Mathematical functions

Trigonometric functions

sin	sine of a Generalizes a scalar Op to tensors.
cos	cosine of a Generalizes a scalar Op to tensors.
tan	tangent of a Generalizes a scalar Op to tensors.
arcsin	arcsine of a Generalizes a scalar Op to tensors.
arccos	arccosine of a Generalizes a scalar Op to tensors.
arctan	arctangent of a Generalizes a scalar Op to tensors.
arctan2	arctangent of a / b Generalizes a scalar Op to tensors.
deg2rad	convert degree a to radian Generalizes a scalar Op to tensors.
rad2deg	convert radian a to degree Generalizes a scalar Op to tensors.

Hyperbolic functions

sinh	hyperbolic sine of a Generalizes a scalar Op to tensors.
cosh	hyperbolic cosine of a Generalizes a scalar Op to tensors.
tanh	hyperbolic tangent of a Generalizes a scalar Op to tensors.
arcsinh	hyperbolic arc sine of a Generalizes a scalar Op to tensors.
arccosh	hyperbolic arc cosine of a Generalizes a scalar Op to tensors.
arctanh	hyperbolic arc tangent of a Generalizes a scalar Op to tensors.

Rounding

round(a[, mode])	round_mode(a) with mode in [half_away_from_zero, half_to_even].
iround(a[, mode])	cast(round(a,mode),'int64')
floor	floor of a Generalizes a scalar Op to tensors.
ceil	ceiling of a Generalizes a scalar Op to tensors.
trunc	trunc of a Generalizes a scalar Op to tensors.
round_half_to_even	round_half_to_even(a) Generalizes a scalar Op to tensors.
round_half_away_from_zero	round_half_away_from_zero(a) Generalizes a scalar Op to tensors.

Sums, products, differences

sum(input[, axis, dtype, keepdims, acc_dtype])	Computes the sum along the given axis(es) of a tensor input.
cumsum(x[, axis])	Return the cumulative sum of the elements along a given axis.
prod(input[, axis, dtype, keepdims, ...])	Computes the product along the given axis(es) of a tensor input.
cumprod(x[, axis])	Return the cumulative product of the elements along a given axis.
diff(x[, n, axis])	Calculate the n-th order discrete difference along the given axis.

Exponents and logarithms

exp	e^`a` Generalizes a scalar Op to tensors.
expm1	e^`a` - 1 Generalizes a scalar Op to tensors.
exp2	2^`a` Generalizes a scalar Op to tensors.
log	base e logarithm of a Generalizes a scalar Op to tensors.
log2	base 2 logarithm of a Generalizes a scalar Op to tensors.
log10	base 10 logarithm of a Generalizes a scalar Op to tensors.
log1p	log(1+a) Generalizes a scalar Op to tensors.
logaddexp(*xs)	Logarithm of the sum of exponentiations of the inputs.
square	square of a Generalizes a scalar Op to tensors.
sqrt	square root of a Generalizes a scalar Op to tensors.
power(x, y)

