-wise.	Ascend
mindspore.mint.bitwise_xor	Returns bitwise xor of two tensors element-wise.	Ascend
mindspore.mint.ceil	Rounds a tensor up to the closest integer element-wise.	Ascend
mindspore.mint.clamp	Clamps tensor values between the specified minimum value and maximum value.	Ascend
mindspore.mint.cos	Computes cosine of input element-wise.	Ascend
mindspore.mint.cosh	Computes hyperbolic cosine of input element-wise.	Ascend
mindspore.mint.cross	Compute the cross product of two input tensors along the specified dimension.	Ascend
mindspore.mint.diff	Computes the n-th forward difference along the given dimension.	Ascend
mindspore.mint.div	Divides each element of the input by the corresponding element of the other .	Ascend
mindspore.mint.divide	Alias for mindspore.mint.div() .	Ascend
mindspore.mint.erf	Compute the Gauss error of input tensor element-wise.	Ascend
mindspore.mint.erfc	Compute the complementary error function of input tensor element-wise.	Ascend
mindspore.mint.erfinv	Compute the inverse error of input tensor element-wise.	Ascend
mindspore.mint.exp	Compute exponential of the input tensor element-wise.	Ascend
mindspore.mint.exp2	Calculates the base-2 exponent of the Tensor input element by element.	Ascend
mindspore.mint.expm1	Compute exponential of the input tensor, then minus 1, element-wise.	Ascend
mindspore.mint.fix	Alias for mindspore.mint.trunc() .	Ascend
mindspore.mint.float_power	Computes input to the power of exponent element-wise in double precision, and always returns a mindspore.float64 tensor.	Ascend
mindspore.mint.floor	Rounds a tensor down to the closest integer element-wise.	Ascend
mindspore.mint.fmod	Computes the floating-point remainder of the division operation input/other.	Ascend
mindspore.mint.frac	Calculates the fractional part of each element in the input.	Ascend
mindspore.mint.lerp	Perform a linear interpolation of two tensors input and end based on a float or tensor weight.	Ascend
mindspore.mint.log	Compute the natural logarithm of the input tensor element-wise.	Ascend
mindspore.mint.log1p	Compute the natural logarithm of (tensor + 1) element-wise.	Ascend
mindspore.mint.log2	Returns the logarithm to the base 2 of a tensor element-wise.	Ascend
mindspore.mint.log10	Returns the logarithm to the base 10 of a tensor element-wise.	Ascend
mindspore.mint.logaddexp	Computes the logarithm of the sum of exponentiations of the inputs.	Ascend
mindspore.mint.logaddexp2	Logarithm of the sum of exponentiations of the inputs in base of 2.	Ascend
mindspore.mint.logical_and	Compute the "logical AND" of two tensors element-wise.	Ascend
mindspore.mint.logical_not	Compute the "logical NOT" of the input tensor element-wise.	Ascend
mindspore.mint.logical_or	Compute the "logical OR" of two tensors element-wise.	Ascend
mindspore.mint.logical_xor	Compute the "logical XOR" of two tensors element-wise.	Ascend
mindspore.mint.mul	Multiply other value by input Tensor.	Ascend
mindspore.mint.mv	Multiply matrix input and vector vec.	Ascend
mindspore.mint.nansum	Computes sum of input over a given dimension, treating NaNs as zero.	Ascend
mindspore.mint.nan_to_num	Replace the NaN, positive infinity and negative infinity values in input with the specified values in nan, posinf and neginf respectively.	Ascend
mindspore.mint.neg	Returns a tensor with negative values of the input tensor element-wise.	Ascend
mindspore.mint.negative	Alias for mindspore.mint.neg() .	Ascend
mindspore.mint.pow	Calculates the exponent power of each element in input.	Ascend
mindspore.mint.polar	Converts polar coordinates to Cartesian coordinates.	Ascend
mindspore.mint.ravel	Expand the multidimensional Tensor into 1D along the 0 axis direction.	Ascend
mindspore.mint.reciprocal	Returns reciprocal of a tensor element-wise.	Ascend
mindspore.mint.remainder	Computes the remainder of input divided by other element-wise.	Ascend
mindspore.mint.roll	Roll the elements of a tensor along a dimension.	Ascend
mindspore.mint.round	Round elements of input to the nearest integer.	Ascend
mindspore.mint.rsqrt	Compute reciprocal of square root of input tensor element-wise.	Ascend
mindspore.mint.sigmoid	Computes Sigmoid of input element-wise.	Ascend
mindspore.mint.sign	Return an element-wise indication of the sign of a number.	Ascend
mindspore.mint.sin	Compute sine of the input tensor element-wise.	Ascend
mindspore.mint.sinc	Compute the normalized sinc of input.	Ascend
mindspore.mint.sinh	Compute hyperbolic sine of the input element-wise.	Ascend
mindspore.mint.softmax	Alias for mindspore.mint.nn.functional.softmax().	Ascend
mindspore.mint.sqrt	Returns sqrt of a tensor element-wise.	Ascend
mindspore.mint.square	Return square of a tensor element-wise.	Ascend
mindspore.mint.sub	Subtracts scaled other value from self Tensor.	Ascend
mindspore.mint.t	Transpose the input tensor.	Ascend
mindspore.mint.tan	Computes tangent of input element-wise.	Ascend
mindspore.mint.tanh	Computes hyperbolic tangent of input element-wise.	Ascend
mindspore.mint.trunc	Returns a tensor with the truncated integer values of the elements of the input tensor.	Ascend
mindspore.mint.xlogy	Computes the first input multiplied by the logarithm of second input element-wise.	Ascend

