clj-nd4j.ndarray — io.github.jeremylcfr/clj-nd4j 0.1.0-SNAPSHOT

->nd-array clj

(->nd-array obj)

Like nd-array constructor but also works for already existing NDArray instances (do nothing). Useful mostly internally to allow flexibility but can be used to design external generic functions as well

->nilable-nd-array clj

(->nilable-nd-array obj)

Internal usage. Also deal with nil

abs clj

(abs a__2098__auto__)

abs! clj

(abs! a__2104__auto__)

acos clj

(acos a__2098__auto__)

acos! clj

(acos! a__2104__auto__)

arg-max clj

(arg-max a dimension)

Returns a NDArray of maximal values along dimension

arg-min clj

(arg-min a dimension)

Returns a NDArray of muinimal values along dimension

asin clj

(asin a__2098__auto__)

asin! clj

(asin! a__2104__auto__)

atan clj

(atan a__2098__auto__)

atan! clj

(atan! a__2104__auto__)

atanh clj

(atanh a__2098__auto__)

atanh! clj

(atanh! a__2104__auto__)

ceil clj

(ceil a__2098__auto__)

ceil! clj

(ceil! a__2104__auto__)

clone clj

(clone nd-array)

Clones a nd-array. Useful for immutability

cos clj

(cos a__2098__auto__)

cos! clj

(cos! a__2104__auto__)

cosh clj

(cosh a__2098__auto__)

cosh! clj

(cosh! a__2104__auto__)

cosine-distance clj

(cosine-distance a1 a2)

cotransform clj

(cotransform kind a1 a2)

(cotransform kind arg1 a1 a2)

(cotransform kind arg1 arg2 a1 a2)

Transforms a vector using an element-wise operator. This operation preserves input. See transform! for destructive operations. Slower than directly calling the specific operation, use this when needed. Usage (NDArray representation simplified for readibility, this is not a vector !) : (transform :ln [1.0 2.0 3.0]) ~ (ln [1.0 2.0 3.0])

[0.0 0.693.. 1.098..]

cotransform! clj

(cotransform! kind a1 a2)

(cotransform! kind arg1 a1 a2)

(cotransform! kind arg1 arg2 a1 a2)

Transforms a vector using an element-wise operator. This is a destructive operation, i.e. the input is altered to target value. Use 'transform' for immutability. Slower than directly calling the specific operation, use this when needed. Usage (NDArray representation simplified for readibility, this is not a vector !) : (transform! :ln [1.0 2.0 3.0]) ~ (ln! [1.0 2.0 3.0])

[0.0 0.693.. 1.098..]

defcooperator clj macro

(defcooperator fn-name static-name)

defcooperator! clj macro

(defcooperator! fn-name static-name)

defoperator clj macro

(defoperator fn-name static-name)

defoperator! clj macro

(defoperator! fn-name static-name)

delu clj

(delu a__2098__auto__)

delu! clj

(delu! a__2104__auto__)

dhard-tanh clj

(dhard-tanh a__2098__auto__)

dhard-tanh! clj

(dhard-tanh! a__2104__auto__)

distance clj

(distance a1 a2)

(distance kind a1 a2)

Computes distance between two NDArray instances using euclidean distance by default (two arities) or the selected method. Methods are : :abs (n1), :euclidean (n2), :cos, :manhattan, :jaccard, :hamming Usage : (distance nd1 nd2) (distance :method nd1 nd2)

dleaky-relu clj

(dleaky-relu a__2098__auto__)

dleaky-relu! clj

(dleaky-relu! a__2104__auto__)

dsigmoid clj

(dsigmoid a__2098__auto__)

dsigmoid! clj

(dsigmoid! a__2104__auto__)

dsoftsign clj

(dsoftsign a__2098__auto__)

dsoftsign! clj

(dsoftsign! a__2104__auto__)

elu clj

(elu a__2098__auto__)

elu! clj

(elu! a__2104__auto__)

exp clj

(exp a__2098__auto__)

exp! clj

(exp! a__2104__auto__)

floor clj

(floor a__2098__auto__)

floor! clj

(floor! a__2104__auto__)

get-column clj

(get-column nd-array n)

Gets nth column of the given nd-array

get-double clj

(get-double nd-array n)

Gets the nth leaf element of the flattended nd-array (line first) as a double

get-in-double clj

(get-in-double nd-array indexes)

Gets the leaf element at specified location, following the standard java specification as a double

get-in-scalar clj

(get-in-scalar nd-array indexes)

Gets the leaf element at specified location, following the standard java specification as a nd-array (0d-array)

get-row clj

(get-row nd-array n)

Gets nth row of the given nd-array

get-scalar clj

(get-scalar nd-array n)

Gets the nth leaf element of the flattended nd-array (line first) as a scalar nd-array (0d-array)

get-shape clj

(get-shape nd-array)

Gets shape of a NDArray as long[]

hamming-distance clj

(hamming-distance a1 a2)

hard-tanh clj

(hard-tanh a__2098__auto__)

hard-tanh! clj

(hard-tanh! a__2104__auto__)

jaccard-distance clj

(jaccard-distance a1 a2)

leaky-relu clj

(leaky-relu a__2098__auto__)

leaky-relu! clj

(leaky-relu! a__2104__auto__)

ln clj

(ln a__2098__auto__)

ln! clj

(ln! a__2104__auto__)

load-nd-array clj

(load-nd-array path)