Special functions

erfc	complementary error function Generalizes a scalar Op to tensors.
erfcx	scaled complementary error function Generalizes a scalar Op to tensors.
erfcinv	inverse complementary error function Generalizes a scalar Op to tensors.
erfinv	inverse error function Generalizes a scalar Op to tensors.
owens_t	owens t function Generalizes a scalar Op to tensors.
gamma	gamma function Generalizes a scalar Op to tensors.
gammaln	log gamma function Generalizes a scalar Op to tensors.
psi	derivative of log gamma function Generalizes a scalar Op to tensors.
tri_gamma	second derivative of the log gamma function Generalizes a scalar Op to tensors.
chi2sf	chi squared survival function Generalizes a scalar Op to tensors.
gammainc	Regularized lower gamma function Generalizes a scalar Op to tensors.
gammaincc	Regularized upper gamma function Generalizes a scalar Op to tensors.
gammau	Upper incomplete gamma function.
hyp2f1	Gaussian hypergeometric function.
hyp2f1_der	Derivatives for Gaussian hypergeometric function.
j0	Bessel function of the first kind of order 0.
j1	Bessel function of the first kind of order 1.
jv	Bessel function of the first kind of order v (real).
i0	Modified Bessel function of the first kind of order 0.
i1	Modified Bessel function of the first kind of order 1.
iv	Modified Bessel function of the first kind of order v (real).
sigmoid	Logistic sigmoid function (1 / (1 + exp(-x)), also known as expit or inverse logit Generalizes a scalar Op to tensors.
softplus	Compute log(1 + exp(x)), also known as softplus or log1pexp Generalizes a scalar Op to tensors.
special.softmax(c[, axis])
special.log_softmax(c[, axis])
extra_ops.to_one_hot(y, nb_class[, dtype])	Return a matrix where each row correspond to the one hot encoding of each element in y.
special.poch(z, m)	Compute the Pochhammer/rising factorial.
special.factorial(n)	Compute the factorial.
log1mexp	Compute log(1 - exp(x)), also known as log1mexp Generalizes a scalar Op to tensors.
betainc	Regularized incomplete beta function Generalizes a scalar Op to tensors.
logsumexp(x[, axis, keepdims])	Compute the log of the sum of exponentials of input elements.
xlogx.xlogx	Generalizes a scalar Op to tensors.
xlogx.xlogy0	Generalizes a scalar Op to tensors.

Arithmetic operations

add	elementwise addition Generalizes a scalar Op to tensors.
reciprocal	1.0/a Generalizes a scalar Op to tensors.
power(x, y)
mod	elementwise modulo Generalizes a scalar Op to tensors.
abs	|`a`| Generalizes a scalar Op to tensors.
neg	-a Generalizes a scalar Op to tensors.
pow	elementwise power Generalizes a scalar Op to tensors.
sgn	sign of a Generalizes a scalar Op to tensors.
mul	elementwise multiplication Generalizes a scalar Op to tensors.
sub	elementwise subtraction Generalizes a scalar Op to tensors.
int_div	elementwise [floor] division (inverse of multiplication) Generalizes a scalar Op to tensors.
ceil_intdiv(a, b)	Safely compute ceil(float_division(a, b)).
true_div	elementwise [true] division (inverse of multiplication) Generalizes a scalar Op to tensors.
divmod(x, y)	Element-wise divmod, using floor_divide and mod_check.

>>> a, b = at.itensor3(), at.itensor3() # example inputs
>>> a + 3      # at.add(a, 3) -> itensor3
>>> 3 - a      # at.sub(3, a)
>>> a * 3.5    # at.mul(a, 3.5) -> ftensor3 or dtensor3 (depending on casting)
>>> 2.2 / a    # at.truediv(2.2, a)
>>> 2.2 // a   # at.intdiv(2.2, a)
>>> 2.2**a     # at.pow(2.2, a)
>>> b % a      # at.mod(b, a)

Extrema

maximum	elemwise maximum.
minimum	elemwise minimum.
topk(x, kth[, axis, sorted, idx_dtype])	Returns the k-largest elements along an axis.
argtopk(x, kth[, axis, sorted, idx_dtype])	Returns the indices of k-largest elements along an axis.
topk_and_argtopk(x, kth[, axis, sorted, ...])	Returns the results of both topk() and argtopk() in one Op.

Complex numbers

angle	Return polar-coordinate angle of complex-valued tensor z Generalizes a scalar Op to tensors.
real	Return real component of complex-valued tensor z Generalizes a scalar Op to tensors.
imag	Return imaginary component of complex-valued tensor z Generalizes a scalar Op to tensors.
conjugate(x)
complex_from_polar	Return complex-valued tensor from polar coordinate specification.
complex	Return complex-valued tensor with real and imag components Generalizes a scalar Op to tensors.

Miscellaneous

clip	Clip x to be between min and max.