Functions | |
Node | assign (NodeInput new_value, str variable_id, Optional[str] name=None) |
Return a node which produces the Assign operation. More... | |
Node | broadcast (NodeInput data, NodeInput target_shape, Optional[NodeInput] axes_mapping=None, str broadcast_spec="NUMPY", Optional[str] name=None) |
Create a node which broadcasts the input node's values along specified axes to a desired shape. More... | |
Node | bucketize (Node data, NodeInput buckets, str output_type="i64", bool with_right_bound=True, Optional[str] name=None) |
Return a node which produces the Bucketize operation. More... | |
Node | cum_sum (NodeInput arg, NodeInput axis, bool exclusive=False, bool reverse=False, Optional[str] name=None) |
Construct a cumulative summation operation. More... | |
Node | embedding_bag_offsets_sum (Node emb_table, NodeInput indices, NodeInput offsets, Optional[NodeInput] default_index=None, Optional[NodeInput] per_sample_weights=None, Optional[str] name=None) |
Return a node which performs sums of bags of embeddings without the intermediate embeddings. More... | |
Node | embedding_bag_packed_sum (NodeInput emb_table, NodeInput indices, Optional[NodeInput] per_sample_weights=None, Optional[str] name=None) |
Return an EmbeddingBagPackedSum node. More... | |
Node | embedding_segments_sum (Node emb_table, NodeInput indices, NodeInput segment_ids, Optional[NodeInput] num_segments=None, Optional[NodeInput] default_index=None, Optional[NodeInput] per_sample_weights=None, Optional[str] name=None) |
Return an EmbeddingSegmentsSum node. More... | |
Node | extract_image_patches (NodeInput image, TensorShape sizes, List[int] strides, TensorShape rates, str auto_pad, Optional[str] name=None) |
Return a node which produces the ExtractImagePatches operation. More... | |
Node | gru_cell (NodeInput X, NodeInput initial_hidden_state, NodeInput W, NodeInput R, NodeInput B, int hidden_size, List[str] activations=None, List[float] activations_alpha=None, List[float] activations_beta=None, float clip=0.0, bool linear_before_reset=False, Optional[str] name=None) |
Perform GRUCell operation on the tensor from input node. More... | |
Node | non_max_suppression (NodeInput boxes, NodeInput scores, Optional[NodeInput] max_output_boxes_per_class=None, Optional[NodeInput] iou_threshold=None, Optional[NodeInput] score_threshold=None, str box_encoding="corner", bool sort_result_descending=True, str output_type="i64", Optional[str] name=None) |
Return a node which performs NonMaxSuppression. More... | |
Node | non_zero (NodeInput data, str output_type="i64", Optional[str] name=None) |
Return the indices of the elements that are non-zero. More... | |
Node | read_value (NodeInput init_value, str variable_id, Optional[str] name=None) |
Return a node which produces the Assign operation. More... | |
Node | rnn_cell (NodeInput X, NodeInput initial_hidden_state, NodeInput W, NodeInput R, NodeInput B, int hidden_size, List[str] activations, List[float] activations_alpha, List[float] activations_beta, float clip=0.0, Optional[str] name=None) |
Perform RNNCell operation on tensor from input node. More... | |
Node | roi_align (NodeInput data, NodeInput rois, NodeInput batch_indices, int pooled_h, int pooled_w, int sampling_ratio, float spatial_scale, str mode, Optional[str] name=None) |
Return a node which performs ROIAlign. More... | |
Node | scatter_elements_update (NodeInput data, NodeInput indices, NodeInput updates, NodeInput axis, Optional[str] name=None) |
Return a node which produces a ScatterElementsUpdate operation. More... | |
Node | scatter_update (Node data, NodeInput indices, NodeInput updates, NodeInput axis, Optional[str] name=None) |
Return a node which produces a ScatterUpdate operation. More... | |
Node | shape_of (NodeInput data, str output_type="i64", Optional[str] name=None) |
Return a node which produces a tensor containing the shape of its input data. More... | |
Node | shuffle_channels (Node data, int axis, int groups, Optional[str] name=None) |
Perform permutation on data in the channel dimension of the input tensor. More... | |
Node | topk (NodeInput data, NodeInput k, int axis, str mode, str sort, str index_element_type="i32", Optional[str] name=None) |
Return a node which performs TopK. More... | |
Node ngraph.opset3.ops.assign | ( | NodeInput | new_value, |
str | variable_id, | ||
Optional[str] | name = None |
||
) |
Return a node which produces the Assign operation.
@param new_value: Node producing a value to be assigned to a variable. @param variable_id: Id of a variable to be updated. @param name: Optional name for output node. @return Assign node
Node ngraph.opset3.ops.broadcast | ( | NodeInput | data, |
NodeInput | target_shape, | ||
Optional[NodeInput] | axes_mapping = None , |
||
str | broadcast_spec = "NUMPY" , |
||
Optional[str] | name = None |
||
) |
Create a node which broadcasts the input node's values along specified axes to a desired shape.
