relu

Rectified linear unit activation function.

relu6

Rectified Linear Unit 6 activation function.

sigmoid(x)

Sigmoid activation function.

softplus(x)

Softplus activation function.

sparse_plus(x)

Sparse plus function.

sparse_sigmoid(x)

Sparse sigmoid activation function.

soft_sign(x)

Soft-sign activation function.

silu(x)

SiLU (aka swish) activation function.

swish(x)

SiLU (aka swish) activation function.

log_sigmoid(x)

Log-sigmoid activation function.

leaky_relu(x[, negative_slope])

Leaky rectified linear unit activation function.

hard_sigmoid(x)

Hard Sigmoid activation function.

hard_silu(x)

Hard SiLU (swish) activation function

hard_swish(x)

Hard SiLU (swish) activation function

hard_tanh(x)

Hard\(\mathrm{tanh}\) activation function.

tanh(x, /)

Calculate element-wise hyperbolic tangent of input.

elu(x[, alpha])

Exponential linear unit activation function.

celu(x[, alpha])

Continuously-differentiable exponential linear unit activation.

selu(x)

Scaled exponential linear unit activation.

gelu(x[, approximate])

Gaussian error linear unit activation function.

glu(x[, axis])

Gated linear unit activation function.

squareplus(x[, b])

Squareplus activation function.

mish(x)

Mish activation function.

identity(x)

Identity activation function.

softmax(x[, axis, where])

Softmax function.

log_softmax(x[, axis, where])

Log-Softmax function.

logmeanexp(x[, axis, where, keepdims])

Log mean exp.

logsumexp()

Log-sum-exp reduction.

standardize(x[, axis, mean, variance, ...])

Standardizes input to zero mean and unit variance.

one_hot(x, num_classes, *[, dtype, axis])

One-hot encodes the given indices.

dot_product_attention()

Scaled dot product attention function.

scaled_matmul(lhs, rhs, lhs_scales, rhs_scales)

Scaled matrix multiplication function.

get_scaled_dot_general_config(mode[, ...])

Get quantization configs for scaled_dot_general.

scaled_dot_general(lhs, rhs, dimension_numbers)

Scaled dot general operation.

log1mexp

Numerically stable calculation of\(\log(1 - \exp(-x))\).