ngraph

ngraph module namespace, exposing factory functions for all ops and other classes.

Functions

abs(node[, name])	Return node which applies f(x) = abs(x) to the input node element-wise.
absolute(node[, name])	Return node which applies f(x) = abs(x) to the input node element-wise.
acos(node[, name])	Apply inverse cosine function on the input node element-wise.
acosh(node[, name])	Apply hyperbolic inverse cosine function on the input node element-wise.
adaptive_avg_pool(data, output_shape)	Return a node which performs AdaptiveAvgPool operation.
adaptive_max_pool(data, output_shape[, ...])	Return a node which performs AdaptiveMaxPool operation.
add(left_node, right_node[, auto_broadcast, ...])	Return node which applies f(x) = A+B to the input nodes element-wise.
asin(node[, name])	Apply inverse sine function on the input node element-wise.
asinh(node[, name])	Apply hyperbolic inverse sinus function on the input node element-wise.
assign(new_value, variable_id[, name])	Return a node which produces the Assign operation.
atan(node[, name])	Apply inverse tangent function on the input node element-wise.
atanh(node[, name])	Apply hyperbolic inverse tangent function on the input node element-wise.
avg_pool(data_batch, strides, pads_begin, ...)	Return average pooling node.
batch_norm_inference(data, gamma, beta, ...)	Perform layer normalizes a input tensor by mean and variance with appling scale and offset.
batch_to_space(data, block_shape, ...[, name])	Perform BatchToSpace operation on the input tensor.
binary_convolution(data, filters, strides, ...)	Create node performing convolution with binary weights, binary input and integer output.
broadcast(data, target_shape[, ...])	Create a node which broadcasts the input node's values along specified axes to a desired shape.
bucketize(data, buckets[, output_type, ...])	Return a node which produces the Bucketize operation.
ceil(node[, name])	Return node which applies ceiling to the input node element-wise.
ceiling(node[, name])	Return node which applies ceiling to the input node element-wise.
clamp(data, min_value, max_value[, name])	Perform clamp element-wise on data from input node.
concat(nodes, axis[, name])	Concatenate input nodes into single new node along specified axis.
constant(value[, dtype, name])	Create a Constant node from provided value.
convert(data, destination_type[, name])	Return node which casts input node values to specified type.
convert_like(data, like[, name])	Return node which casts data node values to the type of another node.
convolution(data, filters, strides, ...[, ...])	Return node performing batched convolution operation.
convolution_backprop_data(data, filters, strides)	Create node performing a batched-convolution backprop data operation.
cos(node[, name])	Apply cosine function on the input node element-wise.
cosh(node[, name])	Apply hyperbolic cosine function on the input node element-wise.
ctc_greedy_decoder(data, sequence_mask[, ...])	Perform greedy decoding on the logits given in input (best path).
ctc_greedy_decoder_seq_len(data, sequence_length)	Return a node which performs CTCGreedyDecoderSeqLen.
ctc_loss(logits, logit_length, labels, ...)	Return a node which performs CTCLoss.
cum_sum(arg, axis[, exclusive, reverse, name])	Construct a cumulative summation operation.
cumsum(arg, axis[, exclusive, reverse, name])	Construct a cumulative summation operation.
deformable_convolution(data, offsets, ...[, ...])	Return a node which performs deformable convolution operation.
deformable_psroi_pooling(feature_maps, ...)	Return node performing DeformablePSROIPooling operation.
depth_to_space(node, mode[, block_size, name])	Rearranges input tensor from depth into blocks of spatial data.
detection_output(box_logits, class_preds, ...)	Generate the detection output using information on location and confidence predictions.
dft(data, axes[, signal_size])	Return a node which performs DFT operation.
divide(left_node, right_node[, ...])	Return node which applies f(x) = A/B to the input nodes element-wise.
einsum(inputs, equation)	Return a node which performs Einsum operation.
elu(data, alpha[, name])	Perform Exponential Linear Unit operation element-wise on data from input node.
embedding_bag_offsets_sum(emb_table, ...[, ...])	Return a node which performs sums of bags of embeddings without the intermediate embeddings.
embedding_bag_packed_sum(emb_table, indices)	Return an EmbeddingBagPackedSum node.
embedding_segments_sum(emb_table, indices, ...)	Return an EmbeddingSegmentsSum node.