@param data: The node with input tensor data. @param target_shape: The node with a new shape we want to broadcast tensor to. @param axes_mapping: The node with a axis positions (0-based) in the result that are being broadcast. @param broadcast_spec: The type of broadcasting that specifies mapping of input tensor axes to output shape axes. Range of values: NUMPY, EXPLICIT, BIDIRECTIONAL. @param name: Optional new name for output node. @return New node with broadcast shape.
Node ngraph.opset3.ops.bucketize | ( | Node | data, |
NodeInput | buckets, | ||
str | output_type = "i64" , |
||
bool | with_right_bound = True , |
||
Optional[str] | name = None |
||
) |
Return a node which produces the Bucketize operation.
@param data: Input data to bucketize @param buckets: 1-D of sorted unique boundaries for buckets @param output_type: Output tensor type, "i64" or "i32", defaults to i64 @param with_right_bound: indicates whether bucket includes the right or left edge of interval. default true = includes right edge @param name: Optional name for output node. @return Bucketize node
Node ngraph.opset3.ops.cum_sum | ( | NodeInput | arg, |
NodeInput | axis, | ||
bool | exclusive = False , |
||
bool | reverse = False , |
||
Optional[str] | name = None |
||
) |
Construct a cumulative summation operation.
@param arg: The tensor to be summed. @param axis: zero dimension tensor specifying axis position along which sum will be performed. @param exclusive: if set to true, the top element is not included @param reverse: if set to true, will perform the sums in reverse direction @return New node performing the operation
Node ngraph.opset3.ops.embedding_bag_offsets_sum | ( | Node | emb_table, |
NodeInput | indices, | ||
NodeInput | offsets, | ||
Optional[NodeInput] | default_index = None , |
||
Optional[NodeInput] | per_sample_weights = None , |
||
Optional[str] | name = None |
||
) |
Return a node which performs sums of bags of embeddings without the intermediate embeddings.
@param emb_table: Tensor containing the embedding lookup table. @param indices: Tensor with indices. @param offsets: Tensor containing the starting index positions of each bag in indices. @param per_sample_weights: Tensor with weights for each sample. @param default_index: Scalar containing default index in embedding table to fill empty bags. @param name: Optional name for output node. @return The new node which performs EmbeddingBagOffsetsSum
Node ngraph.opset3.ops.embedding_bag_packed_sum | ( | NodeInput | emb_table, |
NodeInput | indices, | ||
Optional[NodeInput] | per_sample_weights = None , |
||
Optional[str] | name = None |
||
) |
Return an EmbeddingBagPackedSum node.
EmbeddingSegmentsSum constructs an output tensor by replacing every index in a given input tensor with a row (from the weights matrix) at that index @param emb_table: Tensor containing the embedding lookup table. @param indices: Tensor with indices. @param per_sample_weights: Weights to be multiplied with embedding table. @param name: Optional name for output node. @return EmbeddingBagPackedSum node
Node ngraph.opset3.ops.embedding_segments_sum | ( | Node | emb_table, |
NodeInput | indices, | ||
NodeInput | segment_ids, | ||
Optional[NodeInput] | num_segments = None , |
||
Optional[NodeInput] | default_index = None , |
||
Optional[NodeInput] | per_sample_weights = None , |
||
Optional[str] | name = None |
||
) |
Return an EmbeddingSegmentsSum node.
EmbeddingSegmentsSum constructs an output tensor by replacing every index in a given input tensor with a row (from the weights matrix) at that index @param emb_table: Tensor containing the embedding lookup table. @param indices: Tensor with indices. @param segment_ids: Tensor with indices into the output Tensor @param num_segments: Tensor with number of segments. @param default_index: Scalar containing default index in embedding table to fill empty bags. @param per_sample_weights: Weights to be multiplied with embedding table. @param name: Optional name for output node. @return EmbeddingSegmentsSum node
Node ngraph.opset3.ops.extract_image_patches | ( | NodeInput | image, |
TensorShape | sizes, | ||
List[int] | strides, | ||
TensorShape | rates, | ||
str | auto_pad, | ||
Optional[str] | name = None |
||
) |
Return a node which produces the ExtractImagePatches operation.
@param image: 4-D Input data to extract image patches. @param sizes: Patch size in the format of [size_rows, size_cols]. @param strides: Patch movement stride in the format of [stride_rows, stride_cols] @param rates: Element seleciton rate for creating a patch. @param auto_pad: Padding type. @param name: Optional name for output node. @return ExtractImagePatches node
Node ngraph.opset3.ops.gru_cell | ( | NodeInput | X, |
NodeInput | initial_hidden_state, | ||
NodeInput | W, | ||
NodeInput | R, | ||
NodeInput | B, | ||
int | hidden_size, | ||
List[str] | activations = None , |
||
List[float] | activations_alpha = None , |
||
List[float] | activations_beta = None , |
||
float | clip = 0.0 , |
||
bool | linear_before_reset = False , |
||
Optional[str] | name = None |
||
) |
Perform GRUCell operation on the tensor from input node.
