Index

Modules: accessors, accessors_macros_read, accessors_macros_syntax, accessors_macros_write, accuracy_score, aggregate, algebra, algorithms, align_unroller, ast_utils, autograd, autograd_common, auxiliary_blas, auxiliary_lapack, blas_l3_gemm, blis, common_error_functions, compiler_optim_hints, complex, conv, conv2D, cpuinfo_x86, cross_entropy_losses, cublas, cuda, cuda_global_state, cudnn, cudnn_conv_interface, data_structure, datatypes, dbscan, decomposition, decomposition_lapack, decomposition_rand, deprecate, display, display_cuda, distances, distributions, dynamic_stack_arrays, einsum, embedding, exporting, filling_data, flatten, foreach, foreach_common, foreach_staged, functional, gates_basic, gates_blas, gates_hadamard, gates_reduce, gates_shapeshifting_concat_split, gates_shapeshifting_views, gcn, gemm, gemm_packing, gemm_prepacked, gemm_tiling, gemm_ukernel_avx, gemm_ukernel_avx2, gemm_ukernel_avx512, gemm_ukernel_avx_fma, gemm_ukernel_dispatch, gemm_ukernel_generator, gemm_ukernel_generic, gemm_ukernel_sse, gemm_ukernel_sse2, gemm_ukernel_sse4_1, gemm_utils, global_config, gru, higher_order_applymap, higher_order_foldreduce, imdb, incl_accessors_cuda, incl_higher_order_cuda, incl_kernels_cuda, init, init_colmajor, init_copy_cpu, init_copy_cuda, init_cpu, init_cuda, init_opencl, initialization, io, io_csv, io_hdf5, io_image, io_npy, io_stream_readers, kde, kdtree, kmeans, lapack, least_squares, least_squares_lapack, linear, linear_algebra, linear_systems, math_functions, math_ops_fusion, maxpool2D, mean_square_error_loss, memory, memory_optimization_hints, ml, mnist, naive_l2_gemv, neighbors, nested_containers, nlp, nn, nn_dsl, nn_primitives, nnp_activation, nnp_conv2d_cudnn, nnp_convolution, nnp_embedding, nnp_gru, nnp_linear, nnp_maxpooling, nnp_numerical_gradient, nnp_sigmoid_cross_entropy, nnp_softmax, nnp_softmax_cross_entropy, nnpack, nnpack_interface, opencl_backend, opencl_global_state, openmp, operators_blas_l1, operators_blas_l1_cuda, operators_blas_l1_opencl, operators_blas_l2l3, operators_blas_l2l3_cuda, operators_blas_l2l3_opencl, operators_broadcasted, operators_broadcasted_cuda, operators_broadcasted_opencl, operators_comparison, operators_logical, optim_ops_fusion, optimizers, overload, p_accessors, p_accessors_macros_desugar, p_accessors_macros_read, p_accessors_macros_write, p_activation, p_checks, p_complex, p_display, p_empty_tensors, p_init_cuda, p_init_opencl, p_kernels_interface_cuda, p_kernels_interface_opencl, p_logsumexp, p_nnp_checks, p_nnp_types, p_operator_blas_l2l3, p_shapeshifting, pca, relu, selectors, sequninit, shapeshifting, shapeshifting_cuda, shapeshifting_opencl, sigmoid, simd, softmax, solve_lapack, special_matrices, stats, std_version_types, syntactic_sugar, tanh, tensor, tensor_compare_helper, tensor_cuda, tensor_opencl, tokenizers, triangular, ufunc, util.

API symbols

`!=.`:

• operators_comparison: proc `!=.`[T](t: Tensor[T]; value: T): Tensor[bool]
• operators_comparison: proc `!=.`[T](a, b: Tensor[T]): Tensor[bool]
• operators_comparison: template `!=.`[T](value: T; t: Tensor[T]): Tensor[bool]

`$`:

• display: proc `$`[T](t: Tensor[T]): string
• display_cuda: proc `$`[T](t: CudaTensor[T]): string
• dynamic_stack_arrays: proc `$`(a: DynamicStackArray): string
• pca: proc `$`(pca: PCA_Detailed): string

`&`:

• dynamic_stack_arrays: proc `&`(a, b: DynamicStackArray): DynamicStackArray
• dynamic_stack_arrays: proc `&`[T](a: DynamicStackArray[T]; value: T): DynamicStackArray[T]

`*.=`:

• operators_broadcasted: proc `*.=`[T: SomeNumber](t: var Tensor[Complex64]; val: T)
• operators_broadcasted: proc `*.=`[T: SomeNumber | Complex[float32] | Complex[float64]](t: var Tensor[T]; val: T)
• operators_broadcasted: proc `*.=`[T: SomeNumber | Complex[float32] | Complex[float64]](a: var Tensor[T]; b: Tensor[T])
• operators_broadcasted_cuda: proc `*.=`[T: SomeFloat](a: var CudaTensor[T]; b: CudaTensor[T])

`*.`:

• gates_hadamard: proc `*.`[TT](a, b: Variable[TT]): Variable[TT]
• operators_broadcasted: proc `*.`[T: SomeNumber](val: T; t: Tensor[Complex64]): Tensor[Complex64]
• operators_broadcasted: proc `*.`[T: SomeNumber | Complex[float32] | Complex[float64]](val: T; t: Tensor[T]): Tensor[ T]
• operators_broadcasted: proc `*.`[T: SomeNumber](t: Tensor[Complex64]; val: T): Tensor[Complex64]
• operators_broadcasted: proc `*.`[T: SomeNumber | Complex[float32] | Complex[float64]](t: Tensor[T]; val: T): Tensor[ T]
• operators_broadcasted: proc `*.`[T: SomeNumber | Complex[float32] | Complex[float64]](a, b: Tensor[T]): Tensor[ T]
• operators_broadcasted_cuda: proc `*.`[T: SomeFloat](a, b: CudaTensor[T]): CudaTensor[T]
• operators_broadcasted_opencl: proc `*.`[T: SomeFloat](a, b: ClTensor[T]): ClTensor[T]

`*=`:

• operators_blas_l1: proc `*=`[T: SomeNumber | Complex[float32] | Complex[float64]](t: var Tensor[T]; a: T)
• operators_blas_l1_cuda: proc `*=`[T: SomeFloat](t: var CudaTensor[T]; a: T)

`*`:

• gates_blas: proc `*`[TT](a, b: Variable[TT]): Variable[TT]
• operators_blas_l1: proc `*`[T: SomeNumber | Complex[float32] | Complex[float64]](a: T; t: Tensor[T]): Tensor[ T]
• operators_blas_l1: proc `*`[T: SomeNumber | Complex[float32] | Complex[float64]](t: Tensor[T]; a: T): Tensor[ T]
• operators_blas_l1_cuda: proc `*`[T: SomeFloat](t: CudaTensor[T]; a: T): CudaTensor[T]
• operators_blas_l1_cuda: proc `*`[T: SomeFloat](a: T; t: CudaTensor[T]): CudaTensor[T]
• operators_blas_l2l3: proc `*`[T: Complex[float32] or Complex[float64]](a, b: Tensor[T]): Tensor[T]
• operators_blas_l2l3: proc `*`[T: SomeNumber](a, b: Tensor[T]): Tensor[T]
• operators_blas_l2l3_cuda: proc `*`[T: SomeFloat](a, b: CudaTensor[T]): CudaTensor[T]
• operators_blas_l2l3_opencl: proc `*`[T: SomeFloat](a, b: ClTensor[T]): ClTensor[T]

`+.=`:

• operators_broadcasted: proc `+.=`[T: SomeNumber](t: var Tensor[Complex64]; val: T)
• operators_broadcasted: proc `+.=`[T: SomeNumber | Complex[float32] | Complex[float64]](t: var Tensor[T]; val: T)
• operators_broadcasted: proc `+.=`[T: SomeNumber | Complex[float32] | Complex[float64]](a: var Tensor[T]; b: Tensor[T])
• operators_broadcasted_cuda: proc `+.=`[T: SomeFloat](a: var CudaTensor[T]; b: CudaTensor[T])
• operators_broadcasted_cuda: proc `+.=`[T: SomeFloat](t: var CudaTensor[T]; val: T)

`+.`:

• operators_broadcasted: proc `+.`[T: SomeNumber](val: T; t: Tensor[Complex64]): Tensor[Complex64]
• operators_broadcasted: proc `+.`[T: SomeNumber | Complex[float32] | Complex[float64]](val: T; t: Tensor[T]): Tensor[ T]
• operators_broadcasted: proc `+.`[T: SomeNumber](t: Tensor[Complex64]; val: T): Tensor[Complex64]
• operators_broadcasted: proc `+.`[T: SomeNumber | Complex[float32] | Complex[float64]](t: Tensor[T]; val: T): Tensor[ T]
• operators_broadcasted: proc `+.`[T: SomeNumber | Complex[float32] | Complex[float64]](a, b: Tensor[T]): Tensor[ T]
• operators_broadcasted_cuda: proc `+.`[T: SomeFloat](a, b: CudaTensor[T]): CudaTensor[T]
• operators_broadcasted_cuda: proc `+.`[T: SomeFloat](t: CudaTensor[T]; val: T): CudaTensor[T]
• operators_broadcasted_cuda: proc `+.`[T: SomeFloat](val: T; t: CudaTensor[T]): CudaTensor[T]
• operators_broadcasted_opencl: proc `+.`[T: SomeFloat](a, b: ClTensor[T]): ClTensor[T]

`+=`:

• operators_blas_l1: proc `+=`[T: SomeNumber | Complex[float32] | Complex[float64]](a: var Tensor[T]; b: Tensor[T])
• operators_blas_l1_cuda: proc `+=`[T: SomeFloat](a: var CudaTensor[T]; b: CudaTensor[T])
• operators_blas_l1_opencl: proc `+=`(dst`gensym286: var ClTensor[float32]; src`gensym286: ClTensor[float32])
• operators_blas_l1_opencl: proc `+=`(dst`gensym286: var ClTensor[float32]; src`gensym286: ClTensor[float32])
• operators_blas_l1_opencl: proc `+=`(dst`gensym329: var ClTensor[float64]; src`gensym329: ClTensor[float64])

`+`:

• gates_basic: proc `+`[TT](a, b: Variable[TT]): Variable[TT]
• gemm_utils: proc `+`(p: ptr; offset: int): type(p)
• operators_blas_l1: proc `+`[T: SomeNumber | Complex[float32] | Complex[float64]](val: T; a: Tensor[T]): Tensor[ T]
• operators_blas_l1: proc `+`[T: SomeNumber | Complex[float32] | Complex[float64]](a: Tensor[T]; val: T): Tensor[ T]
• operators_blas_l1: proc `+`[T: SomeNumber | Complex[float32] | Complex[float64]](a, b: Tensor[T]): Tensor[ T]
• operators_blas_l1_cuda: proc `+`[T: SomeFloat](a, b: CudaTensor[T]): CudaTensor[T]
• operators_blas_l1_opencl: proc `+`(a`gensym30, b`gensym30: ClTensor[float32]): ClTensor[float32]
• operators_blas_l1_opencl: proc `+`(a`gensym30, b`gensym30: ClTensor[float32]): ClTensor[float32]
• operators_blas_l1_opencl: proc `+`(a`gensym101, b`gensym101: ClTensor[float64]): ClTensor[float64]

`-.=`:

• operators_broadcasted: proc `-.=`[T: SomeNumber](t: var Tensor[Complex64]; val: T)
• operators_broadcasted: proc `-.=`[T: SomeNumber | Complex[float32] | Complex[float64]](t: var Tensor[T]; val: T)
• operators_broadcasted: proc `-.=`[T: SomeNumber | Complex[float32] | Complex[float64]](a: var Tensor[T]; b: Tensor[T])
• operators_broadcasted_cuda: proc `-.=`[T: SomeFloat](a: var CudaTensor[T]; b: CudaTensor[T])
• operators_broadcasted_cuda: proc `-.=`[T: SomeFloat](t: var CudaTensor[T]; val: T)

`-.`:

• operators_broadcasted: proc `-.`[T: SomeNumber](val: T; t: Tensor[Complex64]): Tensor[Complex64]
• operators_broadcasted: proc `-.`[T: SomeNumber | Complex[float32] | Complex[float64]](val: T; t: Tensor[T]): Tensor[ T]
• operators_broadcasted: proc `-.`[T: SomeNumber](t: Tensor[Complex64]; val: T): Tensor[Complex64]
• operators_broadcasted: proc `-.`[T: SomeNumber | Complex[float32] | Complex[float64]](t: Tensor[T]; val: T): Tensor[ T]
• operators_broadcasted: proc `-.`[T: SomeNumber | Complex[float32] | Complex[float64]](a, b: Tensor[T]): Tensor[ T]
• operators_broadcasted_cuda: proc `-.`[T: SomeFloat](a, b: CudaTensor[T]): CudaTensor[T]
• operators_broadcasted_cuda: proc `-.`[T: SomeFloat](t: CudaTensor[T]; val: T): CudaTensor[T]
• operators_broadcasted_cuda: proc `-.`[T: SomeFloat](val: T; t: CudaTensor[T]): CudaTensor[T]
• operators_broadcasted_opencl: proc `-.`[T: SomeFloat](a, b: ClTensor[T]): ClTensor[T]

`-=`:

• operators_blas_l1: proc `-=`[T: SomeNumber | Complex[float32] | Complex[float64]](a: var Tensor[T]; b: Tensor[T])
• operators_blas_l1_cuda: proc `-=`[T: SomeFloat](a: var CudaTensor[T]; b: CudaTensor[T])
• operators_blas_l1_opencl: proc `-=`(dst`gensym372: var ClTensor[float32]; src`gensym372: ClTensor[float32])
• operators_blas_l1_opencl: proc `-=`(dst`gensym372: var ClTensor[float32]; src`gensym372: ClTensor[float32])
• operators_blas_l1_opencl: proc `-=`(dst`gensym415: var ClTensor[float64]; src`gensym415: ClTensor[float64])

`-`:

• gates_basic: proc `-`[TT](a, b: Variable[TT]): Variable[TT]
• math_functions: proc `-`[T: SomeNumber](t: Tensor[T]): Tensor[T]
• operators_blas_l1: proc `-`[T: SomeNumber | Complex[float32] | Complex[float64]](val: T; a: Tensor[T]): Tensor[ T]
• operators_blas_l1: proc `-`[T: SomeNumber | Complex[float32] | Complex[float64]](a: Tensor[T]; val: T): Tensor[ T]
• operators_blas_l1: proc `-`[T: SomeNumber | Complex[float32] | Complex[float64]](a, b: Tensor[T]): Tensor[ T]
• operators_blas_l1_cuda: proc `-`[T: SomeFloat](a, b: CudaTensor[T]): CudaTensor[T]
• operators_blas_l1_opencl: proc `-`(a`gensym168, b`gensym168: ClTensor[float32]): ClTensor[float32]
• operators_blas_l1_opencl: proc `-`(a`gensym168, b`gensym168: ClTensor[float32]): ClTensor[float32]
• operators_blas_l1_opencl: proc `-`(a`gensym227, b`gensym227: ClTensor[float64]): ClTensor[float64]

`.!=`:

• operators_comparison: proc `.!=`[T](a, b: Tensor[T]): Tensor[bool]

`.*`:

• operators_broadcasted: proc `.*`[T](val: T; t: Tensor[Complex64]): Tensor[Complex64]
• operators_broadcasted: proc `.*`[T](val: T; t: Tensor[T]): Tensor[T]
• operators_broadcasted: proc `.*`[T](t: Tensor[Complex64]; val: T): Tensor[Complex64]
• operators_broadcasted: proc `.*`[T](t: Tensor[T]; val: T): Tensor[T]
• operators_broadcasted: proc `.*`[T](a, b: Tensor[T]): Tensor[T]
• operators_broadcasted_cuda: proc `.*`[T](a, b: CudaTensor[T]): CudaTensor[T]
• operators_broadcasted_opencl: proc `.*`[T](a, b: ClTensor[T]): ClTensor[T]

`.+`:

• operators_broadcasted: proc `.+`[T](val: T; t: Tensor[Complex64]): Tensor[Complex64]
• operators_broadcasted: proc `.+`[T](val: T; t: Tensor[T]): Tensor[T]
• operators_broadcasted: proc `.+`[T](t: Tensor[Complex64]; val: T): Tensor[Complex64]
• operators_broadcasted: proc `.+`[T](t: Tensor[T]; val: T): Tensor[T]
• operators_broadcasted: proc `.+`[T](a, b: Tensor[T]): Tensor[T]
• operators_broadcasted_cuda: proc `.+`[T](a, b: CudaTensor[T]): CudaTensor[T]
• operators_broadcasted_opencl: proc `.+`[T](a, b: ClTensor[T]): ClTensor[T]

`.-`:

• operators_broadcasted: proc `.-`[T](val: T; t: Tensor[Complex64]): Tensor[Complex64]
• operators_broadcasted: proc `.-`[T](val: T; t: Tensor[T]): Tensor[T]
• operators_broadcasted: proc `.-`[T](t: Tensor[Complex64]; val: T): Tensor[Complex64]
• operators_broadcasted: proc `.-`[T](t: Tensor[T]; val: T): Tensor[T]
• operators_broadcasted: proc `.-`[T](a, b: Tensor[T]): Tensor[T]
• operators_broadcasted_cuda: proc `.-`[T](a, b: CudaTensor[T]): CudaTensor[T]
• operators_broadcasted_opencl: proc `.-`[T](a, b: ClTensor[T]): ClTensor[T]

`...`:

• accessors_macros_syntax: const `...`

`..<`:

• accessors_macros_syntax: proc `..<`(a: int; s: Step): SteppedSlice

`..^`:

• accessors_macros_syntax: proc `..^`(a: int; s: Step): SteppedSlice

`..`:

• accessors_macros_syntax: proc `..`(a: int; s: Step): SteppedSlice

`./`:

• operators_broadcasted: proc `./`[T](val: T; t: Tensor[Complex64]): Tensor[Complex64]
• operators_broadcasted: proc `./`[T](val: T; t: Tensor[T]): Tensor[T]
• operators_broadcasted: proc `./`[T](t: Tensor[Complex64]; val: T): Tensor[Complex64]
• operators_broadcasted: proc `./`[T](t: Tensor[T]; val: T): Tensor[T]
• operators_broadcasted: proc `./`[T](a, b: Tensor[T]): Tensor[T]
• operators_broadcasted_cuda: proc `./`[T](a, b: CudaTensor[T]): CudaTensor[T]
• operators_broadcasted_opencl: proc `./`[T](a, b: ClTensor[T]): ClTensor[T]

`.<=`:

• operators_comparison: proc `.<=`[T](a, b: Tensor[T]): Tensor[bool]

`.<`:

• operators_comparison: proc `.<`[T](a, b: Tensor[T]): Tensor[bool]

`.=*`:

• operators_broadcasted: proc `.=*`[T](t: var Tensor[Complex64]; val: T)
• operators_broadcasted_cuda: proc `.=*`[T](a: var CudaTensor[T]; b: CudaTensor[T])

`.=+`:

• operators_broadcasted: proc `.=+`[T](t: var Tensor[Complex64]; val: T)
• operators_broadcasted_cuda: proc `.=+`[T](a: var CudaTensor[T]; b: CudaTensor[T])

`.=-`:

• operators_broadcasted: proc `.=-`[T](t: var Tensor[Complex64]; val: T)
• operators_broadcasted_cuda: proc `.=-`[T](a: var CudaTensor[T]; b: CudaTensor[T])

`.=/`:

• operators_broadcasted: proc `.=/`[T](t: var Tensor[Complex64]; val: T)
• operators_broadcasted_cuda: proc `.=/`[T](a: var CudaTensor[T]; b: CudaTensor[T])

`.==`:

• operators_comparison: proc `.==`[T](a, b: Tensor[T]): Tensor[bool]

`.>=`:

• operators_comparison: proc `.>=`[T](a, b: Tensor[T]): Tensor[bool]

`.>`:

• operators_comparison: proc `.>`[T](a, b: Tensor[T]): Tensor[bool]

`.^=`:

• operators_broadcasted: proc `.^=`[T](t: var Tensor[Complex64]; val: T)

`.^`:

• operators_broadcasted: proc `.^`[T](val: T; t: Tensor[Complex64]): Tensor[Complex64]
• operators_broadcasted: proc `.^`[T](base: T; t: Tensor[T]): Tensor[T]
• operators_broadcasted: proc `.^`[T](t: Tensor[Complex64]; val: T): Tensor[Complex64]
• operators_broadcasted: proc `.^`[T](t: Tensor[T]; exponent: T): Tensor[T]

`/.=`:

• operators_broadcasted: proc `/.=`[T: SomeNumber](t: var Tensor[Complex64]; val: T)
• operators_broadcasted: proc `/.=`[T: SomeNumber | Complex[float32] | Complex[float64]](t: var Tensor[T]; val: T)
• operators_broadcasted: proc `/.=`[T: SomeNumber | Complex[float32] | Complex[float64]](a: var Tensor[T]; b: Tensor[T])
• operators_broadcasted_cuda: proc `/.=`[T: SomeFloat](a: var CudaTensor[T]; b: CudaTensor[T])

`/.`:

• operators_broadcasted: proc `/.`[T: SomeNumber](val: T; t: Tensor[Complex64]): Tensor[Complex64]
• operators_broadcasted: proc `/.`[T: SomeNumber | Complex[float32] | Complex[float64]](val: T; t: Tensor[T]): Tensor[ T]
• operators_broadcasted: proc `/.`[T: SomeNumber](t: Tensor[Complex64]; val: T): Tensor[Complex64]
• operators_broadcasted: proc `/.`[T: SomeNumber | Complex[float32] | Complex[float64]](t: Tensor[T]; val: T): Tensor[ T]
• operators_broadcasted: proc `/.`[T: SomeNumber | Complex[float32] | Complex[float64]](a, b: Tensor[T]): Tensor[ T]
• operators_broadcasted_cuda: proc `/.`[T: SomeFloat](a, b: CudaTensor[T]): CudaTensor[T]
• operators_broadcasted_cuda: proc `/.`[T: SomeFloat](val: T; t: CudaTensor[T]): CudaTensor[T]
• operators_broadcasted_opencl: proc `/.`[T: SomeFloat](a, b: ClTensor[T]): ClTensor[T]

`/=`:

• operators_blas_l1: proc `/=`[T: SomeFloat | Complex[float32] | Complex[float64]](t: var Tensor[T]; a: T)
• operators_blas_l1: proc `/=`[T: SomeInteger](t: var Tensor[T]; a: T)
• operators_blas_l1_cuda: proc `/=`[T: SomeFloat](t: var CudaTensor[T]; a: T)

`/`:

• operators_blas_l1: proc `/`[T: SomeNumber | Complex[float32] | Complex[float64]](t: Tensor[T]; a: T): Tensor[ T]
• operators_blas_l1_cuda: proc `/`[T: SomeFloat](t: CudaTensor[T]; val: T): CudaTensor[T]

`<.`:

• operators_comparison: proc `<.`[T](t: Tensor[T]; value: T): Tensor[bool]
• operators_comparison: proc `<.`[T](a, b: Tensor[T]): Tensor[bool]
• operators_comparison: template `<.`[T](value: T; t: Tensor[T]): Tensor[bool]

`<=.`:

• operators_comparison: proc `<=.`[T](t: Tensor[T]; value: T): Tensor[bool]
• operators_comparison: proc `<=.`[T](a, b: Tensor[T]): Tensor[bool]
• operators_comparison: template `<=.`[T](value: T; t: Tensor[T]): Tensor[bool]

`<`:

• tensor_compare_helper: proc `<`[T](s1, s2: Tensor[T]): bool

`==.`:

• operators_comparison: proc `==.`[T](t: Tensor[T]; value: T): Tensor[bool]
• operators_comparison: proc `==.`[T](a, b: Tensor[T]): Tensor[bool]
• operators_comparison: template `==.`[T](value: T; t: Tensor[T]): Tensor[bool]

`==`:

• dynamic_stack_arrays: proc `==`(a, s: DynamicStackArray): bool
• dynamic_stack_arrays: proc `==`[T](a: DynamicStackArray[T]; s: openArray[T]): bool
• operators_comparison: proc `==`[T](a, b: Tensor[T]): bool

`>.`:

• operators_comparison: proc `>.`[T](t: Tensor[T]; value: T): Tensor[bool]
• operators_comparison: proc `>.`[T](a, b: Tensor[T]): Tensor[bool]
• operators_comparison: template `>.`[T](value: T; t: Tensor[T]): Tensor[bool]

`>=.`:

• operators_comparison: proc `>=.`[T](t: Tensor[T]; value: T): Tensor[bool]
• operators_comparison: proc `>=.`[T](a, b: Tensor[T]): Tensor[bool]
• operators_comparison: template `>=.`[T](value: T; t: Tensor[T]): Tensor[bool]

`@`:

• dynamic_stack_arrays: proc `@`[T](a: DynamicStackArray[T]): seq[T]

`[]=`:

• accessors_macros_write: macro `[]=`[T](t: var Tensor[T]; args: varargs[untyped]): untyped
• accessors_macros_write: template `[]=`[T](t: Tensor[T]; args: varargs[untyped]): untyped
• datatypes: template `[]=`[T](v: RawMutableView[T]; idx: int; val: T)
• dynamic_stack_arrays: proc `[]=`[T](a: var DynamicStackArray[T]; idx: Index; v: T)
• gemm_utils: template `[]=`[T](view: MatrixView[T]; row, col: Natural; value: T)

`[]`:

• accessors_macros_read: macro `[]`[T](t: AnyTensor[T]; args: varargs[untyped]): untyped
• datatypes: template `[]`[T](v: RawImmutableView[T]; idx: int): T
• datatypes: template `[]`[T](v: RawMutableView[T]; idx: int): var T
• dynamic_stack_arrays: proc `[]`[T](a: DynamicStackArray[T]; idx: Index): T
• dynamic_stack_arrays: proc `[]`[T](a: var DynamicStackArray[T]; idx: Index): var T
• dynamic_stack_arrays: proc `[]`[T](a: DynamicStackArray[T]; slice: Slice[int]): DynamicStackArray[T]
• gates_shapeshifting_views: template `[]`[TT](v: Variable[TT]; args: varargs[untyped]): Variable[TT]
• gemm_utils: template `[]`[T](view: MatrixView[T]; row, col: Natural): T

`^.=`:

• operators_broadcasted: proc `^.=`[T: SomeNumber](t: var Tensor[Complex64]; val: T)
• operators_broadcasted: proc `^.=`[T: SomeFloat | Complex[float32] | Complex[float64]](t: var Tensor[T]; exponent: T)

`^.`:

• operators_broadcasted: proc `^.`[T: SomeNumber](val: T; t: Tensor[Complex64]): Tensor[Complex64]
• operators_broadcasted: proc `^.`[T: SomeFloat | Complex[float32] | Complex[float64]](base: T; t: Tensor[T]): Tensor[ T]
• operators_broadcasted: proc `^.`[T: SomeNumber](t: Tensor[Complex64]; val: T): Tensor[Complex64]
• operators_broadcasted: proc `^.`[T: SomeFloat | Complex[float32] | Complex[float64]](t: Tensor[T]; exponent: T): Tensor[T]

`^`:

• accessors_macros_syntax: proc `^`(s: Slice): SteppedSlice
• accessors_macros_syntax: proc `^`(s: SteppedSlice): SteppedSlice

`_`:

• accessors_macros_syntax: const `_`

`and`:

• operators_logical: proc `and`(a, b: Tensor[bool]): Tensor[bool]

`div`:

• operators_blas_l1: proc `div`[T: SomeInteger](t: Tensor[T]; a: T): Tensor[T]

`mod`:

• operators_blas_l1: proc `mod`[T: SomeNumber](val: T; t: Tensor[T]): Tensor[T]
• operators_blas_l1: proc `mod`[T: SomeNumber](t: Tensor[T]; val: T): Tensor[T]
• operators_broadcasted: proc `mod`[T: SomeNumber](a, b: Tensor[T]): Tensor[T]

`not`:

• operators_logical: proc `not`(a: Tensor[bool]): Tensor[bool]

`or`:

• operators_logical: proc `or`(a, b: Tensor[bool]): Tensor[bool]

`xor`:

• operators_logical: proc `xor`(a, b: Tensor[bool]): Tensor[bool]

`|+`:

• accessors_macros_syntax: proc `|+`(b, step: int): Step
• accessors_macros_syntax: proc `|+`(s: Slice[int]; step: int): SteppedSlice
• accessors_macros_syntax: proc `|+`(ss: SteppedSlice; step: int): SteppedSlice

`|-`:

• accessors_macros_syntax: proc `|-`(b, step: int): Step
• accessors_macros_syntax: proc `|-`(s: Slice[int]; step: int): SteppedSlice
• accessors_macros_syntax: proc `|-`(ss: SteppedSlice; step: int): SteppedSlice

`|`:

• accessors_macros_syntax: proc `|`(b, step: int): Step
• accessors_macros_syntax: proc `|`(s: Slice[int]; step: int): SteppedSlice
• accessors_macros_syntax: proc `|`(ss: SteppedSlice; step: int): SteppedSlice

AAt:

• auxiliary_blas: SyrkKind.AAt

abs:

• math_functions: proc abs(t: Tensor[Complex[float32]]): Tensor[float32]
• math_functions: proc abs(t: Tensor[Complex[float64]]): Tensor[float64]
• math_functions: proc abs[T: SomeNumber](t: Tensor[T]): Tensor[T]

absolute_error:

• common_error_functions: proc absolute_error[T: SomeFloat](y, y_true: T | Complex[T]): T
• common_error_functions: proc absolute_error[T: SomeFloat](y, y_true: Tensor[T] | Tensor[Complex[T]]): Tensor[ T]

accuracy_score:

• accuracy_score: proc accuracy_score[T](y_pred, y_true: Tensor[T]): float

Adam:

• optimizers: object Adam

add:

• dynamic_stack_arrays: proc add[T](a: var DynamicStackArray[T]; value: T)

AddGate:

• gates_basic: type AddGate

address:

• decomposition_lapack: template address(x: typed): untyped

advanceStridedIteration:

• p_accessors: template advanceStridedIteration(coord, backstrides, iter_pos, t, iter_offset, iter_size: typed): untyped

align_raw_data:

• memory: proc align_raw_data(T: typedesc; p: pointer): ptr UncheckedArray[T:type]

all:

• aggregate: proc all[T](t: Tensor[T]): bool

allocCpuStorage:

• datatypes: proc allocCpuStorage[T](storage: var CpuStorage[T]; size: int)

almostEqual:

• math_functions: proc almostEqual[T: SomeFloat | Complex32 | Complex64](t1, t2: Tensor[T]; unitsInLastPlace: Natural = 4): Tensor[bool]

any:

• aggregate: proc any[T](t: Tensor[T]): bool

AnyMetric:

• distances: type AnyMetric

AnyTensor:

• data_structure: type AnyTensor

append:

• shapeshifting: proc append[T](t: Tensor[T]; values: Tensor[T]): Tensor[T]
• shapeshifting: proc append[T](t: Tensor[T]; values: varargs[T]): Tensor[T]

apply:

• higher_order_applymap: proc apply[T: KnownSupportsCopyMem](t: var Tensor[T]; f: proc (x: var T))
• higher_order_applymap: proc apply[T](t: var Tensor[T]; f: T -> T)

apply2:

• higher_order_applymap: proc apply2[T: KnownSupportsCopyMem; U](a: var Tensor[T]; f: proc (x: var T; y: T); b: Tensor[U])
• higher_order_applymap: proc apply2[T: not KnownSupportsCopyMem; U](a: var Tensor[T]; f: proc (x: var T; y: T); b: Tensor[U])

apply2_inline:

• higher_order_applymap: template apply2_inline[T: KnownSupportsCopyMem; U](dest: var Tensor[T]; src: Tensor[U]; op: untyped): untyped
• higher_order_applymap: template apply2_inline[T: not KnownSupportsCopyMem; U](dest: var Tensor[T]; src: Tensor[U]; op: untyped): untyped

apply3_inline:

• higher_order_applymap: template apply3_inline[T: KnownSupportsCopyMem; U, V](dest: var Tensor[T]; src1: Tensor[U]; src2: Tensor[V]; op: untyped): untyped

apply_inline:

• higher_order_applymap: template apply_inline[T: KnownSupportsCopyMem](t: var Tensor[T]; op: untyped): untyped
• higher_order_applymap: template apply_inline[T: not KnownSupportsCopyMem](t: var Tensor[T]; op: untyped): untyped

arange:

• init_cpu: proc arange[T: SomeNumber](start, stop, step: T): Tensor[T]
• init_cpu: template arange[T: SomeNumber](stop: T): Tensor[T]
• init_cpu: template arange[T: SomeNumber](start, stop: T): Tensor[T]

arccos:

• ufunc: proc arccos[T](t`gensym9: Tensor[T]): Tensor[T]
• ufunc: proc arccos[T](t`gensym9: Tensor[T]): Tensor[T]

arccosh:

• ufunc: proc arccosh[T](t`gensym12: Tensor[T]): Tensor[T]
• ufunc: proc arccosh[T](t`gensym12: Tensor[T]): Tensor[T]

arcsin:

• ufunc: proc arcsin[T](t`gensym10: Tensor[T]): Tensor[T]
• ufunc: proc arcsin[T](t`gensym10: Tensor[T]): Tensor[T]

arcsinh:

• ufunc: proc arcsinh[T](t`gensym13: Tensor[T]): Tensor[T]
• ufunc: proc arcsinh[T](t`gensym13: Tensor[T]): Tensor[T]

arctan:

• ufunc: proc arctan[T](t`gensym11: Tensor[T]): Tensor[T]
• ufunc: proc arctan[T](t`gensym11: Tensor[T]): Tensor[T]

arctanh:

• ufunc: proc arctanh[T](t`gensym14: Tensor[T]): Tensor[T]
• ufunc: proc arctanh[T](t`gensym14: Tensor[T]): Tensor[T]

argmax:

• aggregate: proc argmax[T](arg: Tensor[T]; axis: int): Tensor[int]

argmax_max:

• aggregate: proc argmax_max[T: SomeNumber](arg: Tensor[T]; axis: int): tuple[ indices: Tensor[int], maxes: Tensor[T]]

argmin:

• aggregate: proc argmin[T](arg: Tensor[T]; axis: int): Tensor[int]

argmin_min:

• aggregate: proc argmin_min[T: SomeNumber](arg: Tensor[T]; axis: int): tuple[ indices: Tensor[int], mins: Tensor[T]]

argsort:

• algorithms: proc argsort[T](t: Tensor[T]; order = SortOrder.Ascending; toCopy = false): Tensor[ int]

ArrayOfSlices:

• accessors_macros_syntax: type ArrayOfSlices

asContiguous:

• shapeshifting: proc asContiguous[T](t: Tensor[T]; layout: OrderType = rowMajor; force: bool = false): Tensor[ T]
• shapeshifting_cuda: proc asContiguous[T: SomeFloat](t: CudaTensor[T]; layout: OrderType = colMajor; force: bool = false): CudaTensor[T]

asCudnnType:

• cudnn: template asCudnnType[T: SomeFloat](typ: typedesc[T]): cudnnDataType_t

assume_aligned:

• compiler_optim_hints: template assume_aligned[T](data: ptr T; alignment: static int = LASER_MEM_ALIGN): ptr T
• memory_optimization_hints: template assume_aligned[T](data: ptr T; n: csize_t): ptr T

asType:

• ufunc: proc asType[T: SomeNumber; U: Complex](t: Tensor[T]; typ: typedesc[U]): Tensor[U]
• ufunc: proc asType[T; U: not Complex](t: Tensor[T]; typ: typedesc[U]): Tensor[U]

at:

• syntactic_sugar: template at[T](t: Tensor[T]; args: varargs[untyped]): untyped

AtA:

• auxiliary_blas: SyrkKind.AtA

atAxisIndex:

• accessors: proc atAxisIndex[T](t: Tensor[T]; axis, idx: int; length = 1): Tensor[T]

atContiguousIndex:

• accessors: proc atContiguousIndex[T](t: Tensor[T]; idx: int): T
• accessors: proc atContiguousIndex[T](t: var Tensor[T]; idx: int): var T

atIndex:

• p_accessors: proc atIndex[T](t: Tensor[T]; idx: varargs[int]): T
• p_accessors: proc atIndex[T](t: var Tensor[T]; idx: varargs[int]): var T

atIndexMut:

• p_accessors: proc atIndexMut[T](t: var Tensor[T]; idx: varargs[int]; val: T)

at_mut:

• syntactic_sugar: template at_mut[T](t: var Tensor[T]; args: varargs[untyped]): untyped

attachGC:

• openmp: template attachGC(): untyped

axis:

• accessors: iterator axis[T](t: Tensor[T]; axis: int): Tensor[T]
• accessors: iterator axis[T](t: Tensor[T]; axis, offset, size: int): Tensor[T]

backprop:

• autograd_common: proc backprop[TT](v: Variable[TT])

Backward:

• autograd_common: type Backward

bc:

• shapeshifting: template bc(t: (Tensor | SomeNumber); shape: Metadata): untyped
• shapeshifting: template bc(t: (Tensor | SomeNumber); shape: varargs[int]): untyped

blasMM_C_eq_aAB_p_bC:

• p_operator_blas_l2l3: proc blasMM_C_eq_aAB_p_bC[T: SomeFloat | Complex[float32] | Complex[float64]]( alpha: T; a, b: Tensor[T]; beta: T; c: var Tensor[T])

blasMV_y_eq_aAx_p_by:

• p_operator_blas_l2l3: proc blasMV_y_eq_aAx_p_by[T: SomeFloat | Complex[float32] | Complex[float64]]( alpha: T; a, x: Tensor[T]; beta: T; y: var Tensor[T])

box:

• distributions: proc box(x: float): float
• distributions: template box[T](t`gensym0: Tensor[T]): Tensor[float]

boxKernel:

• kde: proc boxKernel(x`gensym1, x_i`gensym1, bw`gensym1: float): float

broadcast:

• shapeshifting: proc broadcast[T: SomeNumber](val: T; shape: Metadata): Tensor[T]
• shapeshifting: proc broadcast[T: SomeNumber](val: T; shape: varargs[int]): Tensor[T]
• shapeshifting: proc broadcast[T](t: Tensor[T]; shape: Metadata): Tensor[T]
• shapeshifting: proc broadcast[T](t: Tensor[T]; shape: varargs[int]): Tensor[T]
• shapeshifting_cuda: proc broadcast(t: CudaTensor; shape: Metadata): CudaTensor
• shapeshifting_cuda: proc broadcast(t: CudaTensor; shape: varargs[int]): CudaTensor

broadcast2:

• shapeshifting: proc broadcast2[T](a, b: Tensor[T]): tuple[a, b: Tensor[T]]
• shapeshifting_cuda: proc broadcast2[T](a, b: CudaTensor[T]): tuple[a, b: CudaTensor[T]]
• shapeshifting_opencl: proc broadcast2[T](a, b: ClTensor[T]): tuple[a, b: ClTensor[T]]

broadcast2Impl:

• p_shapeshifting: proc broadcast2Impl[T](a, b: AnyTensor[T]; result: var tuple[a, b: AnyTensor[T]])

broadcastImpl:

• p_shapeshifting: proc broadcastImpl(t: var AnyTensor; shape: varargs[int] | Metadata)

cbrt:

• ufunc: proc cbrt[T](t`gensym4: Tensor[T]): Tensor[T]
• ufunc: proc cbrt[T](t`gensym4: Tensor[T]): Tensor[T]

ceil:

• ufunc: proc ceil[T](t`gensym26: Tensor[T]): Tensor[T]
• ufunc: proc ceil[T](t`gensym26: Tensor[T]): Tensor[T]

check_axis_index:

• p_checks: proc check_axis_index(t: AnyTensor; axis, index, len: Natural)

check_concat:

• p_checks: proc check_concat(t1, t2: Tensor; axis: int)

check_contiguous_index:

• p_checks: proc check_contiguous_index(t: Tensor; idx: int)

check_ctx:

• autograd_common: proc check_ctx(a, b: Variable)

check_dot_prod:

• p_checks: proc check_dot_prod(a, b: AnyTensor)

check_elementwise:

• p_checks: proc check_elementwise[T, U](a: ClTensor[T]; b: ClTensor[U])
• p_checks: proc check_elementwise[T, U](a: CudaTensor[T]; b: CudaTensor[U])
• p_checks: proc check_elementwise[T, U](a: Tensor[T]; b: Tensor[U])

check_index:

• p_checks: proc check_index(t: Tensor; idx: varargs[int])

check_input_target:

• p_nnp_checks: proc check_input_target[T](input, target: Tensor[T])

check_matmat:

• p_checks: proc check_matmat(a, b: AnyTensor)

check_matvec:

• p_checks: proc check_matvec(a, b: AnyTensor)

check_nested_elements:

• p_checks: proc check_nested_elements(shape: Metadata; len: int)

check_reshape:

• p_checks: proc check_reshape(t: AnyTensor; new_shape: Metadata)
• p_checks: proc check_reshape(t: AnyTensor; new_shape: varargs[int])

check_shape:

• p_checks: proc check_shape(a: Tensor; b: Tensor | openArray; relaxed_rank1_check: static[bool] = false)

check_size:

• p_checks: proc check_size[T, U](a: Tensor[T]; b: Tensor[U])

check_squeezeAxis:

• p_checks: proc check_squeezeAxis(t: AnyTensor; axis: int)

check_start_end:

• p_checks: proc check_start_end(a, b: int; dim_size: int)

check_steps:

• p_checks: proc check_steps(a, b, step: int)

check_unsqueezeAxis:

• p_checks: proc check_unsqueezeAxis(t: AnyTensor; axis: int)

chunk:

• gates_shapeshifting_concat_split: proc chunk[TT](v: Variable[TT]; nb_chunks: Positive; axis: Natural): seq[Variable[TT]]
• shapeshifting: proc chunk[T](t: Tensor[T]; nb_chunks: Positive; axis: Natural): seq[Tensor[T]]

ChunkSplitGate:

• gates_shapeshifting_concat_split: type ChunkSplitGate

circulant:

• special_matrices: proc circulant[T](t: Tensor[T]; axis = -1; step = 1): Tensor[T]

clamp:

• math_functions: proc clamp[T](t: Tensor[T]; min, max: T): Tensor[T]

classify:

• math_functions: proc classify[T: SomeFloat](t: Tensor[T]): Tensor[FloatClass]

clContext0:

• opencl_global_state: let clContext0

clDevice0:

• opencl_global_state: let clDevice0

clMalloc:

• opencl_backend: proc clMalloc[T](size: Natural): ptr UncheckedArray[T]

clone:

• init_copy_cpu: proc clone[T](t: Tensor[T]; layout: OrderType = rowMajor): Tensor[T]
• init_copy_cuda: proc clone[T](t: CudaTensor[T]): CudaTensor[T]
• kdtree: proc clone[T](kd: KDTree[T]): KDTree[T]
• kdtree: proc clone[T](n: Node[T]): Node[T]

clQueue0:

• opencl_global_state: let clQueue0

ClStorage:

• data_structure: object ClStorage

ClTensor:

• data_structure: object ClTensor

col2im:

• conv: proc col2im[T](input: Tensor[T]; channels, height, width: int; kernel_size: Size2D; padding: Size2D = (0, 0); stride: Size2D = (1, 1)): Tensor[T]

complex:

• complex: proc complex[T: SomeNumber](re: Tensor[T]): auto
• complex: proc complex[T: SomeNumber](re: Tensor[T]; im: Tensor[T]): auto

Complex32:

• p_complex: converter Complex32[T: SomeNumber](x: T): Complex[float32]

Complex64:

• p_complex: converter Complex64[T: SomeNumber](x: T): Complex[float64]

complex_imag:

• complex: proc complex_imag[T: SomeNumber](im: Tensor[T]): auto

concat:

• dynamic_stack_arrays: proc concat[T](dsas: varargs[DynamicStackArray[T]]): DynamicStackArray[T]
• shapeshifting: proc concat[T](t_list: varargs[Tensor[T]]; axis: int): Tensor[T]

concatMap:

• functional: proc concatMap[T](s: seq[T]; f: proc (ss: T): string): string

conjugate:

• complex: proc conjugate[T: Complex32 | Complex64](t: Tensor[T]): Tensor[T]

contains:

• algorithms: proc contains[T](t: Tensor[T]; item: T): bool

Context:

• autograd_common: type Context

contiguousImpl:

• p_shapeshifting: proc contiguousImpl[T](t: Tensor[T]; layout: OrderType; result: var Tensor[T])

Conv2D:

• conv2D: object Conv2D

conv2d:

• conv2D: proc conv2d[TT](input, weight: Variable[TT]; bias: Variable[TT] = nil; padding: Size2D = (0, 0); stride: Size2D = (1, 1)): Variable[TT]
• nnp_conv2d_cudnn: proc conv2d[T: SomeFloat](input, kernel, bias: CudaTensor[T]; padding: SizeHW = [0, 0]; strides, dilation: SizeHW = [1, 1]): CudaTensor[T]
• nnp_convolution: proc conv2d[T](input, weight, bias: Tensor[T]; padding: Size2D = (0, 0); stride: Size2D = (1, 1); algorithm = Conv2DAlgorithm.Im2ColGEMM): Tensor[ T]

Conv2DAlgorithm:

• nnp_convolution: enum Conv2DAlgorithm

conv2d_backward:

• nnp_conv2d_cudnn: proc conv2d_backward[T: SomeFloat](input, kernel, bias: CudaTensor[T]; padding: SizeHW = [0, 0]; strides, dilation: SizeHW = [1, 1]; grad_output: CudaTensor[T]; grad_input, grad_kernel, grad_bias: var CudaTensor[T])
• nnp_convolution: proc conv2d_backward[T](input, weight, bias: Tensor[T]; padding: Size2D; stride: Size2D; grad_output: Tensor[T]; grad_input, grad_weight, grad_bias: var Tensor[T]; algorithm = Conv2DAlgorithm.Im2ColGEMM)

Conv2DGate:

• conv2D: type Conv2DGate

ConvAlgoSpace:

• cudnn_conv_interface: object ConvAlgoSpace

conv_bwd_data_algo_workspace:

• cudnn_conv_interface: proc conv_bwd_data_algo_workspace[T: SomeFloat]( srcTensorDesc: cudnnTensorDescriptor_t; gradOutputTensorDesc: cudnnTensorDescriptor_t; kernelDesc: cudnnFilterDescriptor_t; convDesc: cudnnConvolutionDescriptor_t; gradInputTensorDesc: cudnnTensorDescriptor_t): ConvAlgoSpace[T, cudnnConvolutionBwdDataAlgo_t]

conv_bwd_kernel_algo_workspace:

• cudnn_conv_interface: proc conv_bwd_kernel_algo_workspace[T: SomeFloat]( srcTensorDesc: cudnnTensorDescriptor_t; gradOutputTensorDesc: cudnnTensorDescriptor_t; gradKernelDesc: cudnnFilterDescriptor_t; convDesc: cudnnConvolutionDescriptor_t): ConvAlgoSpace[T, cudnnConvolutionBwdFilterAlgo_t]

ConvConfig:

• cudnn_conv_interface: object ConvConfig

convolve:

• math_functions: proc convolve[T: SomeNumber | Complex32 | Complex64](t1, t2: Tensor[T]; mode = ConvolveMode.full): Tensor[T]

ConvolveMode:

• math_functions: enum ConvolveMode

convOutDims:

• cudnn_conv_interface: proc convOutDims(input, kernel: CudaTensor; padding, strides, dilation: SizeHW): Metadata

copyFrom:

• dynamic_stack_arrays: proc copyFrom(a: var DynamicStackArray; s: DynamicStackArray)
• dynamic_stack_arrays: proc copyFrom(a: var DynamicStackArray; s: varargs[int])

copy_from:

• filling_data: proc copy_from[T](dst: var Tensor[T]; src: Tensor[T])

copyFrom:

• initialization: proc copyFrom[T](dst: var Tensor[T]; src: Tensor[T])

copyFromRaw:

• initialization: proc copyFromRaw[T](dst: var Tensor[T]; buffer: ptr T; len: Natural)

copySign:

• math_functions: proc copySign[T: SomeFloat](t1, t2: Tensor[T]): Tensor[T]

correlate:

• math_functions: proc correlate[T: Complex32 | Complex64](t1, t2: Tensor[T]; mode = CorrelateMode.valid): Tensor[T]
• math_functions: proc correlate[T: SomeNumber](t1, t2: Tensor[T]; mode = CorrelateMode.valid): Tensor[ T]

CorrelateMode:

• math_functions: type CorrelateMode

cos:

• ufunc: proc cos[T](t`gensym15: Tensor[T]): Tensor[T]
• ufunc: proc cos[T](t`gensym15: Tensor[T]): Tensor[T]

cosh:

• ufunc: proc cosh[T](t`gensym16: Tensor[T]): Tensor[T]
• ufunc: proc cosh[T](t`gensym16: Tensor[T]): Tensor[T]

covariance_matrix:

• stats: proc covariance_matrix[T: SomeFloat](x, y: Tensor[T]): Tensor[T]

cpu:

• init_cuda: proc cpu[T: SomeFloat](t: CudaTensor[T]): Tensor[T]
• init_opencl: proc cpu[T: SomeFloat](t: ClTensor[T]): Tensor[T]

CPUFeatureX86:

• gemm_tiling: enum CPUFeatureX86

CpuStorage:

• datatypes: type CpuStorage

cpuStorageFromBuffer:

• datatypes: proc cpuStorageFromBuffer[T: KnownSupportsCopyMem](storage: var CpuStorage[T]; rawBuffer: pointer; size: int)

create_cache_dirs_if_necessary:

• util: proc create_cache_dirs_if_necessary()

cswap:

• complex: proc cswap[T: Complex32 | Complex64](t: Tensor[T]): Tensor[T]

cublas_axpy:

• cublas: proc cublas_axpy[T: SomeFloat](n: int; alpha: T; x: ptr T; incx: int; y: ptr T; incy: int)

cublas_copy:

• cublas: proc cublas_copy[T: SomeFloat](n: int; x: ptr T; incx: int; y: ptr T; incy: int)

cublas_dot:

• cublas: proc cublas_dot[T: SomeFloat](n: int; x: ptr T; incx: int; y: ptr T; incy: int; output: ptr T)

cublas_geam:

• cublas: proc cublas_geam[T: SomeFloat](transa, transb: cublasOperation_t; m, n: int; alpha: T; A: ptr T; lda: int; beta: T; B: ptr T; ldb: int; C: ptr T; ldc: int)

cublas_gemm:

• cublas: proc cublas_gemm[T: SomeFloat](transa, transb: cublasOperation_t; m, n, k: int; alpha: T; A: ptr T; lda: int; B: ptr T; ldb: int; beta: T; C: ptr T; ldc: int)

cublas_gemmStridedBatched:

• cublas: proc cublas_gemmStridedBatched[T: SomeFloat](transa, transb: cublasOperation_t; m, n, k: int; alpha: T; A: ptr T; lda: int; strideA: int; B: ptr T; ldb: int; strideB: int; beta: T; C: ptr T; ldc: int; strideC: int; batchCount: int)

cublas_gemv:

• cublas: proc cublas_gemv[T: SomeFloat](trans: cublasOperation_t; m, n: int; alpha: T; A: ptr T; lda: int; x: ptr T; incx: int; beta: T; y: ptr T; incy: int)

cublasHandle0:

• cuda_global_state: let cublasHandle0

cublas_scal:

• cublas: proc cublas_scal[T: SomeFloat](n: int; alpha: T; x: ptr T; incx: int)

cuda:

• init_cuda: proc cuda[T: SomeFloat](t: Tensor[T]): CudaTensor[T]

cuda_assign_call:

• p_kernels_interface_cuda: template cuda_assign_call[T: SomeFloat](kernel_name: untyped; destination: var CudaTensor[T]; source: CudaTensor[T]): untyped

cuda_assign_glue:

• p_kernels_interface_cuda: template cuda_assign_glue(kernel_name, op_name: string; binding_name: untyped): untyped

cuda_assignscal_call:

• p_kernels_interface_cuda: template cuda_assignscal_call[T: SomeFloat](kernel_name: untyped; destination: var CudaTensor[T]; val: T): untyped

cuda_assignscal_glue:

• p_kernels_interface_cuda: template cuda_assignscal_glue(kernel_name, op_name: string; binding_name: untyped): untyped

cuda_binary_call:

• p_kernels_interface_cuda: template cuda_binary_call[T: SomeFloat](kernel_name: untyped; destination: var CudaTensor[T]; a, b: CudaTensor[T]): untyped

cuda_binary_glue:

• p_kernels_interface_cuda: template cuda_binary_glue(kernel_name, op_name: string; binding_name: untyped): untyped

CUDA_HOF_BPG:

• global_config: const CUDA_HOF_BPG

CUDA_HOF_TPB:

• global_config: const CUDA_HOF_TPB

cuda_lscal_call:

• p_kernels_interface_cuda: template cuda_lscal_call[T: SomeFloat](kernel_name: untyped; destination: var CudaTensor[T]; alpha: T; source: CudaTensor[T]): untyped

cuda_lscal_glue:

• p_kernels_interface_cuda: template cuda_lscal_glue(kernel_name, op_name: string; binding_name: untyped): untyped

cudaMalloc:

• cuda: proc cudaMalloc[T](size: Natural): ptr T

cuda_rscal_call:

• p_kernels_interface_cuda: template cuda_rscal_call[T: SomeFloat](kernel_name: untyped; destination: var CudaTensor[T]; source: CudaTensor[T]; beta: T): untyped

cuda_rscal_glue:

• p_kernels_interface_cuda: template cuda_rscal_glue(kernel_name, op_name: string; binding_name: untyped): untyped

CudaStorage:

• data_structure: object CudaStorage

cudaStream0:

• cuda_global_state: let cudaStream0

CudaTensor:

• data_structure: object CudaTensor

cudnnHandle0:

• cudnn: let cudnnHandle0

cumprod:

• aggregate: proc cumprod[T](arg: Tensor[T]; axis: int = 0): Tensor[T]

cumsum:

• aggregate: proc cumsum[T](arg: Tensor[T]; axis: int = 0): Tensor[T]

CustomMetric:

• distances: object CustomMetric

cvtmask64_u64:

• simd: proc cvtmask64_u64(a: mmask64): uint64

data=:

• data_structure: proc data=[T](t: var Tensor[T]; s: seq[T])

dataArray:

• data_structure: proc dataArray[T: KnownSupportsCopyMem](t: Tensor[T]): ptr UncheckedArray[T]
• data_structure: proc dataArray[T: not KnownSupportsCopyMem](t: Tensor[T]): ptr UncheckedArray[T]

dbscan:

• dbscan: proc dbscan[T: SomeFloat](X: Tensor[T]; eps: float; minSamples: int; metric: typedesc[AnyMetric] = Euclidean; p = 2.0): seq[int]

deallocCl:

• opencl_backend: proc deallocCl[T](p: ref [ptr UncheckedArray[T]])

deallocCuda:

• cuda: proc deallocCuda[T](p: ref [ptr T])

deepCopy:

• initialization: proc deepCopy[T](dst: var Tensor[T]; src: Tensor[T])

degToRad:

• ufunc: proc degToRad[T](t`gensym29: Tensor[T]): Tensor[T]
• ufunc: proc degToRad[T](t`gensym29: Tensor[T]): Tensor[T]

delete:

• dynamic_stack_arrays: proc delete(a: var DynamicStackArray; index: int)

desugar:

• p_accessors_macros_desugar: macro desugar(args: untyped): void

detachGC:

• openmp: template detachGC(): untyped

diag:

• special_matrices: proc diag[T](d: Tensor[T]; k = 0; anti = false): Tensor[T]

diagonal:

• special_matrices: proc diagonal[T](a: Tensor[T]; k = 0; anti = false): Tensor[T]

diff_discrete:

• aggregate: proc diff_discrete[T](arg: Tensor[T]; n = 1; axis: int = -1): Tensor[T]

disp2d:

• p_display: proc disp2d[T](t: Tensor[T]; alignBy = 6; alignSpacing = 3; precision = -1): string

distance:

• distances: proc distance(metric: typedesc[Euclidean]; v, w: Tensor[float]; squared: static bool = false): float
• distances: proc distance(metric: typedesc[Jaccard]; v, w: Tensor[float]): float
• distances: proc distance(metric: typedesc[Manhattan]; v, w: Tensor[float]): float
• distances: proc distance(metric: typedesc[Minkowski]; v, w: Tensor[float]; p = 2.0; squared: static bool = false): float

distanceMatrix:

• distances: proc distanceMatrix(metric: typedesc[AnyMetric]; x, y: Tensor[float]; p = 2.0; squared: static bool = false): Tensor[float]

dot:

• operators_blas_l1: proc dot[T: SomeFloat](a, b: Tensor[T]): T
• operators_blas_l1: proc dot[T: SomeInteger](a, b: Tensor[T]): T
• operators_blas_l1_cuda: proc dot[T: SomeFloat](a, b: CudaTensor[T]): T
• operators_blas_l1_opencl: proc dot(a`gensym0, b`gensym0: ClTensor[float32]): float32
• operators_blas_l1_opencl: proc dot(a`gensym15, b`gensym15: ClTensor[float64]): float64

dualStridedIteration:

• p_accessors: template dualStridedIteration(strider: IterKind; t1, t2, iter_offset, iter_size: typed): untyped

dualStridedIterationYield:

• p_accessors: template dualStridedIterationYield(strider: IterKind; t1data, t2data, i, t1_iter_pos, t2_iter_pos: typed)

DynamicStackArray:

• dynamic_stack_arrays: object DynamicStackArray

einsum:

• einsum: macro einsum(tensors: varargs[typed]; stmt: untyped): untyped

Ellipsis:

• accessors_macros_syntax: object Ellipsis

elwise_div:

• math_functions: proc elwise_div[T: SomeFloat](a, b: Tensor[T]): Tensor[T]
• math_functions: proc elwise_div[T: SomeInteger](a, b: Tensor[T]): Tensor[T]

elwise_mul:

• math_functions: proc elwise_mul[T](a, b: Tensor[T]): Tensor[T]

Embedding:

• embedding: object Embedding

embedding:

• embedding: proc embedding[TT; Idx: VocabIdx](input_vocab_id: Tensor[Idx]; weight: Variable[TT]; padding_idx: Idx = -1; scale_grad_by_freq: static[bool] = false): Variable[ TT]
• nnp_embedding: proc embedding[T; Idx: byte or char or SomeInteger](vocab_id: Tensor[Idx]; weight: Tensor[T]): Tensor[T]

embedding_backward:

• nnp_embedding: proc embedding_backward[T; Idx: byte or char or SomeInteger](dWeight: var Tensor[T]; vocab_id: Tensor[Idx]; dOutput: Tensor[T]; padding_idx: Idx; scale_grad_by_freq: static[bool] = false)

EmbeddingGate:

• embedding: type EmbeddingGate

enumerate:

• accessors: iterator enumerate[T](t: Tensor[T]): (int, T)
• accessors: iterator enumerate[T](t: Tensor[T]; offset, size: int): (int, T)

enumerateAxis:

• accessors: iterator enumerateAxis[T](t: Tensor[T]; axis: int): (int, Tensor[T])
• accessors: iterator enumerateAxis[T](t: Tensor[T]; axis, offset, size: int): (int, Tensor[T])

enumerateZip:

• accessors: iterator enumerateZip[T, U](t1: Tensor[T]; t2: Tensor[U]): (int, T, U)
• accessors: iterator enumerateZip[T, U](t1: Tensor[T]; t2: Tensor[U]; offset, size: int): (int, T, U)
• accessors: iterator enumerateZip[T, U, V](t1: Tensor[T]; t2: Tensor[U]; t3: Tensor[V]): (int, T, U, V)
• accessors: iterator enumerateZip[T, U, V](t1: Tensor[T]; t2: Tensor[U]; t3: Tensor[V]; offset, size: int): (int, T, U, V)

epanechnikov:

• distributions: proc epanechnikov(x: float): float
• distributions: template epanechnikov[T](t`gensym3: Tensor[T]): Tensor[float]

epanechnikovKernel:

• kde: proc epanechnikovKernel(x`gensym4, x_i`gensym4, bw`gensym4: float): float

erf:

• ufunc: proc erf[T](t`gensym21: Tensor[T]): Tensor[T]
• ufunc: proc erf[T](t`gensym21: Tensor[T]): Tensor[T]

erfc:

• ufunc: proc erfc[T](t`gensym22: Tensor[T]): Tensor[T]
• ufunc: proc erfc[T](t`gensym22: Tensor[T]): Tensor[T]

Euclidean:

• distances: object Euclidean

exch_dim:

• p_shapeshifting: proc exch_dim[T](t: Tensor[T]; dim1, dim2: int): Tensor[T]

exp:

• ufunc: proc exp[T](t`gensym8: Tensor[T]): Tensor[T]
• ufunc: proc exp[T](t`gensym8: Tensor[T]): Tensor[T]

expm1:

• math_ops_fusion: proc expm1(x: float32): float32
• math_ops_fusion: proc expm1(x: float64): float64

export_tensor:

• exporting: proc export_tensor[T](t: Tensor[T]): tuple[shape: seq[int], strides: seq[int], data: seq[T]]

extract_cpu_simd:

• gemm_tiling: macro extract_cpu_simd(ukernel: static MicroKernel): untyped

extract_c_unit_stride:

• gemm_tiling: macro extract_c_unit_stride(ukernel: static MicroKernel): untyped

extract_mr:

• gemm_tiling: macro extract_mr(ukernel: static MicroKernel): untyped

extract_nb_scalars:

• gemm_tiling: macro extract_nb_scalars(ukernel: static MicroKernel): untyped

extract_nb_vecs_nr:

• gemm_tiling: macro extract_nb_vecs_nr(ukernel: static MicroKernel): untyped

extract_nr:

• gemm_tiling: macro extract_nr(ukernel: static MicroKernel): untyped

extract_pt:

• gemm_tiling: macro extract_pt(ukernel: static MicroKernel): untyped

eye:

• special_matrices: proc eye[T](shape: varargs[int]): Tensor[T]

fac:

• ufunc: proc fac[T](t`gensym1: Tensor[T]): Tensor[T]
• ufunc: proc fac[T](t`gensym1: Tensor[T]): Tensor[T]

fallbackMM_C_eq_aAB_p_bC:

• p_operator_blas_l2l3: proc fallbackMM_C_eq_aAB_p_bC[T: SomeInteger](alpha: T; a, b: Tensor[T]; beta: T; c: var Tensor[T])

FancyIndex:

• p_accessors_macros_read: FancySelectorKind.FancyIndex

FancyMaskAxis:

• p_accessors_macros_read: FancySelectorKind.FancyMaskAxis

FancyMaskFull:

• p_accessors_macros_read: FancySelectorKind.FancyMaskFull

FancyNone:

• p_accessors_macros_read: FancySelectorKind.FancyNone

FancySelectorKind:

• p_accessors_macros_read: enum FancySelectorKind

FancyUnknownAxis:

• p_accessors_macros_read: FancySelectorKind.FancyUnknownAxis

FancyUnknownFull:

• p_accessors_macros_read: FancySelectorKind.FancyUnknownFull

flatIter:

• nested_containers: iterator flatIter[T](s: openArray[T]): auto
• nested_containers: iterator flatIter(s: string): string

Flatten:

• flatten: object Flatten

flatten:

• gates_shapeshifting_views: proc flatten[TT](a: Variable[TT]): Variable[TT]
• shapeshifting: proc flatten(t: Tensor): Tensor

floor:

• ufunc: proc floor[T](t`gensym25: Tensor[T]): Tensor[T]
• ufunc: proc floor[T](t`gensym25: Tensor[T]): Tensor[T]

floorMod:

• math_functions: proc floorMod[T: SomeNumber](val: T; t: Tensor[T]): Tensor[T]
• math_functions: proc floorMod[T: SomeNumber](t: Tensor[T]; val: T): Tensor[T]
• math_functions: proc floorMod[T: SomeNumber](t1, t2: Tensor[T]): Tensor[T]

fold:

• higher_order_foldreduce: proc fold[U, T](arg: Tensor[U]; start_val: T; f: (T, U) -> T): T
• higher_order_foldreduce: proc fold[U, T](arg: Tensor[U]; start_val: Tensor[T]; f: (Tensor[T], Tensor[U]) -> Tensor[T]; axis: int): Tensor[T]

fold_axis_inline:

• higher_order_foldreduce: template fold_axis_inline[T](arg: Tensor[T]; accumType: typedesc; fold_axis: int; op_initial, op_middle, op_final: untyped): untyped

fold_inline:

• higher_order_foldreduce: template fold_inline[T](arg: Tensor[T]; op_initial, op_middle, op_final: untyped): untyped

forEach:

• foreach: macro forEach(args: varargs[untyped]): untyped

forEachContiguous:

• foreach: macro forEachContiguous(args: varargs[untyped]): untyped

forEachContiguousSerial:

• foreach: macro forEachContiguousSerial(args: varargs[untyped]): untyped

forEachSerial:

• foreach: macro forEachSerial(args: varargs[untyped]): untyped

forEachStaged:

• foreach_staged: macro forEachStaged(args: varargs[untyped]): untyped

forEachStrided:

• foreach: macro forEachStrided(args: varargs[untyped]): untyped

forEachStridedSerial:

• foreach: macro forEachStridedSerial(args: varargs[untyped]): untyped

forward:

• conv2D: proc forward[T](self: Conv2D[T]; input: Variable[Tensor[T]]): Variable[Tensor[T]]
• embedding: proc forward[T; Idx: VocabIdx](self: Embedding[T]; input: Tensor[Idx]): Variable[ AnyTensor[T]]
• flatten: proc forward[T](self: Flatten[T]; input: Variable[Tensor[T]]): Variable[Tensor[T]]
• gcn: proc forward[T](self: GCNLayer[T]; input, adjacency: Variable[Tensor[T]]): Variable[ Tensor[T]]
• gru: proc forward[T](self: GRULayer[T]; input, hidden0: Variable): tuple[ output, hiddenN: Variable]
• linear: proc forward[T](self: Linear[T]; input: Variable[Tensor[T]]): Variable[Tensor[T]]
• maxpool2D: proc forward[T](self: MaxPool2D[T]; input: Variable[Tensor[T]]): Variable[Tensor[T]]

frobenius_inner_prod:

• lapack: proc frobenius_inner_prod[T](a, b: Tensor[T]): T

fromBuffer:

• initialization: proc fromBuffer[T](rawBuffer: pointer; shape: varargs[int]): Tensor[T]
• initialization: proc fromBuffer[T](rawBuffer: pointer; shape: varargs[int]; layout: static OrderType): Tensor[ T]
• initialization: proc fromBuffer[T](rawBuffer: ptr UncheckedArray[T]; shape: varargs[int]): Tensor[T]
• initialization: proc fromBuffer[T](rawBuffer: ptr UncheckedArray[T]; shape: varargs[int]; layout: static OrderType): Tensor[T]

full:

• math_functions: ConvolveMode.full

gamma:

• ufunc: proc gamma[T](t`gensym23: Tensor[T]): Tensor[T]
• ufunc: proc gamma[T](t`gensym23: Tensor[T]): Tensor[T]

Gate:

• autograd_common: type Gate

gauss:

• distributions: proc gauss[T](x, mean, sigma: T; norm = false): float
• distributions: proc gauss[T](x: Tensor[T]; mean, sigma: T; norm = false): Tensor[float]

gaussKernel:

• kde: proc gaussKernel(x, x_i, bw: float): float

gcn:

• gcn: proc gcn[TT](input, adjacency, weight: Variable[TT]; bias: Variable[TT] = nil): Variable[ TT]

GCNGate:

• gcn: type GCNGate

GCNLayer:

• gcn: object GCNLayer

gebb_ukernel:

• gemm_ukernel_dispatch: proc gebb_ukernel[T; ukernel: static MicroKernel](kc: int; alpha: T; packedA, packedB: ptr UncheckedArray[T]; beta: T; vC: MatrixView[T])

gebb_ukernel_edge:

• gemm_ukernel_dispatch: proc gebb_ukernel_edge[T; ukernel: static MicroKernel](mr, nr, kc: int; alpha: T; packedA, packedB: ptr UncheckedArray[T]; beta: T; vC: MatrixView[T])

gebb_ukernel_edge_epilogue:

• gemm_ukernel_generic: proc gebb_ukernel_edge_epilogue[MR, NR: static int; T](alpha: T; AB: ptr array[MR, array[NR, T]]; beta: T; vC: MatrixView[T]; mr, nr: int)

gebb_ukernel_edge_fallback:

• gemm_ukernel_generic: proc gebb_ukernel_edge_fallback[T; ukernel: static MicroKernel](mr, nr, kc: int; alpha: T; packedA, packedB: ptr UncheckedArray[T]; beta: T; vC: MatrixView[T])

gebb_ukernel_edge_float32_x86_AVX:

• gemm_ukernel_avx: proc gebb_ukernel_edge_float32_x86_AVX[ukernel: static MicroKernel]( mr`gensym3, nr`gensym3, kc`gensym3: int; alpha`gensym3: float32; packedA`gensym3, packedB`gensym3: ptr UncheckedArray[float32]; beta`gensym3: float32; vC`gensym3: MatrixView[float32])

gebb_ukernel_edge_float32_x86_AVX512:

• gemm_ukernel_avx512: proc gebb_ukernel_edge_float32_x86_AVX512[ukernel: static MicroKernel]( mr`gensym3, nr`gensym3, kc`gensym3: int; alpha`gensym3: float32; packedA`gensym3, packedB`gensym3: ptr UncheckedArray[float32]; beta`gensym3: float32; vC`gensym3: MatrixView[float32])

gebb_ukernel_edge_float32_x86_AVX_FMA:

• gemm_ukernel_avx_fma: proc gebb_ukernel_edge_float32_x86_AVX_FMA[ukernel: static MicroKernel]( mr`gensym3, nr`gensym3, kc`gensym3: int; alpha`gensym3: float32; packedA`gensym3, packedB`gensym3: ptr UncheckedArray[float32]; beta`gensym3: float32; vC`gensym3: MatrixView[float32])

gebb_ukernel_edge_float32_x86_SSE:

• gemm_ukernel_sse: proc gebb_ukernel_edge_float32_x86_SSE[ukernel: static MicroKernel]( mr`gensym3, nr`gensym3, kc`gensym3: int; alpha`gensym3: float32; packedA`gensym3, packedB`gensym3: ptr UncheckedArray[float32]; beta`gensym3: float32; vC`gensym3: MatrixView[float32])

gebb_ukernel_edge_float64_x86_AVX:

• gemm_ukernel_avx: proc gebb_ukernel_edge_float64_x86_AVX[ukernel: static MicroKernel]( mr`gensym7, nr`gensym7, kc`gensym7: int; alpha`gensym7: float64; packedA`gensym7, packedB`gensym7: ptr UncheckedArray[float64]; beta`gensym7: float64; vC`gensym7: MatrixView[float64])

gebb_ukernel_edge_float64_x86_AVX512:

• gemm_ukernel_avx512: proc gebb_ukernel_edge_float64_x86_AVX512[ukernel: static MicroKernel]( mr`gensym7, nr`gensym7, kc`gensym7: int; alpha`gensym7: float64; packedA`gensym7, packedB`gensym7: ptr UncheckedArray[float64]; beta`gensym7: float64; vC`gensym7: MatrixView[float64])

gebb_ukernel_edge_float64_x86_AVX_FMA:

• gemm_ukernel_avx_fma: proc gebb_ukernel_edge_float64_x86_AVX_FMA[ukernel: static MicroKernel]( mr`gensym7, nr`gensym7, kc`gensym7: int; alpha`gensym7: float64; packedA`gensym7, packedB`gensym7: ptr UncheckedArray[float64]; beta`gensym7: float64; vC`gensym7: MatrixView[float64])

gebb_ukernel_edge_float64_x86_SSE2:

• gemm_ukernel_sse2: proc gebb_ukernel_edge_float64_x86_SSE2[ukernel: static MicroKernel]( mr`gensym3, nr`gensym3, kc`gensym3: int; alpha`gensym3: float64; packedA`gensym3, packedB`gensym3: ptr UncheckedArray[float64]; beta`gensym3: float64; vC`gensym3: MatrixView[float64])

gebb_ukernel_edge_int32_x86_AVX2:

• gemm_ukernel_avx2: proc gebb_ukernel_edge_int32_x86_AVX2[ukernel: static MicroKernel]( mr`gensym3, nr`gensym3, kc`gensym3: int; alpha`gensym3: int32; packedA`gensym3, packedB`gensym3: ptr UncheckedArray[int32]; beta`gensym3: int32; vC`gensym3: MatrixView[int32])

gebb_ukernel_edge_int32_x86_AVX512:

• gemm_ukernel_avx512: proc gebb_ukernel_edge_int32_x86_AVX512[ukernel: static MicroKernel]( mr`gensym11, nr`gensym11, kc`gensym11: int; alpha`gensym11: int32; packedA`gensym11, packedB`gensym11: ptr UncheckedArray[int32]; beta`gensym11: int32; vC`gensym11: MatrixView[int32])

gebb_ukernel_edge_int32_x86_SSE2:

• gemm_ukernel_sse2: proc gebb_ukernel_edge_int32_x86_SSE2[ukernel: static MicroKernel]( mr`gensym7, nr`gensym7, kc`gensym7: int; alpha`gensym7: int32; packedA`gensym7, packedB`gensym7: ptr UncheckedArray[int32]; beta`gensym7: int32; vC`gensym7: MatrixView[int32])

gebb_ukernel_edge_int32_x86_SSE4_1:

• gemm_ukernel_sse4_1: proc gebb_ukernel_edge_int32_x86_SSE4_1[ukernel: static MicroKernel]( mr`gensym3, nr`gensym3, kc`gensym3: int; alpha`gensym3: int32; packedA`gensym3, packedB`gensym3: ptr UncheckedArray[int32]; beta`gensym3: int32; vC`gensym3: MatrixView[int32])

gebb_ukernel_edge_int64_x86_AVX512:

• gemm_ukernel_avx512: proc gebb_ukernel_edge_int64_x86_AVX512[ukernel: static MicroKernel]( mr`gensym15, nr`gensym15, kc`gensym15: int; alpha`gensym15: int64; packedA`gensym15, packedB`gensym15: ptr UncheckedArray[int64]; beta`gensym15: int64; vC`gensym15: MatrixView[int64])

gebb_ukernel_edge_int64_x86_SSE2:

• gemm_ukernel_sse2: proc gebb_ukernel_edge_int64_x86_SSE2[ukernel: static MicroKernel]( mr`gensym11, nr`gensym11, kc`gensym11: int; alpha`gensym11: int64; packedA`gensym11, packedB`gensym11: ptr UncheckedArray[int64]; beta`gensym11: int64; vC`gensym11: MatrixView[int64])

gebb_ukernel_epilogue_fallback:

• gemm_ukernel_generic: proc gebb_ukernel_epilogue_fallback[MR, NR: static int; T](alpha: T; AB: ptr array[MR, array[NR, T]]; beta: T; vC: MatrixView[T])

gebb_ukernel_fallback:

• gemm_ukernel_generic: proc gebb_ukernel_fallback[T; ukernel: static MicroKernel](kc: int; alpha: T; packedA, packedB: ptr UncheckedArray[T]; beta: T; vC: MatrixView[T])

gebb_ukernel_float32_x86_AVX:

• gemm_ukernel_avx: proc gebb_ukernel_float32_x86_AVX[ukernel: static MicroKernel](kc`gensym2: int; alpha`gensym2: float32; packedA`gensym2, packedB`gensym2: ptr UncheckedArray[float32]; beta`gensym2: float32; vC`gensym2: MatrixView[float32])

gebb_ukernel_float32_x86_AVX512:

• gemm_ukernel_avx512: proc gebb_ukernel_float32_x86_AVX512[ukernel: static MicroKernel](kc`gensym2: int; alpha`gensym2: float32; packedA`gensym2, packedB`gensym2: ptr UncheckedArray[float32]; beta`gensym2: float32; vC`gensym2: MatrixView[float32])

gebb_ukernel_float32_x86_AVX_FMA:

• gemm_ukernel_avx_fma: proc gebb_ukernel_float32_x86_AVX_FMA[ukernel: static MicroKernel](kc`gensym2: int; alpha`gensym2: float32; packedA`gensym2, packedB`gensym2: ptr UncheckedArray[float32]; beta`gensym2: float32; vC`gensym2: MatrixView[float32])

gebb_ukernel_float32_x86_SSE:

• gemm_ukernel_sse: proc gebb_ukernel_float32_x86_SSE[ukernel: static MicroKernel](kc`gensym2: int; alpha`gensym2: float32; packedA`gensym2, packedB`gensym2: ptr UncheckedArray[float32]; beta`gensym2: float32; vC`gensym2: MatrixView[float32])

gebb_ukernel_float64_x86_AVX:

• gemm_ukernel_avx: proc gebb_ukernel_float64_x86_AVX[ukernel: static MicroKernel](kc`gensym6: int; alpha`gensym6: float64; packedA`gensym6, packedB`gensym6: ptr UncheckedArray[float64]; beta`gensym6: float64; vC`gensym6: MatrixView[float64])

gebb_ukernel_float64_x86_AVX512:

• gemm_ukernel_avx512: proc gebb_ukernel_float64_x86_AVX512[ukernel: static MicroKernel](kc`gensym6: int; alpha`gensym6: float64; packedA`gensym6, packedB`gensym6: ptr UncheckedArray[float64]; beta`gensym6: float64; vC`gensym6: MatrixView[float64])

gebb_ukernel_float64_x86_AVX_FMA:

• gemm_ukernel_avx_fma: proc gebb_ukernel_float64_x86_AVX_FMA[ukernel: static MicroKernel](kc`gensym6: int; alpha`gensym6: float64; packedA`gensym6, packedB`gensym6: ptr UncheckedArray[float64]; beta`gensym6: float64; vC`gensym6: MatrixView[float64])

gebb_ukernel_float64_x86_SSE2:

• gemm_ukernel_sse2: proc gebb_ukernel_float64_x86_SSE2[ukernel: static MicroKernel](kc`gensym2: int; alpha`gensym2: float64; packedA`gensym2, packedB`gensym2: ptr UncheckedArray[float64]; beta`gensym2: float64; vC`gensym2: MatrixView[float64])

gebb_ukernel_int32_x86_AVX2:

• gemm_ukernel_avx2: proc gebb_ukernel_int32_x86_AVX2[ukernel: static MicroKernel](kc`gensym2: int; alpha`gensym2: int32; packedA`gensym2, packedB`gensym2: ptr UncheckedArray[int32]; beta`gensym2: int32; vC`gensym2: MatrixView[int32])

gebb_ukernel_int32_x86_AVX512:

• gemm_ukernel_avx512: proc gebb_ukernel_int32_x86_AVX512[ukernel: static MicroKernel](kc`gensym10: int; alpha`gensym10: int32; packedA`gensym10, packedB`gensym10: ptr UncheckedArray[int32]; beta`gensym10: int32; vC`gensym10: MatrixView[int32])

gebb_ukernel_int32_x86_SSE2:

• gemm_ukernel_sse2: proc gebb_ukernel_int32_x86_SSE2[ukernel: static MicroKernel](kc`gensym6: int; alpha`gensym6: int32; packedA`gensym6, packedB`gensym6: ptr UncheckedArray[int32]; beta`gensym6: int32; vC`gensym6: MatrixView[int32])

gebb_ukernel_int32_x86_SSE4_1:

• gemm_ukernel_sse4_1: proc gebb_ukernel_int32_x86_SSE4_1[ukernel: static MicroKernel](kc`gensym2: int; alpha`gensym2: int32; packedA`gensym2, packedB`gensym2: ptr UncheckedArray[int32]; beta`gensym2: int32; vC`gensym2: MatrixView[int32])

gebb_ukernel_int64_x86_AVX512:

• gemm_ukernel_avx512: proc gebb_ukernel_int64_x86_AVX512[ukernel: static MicroKernel](kc`gensym14: int; alpha`gensym14: int64; packedA`gensym14, packedB`gensym14: ptr UncheckedArray[int64]; beta`gensym14: int64; vC`gensym14: MatrixView[int64])

gebb_ukernel_int64_x86_SSE2:

• gemm_ukernel_sse2: proc gebb_ukernel_int64_x86_SSE2[ukernel: static MicroKernel](kc`gensym10: int; alpha`gensym10: int64; packedA`gensym10, packedB`gensym10: ptr UncheckedArray[int64]; beta`gensym10: int64; vC`gensym10: MatrixView[int64])

gebp_mkernel:

• gemm: proc gebp_mkernel[T; ukernel: static MicroKernel](mc, nc, kc: int; alpha: T; packA, packB: ptr UncheckedArray[T]; beta: T; mcncC: MatrixView[T])

gelsd:

• least_squares_lapack: proc gelsd[T: SomeFloat](a, b: Tensor[T]; solution, residuals: var Tensor[T]; singular_values: var Tensor[T]; matrix_rank: var int; rcond = -1.T)

gemm:

• operators_blas_l2l3: proc gemm[T: SomeFloat | Complex](alpha: T; A, B: Tensor[T]; beta: T; C: var Tensor[T])
• operators_blas_l2l3: proc gemm[T: SomeInteger](alpha: T; A, B: Tensor[T]; beta: T; C: var Tensor[T])
• operators_blas_l2l3: proc gemm[T: SomeNumber](A, B: Tensor[T]; C: var Tensor[T])

gemm_nn_fallback:

• blas_l3_gemm: proc gemm_nn_fallback[T](m, n, k: int; alpha: T; A: seq[T]; offA: int; incRowA, incColA: int; B: seq[T]; offB: int; incRowB, incColB: int; beta: T; C: var seq[T]; offC: int; incRowC, incColc: int)

gemm_packed:

• gemm_prepacked: proc gemm_packed[T: SomeNumber](M, N, K: int; alpha: T; packedA: ptr (T or UncheckedArray[T]); packedB: ptr (T or UncheckedArray[T]); beta: T; C: ptr (T or UncheckedArray[T]); rowStrideC, colStrideC: int)

gemm_prepackA:

• gemm_prepacked: proc gemm_prepackA[T](dst_packedA: ptr (T or UncheckedArray[T]); M, N, K: int; src_A: ptr T; rowStrideA, colStrideA: int)

gemm_prepackA_mem_required:

• gemm_prepacked: proc gemm_prepackA_mem_required(T: typedesc; M, N, K: int): int

gemm_prepackA_mem_required_impl:

• gemm_prepacked: proc gemm_prepackA_mem_required_impl(ukernel: static MicroKernel; T: typedesc; M, N, K: int): int

gemm_prepackB:

• gemm_prepacked: proc gemm_prepackB[T](dst_packedB: ptr (T or UncheckedArray[T]); M, N, K: int; src_B: ptr T; rowStrideB, colStrideB: int)

gemm_prepackB_mem_required:

• gemm_prepacked: proc gemm_prepackB_mem_required(T: type; M, N, K: int): int

gemm_prepackB_mem_required_impl:

• gemm_prepacked: proc gemm_prepackB_mem_required_impl(ukernel: static MicroKernel; T: typedesc; M, N, K: int): int

gemm_strided:

• gemm: proc gemm_strided[T: SomeNumber and not (uint32 | uint64 | uint | int)](M, N, K: int; alpha: T; A: ptr T; rowStrideA, colStrideA: int; B: ptr T; rowStrideB, colStrideB: int; beta: T; C: ptr T; rowStrideC, colStrideC: int)
• gemm: proc gemm_strided[T: uint32 | uint64 | uint | int](M, N, K: int; alpha: T; A: ptr T; rowStrideA, colStrideA: int; B: ptr T; rowStrideB, colStrideB: int; beta: T; C: ptr T; rowStrideC, colStrideC: int)

gemv:

• operators_blas_l2l3: proc gemv[T: SomeFloat | Complex](alpha: T; A: Tensor[T]; x: Tensor[T]; beta: T; y: var Tensor[T])
• operators_blas_l2l3: proc gemv[T: SomeInteger](alpha: T; A: Tensor[T]; x: Tensor[T]; beta: T; y: var Tensor[T])

gen_cl_apply2:

• p_kernels_interface_opencl: template gen_cl_apply2(kern_name, ctype, op: string): string

gen_cl_apply3:

• p_kernels_interface_opencl: template gen_cl_apply3(kern_name, ctype, op: string): string

genClInfixOp:

• p_kernels_interface_opencl: template genClInfixOp(T: typedesc; ctype: string; procName: untyped; cName: string; cInfixOp: string; exported: static[bool] = true): untyped

genClInPlaceOp:

• p_kernels_interface_opencl: template genClInPlaceOp(T: typedesc; ctype: string; procName: untyped; cName: string; cInfixOp: string; exported: static[bool] = true): untyped

geomspace:

• init_cpu: proc geomspace[T: SomeFloat](start, stop: T; num: int; endpoint = true): Tensor[float]

geqrf:

• decomposition_lapack: proc geqrf[T: SupportedDecomposition](Q: var Tensor[T]; tau: var seq[T]; scratchspace: var seq[T])

gesdd:

• decomposition_lapack: proc gesdd[T: SupportedDecomposition; X: SupportedDecomposition](a: var Tensor[T]; U: var Tensor[T]; S: var Tensor[X]; Vh: var Tensor[T]; scratchspace: var seq[T])

gesv:

• solve_lapack: proc gesv[T: SomeFloat](a, b: var Tensor[T]; pivot_indices: var seq[int32])

get_cache_dir:

• util: proc get_cache_dir(): string

getContiguousIndex:

• p_accessors: proc getContiguousIndex[T](t: Tensor[T]; idx: int): int

get_data_ptr:

• data_structure: proc get_data_ptr[T](t: CudaTensor[T] or ClTensor[T]): ptr T
• data_structure: proc get_data_ptr[T: not KnownSupportsCopyMem](t: AnyTensor[T]): ptr T
• data_structure: proc get_data_ptr[T: KnownSupportsCopyMem](t: Tensor[T]): ptr T

getFancySelector:

• p_accessors_macros_read: proc getFancySelector(ast: NimNode; axis: var int; selector: var NimNode): FancySelectorKind

getIndex:

• p_accessors: proc getIndex[T](t: Tensor[T]; idx: varargs[int]): int

get_num_tiles:

• gemm_tiling: proc get_num_tiles(dim_size, tile_size: int): int

get_offset_ptr:

• data_structure: proc get_offset_ptr[T](t: CudaTensor[T] or ClTensor[T]): ptr T
• data_structure: proc get_offset_ptr[T: not KnownSupportsCopyMem](t: AnyTensor[T]): ptr T
• data_structure: proc get_offset_ptr[T: KnownSupportsCopyMem](t: Tensor[T]): ptr T

getrf:

• decomposition_lapack: proc getrf[T: SupportedDecomposition](lu: var Tensor[T]; pivot_indices: var seq[int32])

getShape:

• nested_containers: proc getShape[T](s: openArray[T]; parent_shape = Metadata()): Metadata
• nested_containers: proc getShape(s: string; parent_shape = Metadata()): Metadata

getSubType:

• ast_utils: macro getSubType(TT: typedesc): untyped

gru:

• gru: proc gru[TT](input, hidden0: Variable[TT]; W3s0, W3sN, U3s: Variable[TT]; bW3s, bU3s: Variable[TT]): tuple[output, hiddenN: Variable[TT]]

gru_backward:

• nnp_gru: proc gru_backward[T: SomeFloat](dInput, dHidden0, dW3s0, dW3sN, dU3s, dbW3s, dbU3s: var Tensor[ T]; dOutput, dHiddenN: Tensor[T]; cached_inputs: seq[Tensor[T]]; cached_hiddens: seq[seq[Tensor[T]]]; W3s0, W3sN, U3s, rs, zs, ns, Uhs: Tensor[T])

gru_cell_backward:

• nnp_gru: proc gru_cell_backward[T: SomeFloat](dx, dh, dW3, dU3, dbW3, dbU3: var Tensor[T]; dnext: Tensor[T]; x, h, W3, U3: Tensor[T]; r, z, n, Uh: Tensor[T])

gru_cell_forward:

• nnp_gru: proc gru_cell_forward[T: SomeFloat](input, W3, U3, bW3, bU3: Tensor[T]; r, z, n, Uh, hidden: var Tensor[T])

gru_cell_inference:

• nnp_gru: proc gru_cell_inference[T: SomeFloat](input: Tensor[T]; W3, U3, bW3, bU3: Tensor[T]; hidden: var Tensor[T])

gru_forward:

• nnp_gru: proc gru_forward[T: SomeFloat](input: Tensor[T]; W3s0, W3sN: Tensor[T]; U3s, bW3s, bU3s: Tensor[T]; rs, zs, ns, Uhs: var Tensor[T]; output, hidden: var Tensor[T]; cached_inputs: var seq[Tensor[T]]; cached_hiddens: var seq[seq[Tensor[T]]])

GRUGate:

• gru: type GRUGate

gru_inference:

• nnp_gru: proc gru_inference[T: SomeFloat](input: Tensor[T]; W3s0, W3sN: Tensor[T]; U3s, bW3s, bU3s: Tensor[T]; output, hidden: var Tensor[T])

GRULayer:

• gru: object GRULayer

HadamardGate:

• gates_hadamard: type HadamardGate

hankel:

• special_matrices: proc hankel[T](c: Tensor[T]): Tensor[T]
• special_matrices: proc hankel[T](c, r: Tensor[T]): Tensor[T]

has3DNow:

• cpuinfo_x86: proc has3DNow(): bool

has3DNowEnhanced:

• cpuinfo_x86: proc has3DNowEnhanced(): bool

hasAbm:

• cpuinfo_x86: proc hasAbm(): bool

hasAdx:

• cpuinfo_x86: proc hasAdx(): bool

hasAes:

• cpuinfo_x86: proc hasAes(): bool

hasAmdv:

• cpuinfo_x86: proc hasAmdv(): bool

hasAvx:

• cpuinfo_x86: proc hasAvx(): bool

hasAvx2:

• cpuinfo_x86: proc hasAvx2(): bool

hasAvx512bfloat16:

• cpuinfo_x86: proc hasAvx512bfloat16(): bool

hasAvx512bitalg:

• cpuinfo_x86: proc hasAvx512bitalg(): bool

hasAvx512bw:

• cpuinfo_x86: proc hasAvx512bw(): bool

hasAvx512cd:

• cpuinfo_x86: proc hasAvx512cd(): bool

hasAvx512dq:

• cpuinfo_x86: proc hasAvx512dq(): bool

hasAvx512er:

• cpuinfo_x86: proc hasAvx512er(): bool

hasAvx512f:

• cpuinfo_x86: proc hasAvx512f(): bool

hasAvx512fmaps4:

• cpuinfo_x86: proc hasAvx512fmaps4(): bool

hasAvx512ifma:

• cpuinfo_x86: proc hasAvx512ifma(): bool

hasAvx512pf:

• cpuinfo_x86: proc hasAvx512pf(): bool

hasAvx512vbmi:

• cpuinfo_x86: proc hasAvx512vbmi(): bool

hasAvx512vbmi2:

• cpuinfo_x86: proc hasAvx512vbmi2(): bool

hasAvx512vl:

• cpuinfo_x86: proc hasAvx512vl(): bool

hasAvx512vnni:

• cpuinfo_x86: proc hasAvx512vnni(): bool

hasAvx512vnniw4:

• cpuinfo_x86: proc hasAvx512vnniw4(): bool

hasAvx512vp2intersect:

• cpuinfo_x86: proc hasAvx512vp2intersect(): bool

hasAvx512vpopcntdq:

• cpuinfo_x86: proc hasAvx512vpopcntdq(): bool

hasBmi1:

• cpuinfo_x86: proc hasBmi1(): bool

hasBmi2:

• cpuinfo_x86: proc hasBmi2(): bool

hasCas16B:

• cpuinfo_x86: proc hasCas16B(): bool

hasCas8B:

• cpuinfo_x86: proc hasCas8B(): bool

hasClflush:

• cpuinfo_x86: proc hasClflush(): bool

hasClflushOpt:

• cpuinfo_x86: proc hasClflushOpt(): bool

hasClwb:

• cpuinfo_x86: proc hasClwb(): bool

hasFloat16c:

• cpuinfo_x86: proc hasFloat16c(): bool

hasFma3:

• cpuinfo_x86: proc hasFma3(): bool

hasFma4:

• cpuinfo_x86: proc hasFma4(): bool

hasGfni:

• cpuinfo_x86: proc hasGfni(): bool

hasIntelVtx:

• cpuinfo_x86: proc hasIntelVtx(): bool

hasMmx:

• cpuinfo_x86: proc hasMmx(): bool

hasMmxExt:

• cpuinfo_x86: proc hasMmxExt(): bool

hasMovBigEndian:

• cpuinfo_x86: proc hasMovBigEndian(): bool

hasMpx:

• cpuinfo_x86: proc hasMpx(): bool

hasNxBit:

• cpuinfo_x86: proc hasNxBit(): bool

hasPclmulqdq:

• cpuinfo_x86: proc hasPclmulqdq(): bool

hasPopcnt:

• cpuinfo_x86: proc hasPopcnt(): bool

hasPrefetch:

• cpuinfo_x86: proc hasPrefetch(): bool

hasPrefetchWT1:

• cpuinfo_x86: proc hasPrefetchWT1(): bool

hasRdrand:

• cpuinfo_x86: proc hasRdrand(): bool

hasRdseed:

• cpuinfo_x86: proc hasRdseed(): bool

hasSgx:

• cpuinfo_x86: proc hasSgx(): bool

hasSha:

• cpuinfo_x86: proc hasSha(): bool

hasSimultaneousMultithreading:

• cpuinfo_x86: proc hasSimultaneousMultithreading(): bool

hasSse:

• cpuinfo_x86: proc hasSse(): bool

hasSse2:

• cpuinfo_x86: proc hasSse2(): bool

hasSse3:

• cpuinfo_x86: proc hasSse3(): bool

hasSse41:

• cpuinfo_x86: proc hasSse41(): bool

hasSse42:

• cpuinfo_x86: proc hasSse42(): bool

hasSse4a:

• cpuinfo_x86: proc hasSse4a(): bool

hasSsse3:

• cpuinfo_x86: proc hasSsse3(): bool

hasTsxHle:

• cpuinfo_x86: proc hasTsxHle(): bool

hasTsxRtm:

• cpuinfo_x86: proc hasTsxRtm(): bool

hasType:

• ast_utils: proc hasType(x: NimNode; t: static[string]): bool

hasVaes:

• cpuinfo_x86: proc hasVaes(): bool

hasVpclmulqdq:

• cpuinfo_x86: proc hasVpclmulqdq(): bool

hasX87fpu:

• cpuinfo_x86: proc hasX87fpu(): bool

hasXop:

• cpuinfo_x86: proc hasXop(): bool

high:

• dynamic_stack_arrays: proc high(a: DynamicStackArray): int

hilbert:

• special_matrices: proc hilbert(n: int; T: typedesc[SomeFloat]): Tensor[T]

identity:

• special_matrices: proc identity[T](n: int): Tensor[T]

ijgrid:

• special_matrices: MeshGridIndexing.ijgrid

im2col:

• conv: proc im2col[T](input: Tensor[T]; kernel_size: Size2D; padding: Size2D = (0, 0); stride: Size2D = (1, 1); result: var Tensor[T])

Im2ColGEMM:

• nnp_convolution: Conv2DAlgorithm.Im2ColGEMM

im2colgemm_conv2d:

• conv: proc im2colgemm_conv2d[T](input, kernel, bias: Tensor[T]; padding: Size2D = (0, 0); stride: Size2D = (1, 1)): Tensor[T]

im2colgemm_conv2d_gradient:

• conv: proc im2colgemm_conv2d_gradient[T](input, kernel: Tensor[T]; padding: Size2D = (0, 0); stride: Size2D = (1, 1); grad_output: Tensor[T]; grad_input, grad_weight: var Tensor[T])

imag:

• complex: proc imag[T: SomeFloat](t: Tensor[Complex[T]]): Tensor[T]

imag=:

• complex: proc imag=[T: SomeFloat](t: var Tensor[Complex[T]]; val: T)
• complex: proc imag=[T: SomeFloat](t: var Tensor[Complex[T]]; val: Tensor[T])

implDeprecatedBy:

• deprecate: macro implDeprecatedBy(oldName: untyped; replacement: typed; exported: static bool): untyped

index_fill:

• selectors: proc index_fill[T; Idx: byte or char or SomeInteger](t: var Tensor[T]; axis: int; indices: openArray[Idx]; value: T)
• selectors: proc index_fill[T; Idx: byte or char or SomeInteger](t: var Tensor[T]; axis: int; indices: Tensor[Idx]; value: T)

index_select:

• selectors: proc index_select[T; Idx: byte or char or SomeInteger](t: Tensor[T]; axis: int; indices: openArray[Idx]): Tensor[T]
• selectors: proc index_select[T; Idx: byte or char or SomeInteger](t: Tensor[T]; axis: int; indices: Tensor[Idx]): Tensor[T]

infer_shape:

• p_shapeshifting: proc infer_shape(t: Tensor; new_shape: varargs[int]): seq[int]

init:

• conv2D: proc init[T](ctx: Context[Tensor[T]]; layerType: typedesc[Conv2D[T]]; inShape: seq[int]; outChannels: int; kernelSize: Size2D; padding: Size2D = (0, 0); stride: Size2D = (1, 1)): Conv2D[T]
• embedding: proc init[T](ctx: Context[Tensor[T]]; layerType: typedesc[Embedding[T]]; vocabSize, embedSize: int; paddingIdx: VocabIdx = -1): Embedding[T]
• flatten: proc init[T](ctx: Context[Tensor[T]]; layerType: typedesc[Flatten[T]]; inShape: seq[int]): Flatten[T]
• gcn: proc init[T](ctx: Context[Tensor[T]]; layerType: typedesc[GCNLayer[T]]; numInput, numOutput: int): GCNLayer[T]
• gru: proc init[T](ctx: Context[Tensor[T]]; layerType: typedesc[GRULayer[T]]; numInputFeatures, hiddenSize, layers: int): GRULayer[T]
• linear: proc init[T](ctx: Context[Tensor[T]]; layerType: typedesc[Linear[T]]; numInput, numOutput: int): Linear[T]
• maxpool2D: proc init[T](ctx: Context[Tensor[T]]; layerType: typedesc[MaxPool2D[T]]; inShape: seq[int]; kernelSize, padding, stride: Size2D): MaxPool2D[T]

initForEach:

• foreach_common: proc initForEach(params: NimNode; values, aliases, raw_ptrs: var NimNode; aliases_stmt, raw_ptrs_stmt: var NimNode; test_shapes: var NimNode)

initMetadataArray:

• datatypes: proc initMetadataArray(len: int): Metadata

initSpanSlices:

• accessors_macros_syntax: proc initSpanSlices(len: int): ArrayOfSlices

initStridedIteration:

• p_accessors: template initStridedIteration(coord, backstrides, iter_pos: untyped; t, iter_offset, iter_size: typed): untyped

initTensorMetadata:

• initialization: proc initTensorMetadata(result: var Tensor; size: var int; shape: Metadata; layout: static OrderType = rowMajor)
• initialization: proc initTensorMetadata(result: var Tensor; size: var int; shape: openArray[int]; layout: static OrderType = rowMajor)

insert:

• dynamic_stack_arrays: proc insert[T](a: var DynamicStackArray[T]; value: T; index: int = 0)

inShape:

• conv2D: proc inShape[T](self: Conv2D[T]): seq[int]
• embedding: proc inShape[T](self: Embedding[T]): seq[int]
• flatten: proc inShape[T](self: Flatten[T]): seq[int]
• gcn: proc inShape[T](self: GCNLayer[T]): seq[int]
• linear: proc inShape[T](self: Linear[T]): seq[int]
• maxpool2D: proc inShape[T](self: MaxPool2D[T]): seq[int]

int2bit:

• special_matrices: proc int2bit(value: int; n: int; msbfirst = true): Tensor[bool]
• special_matrices: proc int2bit[T: SomeInteger](t: Tensor[T]; n: int; msbfirst = true): Tensor[bool]

intersection:

• algorithms: proc intersection[T](t1, t2: Tensor[T]): Tensor[T]

iqr:

• aggregate: proc iqr[T](arg: Tensor[T]): float

isAllInt:

• ast_utils: proc isAllInt(slice_args: NimNode): bool

isBool:

• ast_utils: proc isBool(x: NimNode): bool

is_C_contiguous:

• data_structure: proc is_C_contiguous(t: CudaTensor or ClTensor): bool
• datatypes: proc is_C_contiguous(t: Tensor): bool

isContiguous:

• data_structure: proc isContiguous(t: AnyTensor): bool

is_F_contiguous:

• data_structure: proc is_F_contiguous(t: AnyTensor): bool

is_grad_needed:

• autograd_common: proc is_grad_needed(v: Variable): bool

isHypervisorPresent:

• cpuinfo_x86: proc isHypervisorPresent(): bool

isInt:

• ast_utils: proc isInt(x: NimNode): bool

ismember:

• algorithms: proc ismember[T](t1, t2: Tensor[T]): Tensor[bool]

isNaN:

• operators_comparison: proc isNaN(t: Tensor[SomeFloat]): Tensor[bool]
• ufunc: proc isNaN[T](t`gensym2: Tensor[T]): Tensor[T]
• ufunc: proc isNaN[T](t`gensym2: Tensor[T]): Tensor[T]

isNotNaN:

• operators_comparison: proc isNotNaN(t: Tensor[SomeFloat]): Tensor[bool]

isOpenArray:

• ast_utils: proc isOpenArray(x: NimNode): bool

item:

• initialization: proc item[T](t: Tensor[T]): T
• initialization: proc item[T_IN, T_OUT](t: Tensor[T_IN]; _: typedesc[T_OUT]): T_OUT

items:

• accessors: iterator items[T](t: Tensor[T]): T
• accessors: iterator items[T](t: Tensor[T]; offset, size: int): T
• dynamic_stack_arrays: iterator items[T](a: DynamicStackArray[T]): T

IterKind:

• p_accessors: enum IterKind

Iter_Values:

• p_accessors: IterKind.Iter_Values

Jaccard:

• distances: object Jaccard

kaiming_normal:

• init: proc kaiming_normal(shape: varargs[int]; T: type): Tensor[T]

kaiming_uniform:

• init: proc kaiming_uniform(shape: varargs[int]; T: type): Tensor[T]

kde:

• kde: proc kde[T: SomeNumber; U: int | Tensor[SomeNumber] | openArray[SomeNumber]]( t: Tensor[T]; kernel: static KernelFunc; kernelKind = knCustom; adjust: float = 1.0; samples: U = 1000; bw: float = NaN; normalize = false; cutoff: float = NaN; weights: Tensor[T] = newTensor[T](0)): Tensor[float]
• kde: proc kde[T: SomeNumber; U: KernelKind | string; V: int | Tensor[SomeNumber] | openArray[SomeNumber]](t: Tensor[T]; kernel: U = "gauss"; adjust: float = 1.0; samples: V = 1000; bw: float = NaN; normalize = false; weights: Tensor[T] = newTensor[T](0)): Tensor[ float]

KDTree:

• kdtree: type KDTree

kdTree:

• kdtree: proc kdTree[T](data: Tensor[T]; leafSize = 16; copyData = true; balancedTree: static bool = true): KDTree[T]

KernelFunc:

• kde: type KernelFunc

KernelKind:

• kde: enum KernelKind

kmeans:

• kmeans: proc kmeans[T: SomeFloat](x: Tensor[T]; n_clusters = 10; tol: float = 0.0001; n_init = 10; max_iters = 300; seed = 1000; random = false): tuple[ labels: Tensor[int], centroids: Tensor[T], inertia: T]
• kmeans: proc kmeans[T: SomeFloat](x: Tensor[T]; centroids: Tensor[T]): Tensor[int]

knBox:

• kde: KernelKind.knBox

knCustom:

• kde: KernelKind.knCustom

knEpanechnikov:

• kde: KernelKind.knEpanechnikov

knGauss:

• kde: KernelKind.knGauss

KnownSupportsCopyMem:

• datatypes: type KnownSupportsCopyMem

knTriangular:

• kde: KernelKind.knTriangular

knTrig:

• kde: KernelKind.knTrig

LASER_MAXRANK:

• dynamic_stack_arrays: const LASER_MAXRANK

LASER_MEM_ALIGN:

• compiler_optim_hints: const LASER_MEM_ALIGN

laswp:

• auxiliary_lapack: proc laswp(a: var Tensor; pivot_indices: openArray[int32]; pivot_from: static int32)

layoutOnDevice:

• cuda: proc layoutOnDevice[T: SomeFloat](t: CudaTensor[T]): CudaTensorLayout[T]
• opencl_backend: proc layoutOnDevice[T: SomeFloat](t: ClTensor[T]): ClTensorLayout[T]

least_squares_solver:

• least_squares: proc least_squares_solver[T: SomeFloat](a, b: Tensor[T]; rcond = -1.T): tuple[ solution: Tensor[T], residuals: Tensor[T], matrix_rank: int, singular_values: Tensor[T]]

len:

• datatypes: proc len[T](t: Tensor[T]): int

letsGoDeeper:

• ast_utils: template letsGoDeeper()

lgamma:

• ufunc: proc lgamma[T](t`gensym24: Tensor[T]): Tensor[T]
• ufunc: proc lgamma[T](t`gensym24: Tensor[T]): Tensor[T]

Linear:

• linear: object Linear

linear:

• linear: proc linear[TT](input, weight: Variable[TT]; bias: Variable[TT] = nil): Variable[TT]
• nnp_linear: proc linear[T](input, weight: Tensor[T]; output: var Tensor[T])
• nnp_linear: proc linear[T](input, weight: Tensor[T]; bias: Tensor[T]; output: var Tensor[T])

linear_backward:

• nnp_linear: proc linear_backward[T](input, weight, gradOutput: Tensor[T]; gradInput, gradWeight: var Tensor[T])
• nnp_linear: proc linear_backward[T](input, weight, gradOutput: Tensor[T]; gradInput, gradWeight, gradBias: var Tensor[T])

LinearGate:

• linear: type LinearGate

linspace:

• init_cpu: proc linspace[T: SomeNumber](start, stop: T; num: int; endpoint = true): Tensor[float]

ln:

• ufunc: proc ln[T](t`gensym5: Tensor[T]): Tensor[T]
• ufunc: proc ln[T](t`gensym5: Tensor[T]): Tensor[T]

ln1p:

• math_ops_fusion: proc ln1p(x: float32): float32
• math_ops_fusion: proc ln1p(x: float64): float64

load_imdb:

• imdb: proc load_imdb(cache: static bool = true): Imdb

load_mnist:

• mnist: proc load_mnist(cache: static bool = true; fashion_mnist: static bool = false): Mnist

log10:

• ufunc: proc log10[T](t`gensym6: Tensor[T]): Tensor[T]
• ufunc: proc log10[T](t`gensym6: Tensor[T]): Tensor[T]

log2:

• ufunc: proc log2[T](t`gensym7: Tensor[T]): Tensor[T]
• ufunc: proc log2[T](t`gensym7: Tensor[T]): Tensor[T]

logspace:

• init_cpu: proc logspace[T: SomeNumber](start, stop: T; num: int; base = 10.0; endpoint = true): Tensor[ float]

logsumexp:

• p_logsumexp: proc logsumexp[T: SomeFloat](t: Tensor[T]): T

low:

• dynamic_stack_arrays: proc low(a: DynamicStackArray): int

lu_permuted:

• decomposition: proc lu_permuted[T: SupportedDecomposition](a: Tensor[T]): tuple[PL, U: Tensor[T]]

m128:

• simd: object m128

m128d:

• simd: object m128d

m128i:

• simd: object m128i

m256:

• simd: object m256

m256d:

• simd: object m256d

m256i:

• simd: object m256i

m512:

• simd: object m512

m512d:

• simd: object m512d

m512i:

• simd: object m512i

mabs:

• math_functions: proc mabs[T](t: var Tensor[T])

makeKernel:

• kde: template makeKernel(fn: untyped): untyped

makeUniversal:

• ufunc: template makeUniversal(func_name: untyped; docSuffix = "")

makeUniversalLocal:

• ufunc: template makeUniversalLocal(func_name: untyped)

Manhattan:

• distances: object Manhattan

map:

• higher_order_applymap: proc map[T; U](t: Tensor[T]; f: T -> U): Tensor[U]

map2:

• higher_order_applymap: proc map2[T, U; V: KnownSupportsCopyMem](t1: Tensor[T]; f: (T, U) -> V; t2: Tensor[U]): Tensor[ V]
• higher_order_applymap: proc map2[T, U; V: not KnownSupportsCopyMem](t1: Tensor[T]; f: (T, U) -> V; t2: Tensor[U]): Tensor[V]

map2_inline:

• higher_order_applymap: template map2_inline[T, U](t1: Tensor[T]; t2: Tensor[U]; op: untyped): untyped

map3_inline:

• higher_order_applymap: template map3_inline[T, U, V](t1: Tensor[T]; t2: Tensor[U]; t3: Tensor[V]; op: untyped): untyped

map_inline:

• higher_order_applymap: template map_inline[T](t: Tensor[T]; op: untyped): untyped

masked_axis_fill:

• selectors: proc masked_axis_fill[T](t: var Tensor[T]; mask: openArray[bool]; axis: int; value: T or Tensor[T])
• selectors: proc masked_axis_fill[T](t: var Tensor[T]; mask: Tensor[bool]; axis: int; value: T or Tensor[T])

masked_axis_select:

• selectors: proc masked_axis_select[T](t: Tensor[T]; mask: openArray[bool]; axis: int): Tensor[T]
• selectors: proc masked_axis_select[T](t: Tensor[T]; mask: Tensor[bool]; axis: int): Tensor[T]

masked_fill:

• selectors: proc masked_fill[T](t: var Tensor[T]; mask: openArray; value: openArray[T])
• selectors: proc masked_fill[T](t: var Tensor[T]; mask: openArray; value: T)
• selectors: proc masked_fill[T](t: var Tensor[T]; mask: openArray; value: Tensor[T])
• selectors: proc masked_fill[T](t: var Tensor[T]; mask: Tensor[bool]; value: openArray[T])
• selectors: proc masked_fill[T](t: var Tensor[T]; mask: Tensor[bool]; value: T)
• selectors: proc masked_fill[T](t: var Tensor[T]; mask: Tensor[bool]; value: Tensor[T])

masked_fill_along_axis:

• selectors: proc masked_fill_along_axis[T](t: var Tensor[T]; mask: Tensor[bool]; axis: int; value: T)

masked_select:

• selectors: proc masked_select[T](t: Tensor[T]; mask: openArray): Tensor[T]
• selectors: proc masked_select[T](t: Tensor[T]; mask: Tensor[bool]): Tensor[T]

MatMulGate:

• gates_blas: type MatMulGate

MatrixKind:

• linear_systems: enum MatrixKind

MatrixView:

• gemm_utils: object MatrixView

max:

• aggregate: proc max[T](arg: Tensor[T]): T
• aggregate: proc max[T](arg: Tensor[T]; axis: int): Tensor[T]
• dynamic_stack_arrays: proc max[T](a: DynamicStackArray[T]): T
• math_functions: proc max[T: SomeNumber](t1, t2: Tensor[T]): Tensor[T]
• math_functions: proc max[T: SomeNumber](args: varargs[Tensor[T]]): Tensor[T]

MaxPool2D:

• maxpool2D: object MaxPool2D

maxpool2d:

• maxpool2D: proc maxpool2d[TT](input: Variable[TT]; kernel: Size2D; padding: Size2D = (0, 0); stride: Size2D = (1, 1)): Variable[TT]
• nnp_maxpooling: proc maxpool2d[T](input: Tensor[T]; kernel: Size2D; padding: Size2D = (0, 0); stride: Size2D = (1, 1)): tuple[max_indices: Tensor[int], maxpooled: Tensor[T]]

maxpool2d_backward:

• nnp_maxpooling: proc maxpool2d_backward[T](cached_input_shape: openArray[int] | Metadata; cached_max_indices: Tensor[int]; gradOutput: Tensor[T]): Tensor[ T]

MaxPool2DGate:

• maxpool2D: type MaxPool2DGate

MAXRANK:

• global_config: const MAXRANK

mclamp:

• math_functions: proc mclamp[T](t: var Tensor[T]; min, max: T)

mcopySign:

• math_functions: proc mcopySign[T: SomeFloat](t1: var Tensor[T]; t2: Tensor[T])

mean:

• aggregate: proc mean[T: Complex[float32] or Complex[float64]](arg: Tensor[T]): T
• aggregate: proc mean[T: Complex[float32] or Complex[float64]](arg: Tensor[T]; axis: int): Tensor[ T]
• aggregate: proc mean[T: SomeFloat](arg: Tensor[T]): T
• aggregate: proc mean[T: SomeFloat](arg: Tensor[T]; axis: int): Tensor[T]
• aggregate: proc mean[T: SomeInteger](arg: Tensor[T]): T
• aggregate: proc mean[T: SomeInteger](arg: Tensor[T]; axis: int): Tensor[T]
• gates_reduce: proc mean[TT](a: Variable[TT]): Variable[TT]
• gates_reduce: proc mean[TT](a: Variable[TT]; axis: Natural): Variable[TT]

mean_absolute_error:

• common_error_functions: proc mean_absolute_error[T: SomeFloat](y, y_true: Tensor[T] | Tensor[Complex[T]]): T

MeanGate:

• gates_reduce: type MeanGate

mean_relative_error:

• common_error_functions: proc mean_relative_error[T: SomeFloat](y, y_true: Tensor[T] | Tensor[Complex[T]]): T

mean_squared_error:

• common_error_functions: proc mean_squared_error[T: SomeFloat](y, y_true: Tensor[T] | Tensor[Complex[T]]): T

median:

• aggregate: proc median[T](arg: Tensor[T]; isSorted = false): float

melwise_div:

• math_functions: proc melwise_div[T: SomeFloat](a: var Tensor[T]; b: Tensor[T])
• math_functions: proc melwise_div[T: SomeInteger](a: var Tensor[T]; b: Tensor[T])

melwise_mul:

• math_functions: proc melwise_mul[T](a: var Tensor[T]; b: Tensor[T])

menumerate:

• accessors: iterator menumerate[T](t: Tensor[T]): (int, var T)
• accessors: iterator menumerate[T](t: Tensor[T]; offset, size: int): (int, var T)

menumerateZip:

• accessors: iterator menumerateZip[T, U](t1: var Tensor[T]; t2: Tensor[U]): (int, var T, U)
• accessors: iterator menumerateZip[T, U](t1: var Tensor[T]; t2: Tensor[U]; offset, size: int): (int, var T, U)

meshgrid:

• special_matrices: proc meshgrid[T](t_list: varargs[Tensor[T]]; indexing = MeshGridIndexing.xygrid): seq[ Tensor[T]]

MeshGridIndexing:

• special_matrices: enum MeshGridIndexing

Metadata:

• datatypes: type Metadata

MetadataArray:

• datatypes: type MetadataArray

MicroKernel:

• gemm_tiling: object MicroKernel

min:

• aggregate: proc min[T](arg: Tensor[T]): T
• aggregate: proc min[T](arg: Tensor[T]; axis: int): Tensor[T]
• math_functions: proc min[T: SomeNumber](t1, t2: Tensor[T]): Tensor[T]
• math_functions: proc min[T: SomeNumber](args: varargs[Tensor[T]]): Tensor[T]

Minkowski:

• distances: object Minkowski

mitems:

• accessors: iterator mitems[T](t: var Tensor[T]): var T
• accessors: iterator mitems[T](t: var Tensor[T]; offset, size: int): var T
• dynamic_stack_arrays: iterator mitems[T](a: var DynamicStackArray[T]): var T

mkGenBand:

• linear_systems: MatrixKind.mkGenBand

mkGeneral:

• linear_systems: MatrixKind.mkGeneral

mkGenTriDiag:

• linear_systems: MatrixKind.mkGenTriDiag

mkPosDef:

• linear_systems: MatrixKind.mkPosDef

mkPosDefBand:

• linear_systems: MatrixKind.mkPosDefBand

mkPosDefTriDiag:

• linear_systems: MatrixKind.mkPosDefTriDiag

mkSymmetric:

• linear_systems: MatrixKind.mkSymmetric

mm256_add_epi16:

• simd: proc mm256_add_epi16(a, b: m256i): m256i

mm256_add_epi32:

• simd: proc mm256_add_epi32(a, b: m256i): m256i

mm256_add_epi64:

• simd: proc mm256_add_epi64(a, b: m256i): m256i

mm256_add_epi8:

• simd: proc mm256_add_epi8(a, b: m256i): m256i

mm256_add_pd:

• simd: proc mm256_add_pd(a, b: m256d): m256d

mm256_add_ps:

• simd: proc mm256_add_ps(a, b: m256): m256

mm256_and_ps:

• simd: proc mm256_and_ps(a, b: m256): m256

mm256_and_si256:

• simd: proc mm256_and_si256(a, b: m256i): m256i

mm256_castps256_ps128:

• simd: proc mm256_castps256_ps128(a: m256): m128

mm256_castps_si256:

• simd: proc mm256_castps_si256(a: m256): m256i

mm256_castsi256_ps:

• simd: proc mm256_castsi256_ps(a: m256i): m256

mm256_cmpgt_epi32:

• simd: proc mm256_cmpgt_epi32(a, b: m256i): m256i

mm256_cvtepi32_ps:

• simd: proc mm256_cvtepi32_ps(a: m256i): m256

mm256_cvtps_epi32:

• simd: proc mm256_cvtps_epi32(a: m256): m256i

mm256_extractf128_ps:

• simd: proc mm256_extractf128_ps(v: m256; m: cint{lit}): m128

mm256_fmadd_pd:

• simd: proc mm256_fmadd_pd(a, b, c: m256d): m256d

mm256_fmadd_ps:

• simd: proc mm256_fmadd_ps(a, b, c: m256): m256

mm256_i32gather_epi32:

• simd: proc mm256_i32gather_epi32(m: ptr (uint32 or int32); i: m256i; s: int32): m256i

mm256_load_pd:

• simd: proc mm256_load_pd(aligned_mem_addr: ptr float64): m256d

mm256_load_ps:

• simd: proc mm256_load_ps(aligned_mem_addr: ptr float32): m256

mm256_load_si256:

• simd: proc mm256_load_si256(mem_addr: ptr m256i): m256i

mm256_loadu_pd:

• simd: proc mm256_loadu_pd(mem_addr: ptr float64): m256d

mm256_loadu_ps:

• simd: proc mm256_loadu_ps(mem_addr: ptr float32): m256

mm256_loadu_si256:

• simd: proc mm256_loadu_si256(mem_addr: ptr m256i): m256i

mm256_max_ps:

• simd: proc mm256_max_ps(a, b: m256): m256

mm256_min_ps:

• simd: proc mm256_min_ps(a, b: m256): m256

mm256_movemask_epi8:

• simd: proc mm256_movemask_epi8(a: m256i): int32

mm256_mul_epu32:

• simd: proc mm256_mul_epu32(a: m256i; b: m256i): m256i

mm256_mullo_epi16:

• simd: proc mm256_mullo_epi16(a, b: m256i): m256i

mm256_mullo_epi32:

• simd: proc mm256_mullo_epi32(a, b: m256i): m256i

mm256_mul_pd:

• simd: proc mm256_mul_pd(a, b: m256d): m256d

mm256_mul_ps:

• simd: proc mm256_mul_ps(a, b: m256): m256

mm256_or_ps:

• simd: proc mm256_or_ps(a, b: m256): m256

mm256_set1_epi16:

• simd: proc mm256_set1_epi16(a: int16 or uint16): m256i

mm256_set1_epi32:

• simd: proc mm256_set1_epi32(a: int32 or uint32): m256i

mm256_set1_epi64x:

• simd: proc mm256_set1_epi64x(a: int64 or uint64): m256i

mm256_set1_epi8:

• simd: proc mm256_set1_epi8(a: int8 or uint8): m256i

mm256_set1_pd:

• simd: proc mm256_set1_pd(a: float64): m256d

mm256_set1_ps:

• simd: proc mm256_set1_ps(a: float32): m256

mm256_setzero_pd:

• simd: proc mm256_setzero_pd(): m256d

mm256_setzero_ps:

• simd: proc mm256_setzero_ps(): m256

mm256_setzero_si256:

• simd: proc mm256_setzero_si256(): m256i

mm256_shuffle_epi32:

• simd: proc mm256_shuffle_epi32(a: m256i; imm8: cint): m256i

mm256_slli_epi32:

• simd: proc mm256_slli_epi32(a: m256i; count: int32): m256i

mm256_srli_epi32:

• simd: proc mm256_srli_epi32(a: m256i; count: int32): m256i

mm256_srli_epi64:

• simd: proc mm256_srli_epi64(a: m256i; imm8: cint): m256i

mm256_store_pd:

• simd: proc mm256_store_pd(mem_addr: ptr float64; a: m256d)

mm256_store_ps:

• simd: proc mm256_store_ps(mem_addr: ptr float32; a: m256)

mm256_storeu_pd:

• simd: proc mm256_storeu_pd(mem_addr: ptr float64; a: m256d)

mm256_storeu_ps:

• simd: proc mm256_storeu_ps(mem_addr: ptr float32; a: m256)

mm256_storeu_si256:

• simd: proc mm256_storeu_si256(mem_addr: ptr m256i; a: m256i)

mm256_sub_ps:

• simd: proc mm256_sub_ps(a, b: m256): m256

mm512_add_epi16:

• simd: proc mm512_add_epi16(a, b: m512i): m512i

mm512_add_epi32:

• simd: proc mm512_add_epi32(a, b: m512i): m512i

mm512_add_epi64:

• simd: proc mm512_add_epi64(a, b: m512i): m512i

mm512_add_epi8:

• simd: proc mm512_add_epi8(a, b: m512i): m512i

mm512_add_pd:

• simd: proc mm512_add_pd(a, b: m512d): m512d

mm512_add_ps:

• simd: proc mm512_add_ps(a, b: m512): m512

mm512_and_si512:

• simd: proc mm512_and_si512(a, b: m512i): m512i

mm512_castps_si512:

• simd: proc mm512_castps_si512(a: m512): m512i

mm512_castsi512_ps:

• simd: proc mm512_castsi512_ps(a: m512i): m512

mm512_cmpgt_epi32_mask:

• simd: proc mm512_cmpgt_epi32_mask(a, b: m512i): mmask16

mm512_cvtepi32_ps:

• simd: proc mm512_cvtepi32_ps(a: m512i): m512

mm512_cvtps_epi32:

• simd: proc mm512_cvtps_epi32(a: m512): m512i

mm512_fmadd_pd:

• simd: proc mm512_fmadd_pd(a, b, c: m512d): m512d

mm512_fmadd_ps:

• simd: proc mm512_fmadd_ps(a, b, c: m512): m512

mm512_i32gather_epi32:

• simd: proc mm512_i32gather_epi32(i: m512i; m: ptr (uint32 or int32); s: int32): m512i

mm512_load_pd:

• simd: proc mm512_load_pd(aligned_mem_addr: ptr float64): m512d

mm512_load_ps:

• simd: proc mm512_load_ps(aligned_mem_addr: ptr float32): m512

mm512_load_si512:

• simd: proc mm512_load_si512(mem_addr: ptr SomeInteger): m512i

mm512_loadu_pd:

• simd: proc mm512_loadu_pd(mem_addr: ptr float64): m512d

mm512_loadu_ps:

• simd: proc mm512_loadu_ps(mem_addr: ptr float32): m512

mm512_loadu_si512:

• simd: proc mm512_loadu_si512(mem_addr: ptr SomeInteger): m512i

mm512_maskz_set1_epi32:

• simd: proc mm512_maskz_set1_epi32(k: mmask16; a: cint): m512i

mm512_max_ps:

• simd: proc mm512_max_ps(a, b: m512): m512

mm512_min_ps:

• simd: proc mm512_min_ps(a, b: m512): m512

mm512_movepi8_mask:

• simd: proc mm512_movepi8_mask(a: m512i): mmask64

mm512_movm_epi32:

• simd: proc mm512_movm_epi32(a: mmask16): m512i

mm512_mullo_epi32:

• simd: proc mm512_mullo_epi32(a, b: m512i): m512i

mm512_mullo_epi64:

• simd: proc mm512_mullo_epi64(a, b: m512i): m512i

mm512_mul_pd:

• simd: proc mm512_mul_pd(a, b: m512d): m512d

mm512_mul_ps:

• simd: proc mm512_mul_ps(a, b: m512): m512

mm512_or_ps:

• simd: proc mm512_or_ps(a, b: m512): m512

mm512_set1_epi16:

• simd: proc mm512_set1_epi16(a: int16 or uint16): m512i

mm512_set1_epi32:

• simd: proc mm512_set1_epi32(a: int32 or uint32): m512i

mm512_set1_epi64:

• simd: proc mm512_set1_epi64(a: int64 or uint64): m512i

mm512_set1_epi8:

• simd: proc mm512_set1_epi8(a: int8 or uint8): m512i

mm512_set1_pd:

• simd: proc mm512_set1_pd(a: float64): m512d

mm512_set1_ps:

• simd: proc mm512_set1_ps(a: float32): m512

mm512_setzero_pd:

• simd: proc mm512_setzero_pd(): m512d

mm512_setzero_ps:

• simd: proc mm512_setzero_ps(): m512

mm512_setzero_si512:

• simd: proc mm512_setzero_si512(): m512i

mm512_slli_epi32:

• simd: proc mm512_slli_epi32(a: m512i; count: int32): m512i

mm512_srli_epi32:

• simd: proc mm512_srli_epi32(a: m512i; count: int32): m512i

mm512_store_pd:

• simd: proc mm512_store_pd(mem_addr: ptr float64; a: m512d)

mm512_store_ps:

• simd: proc mm512_store_ps(mem_addr: ptr float32; a: m512)

mm512_storeu_pd:

• simd: proc mm512_storeu_pd(mem_addr: ptr float64; a: m512d)

mm512_storeu_ps:

• simd: proc mm512_storeu_ps(mem_addr: ptr float32; a: m512)

mm512_storeu_si512:

• simd: proc mm512_storeu_si512(mem_addr: ptr SomeInteger; a: m512i)

mm512_sub_ps:

• simd: proc mm512_sub_ps(a, b: m512): m512

mm_add_epi16:

• simd: proc mm_add_epi16(a, b: m128i): m128i

mm_add_epi32:

• simd: proc mm_add_epi32(a, b: m128i): m128i

mm_add_epi64:

• simd: proc mm_add_epi64(a, b: m128i): m128i

mm_add_epi8:

• simd: proc mm_add_epi8(a, b: m128i): m128i

mm_add_pd:

• simd: proc mm_add_pd(a, b: m128d): m128d

mm_add_ps:

• simd: proc mm_add_ps(a, b: m128): m128

mm_add_ss:

• simd: proc mm_add_ss(a, b: m128): m128

mm_and_si128:

• simd: proc mm_and_si128(a, b: m128i): m128i

mmask16:

• simd: type mmask16

mmask64:

• simd: type mmask64

mmax:

• math_functions: proc mmax[T: SomeNumber](t1: var Tensor[T]; t2: Tensor[T])
• math_functions: proc mmax[T: SomeNumber](t1: var Tensor[T]; args: varargs[Tensor[T]])

mm_castps_si128:

• simd: proc mm_castps_si128(a: m128): m128i

mm_castsi128_ps:

• simd: proc mm_castsi128_ps(a: m128i): m128

mm_cmpgt_epi32:

• simd: proc mm_cmpgt_epi32(a, b: m128i): m128i

mm_cvtepi32_ps:

• simd: proc mm_cvtepi32_ps(a: m128i): m128

mm_cvtps_epi32:

• simd: proc mm_cvtps_epi32(a: m128): m128i

mm_cvtsi128_si32:

• simd: proc mm_cvtsi128_si32(a: m128i): cint

mm_cvtss_f32:

• simd: proc mm_cvtss_f32(a: m128): float32

mm_extract_epi16:

• simd: proc mm_extract_epi16(a: m128i; imm8: cint): cint

mm_i32gather_epi32:

• simd: proc mm_i32gather_epi32(m: ptr (uint32 or int32); i: m128i; s: int32): m128i

mmin:

• math_functions: proc mmin[T: SomeNumber](t1: var Tensor[T]; t2: Tensor[T])
• math_functions: proc mmin[T: SomeNumber](t1: var Tensor[T]; args: varargs[Tensor[T]])

mm_load_pd:

• simd: proc mm_load_pd(aligned_mem_addr: ptr float64): m128d

mm_load_ps:

• simd: proc mm_load_ps(aligned_mem_addr: ptr float32): m128

mm_load_si128:

• simd: proc mm_load_si128(mem_addr: ptr m128i): m128i

mm_load_ss:

• simd: proc mm_load_ss(aligned_mem_addr: ptr float32): m128

mm_loadu_pd:

• simd: proc mm_loadu_pd(mem_addr: ptr float64): m128d

mm_loadu_ps:

• simd: proc mm_loadu_ps(data: ptr float32): m128

mm_loadu_si128:

• simd: proc mm_loadu_si128(mem_addr: ptr m128i): m128i

mm_max_ps:

• simd: proc mm_max_ps(a, b: m128): m128

mm_max_ss:

• simd: proc mm_max_ss(a, b: m128): m128

mm_min_ps:

• simd: proc mm_min_ps(a, b: m128): m128

mm_min_ss:

• simd: proc mm_min_ss(a, b: m128): m128

mm_movehdup_ps:

• simd: proc mm_movehdup_ps(a: m128): m128

mm_movehl_ps:

• simd: proc mm_movehl_ps(a, b: m128): m128

mm_moveldup_ps:

• simd: proc mm_moveldup_ps(a: m128): m128

mm_movelh_ps:

• simd: proc mm_movelh_ps(a, b: m128): m128

mm_movemask_epi8:

• simd: proc mm_movemask_epi8(a: m128i): int32

mm_mul_epu32:

• simd: proc mm_mul_epu32(a: m128i; b: m128i): m128i

mm_mullo_epi16:

• simd: proc mm_mullo_epi16(a, b: m128i): m128i

mm_mullo_epi32:

• simd: proc mm_mullo_epi32(a, b: m128i): m128i

mm_mul_pd:

• simd: proc mm_mul_pd(a, b: m128d): m128d

mm_mul_ps:

• simd: proc mm_mul_ps(a, b: m128): m128

mm_or_ps:

• simd: proc mm_or_ps(a, b: m128): m128

mm_or_si128:

• simd: proc mm_or_si128(a, b: m128i): m128i

mm_set1_epi16:

• simd: proc mm_set1_epi16(a: int16 or uint16): m128i

mm_set1_epi32:

• simd: proc mm_set1_epi32(a: int32 or uint32): m128i

mm_set1_epi64x:

• simd: proc mm_set1_epi64x(a: int64 or uint64): m128i

mm_set1_epi8:

• simd: proc mm_set1_epi8(a: int8 or uint8): m128i

mm_set1_pd:

• simd: proc mm_set1_pd(a: float64): m128d

mm_set1_ps:

• simd: proc mm_set1_ps(a: float32): m128

mm_set_epi32:

• simd: proc mm_set_epi32(e3, e2, e1, e0: cint): m128i

mm_setzero_pd:

• simd: proc mm_setzero_pd(): m128d

mm_setzero_ps:

• simd: proc mm_setzero_ps(): m128

mm_setzero_si128:

• simd: proc mm_setzero_si128(): m128i

mm_shuffle_epi32:

• simd: proc mm_shuffle_epi32(a: m128i; imm8: cint): m128i

mm_slli_epi32:

• simd: proc mm_slli_epi32(a: m128i; count: int32): m128i

mm_slli_epi64:

• simd: proc mm_slli_epi64(a: m128i; imm8: cint): m128i

mm_srli_epi32:

• simd: proc mm_srli_epi32(a: m128i; count: int32): m128i

mm_srli_epi64:

• simd: proc mm_srli_epi64(a: m128i; imm8: cint): m128i

mm_store_pd:

• simd: proc mm_store_pd(mem_addr: ptr float64; a: m128d)

mm_store_ps:

• simd: proc mm_store_ps(mem_addr: ptr float32; a: m128)

mm_storeu_pd:

• simd: proc mm_storeu_pd(mem_addr: ptr float64; a: m128d)

mm_storeu_ps:

• simd: proc mm_storeu_ps(mem_addr: ptr float32; a: m128)

mm_storeu_si128:

• simd: proc mm_storeu_si128(mem_addr: ptr m128i; a: m128i)

mm_sub_pd:

• simd: proc mm_sub_pd(a, b: m128d): m128d

mm_sub_ps:

• simd: proc mm_sub_ps(a, b: m128): m128

mnegate:

• math_functions: proc mnegate[T: SomeSignedInt | SomeFloat](t: var Tensor[T])

moveaxis:

• shapeshifting: proc moveaxis(t: Tensor; initial: Natural; target: Natural): Tensor

mpairs:

• accessors: iterator mpairs[T](t: var Tensor[T]): (seq[int], var T)
• dynamic_stack_arrays: iterator mpairs[T](a: var DynamicStackArray[T]): (int, var T)

mreciprocal:

• math_functions: proc mreciprocal[T: Complex[float32] or Complex[float64]](t: var Tensor[T])
• math_functions: proc mreciprocal[T: SomeFloat](t: var Tensor[T])

mrelu:

• nnp_activation: proc mrelu[T](t: var Tensor[T])

MSELoss:

• mean_square_error_loss: type MSELoss

mse_loss:

• mean_square_error_loss: proc mse_loss[TT](input: Variable[TT]; target: TT): Variable[TT]

msigmoid:

• nnp_activation: proc msigmoid[T: SomeFloat](t: var Tensor[T])

mtanh:

• nnp_activation: proc mtanh[T: SomeFloat](t: var Tensor[T])

mzip:

• accessors: iterator mzip[T, U](t1: var Tensor[T]; t2: Tensor[U]): (var T, U)
• accessors: iterator mzip[T, U](t1: var Tensor[T]; t2: Tensor[U]; offset, size: int): (var T, U)
• accessors: iterator mzip[T, U, V](t1: var Tensor[T]; t2: Tensor[U]; t3: Tensor[V]): (var T, U, V)
• accessors: iterator mzip[T, U, V](t1: var Tensor[T]; t2: Tensor[U]; t3: Tensor[V]; offset, size: int): ( var T, U, V)

naive_gemv_fallback:

• naive_l2_gemv: proc naive_gemv_fallback[T: SomeInteger](alpha: T; A: Tensor[T]; x: Tensor[T]; beta: T; y: var Tensor[T])

nchw_channels:

• p_nnp_types: proc nchw_channels[T](input: Tensor[T]): int

nchw_height:

• p_nnp_types: proc nchw_height[T](input: Tensor[T]): int

nchw_width:

• p_nnp_types: proc nchw_width[T](input: Tensor[T]): int

nearestNeighbors:

• neighbors: proc nearestNeighbors[T](X: Tensor[T]; eps: float; metric: typedesc[AnyMetric]; p = 2.0; useNaiveNearestNeighbor: static bool = false): seq[ Tensor[int]]

negate:

• math_functions: proc negate[T: SomeSignedInt | SomeFloat](t: Tensor[T]): Tensor[T]

network:

• nn_dsl: macro network(modelName: untyped; config: untyped): untyped

newClStorage:

• opencl_backend: proc newClStorage[T: SomeFloat](length: int): ClStorage[T]

newClTensor:

• p_init_opencl: proc newClTensor[T: SomeFloat](shape: Metadata; layout: OrderType = rowMajor): ClTensor[ T]
• p_init_opencl: proc newClTensor[T: SomeFloat](shape: varargs[int]; layout: OrderType = rowMajor): ClTensor[ T]

newContext:

• autograd_common: proc newContext(TT: typedesc): Context[TT]

newConv2dDesc:

• cudnn_conv_interface: proc newConv2dDesc[T: SomeFloat](padding, strides, dilation: SizeHW): cudnnConvolutionDescriptor_t

newConvAlgoSpace:

• cudnn_conv_interface: proc newConvAlgoSpace[T: SomeFloat](srcTensorDesc: cudnnTensorDescriptor_t; kernelDesc: cudnnFilterDescriptor_t; convDesc: cudnnConvolutionDescriptor_t; dstTensorDesc: cudnnTensorDescriptor_t): ConvAlgoSpace[ T, cudnnConvolutionFwdAlgo_t]

newCudaStorage:

• cuda: proc newCudaStorage[T: SomeFloat](length: int): CudaStorage[T]

newCudaTensor:

• p_init_cuda: proc newCudaTensor[T: SomeFloat](shape: Metadata; layout: OrderType = colMajor): CudaTensor[ T]
• p_init_cuda: proc newCudaTensor[T: SomeFloat](shape: varargs[int]; layout: OrderType = colMajor): CudaTensor[ T]

newCudnn4DTensorDesc:

• cudnn: proc newCudnn4DTensorDesc[T: SomeFloat](t: CudaTensor[T]): cudnnTensorDescriptor_t

newCudnnConvKernelDesc:

• cudnn_conv_interface: proc newCudnnConvKernelDesc[T: SomeFloat](convKernel: CudaTensor[T]): cudnnFilterDescriptor_t

newDiffs:

• autograd_common: proc newDiffs[TT](num: Natural): SmallDiffs[TT]

newMatrixUninitColMajor:

• init_colmajor: proc newMatrixUninitColMajor[T](M: var Tensor[T]; rows, cols: int)

newParents:

• autograd_common: proc newParents[TT](num: Natural): Parents[TT]

newSeqUninit:

• sequninit: proc newSeqUninit[T](len: Natural): seq[T]

newSGD:

• optimizers: proc newSGD[T](params: varargs[Variable[Tensor[T]]]; learning_rate: T): SGD[Tensor[T]]

newTensor:

• initialization: proc newTensor[T](shape: Metadata): Tensor[T]
• initialization: proc newTensor[T](shape: varargs[int] = [0]): Tensor[T]

newTensorUninit:

• init_cpu: proc newTensorUninit[T](size: int): Tensor[T]
• init_cpu: proc newTensorUninit[T](shape: Metadata): Tensor[T]
• init_cpu: proc newTensorUninit[T](shape: varargs[int] = [0]): Tensor[T]

newTensorWith:

• init_cpu: proc newTensorWith[T](shape: Metadata; value: T): Tensor[T]
• init_cpu: proc newTensorWith[T](shape: varargs[int]; value: T): Tensor[T]

newTiles:

• gemm_tiling: proc newTiles(ukernel: static MicroKernel; T: typedesc; M, N, K: Natural): Tiles[T]

NNPackAuto:

• nnp_convolution: Conv2DAlgorithm.NNPackAuto

nnpack_conv2d:

• nnpack_interface: proc nnpack_conv2d(input, weight, bias: Tensor[float32]; padding, stride: Size2D): Tensor[ float32]

nnpack_conv2d_gradient:

• nnpack_interface: proc nnpack_conv2d_gradient[T](input, weight: Tensor[float32]; padding, stride: Size2D; grad_output: Tensor[T]; grad_input, grad_weight: var Tensor[T])

nnp_activation:

• nnpack: enum nnp_activation

nnp_convolution_algorithm:

• nnpack: enum nnp_convolution_algorithm

nnp_convolution_inference:

• nnpack: proc nnp_convolution_inference(algorithm: nnp_convolution_algorithm; transform_strategy: nnp_convolution_transform_strategy; input_channels: csize_t; output_channels: csize_t; input_size: nnp_size; input_padding: nnp_padding; kernel_size: nnp_size; output_subsampling: nnp_size; input: ptr cfloat; kernel: ptr cfloat; bias: ptr cfloat; output: ptr cfloat; workspace_buffer: pointer; workspace_size: ptr csize_t; activation: nnp_activation; activation_parameters: pointer; threadpool: pthreadpool_t; profile: ptr nnp_profile): nnp_status

nnp_convolution_input_gradient:

• nnpack: proc nnp_convolution_input_gradient(algorithm: nnp_convolution_algorithm; batch_size: csize_t; input_channels: csize_t; output_channels: csize_t; input_size: nnp_size; input_padding: nnp_padding; kernel_size: nnp_size; grad_output: ptr cfloat; kernel: ptr cfloat; grad_input: ptr cfloat; workspace_buffer: pointer = nil; workspace_size: ptr csize_t = nil; activation: nnp_activation = nnp_activation_identity; activation_parameters: pointer = nil; threadpool: pthreadpool_t = nil; profile: ptr nnp_profile = nil): nnp_status

nnp_convolution_kernel_gradient:

• nnpack: proc nnp_convolution_kernel_gradient(algorithm: nnp_convolution_algorithm; batch_size: csize_t; input_channels: csize_t; output_channels: csize_t; input_size: nnp_size; input_padding: nnp_padding; kernel_size: nnp_size; input: ptr cfloat; grad_output: ptr cfloat; grad_kernel: ptr cfloat; workspace_buffer: pointer = nil; workspace_size: ptr csize_t = nil; activation: nnp_activation = nnp_activation_identity; activation_parameters: pointer = nil; threadpool: pthreadpool_t = nil; profile: ptr nnp_profile = nil): nnp_status

nnp_convolution_output:

• nnpack: proc nnp_convolution_output(algorithm: nnp_convolution_algorithm; batch_size: csize_t; input_channels: csize_t; output_channels: csize_t; input_size: nnp_size; input_padding: nnp_padding; kernel_size: nnp_size; input: ptr cfloat; kernel: ptr cfloat; bias: ptr cfloat; output: ptr cfloat; workspace_buffer: pointer = nil; workspace_size: ptr csize_t = nil; activation: nnp_activation = nnp_activation_identity; activation_parameters: pointer = nil; threadpool: pthreadpool_t = nil; profile: ptr nnp_profile = nil): nnp_status

nnp_convolution_transform_strategy:

• nnpack: enum nnp_convolution_transform_strategy

nnp_convolution_transform_strategy_block_based:

• nnpack: const nnp_convolution_transform_strategy_block_based

nnp_convolution_transform_strategy_tuple_based:

• nnpack: const nnp_convolution_transform_strategy_tuple_based

nnp_deinitialize:

• nnpack: proc nnp_deinitialize(): nnp_status

nnp_fully_connected_inference:

• nnpack: proc nnp_fully_connected_inference(input_channels: csize_t; output_channels: csize_t; input: ptr cfloat; kernel: ptr cfloat; output: ptr cfloat; threadpool: pthreadpool_t): nnp_status

nnp_fully_connected_inference_f16f32:

• nnpack: proc nnp_fully_connected_inference_f16f32(input_channels: csize_t; output_channels: csize_t; input: ptr cfloat; kernel: pointer; output: ptr cfloat; threadpool: pthreadpool_t): nnp_status

nnp_fully_connected_output:

• nnpack: proc nnp_fully_connected_output(batch_size: csize_t; input_channels: csize_t; output_channels: csize_t; input: ptr cfloat; kernel: ptr cfloat; output: ptr cfloat; threadpool: pthreadpool_t; profile: ptr nnp_profile): nnp_status

nnp_initialize:

• nnpack: proc nnp_initialize(): nnp_status

nnp_max_pooling_output:

• nnpack: proc nnp_max_pooling_output(batch_size: csize_t; channels: csize_t; input_size: nnp_size; input_padding: nnp_padding; pooling_size: nnp_size; pooling_stride: nnp_size; input: ptr cfloat; output: ptr cfloat; threadpool: pthreadpool_t): nnp_status

nnp_padding:

• nnpack: object nnp_padding

nnp_profile:

• nnpack: object nnp_profile

nnp_relu_input_gradient:

• nnpack: proc nnp_relu_input_gradient(batch_size: csize_t; channels: csize_t; grad_output: ptr cfloat; input: ptr cfloat; grad_input: ptr cfloat; negative_slope: cfloat; threadpool: pthreadpool_t): nnp_status

nnp_relu_output:

• nnpack: proc nnp_relu_output(batch_size: csize_t; channels: csize_t; input: ptr cfloat; output: ptr cfloat; negative_slope: cfloat; threadpool: pthreadpool_t): nnp_status

nnp_size:

• nnpack: object nnp_size

nnp_softmax_output:

• nnpack: proc nnp_softmax_output(batch_size: csize_t; channels: csize_t; input: ptr cfloat; output: ptr cfloat; threadpool: pthreadpool_t): nnp_status

nnp_status:

• nnpack: enum nnp_status

nnp_status_invalid_activation:

• nnpack: const nnp_status_invalid_activation

nnp_status_invalid_activation_parameters:

• nnpack: const nnp_status_invalid_activation_parameters

nnp_status_invalid_output_subsampling:

• nnpack: const nnp_status_invalid_output_subsampling

Node:

• kdtree: type Node

no_grad_mode:

• autograd_common: template no_grad_mode(ctx: Context; body: untyped): untyped

nonzero:

• aggregate: proc nonzero[T](arg: Tensor[T]): Tensor[int]

numberOne:

• p_complex: template numberOne(T: type Complex[float32]): Complex[float32]
• p_complex: template numberOne(T: type Complex[float64]): Complex[float64]
• p_complex: template numberOne(T: type SomeNumber): SomeNumber

numerical_gradient:

• nnp_numerical_gradient: proc numerical_gradient[T: not Tensor](input: T; f: (proc (x: T): T); h = T(0.00001)): T
• nnp_numerical_gradient: proc numerical_gradient[T](input: Tensor[T]; f: (proc (x: Tensor[T]): T); h = T(0.00001)): Tensor[T]

Offset_Values:

• p_accessors: IterKind.Offset_Values

omp_barrier:

• openmp: template omp_barrier(): untyped

omp_chunks:

• openmp: template omp_chunks(omp_size: Natural; chunk_offset, chunk_size: untyped; body: untyped): untyped

omp_critical:

• openmp: template omp_critical(body: untyped): untyped

omp_flush:

• openmp: macro omp_flush(variables: varargs[untyped]): untyped

omp_for:

• openmp: template omp_for(index: untyped; length: Natural; use_simd, nowait: static bool; body: untyped)

OMP_FOR_THRESHOLD:

• global_config: const OMP_FOR_THRESHOLD

omp_get_max_threads:

• openmp: template omp_get_max_threads(): cint

omp_get_nested:

• openmp: template omp_get_nested(): cint

omp_get_num_threads:

• openmp: template omp_get_num_threads(): cint

omp_get_thread_num:

• openmp: template omp_get_thread_num(): cint

omp_master:

• openmp: template omp_master(body: untyped): untyped

OMP_MAX_REDUCE_BLOCKS:

• global_config: const OMP_MAX_REDUCE_BLOCKS

OMP_MEMORY_BOUND_GRAIN_SIZE:

• openmp: const OMP_MEMORY_BOUND_GRAIN_SIZE

OMP_NON_CONTIGUOUS_SCALE_FACTOR:

• openmp: const OMP_NON_CONTIGUOUS_SCALE_FACTOR

omp_parallel:

• openmp: template omp_parallel(body: untyped): untyped

omp_parallel_chunks:

• openmp: template omp_parallel_chunks(length: Natural; chunk_offset, chunk_size: untyped; omp_grain_size: static Natural; body: untyped): untyped

omp_parallel_chunks_default:

• openmp: template omp_parallel_chunks_default(length: Natural; chunk_offset, chunk_size: untyped; body: untyped): untyped

omp_parallel_for:

• openmp: template omp_parallel_for(index: untyped; length: Natural; omp_grain_size: static Natural; use_simd: static bool; body: untyped)

omp_parallel_for_default:

• openmp: template omp_parallel_for_default(index: untyped; length: Natural; body: untyped)

omp_parallel_if:

• openmp: template omp_parallel_if(condition: bool; body: untyped)

omp_set_nested:

• openmp: template omp_set_nested(x: cint)

omp_set_num_threads:

• openmp: template omp_set_num_threads(x: cint)

omp_single:

• openmp: template omp_single(body: untyped): untyped

omp_single_nowait:

• openmp: template omp_single_nowait(body: untyped): untyped

omp_suffix:

• openmp: proc omp_suffix(genNew: static bool = false): string

omp_task:

• openmp: template omp_task(annotation: static string; body: untyped): untyped

omp_taskloop:

• openmp: template omp_taskloop(index: untyped; length: Natural; annotation: static string; body: untyped)

omp_taskwait:

• openmp: template omp_taskwait(): untyped

ones:

• init_cpu: proc ones[T: SomeNumber | Complex[float32] | Complex[float64] | bool](shape: Metadata): Tensor[ T]
• init_cpu: proc ones[T: SomeNumber | Complex[float32] | Complex[float64] | bool]( shape: varargs[int]): Tensor[T]

ones_like:

• init_cpu: proc ones_like[T: SomeNumber | Complex[float32] | Complex[float64] | bool]( t: Tensor[T]): Tensor[T]
• init_cuda: proc ones_like[T: SomeFloat](t: CudaTensor[T]): CudaTensor[T]
• init_opencl: proc ones_like[T: SomeFloat](t: ClTensor[T]): ClTensor[T]

opencl:

• init_opencl: proc opencl[T: SomeFloat](t: Tensor[T]): ClTensor[T]

Optimizer:

• optimizers: type Optimizer

optimizer:

• optimizers: proc optimizer[M, T](model: M; OptimizerKind: typedesc[Adam]; learning_rate: T = T(0.001); beta1: T = T(0.9); beta2: T = T(0.999); eps: T = T(1e-8)): Adam[Tensor[T]]
• optimizers: proc optimizer[M, T](model: M; OptimizerKind: typedesc[SGD]; learning_rate: T): SGD[ Tensor[T]]
• optimizers: proc optimizer[M, T](model: M; OptimizerKind: typedesc[SGDMomentum]; learning_rate: T; momentum: T = T(0.0); decay: T = T(0.0); nesterov = false): SGDMomentum[Tensor[T]]

optimizerAdam:

• optimizers: proc optimizerAdam[M, T](model: M; learning_rate: T; beta1: T = T(0.9); beta2: T = T(0.999); eps: T = T(1e-8)): Adam[Tensor[T]]

optimizerSGD:

• optimizers: proc optimizerSGD[M, T](model: M; learning_rate: T): SGD[Tensor[T]]

optimizerSGDMomentum:

• optimizers: proc optimizerSGDMomentum[M, T](model: M; learning_rate: T; momentum = T(0.0); decay = T(0.0); nesterov = false): SGDMomentum[ Tensor[T]]

orgqr:

• auxiliary_lapack: proc orgqr[T: SomeFloat](rv_q: var Tensor[T]; tau: openArray[T]; scratchspace: var seq[T])

ormqr:

• auxiliary_lapack: proc ormqr[T: SomeFloat](C: var Tensor[T]; Q: Tensor[T]; tau: openArray[T]; side, trans: static char; scratchspace: var seq[T])

outShape:

• conv2D: proc outShape[T](self: Conv2D[T]): seq[int]
• embedding: proc outShape[T](self: Embedding[T]): seq[int]
• flatten: proc outShape[T](self: Flatten[T]): seq[int]
• gcn: proc outShape[T](self: GCNLayer[T]): seq[int]
• linear: proc outShape[T](self: Linear[T]): seq[int]
• maxpool2D: proc outShape[T](self: MaxPool2D[T]): seq[int]

overload:

• overload: macro overload(overloaded_name: untyped; lapack_name: typed{nkSym}): untyped

pack_A_mc_kc:

• gemm_packing: proc pack_A_mc_kc[T; ukernel: static MicroKernel](packedA: ptr UncheckedArray[T]; mc, kc: int; A: MatrixView[T])

pack_B_kc_nc:

• gemm_packing: proc pack_B_kc_nc[T; ukernel: static MicroKernel](packedB: ptr UncheckedArray[T]; kc, nc: int; B: MatrixView[T])

pairs:

• accessors: iterator pairs[T](t: Tensor[T]): (seq[int], T)
• dynamic_stack_arrays: iterator pairs[T](a: DynamicStackArray[T]): (int, T)

pairwiseDistances:

• distances: proc pairwiseDistances(metric: typedesc[AnyMetric]; x, y: Tensor[float]; p = 2.0; squared: static bool = false): Tensor[float]

partitionMNK:

• gemm_tiling: proc partitionMNK(ukernel: static MicroKernel; T: typedesc; M, N, K: Natural): tuple[ mc, nc, kc: int]

Payload:

• autograd_common: object Payload

PayloadKind:

• autograd_common: enum PayloadKind

pca:

• pca: proc pca[T: SomeFloat](X: Tensor[T]; n_components = 2; center: static bool = true; n_oversamples = 5; n_power_iters = 2): tuple[ projected: Tensor[T], components: Tensor[T]]

PCA_Detailed:

• pca: object PCA_Detailed

pca_detailed:

• pca: proc pca_detailed[T: SomeFloat](X: Tensor[T]; n_components = 2; center: static bool = true; n_oversamples = 5; n_power_iters = 2): PCA_Detailed[T]

percentile:

• aggregate: proc percentile[T](arg: Tensor[T]; p: int; isSorted = false): float

permute:

• shapeshifting: proc permute(t: Tensor; dims: varargs[int]): Tensor

permuteImpl:

• p_shapeshifting: proc permuteImpl[T](result: var Tensor[T]; dims: varargs[int])

phase:

• math_functions: proc phase(t: Tensor[Complex[float32]]): Tensor[float32]
• math_functions: proc phase(t: Tensor[Complex[float64]]): Tensor[float64]

pinv:

• algebra: proc pinv[T: Complex32 | Complex64](A: Tensor[T]; rcond = 1e-15): Tensor[T]
• algebra: proc pinv[T: SomeFloat](A: Tensor[T]; rcond = 1e-15): Tensor[T]

pkSeq:

• autograd_common: PayloadKind.pkSeq

pkVar:

• autograd_common: PayloadKind.pkVar

pop:

• ast_utils: proc pop(tree: var NimNode): NimNode

prefetch:

• compiler_optim_hints: template prefetch[T](data: ptr (T or UncheckedArray[T]); rw: static PrefetchRW = Read; locality: static PrefetchLocality = HighTemporalLocality)

PrefetchLocality:

• compiler_optim_hints: enum PrefetchLocality

PrefetchRW:

• compiler_optim_hints: enum PrefetchRW

pretty:

• display: proc pretty[T](t: Tensor[T]; precision = -1): string
• display_cuda: proc pretty[T](t: CudaTensor[T]; precision = -1): string

prettyImpl:

• p_display: proc prettyImpl[T](t: Tensor[T]; inputRank = 0; alignBy = 0; alignSpacing = 4; precision = -1): string

product:

• aggregate: proc product[T](arg: Tensor[T]): T
• aggregate: proc product[T](arg: Tensor[T]; axis: int): Tensor[T]
• dynamic_stack_arrays: proc product[T: SomeNumber](a: DynamicStackArray[T]): T

pthreadpool_t:

• nnpack: type pthreadpool_t

qr:

• decomposition: proc qr[T: SupportedDecomposition](a: Tensor[T]): tuple[Q, R: Tensor[T]]

query:

• kdtree: proc query[T](tree: KDTree[T]; x: Tensor[T]; k = 1; eps = 0.0; metric: typedesc[AnyMetric] = Euclidean; p = 2.0; distanceUpperBound = Inf): tuple[dist: Tensor[T], idx: Tensor[int]]

query_ball_point:

• kdtree: proc query_ball_point[T](tree: KDTree[T]; x: Tensor[T]; radius: float; eps = 0.0; metric: typedesc[AnyMetric] = Euclidean; p = 2.0): tuple[ dist: Tensor[T], idx: Tensor[int]]

radToDeg:

• ufunc: proc radToDeg[T](t`gensym30: Tensor[T]): Tensor[T]
• ufunc: proc radToDeg[T](t`gensym30: Tensor[T]): Tensor[T]

randomNormalTensor:

• init_cpu: proc randomNormalTensor[T: SomeFloat](shape: varargs[int]; mean: T = 0; std: T = 1): Tensor[ T]

randomTensor:

• init_cpu: proc randomTensor[T: bool](shape: varargs[int]): Tensor[T]
• init_cpu: proc randomTensor(shape: varargs[int]; max: int): Tensor[int]
• init_cpu: proc randomTensor[T](shape: varargs[int]; sample_source: openArray[T]): Tensor[T]
• init_cpu: proc randomTensor[T](shape: varargs[int]; slice: Slice[T]): Tensor[T]
• init_cpu: proc randomTensor[T: SomeFloat](shape: varargs[int]; max: T): Tensor[T]

rank:

• data_structure: proc rank[T](t: CudaTensor[T] or ClTensor[T]): range[0 .. LASER_MAXRANK]
• datatypes: proc rank[T](t: Tensor[T]): range[0 .. LASER_MAXRANK]

raw_data_unaligned:

• datatypes: macro raw_data_unaligned(body: untyped): untyped

RawImmutableView:

• datatypes: type RawImmutableView

RawMutableView:

• datatypes: type RawMutableView

read_csv:

• io_csv: proc read_csv[T: SomeNumber | bool | string](csvPath: string; skipHeader = false; separator = ','; quote = '\"'): Tensor[T]

readFloat32BE:

• io_stream_readers: proc readFloat32BE(stream: Stream): float32

readFloat32LE:

• io_stream_readers: proc readFloat32LE(stream: Stream): float32

readFloat64BE:

• io_stream_readers: proc readFloat64BE(stream: Stream): float64

readFloat64LE:

• io_stream_readers: proc readFloat64LE(stream: Stream): float64

read_hdf5:

• io_hdf5: proc read_hdf5[T: SomeNumber](h5f: var H5FileObj; name, group: Option[string]; number: Option[int]): Tensor[T]
• io_hdf5: proc read_hdf5[T: SomeNumber](h5f: var H5FileObj; name, group = ""; number = -1): Tensor[ T]
• io_hdf5: proc read_hdf5[T: SomeNumber](hdf5Path: string; name, group = ""; number = -1): Tensor[ T]

read_image:

• io_image: proc read_image(buffer: seq[byte]): Tensor[uint8]
• io_image: proc read_image(filepath: string): Tensor[uint8]

readInt32BE:

• io_stream_readers: proc readInt32BE(stream: Stream): int32

readInt32LE:

• io_stream_readers: proc readInt32LE(stream: Stream): int32

readInt64BE:

• io_stream_readers: proc readInt64BE(stream: Stream): int64

readInt64LE:

• io_stream_readers: proc readInt64LE(stream: Stream): int64

read_mnist_images:

• mnist: proc read_mnist_images(imgsPath: string): Tensor[uint8]

read_mnist_labels:

• mnist: proc read_mnist_labels(stream: Stream): Tensor[uint8]
• mnist: proc read_mnist_labels(labelsPath: string): Tensor[uint8]

read_npy:

• io_npy: proc read_npy[T: SomeNumber](npyPath: string): Tensor[T]

readUInt16LE:

• io_stream_readers: proc readUInt16LE(stream: Stream): uint16

readUInt32BE:

• io_stream_readers: proc readUInt32BE(stream: Stream): uint32

readUInt32LE:

• io_stream_readers: proc readUInt32LE(stream: Stream): uint32

readUInt64BE:

• io_stream_readers: proc readUInt64BE(stream: Stream): uint64

readUInt64LE:

• io_stream_readers: proc readUInt64LE(stream: Stream): uint64

real:

• complex: proc real[T: SomeFloat](t: Tensor[Complex[T]]): Tensor[T]

real=:

• complex: proc real=[T: SomeFloat](t: var Tensor[Complex[T]]; val: T)
• complex: proc real=[T: SomeFloat](t: var Tensor[Complex[T]]; val: Tensor[T])

reciprocal:

• math_functions: proc reciprocal[T: Complex[float32] or Complex[float64]](t: Tensor[T]): Tensor[T]
• math_functions: proc reciprocal[T: SomeFloat](t: Tensor[T]): Tensor[T]

reduce:

• higher_order_foldreduce: proc reduce[T](arg: Tensor[T]; f: (T, T) -> T): T
• higher_order_foldreduce: proc reduce[T](arg: Tensor[T]; f: (Tensor[T], Tensor[T]) -> Tensor[T]; axis: int): Tensor[ T]

reduce_axis_inline:

• higher_order_foldreduce: template reduce_axis_inline[T](arg: Tensor[T]; reduction_axis: int; op: untyped): untyped

reduce_inline:

• higher_order_foldreduce: template reduce_inline[T](arg: Tensor[T]; op: untyped): untyped

register_node:

• autograd_common: proc register_node[TT](name: static string; gate: Gate[TT]; backward: Backward[TT]; result: Variable[TT] or seq[Variable[TT]]; parents: varargs[Variable[TT]])

relative_error:

• common_error_functions: proc relative_error[T: SomeFloat](y, y_true: T | Complex[T]): T
• common_error_functions: proc relative_error[T: SomeFloat](y, y_true: Tensor[T] | Tensor[Complex[T]]): Tensor[ T]

RelaxedRankOne:

• p_accessors_macros_write: const RelaxedRankOne

relu:

• nnp_activation: proc relu[T](t: Tensor[T]): Tensor[T]
• relu: proc relu[TT](a: Variable[TT]): Variable[TT]

ReluActivation:

• relu: type ReluActivation

relu_backward:

• nnp_activation: proc relu_backward[T](gradient: Tensor[T]; cached_tensor: Tensor[T]): Tensor[T]

repeat_values:

• shapeshifting: proc repeat_values[T](t: Tensor[T]; reps: int; axis = -1): Tensor[T]
• shapeshifting: proc repeat_values[T](t: Tensor[T]; reps: openArray[int]): Tensor[T]
• shapeshifting: proc repeat_values[T](t: Tensor[T]; reps: Tensor[int]): Tensor[T]

replaceNodes:

• ast_utils: proc replaceNodes(ast: NimNode; replacements: NimNode; to_replace: NimNode): NimNode

replaceSymsByIdents:

• ast_utils: proc replaceSymsByIdents(ast: NimNode): NimNode

reshape:

• gates_shapeshifting_views: proc reshape[TT](a: Variable[TT]; shape: Metadata): Variable[TT]
• gates_shapeshifting_views: proc reshape[TT](a: Variable[TT]; shape: varargs[int]): Variable[TT]
• shapeshifting: proc reshape(t: Tensor; new_shape: Metadata): Tensor
• shapeshifting: proc reshape(t: Tensor; new_shape: varargs[int]): Tensor
• shapeshifting_cuda: proc reshape(t: CudaTensor; new_shape: varargs[int]): CudaTensor

ReshapeGate:

• gates_shapeshifting_views: type ReshapeGate

reshapeImpl:

• p_shapeshifting: proc reshapeImpl(t: AnyTensor; new_shape: varargs[int] | Metadata | seq[int]; result: var AnyTensor; infer: static bool)

reshape_infer:

• shapeshifting: proc reshape_infer(t: Tensor; new_shape: varargs[int]): Tensor

reshape_no_copy:

• p_shapeshifting: proc reshape_no_copy(t: AnyTensor; new_shape: varargs[int] | Metadata | seq[int]; result: var AnyTensor; layout: OrderType)

reshape_with_copy:

• p_shapeshifting: proc reshape_with_copy[T](t: Tensor[T]; new_shape: varargs[int] | Metadata | seq[int]; result: var Tensor[T])

returnEmptyIfEmpty:

• p_empty_tensors: macro returnEmptyIfEmpty(tensors: varargs[untyped]): untyped

reversed:

• dynamic_stack_arrays: proc reversed(a: DynamicStackArray): DynamicStackArray
• dynamic_stack_arrays: proc reversed(a: DynamicStackArray; result: var DynamicStackArray)

rewriteTensor_AddMultiply:

• optim_ops_fusion: template rewriteTensor_AddMultiply{ C + `*`(A, B) }[T](A, B, C: Tensor[T]): auto

rewriteTensor_MultiplyAdd:

• optim_ops_fusion: template rewriteTensor_MultiplyAdd{ `*`(A, B) + C }[T](A, B, C: Tensor[T]): auto

rewriteTensor_MultiplyAdd_inplace:

• optim_ops_fusion: template rewriteTensor_MultiplyAdd_inplace{ C += `*`(A, B) }[T](A, B: Tensor[T]; C: var Tensor[T])

rewriteToTensorReshape:

• optim_ops_fusion: template rewriteToTensorReshape{ reshape(toTensor(oa, dummy_bugfix), shape) }(oa: openArray; shape: varargs[int]; dummy_bugfix: static[int]): auto

roll:

• shapeshifting: proc roll[T](t: Tensor[T]; shift: int): Tensor[T]
• shapeshifting: proc roll[T](t: Tensor[T]; shift: int; axis: Natural): Tensor[T]

round:

• ufunc: proc round[T](t`gensym28: Tensor[T]): Tensor[T]
• ufunc: proc round[T](t`gensym28: Tensor[T]): Tensor[T]

round_step_down:

• align_unroller: proc round_step_down(x: Natural; step: static Natural): int

round_step_up:

• align_unroller: proc round_step_up(x: Natural; step: static Natural): int

same:

• math_functions: ConvolveMode.same

set_diagonal:

• special_matrices: proc set_diagonal[T](a: var Tensor[T]; d: Tensor[T]; k = 0; anti = false)

setDiff:

• algorithms: proc setDiff[T](t1, t2: Tensor[T]; symmetric = false): Tensor[T]

setLen:

• dynamic_stack_arrays: proc setLen(a: var DynamicStackArray; len: int)

setZero:

• initialization: proc setZero[T](t: var Tensor[T]; check_contiguous: static bool = true)

SGD:

• optimizers: object SGD

SGDMomentum:

• optimizers: object SGDMomentum

sgn:

• math_functions: proc sgn[T: SomeNumber](t: Tensor[T]): Tensor[int]

shape_to_strides:

• data_structure: proc shape_to_strides(shape: Metadata; layout: OrderType = rowMajor; result: var Metadata)

sigmoid:

• nnp_activation: proc sigmoid[T: SomeFloat](t: Tensor[T]): Tensor[T]
• p_activation: proc sigmoid[T: SomeFloat](x: T): T
• sigmoid: proc sigmoid[TT](a: Variable[TT]): Variable[TT]

SigmoidActivation:

• sigmoid: type SigmoidActivation

sigmoid_backward:

• nnp_activation: proc sigmoid_backward[T](gradient: Tensor[T]; cached_tensor: Tensor[T]): Tensor[T]

sigmoid_cross_entropy:

• cross_entropy_losses: proc sigmoid_cross_entropy[TT](a`gensym0: Variable[TT]; target`gensym0: TT): Variable[ TT]
• nnp_sigmoid_cross_entropy: proc sigmoid_cross_entropy[T](input, target: Tensor[T]): T

sigmoid_cross_entropy_backward:

• nnp_sigmoid_cross_entropy: proc sigmoid_cross_entropy_backward[T](gradient: Tensor[T] or T; cached_tensor: Tensor[T]; target: Tensor[T]): Tensor[ T]

SigmoidCrossEntropyLoss:

• cross_entropy_losses: type SigmoidCrossEntropyLoss

sin:

• ufunc: proc sin[T](t`gensym18: Tensor[T]): Tensor[T]
• ufunc: proc sin[T](t`gensym18: Tensor[T]): Tensor[T]

sinc:

• math_functions: proc sinc[T: SomeFloat](x: T; normalized: static bool = true): T
• math_functions: proc sinc[T: SomeFloat](t: Tensor[T]; normalized: static bool = true): Tensor[T]

sinh:

• ufunc: proc sinh[T](t`gensym17: Tensor[T]): Tensor[T]
• ufunc: proc sinh[T](t`gensym17: Tensor[T]): Tensor[T]

size:

• data_structure: proc size[T](t: CudaTensor[T] or ClTensor[T]): Natural
• datatypes: proc size[T](t: Tensor[T]): Natural

Size2D:

• p_nnp_types: tuple Size2D

SizeHW:

• cudnn_conv_interface: type SizeHW

skipIfEmpty:

• p_empty_tensors: template skipIfEmpty(t: typed): untyped

sliceDispatchImpl:

• p_accessors_macros_read: proc sliceDispatchImpl(result: NimNode; args: NimNode; isRead: bool)

slicer:

• p_accessors_macros_read: proc slicer[T](t: AnyTensor[T]; ellipsis: Ellipsis; slices: openArray[SteppedSlice]): AnyTensor[ T]
• p_accessors_macros_read: proc slicer[T](t: AnyTensor[T]; slices: openArray[SteppedSlice]): AnyTensor[T]
• p_accessors_macros_read: proc slicer[T](t: AnyTensor[T]; slices: openArray[SteppedSlice]; ellipsis: Ellipsis): AnyTensor[ T]
• p_accessors_macros_read: proc slicer[T](t: AnyTensor[T]; slices1: openArray[SteppedSlice]; ellipsis: Ellipsis; slices2: openArray[SteppedSlice]): AnyTensor[T]
• p_accessors_macros_read: proc slicer[T](t: Tensor[T]; slices: ArrayOfSlices): Tensor[T]

slicerImpl:

• p_accessors_macros_read: template slicerImpl[T](result: AnyTensor[T] | var AnyTensor[T]; slices: ArrayOfSlices): untyped

slicerMut:

• p_accessors_macros_write: proc slicerMut[T](t: var Tensor[T]; ellipsis: Ellipsis; slices: openArray[SteppedSlice]; oa: openArray)
• p_accessors_macros_write: proc slicerMut[T](t: var Tensor[T]; ellipsis: Ellipsis; slices: openArray[SteppedSlice]; val: T)
• p_accessors_macros_write: proc slicerMut[T](t: var Tensor[T]; ellipsis: Ellipsis; slices: openArray[SteppedSlice]; t2: Tensor[T])
• p_accessors_macros_write: proc slicerMut[T](t: var Tensor[T]; slices: openArray[SteppedSlice]; ellipsis: Ellipsis; oa: openArray)
• p_accessors_macros_write: proc slicerMut[T](t: var Tensor[T]; slices1: openArray[SteppedSlice]; ellipsis: Ellipsis; slices2: openArray[SteppedSlice]; oa: openArray)
• p_accessors_macros_write: proc slicerMut[T](t: var Tensor[T]; slices1: openArray[SteppedSlice]; ellipsis: Ellipsis; slices2: openArray[SteppedSlice]; val: T)
• p_accessors_macros_write: proc slicerMut[T](t: var Tensor[T]; slices1: openArray[SteppedSlice]; ellipsis: Ellipsis; slices2: openArray[SteppedSlice]; t2: Tensor[T])
• p_accessors_macros_write: proc slicerMut[T](t: var Tensor[T]; slices: openArray[SteppedSlice]; ellipsis: Ellipsis; val: T)
• p_accessors_macros_write: proc slicerMut[T](t: var Tensor[T]; slices: openArray[SteppedSlice]; ellipsis: Ellipsis; t2: Tensor[T])
• p_accessors_macros_write: proc slicerMut[T](t: var Tensor[T]; slices: openArray[SteppedSlice]; oa: openArray)
• p_accessors_macros_write: proc slicerMut[T](t: var Tensor[T]; slices: openArray[SteppedSlice]; val: T)
• p_accessors_macros_write: proc slicerMut[T](t: var Tensor[T]; slices: openArray[SteppedSlice]; t2: Tensor[T])

slice_typed_dispatch:

• p_accessors_macros_read: macro slice_typed_dispatch(t: typed; args: varargs[typed]): untyped

slice_typed_dispatch_mut:

• p_accessors_macros_write: macro slice_typed_dispatch_mut(t: typed; args: varargs[typed]; val: typed): untyped

slice_typed_dispatch_var:

• p_accessors_macros_write: macro slice_typed_dispatch_var(t: typed; args: varargs[typed]): untyped

SmallDiffs:

• autograd_common: type SmallDiffs

softmax:

• nnp_softmax: proc softmax[T](input: Tensor[T]): Tensor[T]
• softmax: proc softmax[TT](a: Variable[TT]): Variable[TT]

SoftmaxActivation:

• softmax: type SoftmaxActivation

softmax_cross_entropy:

• cross_entropy_losses: proc softmax_cross_entropy[TT](a`gensym1: Variable[TT]; target`gensym1: TT): Variable[ TT]
• nnp_softmax_cross_entropy: proc softmax_cross_entropy[T](input, target: Tensor[T]): T

softmax_cross_entropy_backward:

• nnp_softmax_cross_entropy: proc softmax_cross_entropy_backward[T](gradient: Tensor[T] or T; cached_tensor: Tensor[T]; target: Tensor[T]): Tensor[ T]

SoftmaxCrossEntropyLoss:

• cross_entropy_losses: type SoftmaxCrossEntropyLoss

solve:

• linear_systems: proc solve[T: SomeFloat](a, b: Tensor[T]; kind: MatrixKind = mkGeneral): Tensor[T]

sort:

• algorithms: proc sort[T](t: var Tensor[T]; order = SortOrder.Ascending)

sorted:

• algorithms: proc sorted[T](t: Tensor[T]; order = SortOrder.Ascending): Tensor[T]

sparse_softmax_cross_entropy:

• cross_entropy_losses: proc sparse_softmax_cross_entropy[TT; Idx: SomeNumber or byte or char or enum]( a: Variable[TT]; target: Tensor[Idx]): Variable[TT]
• nnp_softmax_cross_entropy: proc sparse_softmax_cross_entropy[T; Idx: SomeNumber or byte or char or enum]( input: Tensor[T]; target: Tensor[Idx]): T

sparse_softmax_cross_entropy_backward:

• nnp_softmax_cross_entropy: proc sparse_softmax_cross_entropy_backward[T; Idx: SomeNumber or byte or char or enum]( gradient: Tensor[T] or T; cached_tensor: Tensor[T]; target: Tensor[Idx]): Tensor[ T]

SparseSoftmaxCrossEntropyLoss:

• cross_entropy_losses: type SparseSoftmaxCrossEntropyLoss

split:

• shapeshifting: proc split[T](t: Tensor[T]; chunk_size: Positive; axis: Natural): seq[Tensor[T]]

sqrt:

• ufunc: proc sqrt[T](t`gensym3: Tensor[T]): Tensor[T]
• ufunc: proc sqrt[T](t`gensym3: Tensor[T]): Tensor[T]

square:

• math_functions: proc square[T](t`gensym101: Tensor[T]): Tensor[T]
• math_functions: proc square[T](x: T): T
• math_functions: proc square[T](x: T): T
• math_functions: proc square[T](t`gensym101: Tensor[T]): Tensor[T]

squared_error:

• common_error_functions: proc squared_error[T: SomeFloat](y, y_true: Tensor[T] | Tensor[Complex[T]]): Tensor[T]
• common_error_functions: proc squared_error[T: SomeFloat](y, y_true: Complex[T]): T
• common_error_functions: proc squared_error[T: SomeFloat](y, y_true: T): T

squeeze:

• gates_shapeshifting_views: proc squeeze[TT](v`gensym0: Variable[TT]; axis`gensym0: Natural): Variable[TT]
• shapeshifting: proc squeeze(t: AnyTensor): AnyTensor
• shapeshifting: proc squeeze(t: Tensor; axis: Natural): Tensor
• shapeshifting_cuda: proc squeeze(t: CudaTensor; axis: int): CudaTensor

squeezeImpl:

• p_shapeshifting: proc squeezeImpl(t: var AnyTensor)
• p_shapeshifting: proc squeezeImpl(t: var AnyTensor; axis: int)

stable_softmax:

• p_logsumexp: proc stable_softmax[T](x, max, sumexp: T): T

stack:

• gates_shapeshifting_concat_split: proc stack[TT](variables: varargs[Variable[TT]]; axis = 0): Variable[TT]
• shapeshifting: proc stack[T](tensors: varargs[Tensor[T]]; axis: Natural = 0): Tensor[T]

std:

• aggregate: proc std[T: SomeFloat](arg: Tensor[T]): T
• aggregate: proc std[T: SomeFloat](arg: Tensor[T]; axis: int): Tensor[T]

Step:

• accessors_macros_syntax: object Step

SteppedSlice:

• accessors_macros_syntax: object SteppedSlice

streaming_max_sumexp:

• p_logsumexp: proc streaming_max_sumexp[T](t: Tensor[T]): tuple[max: T, sumexp: T]
• p_logsumexp: proc streaming_max_sumexp[T](t: Tensor[T]; axis: int): Tensor[ tuple[max: T, sumexp: T]]

stride:

• gemm_utils: proc stride[T](view: MatrixView[T]; row, col: Natural): MatrixView[T]

stridedBodyTemplate:

• foreach_common: template stridedBodyTemplate(): untyped

stridedChunkOffset:

• foreach_common: template stridedChunkOffset(): untyped

stridedCoordsIteration:

• p_accessors: template stridedCoordsIteration(t, iter_offset, iter_size: typed): untyped

stridedIteration:

• p_accessors: template stridedIteration(strider: IterKind; t, iter_offset, iter_size: typed): untyped

stridedIterationYield:

• p_accessors: template stridedIterationYield(strider: IterKind; data, i, iter_pos: typed)

stridedVarsSetup:

• foreach_common: template stridedVarsSetup(): untyped

SubGate:

• gates_basic: type SubGate

sum:

• aggregate: proc sum[T](arg: Tensor[T]): T
• aggregate: proc sum[T](arg: Tensor[T]; axis: int): Tensor[T]
• gates_reduce: proc sum[TT](a: Variable[TT]): Variable[TT]

SumGate:

• gates_reduce: type SumGate

SupportedDecomposition:

• decomposition_lapack: type SupportedDecomposition

svd:

• decomposition: proc svd[T: SupportedDecomposition](A: Tensor[T]): auto
• decomposition: proc svd[T: SupportedDecomposition; U: SomeFloat](A: Tensor[T]; _: typedesc[U]): tuple[ U: Tensor[T], S: Tensor[U], Vh: Tensor[T]]

svd_randomized:

• decomposition_rand: proc svd_randomized[T](A: Tensor[T]; n_components = 2; n_oversamples = 5; n_power_iters = 2): tuple[U, S, Vh: Tensor[T]]

syevr:

• decomposition_lapack: proc syevr[T: SupportedDecomposition](a: var Tensor[T]; uplo: static char; return_eigenvectors: static bool; low_idx: int; high_idx: int; eigenval, eigenvec: var Tensor[T]; scratchspace: var seq[T])

symeig:

• decomposition: proc symeig[T: SupportedDecomposition](a: Tensor[T]; return_eigenvectors: static bool = false; uplo: static char = 'U'): tuple[ eigenval, eigenvec: Tensor[T]]
• decomposition: proc symeig[T: SupportedDecomposition](a: Tensor[T]; return_eigenvectors: static bool = false; uplo: static char = 'U'; slice: HSlice[[type node], [type node]]): tuple[ eigenval, eigenvec: Tensor[T]]

syrk:

• auxiliary_blas: proc syrk[T: SomeFloat](alpha: T; A: Tensor[T]; mul_order: static SyrkKind; beta: T; C: var Tensor[T]; uplo: static char)

SyrkKind:

• auxiliary_blas: enum SyrkKind

tan:

• ufunc: proc tan[T](t`gensym19: Tensor[T]): Tensor[T]
• ufunc: proc tan[T](t`gensym19: Tensor[T]): Tensor[T]

tanh:

• nnp_activation: proc tanh[T: SomeFloat](t: Tensor[T]): Tensor[T]
• tanh: proc tanh[TT](a: Variable[TT]): Variable[TT]
• ufunc: proc tanh[T](t`gensym20: Tensor[T]): Tensor[T]
• ufunc: proc tanh[T](t`gensym20: Tensor[T]): Tensor[T]

TanhActivation:

• tanh: type TanhActivation

tanh_backward:

• nnp_activation: proc tanh_backward[T](gradient: Tensor[T]; cached_tensor: Tensor[T]): Tensor[T]

Tensor:

• datatypes: object Tensor

tile:

• shapeshifting: proc tile[T](t: Tensor[T]; reps: varargs[int]): Tensor[T]

Tiles:

• gemm_tiling: type Tiles

tnInner:

• kdtree: TreeNodeKind.tnInner

tnLeaf:

• kdtree: TreeNodeKind.tnLeaf

toArrayOfSlices:

• accessors_macros_syntax: proc toArrayOfSlices(s: varargs[SteppedSlice]): ArrayOfSlices

toClpointer:

• opencl_backend: proc toClpointer[T](p: ptr T | ptr UncheckedArray[T]): Pmem
• opencl_backend: proc toClpointer[T](p: ClStorage[T]): Pmem
• opencl_backend: proc toClpointer[T](p: ClTensor[T]): Pmem

to_csv:

• io_csv: proc to_csv[T](tensor: Tensor[T]; csvPath: string; separator = ',')

toeplitz:

• special_matrices: proc toeplitz[T](c: Tensor[T]): Tensor[T]
• special_matrices: proc toeplitz[T](c, r: Tensor[T]): Tensor[T]

toFlatSeq:

• exporting: proc toFlatSeq[T](t: Tensor[T]): seq[T]

toHashSet:

• initialization: proc toHashSet[T](t: Tensor[T]): HashSet[T]

toMatrixView:

• gemm_utils: proc toMatrixView[T](data: ptr T; rowStride, colStride: int): MatrixView[T]

toMetadata:

• initialization: proc toMetadata(s: varargs[int]): Metadata
• initialization: template toMetadata(m: Metadata): Metadata

toMetadataArray:

• datatypes: proc toMetadataArray(s: varargs[int]): Metadata

to_ptr:

• gemm_utils: template to_ptr(AB: typed; MR, NR: static int; T: typedesc): untyped

toRawSeq:

• exporting: proc toRawSeq[T](t: Tensor[T]): seq[T]

toSeq1D:

• exporting: proc toSeq1D[T](t: Tensor[T]): seq[T]

toSeq2D:

• exporting: proc toSeq2D[T](t: Tensor[T]): seq[seq[T]]

toSeq3D:

• exporting: proc toSeq3D[T](t: Tensor[T]): seq[seq[seq[T]]]

toSeq4D:

• exporting: proc toSeq4D[T](t: Tensor[T]): seq[seq[seq[seq[T]]]]

toSeq5D:

• exporting: proc toSeq5D[T](t: Tensor[T]): seq[seq[seq[seq[seq[T]]]]]

toTensor:

• initialization: proc toTensor[T](a: openArray[T]): auto
• initialization: proc toTensor[T](a: SomeSet[T]): auto

toUnsafeView:

• initialization: proc toUnsafeView[T: KnownSupportsCopyMem](t: Tensor[T]; aligned: static bool = true): ptr UncheckedArray[ T]

transpose:

• shapeshifting: proc transpose(t: Tensor): Tensor
• shapeshifting_cuda: proc transpose(t: CudaTensor): CudaTensor

TreeNodeKind:

• kdtree: enum TreeNodeKind

tri:

• special_matrices: proc tri[T](shape_ax1, shape_ax0: int; k: static int = 0; upper: static bool = false): Tensor[ T]
• special_matrices: proc tri[T](shape: Metadata; k: static int = 0; upper: static bool = false): Tensor[T]

triangular:

• distributions: proc triangular(x: float): float
• distributions: template triangular[T](t`gensym1: Tensor[T]): Tensor[float]

triangularKernel:

• kde: proc triangularKernel(x`gensym2, x_i`gensym2, bw`gensym2: float): float

trigonometric:

• distributions: proc trigonometric(x: float): float
• distributions: template trigonometric[T](t`gensym2: Tensor[T]): Tensor[float]

trigonometricKernel:

• kde: proc trigonometricKernel(x`gensym3, x_i`gensym3, bw`gensym3: float): float

tril:

• triangular: proc tril[T](a: Tensor[T]; k: static int = 0): Tensor[T]

tril_unit_diag:

• triangular: proc tril_unit_diag[T](a: Tensor[T]): Tensor[T]

tril_unit_diag_mut:

• triangular: proc tril_unit_diag_mut[T](a: var Tensor[T])

tripleStridedIteration:

• p_accessors: template tripleStridedIteration(strider: IterKind; t1, t2, t3, iter_offset, iter_size: typed): untyped

tripleStridedIterationYield:

• p_accessors: template tripleStridedIterationYield(strider: IterKind; t1data, t2data, t3data, i, t1_iter_pos, t2_iter_pos, t3_iter_pos: typed)

triu:

• triangular: proc triu[T](a: Tensor[T]; k: static int = 0): Tensor[T]

trunc:

• ufunc: proc trunc[T](t`gensym27: Tensor[T]): Tensor[T]
• ufunc: proc trunc[T](t`gensym27: Tensor[T]): Tensor[T]

ukernel_generator:

• gemm_ukernel_generator: macro ukernel_generator(simd: static CPUFeatureX86; typ: untyped; vectype: untyped; nb_scalars: static int; simd_setZero: untyped; simd_broadcast_value: untyped; simd_load_aligned: untyped; simd_load_unaligned: untyped; simd_store_unaligned: untyped; simd_mul: untyped; simd_add: untyped; simd_fma: untyped): untyped

ukernel_generic_impl:

• gemm_ukernel_generic: template ukernel_generic_impl()

ukernel_simd_impl:

• gemm_ukernel_generator: macro ukernel_simd_impl(ukernel: static MicroKernel; V: untyped; A, B: untyped; kc: int; simd_setZero, simd_load_aligned, simd_broadcast_value, simd_fma: untyped): untyped

union:

• algorithms: proc union[T](t1, t2: Tensor[T]): Tensor[T]

unique:

• algorithms: proc unique[T](t: Tensor[T]; isSorted = false): Tensor[T]
• algorithms: proc unique[T](t: Tensor[T]; order: SortOrder): Tensor[T]

unsafe_raw_buf:

• datatypes: proc unsafe_raw_buf[T: KnownSupportsCopyMem](t: Tensor[T]; aligned: static bool = true): RawImmutableView[ T]
• datatypes: proc unsafe_raw_buf[T: KnownSupportsCopyMem](t: var Tensor[T]; aligned: static bool = true): RawMutableView[ T]
• datatypes: proc unsafe_raw_buf[T: not KnownSupportsCopyMem](t: Tensor[T]; aligned: static bool = true): ptr UncheckedArray[T]

unsafe_raw_offset:

• datatypes: proc unsafe_raw_offset[T: KnownSupportsCopyMem](t: Tensor[T]; aligned: static bool = true): RawImmutableView[T]
• datatypes: proc unsafe_raw_offset[T: KnownSupportsCopyMem](t: var Tensor[T]; aligned: static bool = true): RawMutableView[T]
• datatypes: proc unsafe_raw_offset[T: not KnownSupportsCopyMem](t: Tensor[T]; aligned: static bool = true): ptr UncheckedArray[T]

unsqueeze:

• gates_shapeshifting_views: proc unsqueeze[TT](v`gensym1: Variable[TT]; axis`gensym1: Natural): Variable[TT]
• shapeshifting: proc unsqueeze(t: Tensor; axis: Natural): Tensor
• shapeshifting_cuda: proc unsqueeze(t: CudaTensor; axis: int): CudaTensor

unsqueezeImpl:

• p_shapeshifting: proc unsqueezeImpl(t: var AnyTensor; axis: int)

unwrap_period:

• aggregate: proc unwrap_period[T: SomeNumber](t: Tensor[T]; discont: T = -1; axis = -1; period: T = default(T)): Tensor[T]

update:

• optimizers: proc update(self: var Adam)
• optimizers: proc update(self: SGD)
• optimizers: proc update(self: var SGDMomentum)

valid:

• math_functions: ConvolveMode.valid

Values:

• p_accessors: IterKind.Values

vander:

• special_matrices: proc vander(order: int = -1; increasing = false): Tensor[float]
• special_matrices: proc vander[T](x: Tensor[T]; order: int = -1; increasing = false): Tensor[float]

vandermonde:

• special_matrices: proc vandermonde(order: int; increasing = true): Tensor[float]
• special_matrices: proc vandermonde[T](x: Tensor[T]; order: int = -1; increasing = true): Tensor[float]
• special_matrices: proc vandermonde[T](x: Tensor[T]; orders: Tensor[SomeNumber]): Tensor[float]

Variable:

• autograd_common: type Variable

variable:

• autograd_common: proc variable[TT](ctx: Context[TT]; value: TT; requires_grad = false): Variable[TT]

variance:

• aggregate: proc variance[T: SomeFloat](arg: Tensor[T]): T
• aggregate: proc variance[T: SomeFloat](arg: Tensor[T]; axis: int): Tensor[T]

whitespaceTokenizer:

• tokenizers: iterator whitespaceTokenizer(input: Tensor[string]): seq[string]

withCompilerOptimHints:

• compiler_optim_hints: template withCompilerOptimHints()

with_diagonal:

• special_matrices: proc with_diagonal[T](a: Tensor[T]; d: Tensor[T]; k = 0; anti = false): Tensor[T]

withMemoryOptimHints:

• memory_optimization_hints: template withMemoryOptimHints()

write_bmp:

• io_image: proc write_bmp(img: Tensor[uint8]; filepath: string)

write_hdf5:

• io_hdf5: proc write_hdf5[T: SomeNumber](h5f: var H5FileObj; t: Tensor[T]; name, group: Option[string])
• io_hdf5: proc write_hdf5[T: SomeNumber](h5f: var H5FileObj; t: Tensor[T]; name, group = "")
• io_hdf5: proc write_hdf5[T: SomeNumber](t: Tensor[T]; hdf5Path: string; name, group = "")

write_jpg:

• io_image: proc write_jpg(img: Tensor[uint8]; filepath: string; quality = 100)

write_npy:

• io_npy: proc write_npy[T: SomeNumber](t: Tensor[T]; npyPath: string)

write_png:

• io_image: proc write_png(img: Tensor[uint8]; filepath: string)

write_tga:

• io_image: proc write_tga(img: Tensor[uint8]; filepath: string)

x86_AVX:

• gemm_tiling: CPUFeatureX86.x86_AVX

x86_AVX2:

• gemm_tiling: CPUFeatureX86.x86_AVX2

x86_AVX512:

• gemm_tiling: CPUFeatureX86.x86_AVX512

x86_AVX_FMA:

• gemm_tiling: CPUFeatureX86.x86_AVX_FMA

x86_Generic:

• gemm_tiling: CPUFeatureX86.x86_Generic

x86only:

• gemm_ukernel_generator: template x86only(): untyped

x86_SSE:

• gemm_tiling: CPUFeatureX86.x86_SSE

x86_SSE2:

• gemm_tiling: CPUFeatureX86.x86_SSE2

x86_SSE4_1:

• gemm_tiling: CPUFeatureX86.x86_SSE4_1

x86_ukernel:

• gemm_tiling: proc x86_ukernel(cpu: CPUFeatureX86; T: typedesc; c_unit_stride: bool): MicroKernel

xavier_normal:

• init: proc xavier_normal(shape: varargs[int]; T: type): Tensor[T]

xavier_uniform:

• init: proc xavier_uniform(shape: varargs[int]; T: type): Tensor[T]

xygrid:

• special_matrices: MeshGridIndexing.xygrid

yann_normal:

• init: proc yann_normal(shape: varargs[int]; T: type): Tensor[T]

yann_uniform:

• init: proc yann_uniform(shape: varargs[int]; T: type): Tensor[T]

zeroGrads:

• optimizers: proc zeroGrads(o: Optimizer)

zeros:

• init_cpu: proc zeros[T: SomeNumber | Complex[float32] | Complex[float64] | bool]( shape: Metadata): Tensor[T]
• init_cpu: proc zeros[T: SomeNumber | Complex[float32] | Complex[float64] | bool]( shape: varargs[int]): Tensor[T]

zeros_like:

• init_cpu: proc zeros_like[T: SomeNumber | Complex[float32] | Complex[float64] | bool]( t: Tensor[T]): Tensor[T]
• init_cuda: proc zeros_like[T: SomeFloat](t: CudaTensor[T]): CudaTensor[T]
• init_opencl: proc zeros_like[T: SomeFloat](t: ClTensor[T]): ClTensor[T]

zip:

• accessors: iterator zip[T, U](t1: Tensor[T]; t2: Tensor[U]): (T, U)
• accessors: iterator zip[T, U](t1: Tensor[T]; t2: Tensor[U]; offset, size: int): (T, U)
• accessors: iterator zip[T, U, V](t1: Tensor[T]; t2: Tensor[U]; t3: Tensor[V]): (T, U, V)
• accessors: iterator zip[T, U, V](t1: Tensor[T]; t2: Tensor[U]; t3: Tensor[V]; offset, size: int): ( T, U, V)
• dynamic_stack_arrays: iterator zip[T, U](a: DynamicStackArray[T]; b: DynamicStackArray[U]): (T, T)

zipAxis:

• accessors: iterator zipAxis[T, U](a: Tensor[T]; b: Tensor[U]; axis: int): tuple[a: Tensor[T], b: Tensor[U]]