Skip to main content

Ctrl+K

Getting Started

Quickstart Guide

Physical Units

Quantity
Math Operations
Standard Units
Constants
Unit Conversion
Temperature Conversions

Unit-aware Operations

Array Creation
NumPy Functions
Einstein Operations
Linear Algebra Functions
Fourier Transform Functions
JAX LAX Functions
Activation Functions
Checking Function Units
Assigning Function Units
JAX Transforms: JIT and vmap with Units
Automatic Differentiation with Units
Sparse Matrices with Units
Matplotlib Integration
Unit-Aware Type Annotations

Advanced Tutorials

Combining and Defining Unit
Mechanism of Quantity
Unit-aware Computation with CustomArray
FAQs

API Documentation

Release Notes
brainunit module
brainunit.typing module
brainunit.autograd module
brainunit.math module
- relu
- relu6
- sigmoid
- softplus
- sparse_plus
- sparse_sigmoid
- soft_sign
- silu
- swish
- log_sigmoid
- leaky_relu
- hard_sigmoid
- hard_silu
- hard_swish
- hard_tanh
- elu
- celu
- selu
- gelu
- glu
- squareplus
- mish
- is_dimensionless
- is_unitless
- get_dim
- get_unit
- get_mantissa
- get_magnitude
- display_in_unit
- maybe_decimal
- check_dims
- check_units
- fail_for_dimension_mismatch
- fail_for_unit_mismatch
- assert_quantity
- get_or_create_dimension
- einreduce
- einrearrange
- einrepeat
- einshape
- einsum
- exprel
- set_exprel_order
- exp
- exp2
- expm1
- log
- log10
- log1p
- log2
- arccos
- arccosh
- arcsin
- arcsinh
- arctan
- arctanh
- cos
- cosh
- sin
- sinc
- sinh
- tan
- tanh
- deg2rad
- rad2deg
- degrees
- radians
- angle
- frexp
- hypot
- arctan2
- logaddexp
- logaddexp2
- corrcoef
- correlate
- cov
- ldexp
- bitwise_not
- invert
- bitwise_and
- bitwise_or
- bitwise_xor
- left_shift
- right_shift
- full
- eye
- identity
- tri
- empty
- ones
- zeros
- full_like
- diag
- tril
- triu
- empty_like
- ones_like
- zeros_like
- fill_diagonal
- array
- asarray
- arange
- linspace
- logspace
- meshgrid
- vander
- tril_indices
- tril_indices_from
- triu_indices
- triu_indices_from
- from_numpy
- as_numpy
- tree_ones_like
- tree_zeros_like
- reciprocal
- prod
- product
- nancumprod
- nanprod
- cumprod
- cumproduct
- var
- nanvar
- cbrt
- square
- sqrt
- multiply
- divide
- power
- cross
- true_divide
- floor_divide
- float_power
- divmod
- convolve
- dot
- multi_dot
- vdot
- vecdot
- inner
- outer
- kron
- matmul
- tensordot
- matrix_power
- row_stack
- concatenate
- stack
- vstack
- hstack
- dstack
- column_stack
- block
- append
- split
- array_split
- dsplit
- hsplit
- vsplit
- atleast_1d
- atleast_2d
- atleast_3d
- broadcast_arrays
- broadcast_to
- reshape
- moveaxis
- transpose
- swapaxes
- tile
- repeat
- flip
- fliplr
- flipud
- roll
- expand_dims
- squeeze
- sort
- max
- min
- amax
- amin
- diagflat
- diagonal
- choose
- ravel
- flatten
- unflatten
- remove_diag
- astype
- real
- imag
- conj
- conjugate
- negative
- positive
- abs
- sum
- nancumsum
- nansum
- cumsum
- ediff1d
- absolute
- fabs
- median
- nanmin
- nanmax
- ptp
- average
- mean
- std
- nanmedian
- nanmean
- nanstd
- diff
- rot90
- intersect1d
- nan_to_num
- percentile
- nanpercentile
- quantile
- nanquantile
- round
- around
- rint