GRUCell represents a single GRU Cell that computes the output using the formula described in the paper: https://arxiv.org/abs/1406.1078 Note this class represents only single *cell* and not whole *layer*. @param X: The input tensor with shape: [batch_size, input_size]. @param initial_hidden_state: The hidden state tensor at current time step with shape: [batch_size, hidden_size]. @param W: The weights for matrix multiplication, gate order: zrh. Shape: [3*hidden_size, input_size]. @param R: The recurrence weights for matrix multiplication. Shape: [3*hidden_size, hidden_size]. @param B: The sum of biases (weight and recurrence). For linear_before_reset set True the shape is [4*hidden_size]. Otherwise the shape is [3*hidden_size]. @param hidden_size: The number of hidden units for recurrent cell. Specifies hidden state size. @param activations: The vector of activation functions used inside recurrent cell. @param activation_alpha: The vector of alpha parameters for activation functions in order respective to activation list. @param activation_beta: The vector of beta parameters for activation functions in order respective to activation list. @param clip: The value defining clipping range [-clip, clip] on input of activation functions. @param linear_before_reset: Flag denotes if the layer behaves according to the modification of GRUCell described in the formula in the ONNX documentation. @param name: Optional output node name. @return The new node performing a GRUCell operation on tensor from input node.
Node ngraph.opset3.ops.non_max_suppression | ( | NodeInput | boxes, |
NodeInput | scores, | ||
Optional[NodeInput] | max_output_boxes_per_class = None , |
||
Optional[NodeInput] | iou_threshold = None , |
||
Optional[NodeInput] | score_threshold = None , |
||
str | box_encoding = "corner" , |
||
bool | sort_result_descending = True , |
||
str | output_type = "i64" , |
||
Optional[str] | name = None |
||
) |
Return a node which performs NonMaxSuppression.
@param boxes: Tensor with box coordinates. @param scores: Tensor with box scores. @param max_output_boxes_per_class: Tensor Specifying maximum number of boxes to be selected per class. @param iou_threshold: Tensor specifying intersection over union threshold @param score_threshold: Tensor specifying minimum score to consider box for the processing. @param box_encoding: Format of boxes data encoding. @param sort_result_descending: Flag that specifies whenever it is necessary to sort selected boxes across batches or not. @param output_type: Output element type. @return The new node which performs NonMaxSuppression
Node ngraph.opset3.ops.non_zero | ( | NodeInput | data, |
str | output_type = "i64" , |
||
Optional[str] | name = None |
||
) |
Return the indices of the elements that are non-zero.
@param data: Input data. @param output_type: Output tensor type. @return The new node which performs NonZero
Node ngraph.opset3.ops.read_value | ( | NodeInput | init_value, |
str | variable_id, | ||
Optional[str] | name = None |
||
) |
Return a node which produces the Assign operation.
@param init_value: Node producing a value to be returned instead of an unassigned variable. @param variable_id: Id of a variable to be read. @param name: Optional name for output node. @return ReadValue node
Node ngraph.opset3.ops.rnn_cell | ( | NodeInput | X, |
NodeInput | initial_hidden_state, | ||
NodeInput | W, | ||
NodeInput | R, | ||
NodeInput | B, | ||
int | hidden_size, | ||
List[str] | activations, | ||
List[float] | activations_alpha, | ||
List[float] | activations_beta, | ||
float | clip = 0.0 , |
||
Optional[str] | name = None |
||
) |
Perform RNNCell operation on tensor from input node.
It follows notation and equations defined as in ONNX standard: https://github.com/onnx/onnx/blob/master/docs/Operators.md#RNN Note this class represents only single *cell* and not whole RNN *layer*. @param X: The input tensor with shape: [batch_size, input_size]. @param initial_hidden_state: The hidden state tensor at current time step with shape: [batch_size, hidden_size]. @param W: The weight tensor with shape: [hidden_size, input_size]. @param R: The recurrence weight tensor with shape: [hidden_size, hidden_size]. @param B: The sum of biases (weight and recurrence) with shape: [hidden_size]. @param hidden_size: The number of hidden units for recurrent cell. Specifies hidden state size. @param activations: The vector of activation functions used inside recurrent cell. @param activation_alpha: The vector of alpha parameters for activation functions in order respective to activation list. @param activation_beta: The vector of beta parameters for activation functions in order respective to activation list. @param clip: The value defining clipping range [-clip, clip] on input of activation functions. @param name: Optional output node name. @return The new node performing a RNNCell operation on tensor from input node.
