All NNabla functions are derived from the nnabla.function.Function class.
nnabla.function
Function
PythonFunction
__init__
nnabla.functions
The nnabla.functions module provides various types of functions listed below. These functions takes input nnabla.Variable (s) as its leading argument(s), followed by options specific to each function.
Note
The functions can also take ~nnabla.NdArray (s) as inputs instead of ~nnabla.Variable (s). It will execute the function operation immediately, and returns ~nnabla.NdArray (s) as output(s) holding output values of the operation. We call this "Imperative Mode" (NdArray + Functions).
affine
convolution
depthwise_convolution
deconvolution
depthwise_deconvolution
deformable_convolution
adaptive_separable_convolution
max_pooling
average_pooling
global_average_pooling
sum_pooling
unpooling
embed
rnn
lstm
gru
multi_head_attention
patch_correlation
sigmoid
swish
tanh
relu
softmax
log_softmax
elu
selu
crelu
celu
gelu
mish
prelu
leaky_relu
relu6
hard_sigmoid
hard_tanh
log_sigmoid
softplus
softsign
tanh_shrink
sinc
batch_normalization
fused_batch_normalization
sync_batch_normalization
mean_subtraction
norm_normalization
clip_by_value
clip_grad_by_value
clip_by_norm
clip_grad_by_norm
layer_normalization
instance_normalization
group_normalization
weight_standardization
weight_normalization
sum
mean
max
min
norm
prod
reduce_sum
reduce_mean
add2
add_n
sub2
mul2
mul_n
div2
pow2
add_scalar
mul_scalar
pow_scalar
r_sub_scalar
r_div_scalar
r_pow_scalar
equal
equal_scalar
greater
greater_equal
greater_equal_scalar
greater_scalar
less
less_equal
less_equal_scalar
less_scalar
logical_and
logical_and_scalar
logical_not
logical_or
logical_or_scalar
logical_xor
logical_xor_scalar
not_equal
not_equal_scalar
sign
minimum2
maximum2
minimum_scalar
maximum_scalar
isnan
isinf
reset_nan
reset_inf
where
constant
arange
abs
exp
log
round
ceil
floor
identity
matrix_diag
matrix_diag_part
batch_matmul
sin
cos
tan
sinh
cosh
tanh
asin
acos
atan
atan2
asinh
acosh
atanh
concatenate
split
stack
slice
gather
gather_nd
scatter_nd
scatter_add
pad
transpose
broadcast
broadcast_to
tile
flip
shift
sort
reshape
one_hot
batch_inv
batch_det
batch_logdet
assign
top_k_data
top_k_grad
pack_padded_sequence
pad_packed_sequence
rand
randint
randn
rand_binomial
rand_beta
rand_gamma
dropout
random_choice
random_crop
random_erase
random_flip
random_shift
image_augmentation
sigmoid_cross_entropy
binary_cross_entropy
softmax_cross_entropy
categorical_cross_entropy
squared_error
absolute_error
huber_loss
epsilon_insensitive_loss
kl_multinomial
interpolate
fft
ifft
stft
istft
affine_grid
warp_by_grid
warp_by_flow
binary_sigmoid
binary_tanh
binary_connect_affine
binary_connect_convolution
binary_weight_affine
binary_weight_convolution
fixed_point_quantize
min_max_quantize
pow2_quantize
prune
inq_affine
inq_convolution
vat_noise
unlink
sink
confusion_matrix
nms_detection2d
top_n_error
binary_error