Reduction Operations

API Name	Description	Supported Platforms
mindspore.mint.amax	Computes the maximum value of of all elements along the specified dim dimension of the input, and retains the dimension based on the keepdim parameter.	Ascend
mindspore.mint.amin	Computes the minimum value of of all elements along the specified dim dimension of the input, and retains the dimension based on the keepdim parameter.	Ascend
mindspore.mint.argmax	Return the indices of the maximum values of a tensor.	Ascend
mindspore.mint.argmin	Return the indices of the minimum values of a tensor across a dimension.	Ascend
mindspore.mint.argsort	Sorts the input tensor along the given dimension in specified order and return the sorted indices.	Ascend
mindspore.mint.all	Tests if all element in input evaluates to True.	Ascend
mindspore.mint.any	Tests if any element in input evaluates to True along the given axes.	Ascend
mindspore.mint.cumprod	Return the cumulative product along the given dimension of the tensor.	Ascend
mindspore.mint.histc	Computes the histogram of a tensor.	Ascend
mindspore.mint.logsumexp	Computes the logarithm of the sum of exponentiations of all elements along the specified dim dimension of the input (with numerical stabilization), and retains the dimension based on the keepdim parameter.	Ascend
mindspore.mint.max	Returns the maximum value of the input tensor.	Ascend
mindspore.mint.mean	Reduces all dimension of a tensor by averaging all elements.	Ascend
mindspore.mint.median	Output the median on the specified dimension dim and its corresponding index.	Ascend
mindspore.mint.min	Returns the minimum value of the input tensor.	Ascend
mindspore.mint.norm	Returns the matrix norm or vector norm of a given tensor.	Ascend
mindspore.mint.prod	Multiplying all elements of input.	Ascend
mindspore.mint.sum	Calculate sum of all elements in Tensor.	Ascend
mindspore.mint.std	Calculates the standard deviation over the dimensions specified by dim.	Ascend
mindspore.mint.std_mean	By default, return the standard deviation and mean of each dimension in Tensor.	Ascend
mindspore.mint.unique	Returns the unique elements of input tensor.	Ascend
mindspore.mint.var	Calculates the variance over the dimensions specified by dim.	Ascend
mindspore.mint.var_mean	By default, return the variance and mean of each dimension in Tensor.	Ascend