equal(left_node, right_node[, ...])	Return node which checks if input nodes are equal element-wise.
erf(node[, name])	Return node which calculates Gauss error function element-wise with given tensor.
exp(node[, name])	Return node which applies exponential function to the input node element-wise.
extract_image_patches(image, sizes, strides, ...)	Return a node which produces the ExtractImagePatches operation.
eye(num_rows, num_columns, diagonal_index, ...)	Return a node which performs eye operation.
fake_quantize(data, input_low, input_high, ...)	Perform an element-wise linear quantization on input data.
floor(node[, name])	Return node which applies floor to the input node element-wise.
floor_mod(left_node, right_node[, ...])	Return node performing element-wise FloorMod (division reminder) with two given tensors.
function_from_cnn(cnn_network)	Get nGraph function from Inference Engine CNN network.
function_to_cnn(ng_function)	Get Inference Engine CNN network from nGraph function.
gather(data, indices, axis[, batch_dims])	Return a node which performs Gather with support of negative indices.
gather_elements(data, indices[, axis, name])	Return a node which performs GatherElements.
gather_nd(data, indices[, batch_dims, name])	Return a node which performs GatherND.
gather_tree(step_ids, parent_idx, ...[, name])	Perform GatherTree operation.
gelu(data, approximation_mode[, name])	Return a node which performs Gelu activation function.
generate_proposals(im_info, anchors, deltas, ...)	Return a node which performs GenerateProposals operation.
greater(left_node, right_node[, ...])	Return node which checks if left input node is greater than the right node element-wise.
greater_equal(left_node, right_node[, ...])	Return node which checks if left node is greater or equal to the right node element-wise.
grid_sample(data, grid, attributes[, name])	Return a node which performs GridSample operation.
grn(data, bias[, name])	Perform Global Response Normalization with L2 norm (across channels only).
group_convolution(data, filters, strides, ...)	Perform Group Convolution operation on data from input node.
group_convolution_backprop_data(data, ...[, ...])	Perform Group Convolution operation on data from input node.
gru_cell(X, initial_hidden_state, W, R, B, ...)	Perform GRUCell operation on the tensor from input node.
gru_sequence(X, initial_hidden_state, ...[, ...])	Return a node which performs GRUSequence operation.
hard_sigmoid(data, alpha, beta[, name])	Perform Hard Sigmoid operation element-wise on data from input node.
hsigmoid(data[, name])	Return a node which performs HSigmoid.
hswish(data[, name])	Return a node which performs HSwish (hard version of Swish).
i420_to_bgr(arg[, arg_u, arg_v, name])	Return a node which performs I420toBGR operation.
i420_to_rgb(arg[, arg_u, arg_v, name])	Return a node which performs I420toRGB operation.
idft(data, axes[, signal_size])	Return a node which performs IDFT operation.
if_op(condition, inputs, bodies, input_desc, ...)	Execute one of the bodies depending on condtion value.
interpolate(image, scales_or_sizes, mode, ...)	Performs the interpolation of the input tensor.
irdft(data, axes[, signal_size])	Return a node which performs IRDFT operation.
is_finite(data[, name])	Performs element-wise mapping from NaN and Infinity to False.
is_inf(data[, attributes, name])	Return a node which performs IsInf operation.
is_nan(data[, name])	Performs element-wise mapping from NaN to True.
less(left_node, right_node[, ...])	Return node which checks if left input node is less than the right node element-wise.
less_equal(left_node, right_node[, ...])	Return node which checks if left input node is less or equal the right node element-wise.
log(node[, name])	Return node which applies natural logarithm to the input node element-wise.
log_softmax(data, axis[, name])	Apply LogSoftmax operation on each element of input tensor.
logical_and(left_node, right_node[, ...])	Return node which perform logical and operation on input nodes element-wise.
logical_not(node[, name])	Return node which applies element-wise logical negation to the input node.
logical_or(left_node, right_node[, ...])	Return node which performs logical OR operation on input nodes element-wise.
logical_xor(left_node, right_node[, ...])	Return node which performs logical XOR operation on input nodes element-wise.
loop(trip_count, execution_condition, ...[, ...])	Perform recurrent execution of the network described in the body, iterating through the data.
lrn(data, axes[, alpha, beta, bias, size, name])	Return a node which performs element-wise Local Response Normalization (LRN) operation.