Load a NDArray from a file written with write-nd-array!

log clj

(log base a)

log! clj

(log! base a)

manhattan-distance clj

(manhattan-distance a1 a2)

mmax clj

(mmax n1__2195__auto__ n2__2196__auto__)

mmax! clj

(mmax! n1__2202__auto__ n2__2203__auto__)

n1-distance clj

(n1-distance a1 a2)

n2-distance clj

(n2-distance a1 a2)

nd-array clj

(nd-array data)

(nd-array shape data)

Creates a nd array (n dimensional array) from data. For predictable input, indicating a shape (dimensions) is wiser. Else, shape will be infered from a walk from every first element unless input is When you indicate a shape, data must be provided flattened, else the structure must be already formatted as a nested structure. The sole exceptions to this rules are when you want to create a matrix or a (1D) array (most common cases). There you can replace shape by a type hint : :array or :matrix. You can also use specialized constructors : matrix and array Example : (nd-array [2 2] [1 2 3 4]) ;; or (->nd-array [[1 2] [3 4]]) or (->nd-array :matrix [[1 2] [3 4]]) ~ [[1 2] [3 4]] (a matrix) (nd-array [4] [1 2 3 4]) ;; (or (->nd-array [1 2 3 4])) ~ [1 2 3 4] (an array)

nd-array? clj

(nd-array? obj)

nd-ones clj

(nd-ones shape)

Creates a NDArray full of ones of input shape. Example : (nd-ones [2 2]) ~ [[1 1] [1 1]]

nd-rand clj

(nd-rand shape)

Creates a NDArray initialized with random values (between 0 and 1) and of input shape.

nd-zeros clj

(nd-zeros shape)

Creates a NDArray full of zeros of input shape. Example : (nd-zeros [2 2]) ~ [[0 0] [0 0]]

negate clj

(negate a__2098__auto__)

negate! clj

(negate! a__2104__auto__)

normalize clj

(normalize a)

NDArray statistical normalization (zero mean and unit variance. Preserves input unlike normalize!

normalize! clj

(normalize! a)

NDArray statistical normalization (zero mean and unit variance. Alters input unlike normalize

pow clj

(pow n1__2195__auto__ n2__2196__auto__)

pow! clj

(pow! n1__2202__auto__ n2__2203__auto__)

relu clj

(relu a__2098__auto__)

relu! clj

(relu! a__2104__auto__)

relu6 clj

(relu6 a__2098__auto__)

relu6! clj

(relu6! a__2104__auto__)

roll-axis clj

(roll-axis nd-array index)

(roll-axis nd-array start end)

round clj

(round a__2098__auto__)

round! clj

(round! a__2104__auto__)

sigmoid clj

(sigmoid a__2098__auto__)

sigmoid! clj

(sigmoid! a__2104__auto__)

sign clj

(sign a__2098__auto__)

sign! clj

(sign! a__2104__auto__)

sin clj

(sin a__2098__auto__)

sin! clj

(sin! a__2104__auto__)

single-coops clj

single-coops! clj

single-ops clj

single-ops! clj

sinh clj

(sinh a__2098__auto__)

sinh! clj

(sinh! a__2104__auto__)

softmax clj

(softmax a__2098__auto__)

softmax! clj

(softmax! a__2104__auto__)

softplus clj

(softplus a__2098__auto__)

softplus! clj

(softplus! a__2104__auto__)

softsign clj

(softsign a__2098__auto__)

softsign! clj

(softsign! a__2104__auto__)

sqrt clj

(sqrt a__2098__auto__)

sqrt! clj

(sqrt! a__2104__auto__)

stabilize clj

(stabilize k a)

stabilize! clj

(stabilize! k a)

t* clj

(t* left right axis)

tanh clj

(tanh a__2098__auto__)

tanh! clj

(tanh! a__2104__auto__)

transform clj

(transform kind a)

(transform kind arg1 a)

(transform kind arg1 arg2 a)

Transforms a vector using an element-wise operator. This operation preserves input. See transform! for destructive operations. Slower than directly calling the specific operation, use this when needed. Usage (NDArray representation simplified for readibility, this is not a vector !) : (transform :ln [1.0 2.0 3.0]) ~ (ln [1.0 2.0 3.0])

[0.0 0.693.. 1.098..]

transform! clj

(transform! kind a)

(transform! kind a arg1)

(transform! kind a arg1 arg2)

Transforms a vector using an element-wise operator. This is a destructive operation, i.e. the input is altered to target value. Use 'transform' for immutability. Slower than directly calling the specific operation, use this when needed. Usage (NDArray representation simplified for readibility, this is not a vector !) : (transform! :ln [1.0 2.0 3.0]) ~ (ln! [1.0 2.0 3.0])

[0.0 0.693.. 1.098..]

transpose clj

(transpose nd-array)

NDArray transposition

write-nd-array! clj

(write-nd-array! path nd-array)

(write-nd-array! path split nd-array)

Writes a nd-array directly into a text file (.txt or .csv). The main advantage is to avoid the potentially costly conversion to clojure code. By default, the split is ';'