* Added RandomUniform specification. * Extended description of obtaining float numbers. * Added description of obtainin float16 and double values. * Added description of xor symbol. * Small correction. * Small corrections. * Added initial type attribute. * Corrected IR example, removed unnecessary default values. * Small correction. * Added information about type cast between initial and output types. * Removed initial type attribute. * Corrected output_type description. * Corrected minval, maxval description. Corrected IR example. * Apply suggestions from code review Co-authored-by: Tatiana Savina <tatiana.savina@intel.com> * Removed unnecessary paper link. Co-authored-by: Tatiana Savina <tatiana.savina@intel.com>
7.1 KiB
7.1 KiB
Operation Set opset8 Specification
This specification document describes the opset8 operation set supported in OpenVINO™.
Support for each particular operation from the list below depends on the capabilities of an inference plugin
and may vary among different hardware platforms and devices. Examples of operation instances are provided as IR V10 xml
snippets. Such IR is generated by the Model Optimizer. The semantics match corresponding nGraph operation classes
declared in namespace opset8.
Table of Contents
- Abs
- Acos
- Acosh
- AdaptiveAvgPool
- AdaptiveMaxPool
- Add
- Asin
- Asinh
- Assign
- Atan
- Atanh
- AvgPool
- BatchNormInference
- BatchToSpace
- BinaryConvolution
- Broadcast
- Bucketize
- CTCGreedyDecoder
- CTCGreedyDecoderSeqLen
- CTCLoss
- Ceiling
- Clamp
- Concat
- Constant
- Convert
- ConvertLike
- Convolution
- ConvolutionBackpropData
- Cos
- Cosh
- CumSum
- DeformableConvolution
- DeformablePSROIPooling
- DepthToSpace
- DetectionOutput
- DFT
- Divide
- Einsum
- Elu
- EmbeddingBagOffsetsSum
- EmbeddingBagPackedSum
- EmbeddingSegmentsSum
- Equal
- Erf
- Exp
- ExperimentalDetectronDetectionOutput_6
- ExperimentalDetectronGenerateProposalsSingleImage_6
- ExperimentalDetectronPriorGridGenerator_6
- ExperimentalDetectronROIFeatureExtractor_6
- ExperimentalDetectronTopKROIs_6
- ExtractImagePatches
- FakeQuantize
- Floor
- FloorMod
- Gather
- GatherElements
- GatherND_5
- GatherTree
- Gelu
- Greater
- GreaterEqual
- GRN
- GroupConvolution
- GroupConvolutionBackpropData
- GRUCell
- GRUSequence
- HardSigmoid
- HSigmoid
- HSwish
- IDFT
- If
- Interpolate
- Less
- LessEqual
- Log
- LogicalAnd
- LogicalNot
- LogicalOr
- LogicalXor
- LogSoftmax
- Loop
- LRN
- LSTMCell
- LSTMSequence
- MatMul
- MaxPool
- Maximum
- Minimum
- Mish
- Mod
- MVN
- Multiply
- Negative
- NonMaxSuppression
- NonZero
- NormalizeL2
- NotEqual
- OneHot
- Pad
- Parameter
- Power
- PReLU
- PriorBoxClustered
- PriorBox
- Proposal
- PSROIPooling
- RandomUniform
- Range
- ReLU
- ReadValue
- ReduceL1
- ReduceL2
- ReduceLogicalAnd
- ReduceLogicalOr
- ReduceMax
- ReduceMean
- ReduceMin
- ReduceProd
- ReduceSum
- RegionYolo
- ReorgYolo
- Reshape
- Result
- ReverseSequence
- RNNCell
- RNNSequence
- ROIAlign
- ROIPooling
- Roll
- Round
- ScatterElementsUpdate
- ScatterNDUpdate
- ScatterUpdate
- Select
- Selu
- ShapeOf
- ShuffleChannels
- Sigmoid
- Sign
- Sin
- Sinh
- SoftMax
- SoftPlus
- SpaceToBatch
- SpaceToDepth
- Split
- Sqrt
- SquaredDifference
- Squeeze
- StridedSlice
- Subtract
- Swish
- Tan
- Tanh
- TensorIterator
- Tile
- TopK
- Transpose
- Unsqueeze
- VariadicSplit