Comparison Operations

API Name	Description	Supported Platforms
mindspore.mint.allclose	Returns a new Tensor with boolean elements representing if each element of input is “close” to the corresponding element of other.	Ascend
mindspore.mint.argsort	Sorts the input tensor along the given dimension in specified order and return the sorted indices.	Ascend
mindspore.mint.eq	Compute the equivalence of the two inputs element-wise.	Ascend
mindspore.mint.equal	Computes the equivalence between two tensors.	Ascend
mindspore.mint.greater	Compute the value of $input>other$ element-wise.	Ascend
mindspore.mint.greater_equal	Computes the boolean value of $input>=other$ element-wise.	Ascend
mindspore.mint.gt	Compute the value of $input>other$ element-wise.	Ascend
mindspore.mint.isclose	Return a boolean tensor where two tensors are element-wise equal within a tolerance.	Ascend
mindspore.mint.isfinite	Determine which elements are finite for each position.	Ascend
mindspore.mint.isinf	Return a boolean tensor indicating which elements are +/- inifnity.	Ascend
mindspore.mint.isneginf	Determines which elements are -inf for each position.	Ascend
mindspore.mint.le	Compute the value of $input<=other$ element-wise.	Ascend
mindspore.mint.less	Compute the value of $input<other$ element-wise.	Ascend
mindspore.mint.less_equal	Compute the value of $input<=other$ element-wise.	Ascend
mindspore.mint.lt	Alias for mindspore.mint.less() .	Ascend
mindspore.mint.maximum	Compute the maximum of the two input tensors element-wise.	Ascend
mindspore.mint.minimum	Compute the minimum of the two input tensors element-wise.	Ascend
mindspore.mint.ne	Compute the non-equivalence of two inputs element-wise.	Ascend
mindspore.mint.not_equal	Alias for mindspore.mint.ne() .	Ascend
mindspore.mint.topk	Finds values and indices of the k largest or smallest entries along a given dimension.	Ascend
mindspore.mint.sort	Sorts the elements of the input tensor along the given dimension in the specified order.	Ascend

BLAS and LAPACK Operations

API Name	Description	Supported Platforms
mindspore.mint.addbmm	Applies batch matrix multiplication to batch1 and batch2, with a reduced add step and add input to the result.	Ascend
mindspore.mint.addmm	Performs a matrix multiplication of the 2-D matrices mat1 and mat2.	Ascend
mindspore.mint.baddbmm	The result is the sum of the input and a batch matrix-matrix product of matrices in batch1 and batch2.	Ascend
mindspore.mint.bmm	Performs batch matrix-matrix multiplication of two three-dimensional tensors.	Ascend
mindspore.mint.dot	Computes the dot product of two 1D tensor.	Ascend
mindspore.mint.inverse	Compute the inverse of the input matrix.	Ascend
mindspore.mint.matmul	Return the matrix product of two tensors.	Ascend
mindspore.mint.meshgrid	Generates coordinate matrices from given coordinate tensors.	Ascend
mindspore.mint.mm	Returns the matrix product of two arrays.	Ascend
mindspore.mint.outer	Return outer product of input and vec2.	Ascend
mindspore.mint.trace	Returns a new tensor that is the sum of the input main trace.	Ascend

Other Operations

API Name	Description	Supported Platforms
mindspore.mint.broadcast_to	Broadcasts input tensor to a given shape.	Ascend
mindspore.mint.cdist	Computes p-norm distance between each pair of row vectors of two input Tensors.	Ascend
mindspore.mint.cummax	Returns a tuple (values, indices) where values is the cumulative maximum value of input Tensor input along the dimension dim, and indices is the index location of each maximum value.	Ascend
mindspore.mint.cummin	Returns a tuple (values, indices) where values is the cumulative minimum value of input Tensor input along the dimension dim, and indices is the index location of each minimum value.	Ascend
mindspore.mint.cumsum	Computes the cumulative sum of input Tensor along dim.	Ascend
mindspore.mint.diag	If input is a vector (1-D tensor), then returns a 2-D square tensor with the elements of input as the diagonal.	Ascend
mindspore.mint.flatten	Flatten a tensor along dimensions from start_dim to end_dim.	Ascend
mindspore.mint.flip	Reverses elements in a tensor along the given dims.	Ascend
mindspore.mint.repeat_interleave	Repeat elements of a tensor along an axis, like mindspore.numpy.repeat().	Ascend
mindspore.mint.searchsorted	Return the position indices where the elements can be inserted into the input tensor to maintain the increasing order of the input tensor.	Ascend
mindspore.mint.tril	Zero the input tensor above the diagonal specified.	Ascend
mindspore.mint.triangular_solve	Solves a system of equations with a square upper or lower triangular invertible matrix A and multiple right-hand sides b.	Ascend

mindspore.mint.nn

Convolution Layers

API Name	Description	Supported Platforms
mindspore.mint.nn.Conv2d	2D convolution layer.	Ascend
mindspore.mint.nn.Conv3d	3D convolution layer.	Ascend
mindspore.mint.nn.ConvTranspose2d	Applies a 2D transposed convolution operator over an input image composed of several input planes.	Ascend
mindspore.mint.nn.Fold	Combines an array of sliding local blocks into a large containing tensor.	Ascend
mindspore.mint.nn.Unfold	Extracts sliding local blocks from a batched input tensor.	Ascend

