Skip to content
This repository has been archived by the owner on Nov 16, 2023. It is now read-only.

Latest commit

 

History

History
172 lines (168 loc) · 132 KB

operators.md

File metadata and controls

172 lines (168 loc) · 132 KB

Operators Support Table

The following table shows ai.onnx operators from which onnx opset version are currently supported by onnxjs. For example, 4-6, 8+ means ONNX.js currently support opset version 4 to 6, 8 and above.

See Compatibility for a list of the supported platforms.

This file is automatically generated from the def files via this script. Do not modify directly.

Operator Cpu Backend Wasm Backend WebGl Backend
Abs 6+ 6+
Acos 7+ 7+
Acosh 9+
Add 7+ 7+ 7+
And 7+ 7+ 7+
ArgMax 1-10, 11
ArgMin
Asin 7+ 7+
Asinh 9+
Atan 7+ 7+
Atanh 9+
AveragePool 7-9, 10 7-9, 10 7-9, 10
BatchNormalization 7-8, 9+ 7-8, 9+ 7-8, 9+
BitShift
Cast 6-8, 9+
Ceil 6+ 6+
Celu
Clip 6-10 6-10 6-10
Compress
Concat 4-10, 11+ 4-10, 11+
ConcatFromSequence
Constant
ConstantOfShape
Conv 1-10, 11+ 1-10, 11+ 1-10, 11+
ConvInteger
ConvTranspose
Cos 7+ 7+
Cosh 9+
CumSum
DepthToSpace
DequantizeLinear
Det
Div 7+ 7+ 7+
Dropout 7-9, 10-11, 12+ 7-9, 10-11, 12+
DynamicQuantizeLinear
Einsum
Elu 6+ 6+
Equal 7-10, 11+
Erf
Exp 6+ 6+
Expand 8+
EyeLike
Flatten 1-8, 9-10, 11+ 1-8, 9-10, 11+
Floor 6+ 6+
GRU
Gather 1-10, 11+ 1-10, 11+
GatherElements
GatherND
Gemm 7-8, 9-10, 11+ 7-8, 9-10, 11+ 7-8, 9-10, 11+
GlobalAveragePool 1+ 1+ 1+
GlobalLpPool
GlobalMaxPool 1+ 1+ 1+
Greater 7-8, 9+
GreaterOrEqual
HardSigmoid
Hardmax
Identity 1+
If
InstanceNormalization 6+ 6+ 6+
IsInf
IsNaN 9+
LRN 1+
LSTM
LeakyRelu 6+ 6+
Less 7-8, 9+ 7-8, 9+
LessOrEqual
Log 6+ 6+
LogSoftmax
Loop
LpNormalization
LpPool
MatMul 1-8, 9+ 1-8, 9+ 1-8, 9+
MatMulInteger
Max
MaxPool 1-7, 8-9 1-7, 8-9 1-7, 8-9
MaxRoiPool
MaxUnpool
Mean
MeanVarianceNormalization
Min
Mod
Mul 7+ 7+ 7+
Multinomial
Neg 6+ 6+
NegativeLogLikelihoodLoss
NonMaxSuppression
NonZero
Not 1+ 1+
OneHot
Or 7+ 7+ 7+
PRelu 7-8, 9+ 7-8, 9+ 7-8, 9+
Pad 2-10 2-10
Pow 7-11, 12+ 7-11, 12+
QLinearConv
QLinearMatMul
QuantizeLinear
RNN
RandomNormal
RandomNormalLike
RandomUniform
RandomUniformLike
Range
Reciprocal 6+
ReduceL1
ReduceL2
ReduceLogSum 1-10, 11+ 1-10, 11+
ReduceLogSumExp
ReduceMax 1-10, 11, 12+ 1-10, 11, 12+
ReduceMean 1-10, 11+ 1-10, 11+
ReduceMin 1-10, 11, 12+ 1-10, 11, 12+
ReduceProd 1-10, 11+ 1-10, 11+
ReduceSum 1-10, 11+ 1-10, 11+
ReduceSumSquare 1-10, 11+ 1-10, 11+
Relu 6+ 6+
Reshape 5+ 5+
Resize 10, 11+
ReverseSequence
RoiAlign
Round
Scan
Scatter
ScatterElements
ScatterND
Selu
SequenceAt
SequenceConstruct
SequenceEmpty
SequenceErase
SequenceInsert
SequenceLength
Shape 1+
Shrink
Sigmoid 6+ 6+
Sign 9+
Sin 7+ 7+
Sinh 9+
Size
Slice 1-9, 10, 11+ 1-9, 10, 11+
Softmax 1-10, 11+ 1-10, 11+ 1-10, 11+
SoftmaxCrossEntropyLoss
Softplus
Softsign
SpaceToDepth
Split 2-10, 11+
SplitToSequence
Sqrt 6+ 6+
Squeeze 1-10, 11+ 1-10, 11+
StringNormalizer
Sub 7+ 7+ 7+
Sum 6-7, 8+ 6-7, 8+ 6-7, 8+
Tan 7+ 7+
Tanh 6+ 6+
TfIdfVectorizer
ThresholdedRelu
Tile 6+ 6+
TopK
Transpose 1+ 1+
Unique
Unsqueeze 1-10, 11+ 1-10, 11+
Upsample 7-8, 9 7-8, 9
Where
Xor 7+ 7+ 7+