Node ngraph.opset3.ops.roi_align | ( | NodeInput | data, |
NodeInput | rois, | ||
NodeInput | batch_indices, | ||
int | pooled_h, | ||
int | pooled_w, | ||
int | sampling_ratio, | ||
float | spatial_scale, | ||
str | mode, | ||
Optional[str] | name = None |
||
) |
Return a node which performs ROIAlign.
@param data: Input data. @param rois: RoIs (Regions of Interest) to pool over. @param batch_indices: Tensor with each element denoting the index of the corresponding image in the batch. @param pooled_h: Height of the ROI output feature map. @param pooled_w: Width of the ROI output feature map. @param sampling_ratio: Number of bins over height and width to use to calculate each output feature map element. @param spatial_scale: Multiplicative spatial scale factor to translate ROI coordinates. @param mode: Method to perform pooling to produce output feature map elements. @return The new node which performs ROIAlign
Node ngraph.opset3.ops.scatter_elements_update | ( | NodeInput | data, |
NodeInput | indices, | ||
NodeInput | updates, | ||
NodeInput | axis, | ||
Optional[str] | name = None |
||
) |
Return a node which produces a ScatterElementsUpdate operation.
@param data: The input tensor to be updated. @param indices: The tensor with indexes which will be updated. @param updates: The tensor with update values. @param axis: The axis for scatter. @return ScatterElementsUpdate node ScatterElementsUpdate creates a copy of the first input tensor with updated elements specified with second and third input tensors. For each entry in `updates`, the target index in `data` is obtained by combining the corresponding entry in `indices` with the index of the entry itself: the index-value for dimension equal to `axis` is obtained from the value of the corresponding entry in `indices` and the index-value for dimension not equal to `axis` is obtained from the index of the entry itself.
Node ngraph.opset3.ops.scatter_update | ( | Node | data, |
NodeInput | indices, | ||
NodeInput | updates, | ||
NodeInput | axis, | ||
Optional[str] | name = None |
||
) |
Return a node which produces a ScatterUpdate operation.
ScatterUpdate sets new values to slices from data addressed by indices. @param data: The input tensor to be updated. @param indices: The tensor with indexes which will be updated. @param updates: The tensor with update values. @param axis: The axis at which elements will be updated. @return ScatterUpdate node
Node ngraph.opset3.ops.shape_of | ( | NodeInput | data, |
str | output_type = "i64" , |
||
Optional[str] | name = None |
||
) |
Return a node which produces a tensor containing the shape of its input data.
@param data: The tensor containing the input data. @param output_type: Output element type. @return ShapeOf node
Node ngraph.opset3.ops.shuffle_channels | ( | Node | data, |
int | axis, | ||
int | groups, | ||
Optional[str] | name = None |
||
) |
Perform permutation on data in the channel dimension of the input tensor.
@param data: The node with input tensor. @param axis: Channel dimension index in the data tensor. A negative value means that the index should be calculated from the back of the input data shape. @param group: The channel dimension specified by the axis parameter should be split into this number of groups. @param name: Optional output node name. @return The new node performing a permutation on data in the channel dimension of the input tensor. The operation is the equivalent with the following transformation of the input tensor `data` of shape [N, C, H, W]: `data_reshaped` = reshape(`data`, [N, group, C / group, H * W]) `data_trnasposed` = transpose(`data_reshaped`, [0, 2, 1, 3]) `output` = reshape(`data_trnasposed`, [N, C, H, W]) For example: @code{.py} Inputs: tensor of shape [1, 6, 2, 2] data = [[[[ 0., 1.], [ 2., 3.]], [[ 4., 5.], [ 6., 7.]], [[ 8., 9.], [10., 11.]], [[12., 13.], [14., 15.]], [[16., 17.], [18., 19.]], [[20., 21.], [22., 23.]]]] axis = 1 groups = 3 Output: tensor of shape [1, 6, 2, 2] output = [[[[ 0., 1.], [ 2., 3.]], [[ 8., 9.], [10., 11.]], [[16., 17.], [18., 19.]], [[ 4., 5.], [ 6., 7.]], [[12., 13.], [14., 15.]], [[20., 21.], [22., 23.]]]] @endcode
Node ngraph.opset3.ops.topk | ( | NodeInput | data, |
NodeInput | k, | ||
int | axis, | ||
str | mode, | ||
str | sort, | ||
str | index_element_type = "i32" , |
||
Optional[str] | name = None |
||
) |
Return a node which performs TopK.
@param data: Input data. @param k: K. @param axis: TopK Axis. @param mode: Compute TopK largest ('max') or smallest ('min') @param sort: Order of output elements (sort by: 'none', 'index' or 'value') @param index_element_type: Type of output tensor with indices. @return The new node which performs TopK (both indices and values)