- floor
- ceil
- trunc
- fix
- modf
- fmod
- mod
- copysign
- remainder
- maximum
- minimum
- fmax
- fmin
- lcm
- gcd
- trace
- add
- subtract
- nextafter
- promote_dtypes
- interp
- clip
- histogram
- compress
- extract
- take
- select
- where
- unique
- gather
- iscomplexobj
- heaviside
- signbit
- sign
- bincount
- digitize
- get_promote_dtypes
- all
- any
- logical_not
- equal
- not_equal
- greater
- greater_equal
- less
- less_equal
- array_equal
- isclose
- allclose
- logical_and
- logical_or
- logical_xor
- alltrue
- sometrue
- argsort
- argmax
- argmin
- nanargmax
- nanargmin
- argwhere
- nonzero
- flatnonzero
- searchsorted
- count_nonzero
- diag_indices_from
- finfo
- iinfo
- is_quantity
- issubdtype
- result_type
- ndim
- isreal
- isscalar
- isfinite
- isinf
- isnan
- shape
- size
- get_dtype
- is_float
- is_int
- broadcast_shapes
- gradient
- bartlett
- blackman
- hamming
- hanning
- kaiser
- bool
- uint2
- uint4
- uint8
- uint16
- uint32
- uint64
- int2
- int4
- int8
- int16
- int32
- int64
- bfloat16
- float16
- float32
- float64
- complex64
- complex128
- int
- uint
- float
- complex
- single
- double
- csingle
- cdouble
- inexact
- dtype
- e
- pi
- inf
- nan
- euler_gamma
- newaxis
brainunit.linalg module
- dot
- multi_dot
- vdot
- vecdot
- inner
- kron
- matmul
- tensordot
- matrix_power
- cross
- cholesky
- outer
- det
- solve
- tensorsolve
- lstsq
- inv
- pinv
- tensorinv
- qr
- svd
- svdvals
- eig
- eigh
- eigvals
- eigvalsh
- norm
- matrix_norm
- vector_norm
- trace
- diagonal
- matrix_transpose
brainunit.lax module
- acos
- acosh
- asin
- asinh
- atan
- atanh
- collapse
- cumlogsumexp
- bessel_i0e
- bessel_i1e
- digamma
- lgamma
- erf
- erfc
- erf_inv
- logistic
- atan2
- polygamma
- igamma
- igammac
- igamma_grad_a
- random_gamma_grad
- zeta
- betainc
- shift_left
- shift_right_arithmetic
- shift_right_logical
- fft
- collapse
- zeros_like_array
- iota
- broadcasted_iota
- rsqrt
- div
- pow
- integer_pow
- mul
- rem
- batch_matmul
- conv
- conv_transpose
- dot_general
- slice
- dynamic_slice
- dynamic_update_slice
- gather
- index_take
- slice_in_dim
- index_in_dim
- dynamic_slice_ind_dim
- dynamic_index_in_dim
- dynamic_update_slice_in_dim
- dynamic_update_index_in_dim
- sort
- sort_key_val
- neg
- cummax
- cummin
- cumsum
- scatter
- scatter_add
- scatter_sub
- scatter_mul
- scatter_min
- scatter_max
- scatter_apply
- sub
- complex
- pad
- clamp
- convert_element_type
- bitcast_convert_type
- approx_max_k
- approx_min_k
- top_k
- broadcast
- broadcast_in_dim
- broadcast_to_rank
- population_count
- clz
- eq
- ne
- ge
- gt
- le
- lt
- cholesky
- eig
- eigh
- hessenberg
- lu
- qdwh
- qr
- schur
- svd
- tridiagonal
- householder_product
- triangular_solve
- tridiagonal_solve
- reduce
- reduce_precision
- broadcast_shapes
brainunit.fft module
- fft
- rfft
- ifft
- irfft
- fft2
- fftn
- rfft2
- rfftn
- ifft2
- ifftn
- irfft2
- irfftn
- fftfreq
- rfftfreq
- fftshift
- ifftshift
brainunit.sparse module
brainunit.constants module

.rst

hard_swish

hard_swish#

saiunit.math.hard_swish#: alias of hard_silu

Contents

hard_swish

By BrainX Ecosystem

Last updated on May 13, 2026.