lstm_cell(X, initial_hidden_state, ...[, ...])	Return a node which performs LSTMCell operation.
lstm_sequence(X, initial_hidden_state, ...)	Return a node which performs LSTMSequence operation.
matmul(data_a, data_b, transpose_a, transpose_b)	Return the Matrix Multiplication operation.
matrix_nms(boxes, scores[, ...])	Return a node which performs MatrixNms.
max_pool(data, strides, dilations, ...[, ...])	Perform max pooling operation and return both values and indices of the selected elements.
maximum(left_node, right_node[, ...])	Return node which applies the maximum operation to input nodes elementwise.
minimum(left_node, right_node[, ...])	Return node which applies the minimum operation to input nodes elementwise.
mish(data[, name])	Return a node which performs Mish.
mod(left_node, right_node[, auto_broadcast, ...])	Return node performing element-wise division reminder with two given tensors.
multiclass_nms(boxes, scores[, roisnum, ...])	Return a node which performs MulticlassNms.
multiply(left_node, right_node[, ...])	Return node which applies f(x) = A*B to the input nodes elementwise.
mvn(data, axes, normalize_variance, eps, ...)	Return a node which performs MeanVarianceNormalization (MVN).
negative(node[, name])	Return node which applies f(x) = -x to the input node elementwise.
non_max_suppression(boxes, scores[, ...])	Return a node which performs NonMaxSuppression.
non_zero(data[, output_type, name])	Return the indices of the elements that are non-zero.
normalize_l2(data, axes, eps, eps_mode[, name])	Construct an NormalizeL2 operation.
not_equal(left_node, right_node[, ...])	Return node which checks if input nodes are unequal element-wise.
nv12_to_bgr(arg[, arg_uv, name])	Return a node which performs NV12toBGR operation.
nv12_to_rgb(arg[, arg_uv, name])	Return a node which performs NV12toRGB operation.
one_hot(indices, depth, on_value, off_value, ...)	Create node performing one-hot encoding on input data.
pad(arg, pads_begin, pads_end, pad_mode[, ...])	Return a generic padding operation.
parameter(shape[, dtype, name])	Return an ngraph Parameter object.
power(left_node, right_node[, ...])	Return node which perform element-wise exponentiation operation.
prelu(data, slope[, name])	Perform Parametrized Relu operation element-wise on data from input node.
prior_box(layer_shape, image_shape, attrs[, ...])	Generate prior boxes of specified sizes and aspect ratios across all dimensions.
prior_box_clustered(output_size, image_size, ...)	Generate prior boxes of specified sizes normalized to the input image size.
proposal(class_probs, bbox_deltas, ...[, name])	Filter bounding boxes and outputs only those with the highest prediction confidence.
psroi_pooling(input, coords, output_dim, ...)	Return a node which produces a PSROIPooling operation.
random_uniform(output_shape, min_val, ...[, ...])	Return a node which generates sequence of random values from uniform distribution.
range(start, stop, step[, name])	Return a node which produces the Range operation.
rdft(data, axes[, signal_size])	Return a node which performs RDFT operation.
read_value(init_value, variable_id[, name])	Return a node which produces the Assign operation.
reduce_l1(node, reduction_axes[, keep_dims, ...])	L1-reduction operation on input tensor, eliminating the specified reduction axes.
reduce_l2(node, reduction_axes[, keep_dims, ...])	L2-reduction operation on input tensor, eliminating the specified reduction axes.
reduce_logical_and(node, reduction_axes[, ...])	Logical AND reduction operation on input tensor, eliminating the specified reduction axes.
reduce_logical_or(node, reduction_axes[, ...])	Logical OR reduction operation on input tensor, eliminating the specified reduction axes.
reduce_max(node, reduction_axes[, ...])	Max-reduction operation on input tensor, eliminating the specified reduction axes.
reduce_mean(node, reduction_axes[, ...])	Mean-reduction operation on input tensor, eliminating the specified reduction axes.
reduce_min(node, reduction_axes[, ...])	Min-reduction operation on input tensor, eliminating the specified reduction axes.
reduce_prod(node, reduction_axes[, ...])	Product-reduction operation on input tensor, eliminating the specified reduction axes.
reduce_sum(node, reduction_axes[, ...])	Perform element-wise sums of the input tensor, eliminating the specified reduction axes.