Normalization Layers

API Name	Description	Supported Platforms
mindspore.mint.nn.BatchNorm1d	Applies Batch Normalization over a 2D or 3D input as described in the paper Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift .	Ascend
mindspore.mint.nn.BatchNorm2d	Applies Batch Normalization over a 4D input as described in the paper Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift .	Ascend
mindspore.mint.nn.BatchNorm3d	Applies Batch Normalization over a 5D input as described in the paper Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift .	Ascend
mindspore.mint.nn.GroupNorm	Group Normalization over a mini-batch of inputs.	Ascend
mindspore.mint.nn.LayerNorm	Applies Layer Normalization over a mini-batch of inputs.	Ascend
mindspore.mint.nn.SyncBatchNorm	Sync Batch Normalization layer over a N-dimension input.	Ascend

Non-linear Activations (weighted sum, nonlinearity)

API Name	Description	Supported Platforms
mindspore.mint.nn.ELU	Exponential Linear Unit activation function	Ascend
mindspore.mint.nn.GELU	Activation function GELU (Gaussian Error Linear Unit).	Ascend
mindspore.mint.nn.GLU	Computes GLU (Gated Linear Unit activation function) of the input tensor.	Ascend
mindspore.mint.nn.Hardshrink	Applies Hard Shrink activation function element-wise.	Ascend
mindspore.mint.nn.Hardsigmoid	Applies Hard Sigmoid activation function element-wise.	Ascend
mindspore.mint.nn.Hardswish	Applies Hard Swish activation function element-wise.	Ascend
mindspore.mint.nn.LogSigmoid	Applies logsigmoid activation element-wise.	Ascend
mindspore.mint.nn.LogSoftmax	Applies the Log Softmax function to the input tensor on the specified axis.	Ascend
mindspore.mint.nn.Mish	Computes MISH (A Self Regularized Non-Monotonic Neural Activation Function) of input tensors element-wise.	Ascend
mindspore.mint.nn.PReLU	Applies PReLU activation function element-wise.	Ascend
mindspore.mint.nn.ReLU	Applies ReLU (Rectified Linear Unit activation function) element-wise.	Ascend
mindspore.mint.nn.ReLU6	Activation function ReLU6.	Ascend
mindspore.mint.nn.SELU	Activation function SELU (Scaled exponential Linear Unit).	Ascend
mindspore.mint.nn.SiLU	Calculates the SiLU activation function element-wise.	Ascend
mindspore.mint.nn.Sigmoid	Applies sigmoid activation function element-wise.	Ascend
mindspore.mint.nn.Softmax	Applies the Softmax function to an n-dimensional input Tensor.	Ascend
mindspore.mint.nn.Softshrink	Applies the Softshrink function element-wise.	Ascend
mindspore.mint.nn.Tanh	Applies the Tanh function element-wise, returns a new tensor with the hyperbolic tangent of the elements of input.	Ascend

Embedding Layers

API Name	Description	Supported Platforms
mindspore.mint.nn.Embedding	The value in input is used as the index, and the corresponding embedding vector is queried from weight .	Ascend

Linear Layers

API Name	Description	Supported Platforms
mindspore.mint.nn.Linear	The linear connected layer.	Ascend

Dropout Layers

API Name	Description	Supported Platforms
mindspore.mint.nn.Dropout	Dropout layer for the input.	Ascend
mindspore.mint.nn.Dropout2d	During training, randomly zeroes some channels of the input tensor with probability p from a Bernoulli distribution (For a 4-dimensional tensor with a shape of $NCHW$, the channel feature map refers to a 2-dimensional feature map with the shape of $HW$).	Ascend

