package caisar
Legend:
Library
Module
Module type
Parameter
Class
Class type
Library
Module
Module type
Parameter
Class
Class type
Type describing the different operations handled. Those operations are inspired by those defined in the ONNX documentation.
val str_op : operator -> Base.string
val show_shape : shape -> Base.string
type operator_parameters =
| Pool_params of ksize * stride Base.option * pads Base.option * dilations Base.option
| Conv_params of ksize * stride Base.option * pads Base.option * dilations Base.option
| Transpose_params of shape
| RW_Linearized_ReLu_params of Base.bool Base.list Base.list * ((Base.string, Base.float) Base.Hashtbl.t Base.list * Base.int)
val str_op_params : operator_parameters -> Base.string
type ('a, 'b) t = {
id : Base.int;
name : Base.string Base.option;
shape : shape;
operator : operator;
operator_parameters : operator_parameters Base.option;
pred : Base.string Base.list;
succ : Base.string Base.list;
tensor : ('a, 'b) Tensor.t Base.option;
}
Type encapsulating parameters for operations. For Convolutions and Pooling, kernel size, padding, strides For Transpose, shape
val hash : ('a, 'b) t -> Base.int
val create :
id:Base.int ->
name:Base.string Base.option ->
sh:shape ->
op:operator ->
op_p:operator_parameters Base.option ->
pred:Base.string Base.list ->
succ:Base.string Base.list ->
tensor:('a, 'b) Tensor.t Base.option ->
('a, 'b) t
val get_name : ('a, 'b) t -> Base.string
val get_pred_list : ('a, 'b) t -> Base.string Base.list
val get_succ_list : ('a, 'b) t -> Base.string Base.list
val is_data_node : ('a, 'b) t -> Base.bool
val is_input_node : ('a, 'b) t -> Base.bool
val is_output_node : ('a, 'b) t -> Base.bool
val num_neurons : ('a, 'b) t -> Base.int
val show : ('a, 'b) t -> ('a -> Base.string) -> Base.string
sectionYPositions = computeSectionYPositions($el), 10)"
x-init="setTimeout(() => sectionYPositions = computeSectionYPositions($el), 10)"
>