region_yolo(input, coords, classes, num, ...)	Return a node which produces the RegionYolo operation.
relu(node[, name])	Perform rectified linear unit operation on input node element-wise.
reorg_yolo(input, stride[, name])	Return a node which produces the ReorgYolo operation.
reshape(node, output_shape, special_zero[, name])	Return reshaped node according to provided parameters.
result(data[, name])	Return a node which represents an output of a graph (Function).
reverse_sequence(input, seq_lengths, ...[, name])	Return a node which produces a ReverseSequence operation.
rnn_cell(X, initial_hidden_state, W, R, B, ...)	Perform RNNCell operation on tensor from input node.
rnn_sequence(X, initial_hidden_state, ...[, ...])	Return a node which performs RNNSequence operation.
roi_align(data, rois, batch_indices, ...[, ...])	Return a node which performs ROIAlign operation.
roi_pooling(input, coords, output_size, ...)	Return a node which produces an ROIPooling operation.
roll(data, shift, axes)	Return a node which performs Roll operation.
round(data[, mode, name])	Apply Round operation on each element of input tensor.
scatter_elements_update(data, indices, ...)	Return a node which produces a ScatterElementsUpdate operation.
scatter_update(data, indices, updates, axis)	Return a node which produces a ScatterUpdate operation.
select(cond, then_node, else_node[, ...])	Perform an element-wise selection operation on input tensors.
selu(data, alpha, lambda_value[, name])	Perform a Scaled Exponential Linear Unit (SELU) operation on input node element-wise.
shape_of(data[, output_type, name])	Return a node which produces a tensor containing the shape of its input data.
shuffle_channels(data, axis, group[, name])	Perform permutation on data in the channel dimension of the input tensor.
sigmoid(data[, name])	Return a node which applies the sigmoid function element-wise.
sign(node[, name])	Perform element-wise sign operation.
sin(node[, name])	Apply sine function on the input node element-wise.
sinh(node[, name])	Apply hyperbolic sine function on the input node element-wise.
slice(data, start, stop, step[, axes, name])	Return a node which generates Slice operation.
softmax(data, axis[, name])	Apply softmax operation on each element of input tensor.
softplus(data[, name])	Apply SoftPlus operation on each element of input tensor.
softsign(node[, name])	Apply SoftSign operation on the input node element-wise.
space_to_batch(data, block_shape, ...[, name])	Perform SpaceToBatch operation on the input tensor.
space_to_depth(data, mode[, block_size, name])	Perform SpaceToDepth operation on the input tensor.
split(data, axis, num_splits[, name])	Return a node which splits the input tensor into same-length slices.
sqrt(node[, name])	Return node which applies square root to the input node element-wise.
squared_difference(x1, x2[, auto_broadcast, ...])	Perform an element-wise squared difference between two tensors.
squeeze(data, axes[, name])	Perform squeeze operation on input tensor.
strided_slice(data, begin, end, strides, ...)	Return a node which dynamically repeats(replicates) the input data tensor.
subtract(left_node, right_node[, ...])	Return node which applies f(x) = A-B to the input nodes element-wise.
swish(data[, beta, name])	Return a node which performing Swish activation function Swish(x, beta=1.0) = x * sigmoid(x * beta)).
tan(node[, name])	Apply tangent function on the input node element-wise.
tanh(node[, name])	Return node which applies hyperbolic tangent to the input node element-wise.
tensor_iterator(inputs, graph_body, ...[, name])	Perform recurrent execution of the network described in the body, iterating through the data.
tile(data, repeats[, name])	Return a node which dynamically repeats(replicates) the input data tensor.
topk(data, k, axis, mode, sort[, ...])	Return a node which performs TopK.
transpose(data, input_order[, name])	Return a node which transposes the data in the input tensor.
unique(data[, axis, sorted, ...])	Operator which selects and returns unique elements or unique slices of the input tensor.
unsqueeze(data, axes[, name])	Perform unsqueeze operation on input tensor.
variadic_split(data, axis, split_lengths[, name])	Return a node which splits the input tensor into variadic length slices.

Classes

Dimension	ngraph.impl.Dimension wraps ngraph::Dimension
Function	ngraph.impl.Function wraps ngraph::Function
Node	ngraph.impl.Node wraps ngraph::Node
PartialShape	ngraph.impl.PartialShape wraps ngraph::PartialShape