Pooling Layers

API Name	Description	Supported Platforms
mindspore.mint.nn.AdaptiveAvgPool1d	Applies a 1D adaptive average pooling over an input signal composed of several input planes.	Ascend
mindspore.mint.nn.AdaptiveAvgPool2d	Applies a 2D adaptive average pooling over an input signal composed of several input planes.	Ascend
mindspore.mint.nn.AdaptiveAvgPool3d	This operator applies a 3D adaptive average pooling to an input signal composed of multiple input planes.	Ascend
mindspore.mint.nn.AdaptiveMaxPool1d	Applies a 1D adaptive max pooling over an input signal composed of several input planes.	Ascend
mindspore.mint.nn.AvgPool2d	Applies a 2D average pooling over an input Tensor which can be regarded as a composition of 2D input planes.	Ascend
mindspore.mint.nn.AvgPool3d	Applies a 3D average pooling over an input Tensor which can be regarded as a composition of 3D input planes.	Ascend
mindspore.mint.nn.MaxUnpool2d	Computes the inverse of Maxpool2d.	Ascend

Padding Layers

API Name	Description	Supported Platforms
mindspore.mint.nn.ConstantPad1d	Pad the last dimension of input tensor using padding and value.	Ascend
mindspore.mint.nn.ConstantPad2d	Pad the last 2 dimensions of input tensor using padding and value.	Ascend
mindspore.mint.nn.ConstantPad3d	Pad the last 3 dimension of input tensor using padding and value.	Ascend
mindspore.mint.nn.ReflectionPad1d	Pad the last dimension of input tensor using the reflection of the input boundary.	Ascend
mindspore.mint.nn.ReflectionPad2d	Pad the last 2 dimension of input tensor using the reflection of the input boundary.	Ascend
mindspore.mint.nn.ReflectionPad3d	Pad the last 3 dimension of input tensor using the reflection of the input boundary.	Ascend
mindspore.mint.nn.ReplicationPad1d	Pad the last dimension of input tensor using the replication of the input boundary.	Ascend
mindspore.mint.nn.ReplicationPad2d	Pad the last 2 dimension of input tensor using the replication of the input boundary.	Ascend
mindspore.mint.nn.ReplicationPad3d	Pad the last 3 dimension of input tensor using the replication of the input boundary.	Ascend
mindspore.mint.nn.ZeroPad1d	Pad the last dimension of input tensor with 0 using padding.	Ascend
mindspore.mint.nn.ZeroPad2d	Pad the last 2 dimension of input tensor with 0 using padding.	Ascend
mindspore.mint.nn.ZeroPad3d	Pad the last 3 dimension of input tensor with 0 using padding.	Ascend

Loss Functions

API Name	Description	Supported Platforms
mindspore.mint.nn.BCELoss	Compute the binary cross entropy between the true labels and predicted labels.	Ascend
mindspore.mint.nn.BCEWithLogitsLoss	Adds sigmoid activation function to input as logits, and uses this logits to compute binary cross entropy between the logits and the target.	Ascend
mindspore.mint.nn.CrossEntropyLoss	The cross entropy loss between input and target.	Ascend
mindspore.mint.nn.KLDivLoss	Computes the Kullback-Leibler divergence between the input and the target.	Ascend
mindspore.mint.nn.L1Loss	L1Loss is used to calculate the mean absolute error between the predicted value and the target value.	Ascend
mindspore.mint.nn.MSELoss	Calculates the mean squared error between the predicted value and the label value.	Ascend
mindspore.mint.nn.NLLLoss	Gets the negative log likelihood loss between inputs and target.	Ascend
mindspore.mint.nn.SmoothL1Loss	Computes smooth L1 loss, a robust L1 loss.	Ascend

Vision Layer

API Name	Description	Supported Platforms
mindspore.mint.nn.PixelShuffle	Rearrange elements in a tensor according to an upscaling factor.	Ascend
mindspore.mint.nn.Upsample	For details, please refer to mindspore.mint.nn.functional.interpolate().	Ascend

Tools

API Name	Description	Supported Platforms
mindspore.mint.nn.Identity	A placeholder identity operator that returns the same as input.	Ascend

mindspore.mint.nn.functional

Convolution functions

API Name	Description	Supported Platforms
mindspore.mint.nn.functional.conv2d	Applies a 2D convolution over an input tensor.	Ascend
mindspore.mint.nn.functional.conv3d	Applies a 3D convolution over an input tensor.	Ascend
mindspore.mint.nn.functional.conv_transpose2d	Applies a 2D transposed convolution operator over an input image composed of several input planes, sometimes also called deconvolution (although it is not an actual deconvolution).	Ascend
mindspore.mint.nn.functional.fold	Combines an array of sliding local blocks into a large containing tensor.	Ascend
mindspore.mint.nn.functional.unfold	Extracts sliding local blocks from a batched input tensor.	Ascend

