_api_guide_math_en:

Mathematical operation¶

Paddle provides a wealth of mathematical operations. The mathematical operations listed below are all elementwise operations on the target tensor. If the two inputs of the binary operations have different shapes, they will be processed first by broadcast. Some mathematical operations also support mathematical operators, such as: +, -, *, /, etc. Math operators not only support tensors but also scalars.

Unary operation¶

exp¶

Perform an exp operation on each input Tensor element.

API Reference: exp

tanh¶

For the input Tensor, take the tanh value of each element.

API Reference: tanh

sqrt¶

For the input Tensor, take the square root of each element.

API Reference: sqrt

abs¶

For the input Tensor, take the elementwise absolute value.

API Reference: abs

ceil¶

Round up each input Tensor element to the nearest greater integer.

API Reference: ceil

floor¶

Round down each input Tensor element to the nearest less integer.

API Reference: floor

sin¶

For the input Tensor, take the elementwise sin value.

API Reference: sin

cos¶

For input Tensor, take the elementwise cosine value.

API Reference: cos

round¶

Rounding the input Tensor in elementwise order.

API Reference: round

square¶

Square the input Tensor in elementwise order.

API Reference: square

reciprocal¶

For the input Tensor, take the reciprocal in elementwise order.

API Reference: reciprocal

reduce¶

For the input Tensor, it performs reduce operations on the specified axes, including: min, max, sum, mean, product

API Reference: reduce_min reduce_max fluid_layers_reduce_sum reduce_mean reduce_prod

Binary operation¶

elementwise_add¶

Add two Tensor in elementwise order, and the corresponding math operator is + .

API Reference: elementwise_add

elementwise_sub¶

Sub two Tensor in elementwise order, the corresponding math operator is - .

API Reference: elementwise_sub

elementwise_mul¶

Multiply two Tensor in elementwise order, and the corresponding math operator is * .

API Reference: elementwise_mul

elementwise_div¶

Divide two Tensor in elementwise order, and the corresponding math operator is / or // .

API Reference: elementwise_div

elementwise_pow¶

Do power operations on two Tensor in elementwise order, and the corresponding math operator is ** .

API Reference: elementwise_pow

equal¶

Judge whether the two Tensor elements are equal, and the corresponding math operator is == .

API Reference: equal

less_than¶

Judge whether the two Tensor elements satisfy the ‘less than’ relationship, and the corresponding math operator is < .

API Reference: less_than

sum¶

Add two Tensor in elementwise order.

API Reference: sum

elementwise_min¶

Perform min(x, y) operations on two Tensor in elementwise order .

API Reference: elementwise_min

elementwise_max¶

Perform max(x, y) operations on two Tensor in elementwise order .

API Reference: elementwise_max

matmul¶

Perform matrix multiplication operations on two Tensor.

API Reference: matmul