Pooling functions

API Name	Description	Supported Platforms
mindspore.mint.nn.functional.adaptive_avg_pool1d	Performs 1D adaptive average pooling on a multi-plane input signal.	Ascend
mindspore.mint.nn.functional.adaptive_avg_pool2d	Performs 2D adaptive average pooling on a multi-plane input signal.	Ascend
mindspore.mint.nn.functional.adaptive_avg_pool3d	Performs 3D adaptive average pooling on a multi-plane input signal.	Ascend
mindspore.mint.nn.functional.adaptive_max_pool1d	Performs 1D adaptive max pooling on a multi-plane input signal.	Ascend
mindspore.mint.nn.functional.avg_pool1d	Applies a 1D average pooling over an input Tensor which can be regarded as a composition of 1D input planes.	Ascend
mindspore.mint.nn.functional.avg_pool2d	Applies a 2D average pooling over an input Tensor which can be regarded as a composition of 2D input planes.	Ascend
mindspore.mint.nn.functional.avg_pool3d	Applies a 3D average pooling over an input Tensor which can be regarded as a composition of 3D input planes.	Ascend
mindspore.mint.nn.functional.max_pool2d	Performs a 2D max pooling on the input Tensor.	Ascend
mindspore.mint.nn.functional.max_unpool2d	Computes the inverse of max_pool2d.	Ascend

Non-linear activation functions

API Name	Description	Supported Platforms
mindspore.mint.nn.functional.batch_norm	Batch Normalization for input data and updated parameters.	Ascend
mindspore.mint.nn.functional.elu	Exponential Linear Unit activation function	Ascend
mindspore.mint.nn.functional.elu_	Exponential Linear Unit activation function	Ascend
mindspore.mint.nn.functional.gelu	Gaussian Error Linear Units activation function.	Ascend
mindspore.mint.nn.functional.glu	Computes GLU (Gated Linear Unit activation function) of the input tensor.	Ascend
mindspore.mint.nn.functional.group_norm	Group Normalization over a mini-batch of inputs.	Ascend
mindspore.mint.nn.functional.hardshrink	Hard Shrink activation function.	Ascend
mindspore.mint.nn.functional.hardsigmoid	Hard Sigmoid activation function.	Ascend
mindspore.mint.nn.functional.hardswish	Hard Swish activation function.	Ascend
mindspore.mint.nn.functional.layer_norm	Applies the Layer Normalization on the mini-batch input.	Ascend
mindspore.mint.nn.functional.leaky_relu	leaky_relu activation function.	Ascend
mindspore.mint.nn.functional.log_softmax	Applies the Log Softmax function to the input tensor on the specified axis.	Ascend
mindspore.mint.nn.functional.logsigmoid	Applies logsigmoid activation element-wise.	Ascend
mindspore.mint.nn.functional.mish	Computes MISH (A Self Regularized Non-Monotonic Neural Activation Function) of input tensors element-wise.	Ascend
mindspore.mint.nn.functional.prelu	Parametric Rectified Linear Unit activation function.	Ascend
mindspore.mint.nn.functional.relu	Computes ReLU (Rectified Linear Unit activation function) of input tensors element-wise.	Ascend
mindspore.mint.nn.functional.relu6	Computes ReLU (Rectified Linear Unit) upper bounded by 6 of input tensors element-wise.	Ascend
mindspore.mint.nn.functional.relu_	ReLuComputes ReLU (Rectified Linear Unit activation function) inplace of input tensors element-wise.	Ascend
mindspore.mint.nn.functional.selu	Activation function SELU (Scaled exponential Linear Unit).	Ascend
mindspore.mint.nn.functional.sigmoid	Computes Sigmoid of input element-wise.	Ascend
mindspore.mint.nn.functional.silu	Computes Sigmoid Linear Unit of input element-wise, also known as Swish function.	Ascend
mindspore.mint.nn.functional.softmax	Applies the Softmax operation to the input tensor on the specified axis.	Ascend
mindspore.mint.nn.functional.softplus	Applies softplus function to input element-wise.	Ascend
mindspore.mint.nn.functional.softshrink	Soft Shrink activation function.	Ascend
mindspore.mint.nn.functional.tanh	Computes hyperbolic tangent of input element-wise.	Ascend

Normalization functions

API Name	Description	Supported Platforms
mindspore.mint.nn.functional.normalize	Perform normalization of inputs over specified dimension	Ascend

Linear functions

API Name	Description	Supported Platforms
mindspore.mint.nn.functional.linear	Applies the dense connected operation to the input.	Ascend

Dropout functions

API Name	Description	Supported Platforms
mindspore.mint.nn.functional.dropout	During training, randomly zeroes some of the elements of the input tensor with probability p from a Bernoulli distribution.	Ascend
mindspore.mint.nn.functional.dropout2d	During training, randomly zeroes some channels of the input tensor with probability p from a Bernoulli distribution (For a 4-dimensional tensor with a shape of $(N,C,H,W)$, the channel feature map refers to a 2-dimensional feature map with the shape of $(H,W)$).	Ascend

Sparse functions

API Name	Description	Supported Platforms
mindspore.mint.nn.functional.embedding	Retrieve the word embeddings in weight using indices specified in input.	Ascend
mindspore.mint.nn.functional.one_hot	Computes a one-hot tensor.	Ascend

Loss Functions

API Name	Description	Supported Platforms
mindspore.mint.nn.functional.cross_entropy	The cross entropy loss between input and target.	Ascend
mindspore.mint.nn.functional.binary_cross_entropy	Computes the binary cross entropy(Measure the difference information between two probability distributions) between predictive value input and target value target.	Ascend
mindspore.mint.nn.functional.binary_cross_entropy_with_logits	Adds sigmoid activation function to input as logits, and uses this logits to compute binary cross entropy between the logits and the target.	Ascend
mindspore.mint.nn.functional.kl_div	Computes the Kullback-Leibler divergence between the input and the target.	Ascend
mindspore.mint.nn.functional.l1_loss	Calculate the mean absolute error between the input value and the target value.	Ascend
mindspore.mint.nn.functional.mse_loss	Calculates the mean squared error between the predicted value and the label value.	Ascend
mindspore.mint.nn.functional.nll_loss	Gets the negative log likelihood loss between input and target.	Ascend
mindspore.mint.nn.functional.smooth_l1_loss	Computes smooth L1 loss, a robust L1 loss.	Ascend

Vision functions

API Name	Description	Supported Platforms
mindspore.mint.nn.functional.interpolate	Samples the input Tensor to the given size or scale_factor by using one of the interpolate algorithms.	Ascend
mindspore.mint.nn.functional.grid_sample	Given an input and a flow-field grid, computes the output using input values and pixel locations from grid.	Ascend
mindspore.mint.nn.functional.pad	Pads the input tensor according to the pad.	Ascend
mindspore.mint.nn.functional.pixel_shuffle	Rearrange elements in a tensor according to an upscaling factor.	Ascend

mindspore.mint.optim

API Name	Description	Supported Platforms
mindspore.mint.optim.Adam	Implements Adaptive Moment Estimation (Adam) algorithm.	Ascend
mindspore.mint.optim.AdamW	Implements Adam Weight Decay algorithm.	Ascend
mindspore.mint.optim.SGD	Stochastic Gradient Descent optimizer.	Ascend

mindspore.mint.linalg

Inverses

API Name	Description	Supported Platforms
mindspore.mint.linalg.inv	Compute the inverse of the input matrix.	Ascend
mindspore.mint.linalg.matrix_norm	Returns the matrix norm of a given tensor on the specified dimensions.	Ascend
mindspore.mint.linalg.norm	Returns the matrix norm or vector norm of a given tensor.	Ascend
mindspore.mint.linalg.vector_norm	Returns the vector norm of the given tensor on the specified dimensions.	Ascend
mindspore.mint.linalg.qr	Orthogonal decomposition of the input $A=QR$.	Ascend

mindspore.mint.special

Pointwise Operations

API Name	Description	Supported Platforms
mindspore.mint.special.erfc	Compute the complementary error function of input tensor element-wise.	Ascend
mindspore.mint.special.exp2	Calculates the base-2 exponent of the Tensor input element by element.	Ascend
mindspore.mint.special.expm1	Compute exponential of the input tensor, then minus 1, element-wise.	Ascend
mindspore.mint.special.log1p	Compute the natural logarithm of (tensor + 1) element-wise.	Ascend
mindspore.mint.special.log_softmax	Applies the Log Softmax function to the input tensor on the specified axis.	Ascend
mindspore.mint.special.round	Returns half to even of a tensor element-wise.	Ascend
mindspore.mint.special.sinc	Compute the normalized sinc of input.	Ascend

mindspore.mint.distributed

API Name	Description	Supported Platforms
mindspore.mint.distributed.all_gather	Gathers tensors from the specified communication group and returns the tensor list which is all gathered.	Ascend
mindspore.mint.distributed.all_gather_into_tensor	Gathers tensors from the specified communication group and returns the tensor which is all gathered.	Ascend
mindspore.mint.distributed.all_gather_object	Aggregates Python objects in a specified communication group.	Ascend
mindspore.mint.distributed.all_reduce	Reduce tensors across all devices in such a way that all deviceswill get the same final result, returns the tensor which is all reduced.	Ascend
mindspore.mint.distributed.all_to_all	scatter and gather list of tensor to/from all rank according to input/output tensor list.	Ascend
mindspore.mint.distributed.all_to_all_single	scatter and gather input with split size to/from all rank, and return result in a single tensor.	Ascend
mindspore.mint.distributed.barrier	Synchronizes all processes in the specified group.	Ascend
mindspore.mint.distributed.batch_isend_irecv	Batch send and recv tensors asynchronously.	Ascend
mindspore.mint.distributed.broadcast	Broadcasts the tensor to the whole group.	Ascend
mindspore.mint.distributed.broadcast_object_list	Broadcasts the entire group of input Python objects.	Ascend
mindspore.mint.distributed.destroy_process_group	Destroy the user collective communication group.	Ascend
mindspore.mint.distributed.gather	Gathers tensors from the specified communication group.	Ascend
mindspore.mint.distributed.gather_object	Gathers python objects from the whole group in a single process.	Ascend
mindspore.mint.distributed.get_backend	Get the backend of communication process groups.	Ascend
mindspore.mint.distributed.get_global_rank	A function that returns the rank id in the world group corresponding to the rank which id is 'group_rank' in the user group.	Ascend
mindspore.mint.distributed.get_group_rank	Get the rank ID in the specified user communication group corresponding to the rank ID in the world communication group.	Ascend
mindspore.mint.distributed.get_process_group_ranks	Gets the ranks of the specific group and returns the process ranks in the communication group as a list.	Ascend
mindspore.mint.distributed.get_rank	Get the rank ID for the current device in the specified collective communication group.	Ascend
mindspore.mint.distributed.get_world_size	Get the rank size of the specified collective communication group.	Ascend
mindspore.mint.distributed.init_process_group	Init collective communication lib.	Ascend
mindspore.mint.distributed.irecv	Receive tensors from src asynchronously.	Ascend
mindspore.mint.distributed.isend	Send tensors to the specified dest_rank asynchronously.	Ascend
mindspore.mint.distributed.is_available	Checks if distributed module is available.	Ascend
mindspore.mint.distributed.is_initialized	Checks if default process group has been initialized.	Ascend
mindspore.mint.distributed.new_group	Create a new distributed group.	Ascend
mindspore.mint.distributed.P2POp	Object for batch_isend_irecv input, to store information of "isend" and "irecv".	Ascend
mindspore.mint.distributed.recv	Receive tensors from src.	Ascend
mindspore.mint.distributed.reduce	Reduces tensors across the processes in the specified communication group, sends the result to the target dst(global rank), and returns the tensor which is sent to the target process.	Ascend
mindspore.mint.distributed.reduce_scatter	Reduces and scatters tensors from the specified communication group and returns the tensor which is reduced and scattered.	Ascend
mindspore.mint.distributed.reduce_scatter_tensor	Reduces and scatters tensors from the specified communication group and returns the tensor which is reduced and scattered.	Ascend
mindspore.mint.distributed.scatter	Scatter tensor evently across the processes in the specified communication group.	Ascend
mindspore.mint.distributed.scatter_object_list	Scatters picklable objects in scatter_object_input_list to the whole group.	Ascend
mindspore.mint.distributed.send	Send tensors to the specified dest_rank.	Ascend