如何查看TensorRT默认支持的算子(operator)

• TensorRT7.0支持的ONNX算子列表

https://github.com/onnx/onnx-tensorrt/blob/84b5be1d6fc03564f2c0dba85a2ee75bad242c2e/oper
ators.md
 

Operator	Supported?	Restrictions
Abs	Y
Acos	Y
Acosh	Y
Add	Y
And	Y
ArgMax	Y
ArgMin	Y
Asin	Y
Asinh	Y
Atan	Y
Atanh	Y
AveragePool	Y	2D or 3D Pooling only
BatchNormalization	Y
BitShift	N
Cast	Y	Cast is only supported for TRT types
Ceil	Y
Clip	Y	min and max clip values must be an initializer
Compress	N
Concat	Y
ConcatFromSequence N
Constant	Y
ConstantOfShape	Y
Conv	Y	2D or 3D convolutions only
ConvInteger	N
ConvTranspose	Y	2D or 3D deconvolutions only. Weights must be an initializer
Cos	Y
Cosh	Y
CumSum	N
DepthToSpace	Y
DequantizeLinear	Y	Scales and zero-point value must be initializers
Det	N
Div	Y
Dropout	N
Elu	Y
Equal	Y
Erf	Y
Exp	Y
Expand	Y
EyeLike	N
Flatten	Y
Floor	Y
Gather	Y
GatherElements	N
GatherND	N
Gemm	Y
GlobalAveragePool	Y
GlobalLpPool	N
GlobalMaxPool	Y
Greater	Y
GRU	Y
HardSigmoid	Y
Hardmax	N
Identity	Y
If	N
ImageScaler	Y
InstanceNormalization	Y	Scales and biases must be an initializer
IsInf	N
IsNaN	N
LeakyRelu	Y
Less	Y
Log	Y
LogSoftmax	Y
Loop	Y
LRN	Y
LSTM	Y
LpNormalization	N
LpPool	N
MatMul	Y
MatMulInteger	N
Max	Y
MaxPool	Y
MaxRoiPool	N
MaxUnpool	N
Mean	Y
Min	Y
Mod	N
Mul	Y
Multinomial	N
Neg	Y
NonMaxSuppression	N
NonZero	N
Not	Y
OneHot	N
Or	Y
Pad	Y	Zero-padding on last 2 dimensions only
ParametricSoftplus	Y
Pow	Y
PRelu	Y
QLinearConv	N
QLinearMatMul	N
QuantizeLinear	Y	Scales and zero-point value must be initializers
RNN	N
RandomNormal	N
RandomNormalLike	N
RandomUniform	Y
RandomUniformLike	Y
Range	Y	Float inputs are only supported if start, limit and delta inputs are initializers
Reciprocal	N
ReduceL1	Y
ReduceL2	Y
ReduceLogSum	Y
ReduceLogSumExp	Y
ReduceMax	Y
ReduceMean	Y
ReduceMin	Y
ReduceProd	Y
ReduceSum	Y
ReduceSumSquare	Y
Relu	Y
Reshape	Y
Resize	Y	Asymmetric coordinate transformation mode only. Nearest or Linear resizing mode only. "floor" mode only for resize_mode attribute.
ReverseSequence	N
RNN	Y
RoiAlign	N
Round	N
ScaledTanh	Y
Scan	Y
Scatter	N
ScatterElements	N
ScatterND	N
Selu	Y
SequenceAt	N
SequenceConstruct	N
SequenceEmpty	N
SequenceErase	N
SequenceInsert	N
SequenceLength	N
Shape	Y
Shrink	N
Sigmoid	Y
Sign	N
Sin	Y
Sinh	Y
Size	Y
Slice	Y	Slice axes must be an initializer
Softmax	Y
Softplus	Y
Softsign	Y
SpaceToDepth	Y
Split	Y
SplitToSequence	N
Sqrt	Y
Squeeze	Y
StringNormalizer	N
Sub	Y
Sum	Y
Tan	Y
Tanh	Y
TfIdfVectorizer	N
ThresholdedRelu	Y
Tile	Y
TopK	Y
Transpose	Y
Unique	N
Unsqueeze	Y
Upsample	Y
Where	Y
Xor	N

• TensorRT8.2支持的ONNX算子列表

onnx-tensorrt/operators.md at main · onnx/onnx-tensorrt · GitHub

Operator	Supported	Supported Types	Restrictions
Abs	Y	FP32, FP16, INT32
Acos	Y	FP32, FP16
Acosh	Y	FP32, FP16
Add	Y	FP32, FP16, INT32
And	Y	BOOL
ArgMax	Y	FP32, FP16
ArgMin	Y	FP32, FP16
Asin	Y	FP32, FP16
Asinh	Y	FP32, FP16
Atan	Y	FP32, FP16
Atanh	Y	FP32, FP16
AveragePool	Y	FP32, FP16, INT8, INT32	2D or 3D Pooling only
BatchNormalization	Y	FP32, FP16
BitShift	N
Cast	Y	FP32, FP16, INT32, INT8, BOOL
Ceil	Y	FP32, FP16
Celu	Y	FP32, FP16
Clip	Y	FP32, FP16, INT8
Compress	N
Concat	Y	FP32, FP16, INT32, INT8, BOOL
ConcatFromSequence	N
Constant	Y	FP32, FP16, INT32, INT8, BOOL
ConstantOfShape	Y	FP32
Conv	Y	FP32, FP16, INT8	2D or 3D convolutions only. Weights W must be an initailizer
ConvInteger	N
ConvTranspose	Y	FP32, FP16, INT8	2D or 3D deconvolutions only. Weights W must be an initializer
Cos	Y	FP32, FP16
Cosh	Y	FP32, FP16
CumSum	Y	FP32, FP16	axis must be an initializer
DepthToSpace	Y	FP32, FP16, INT32
DequantizeLinear	Y	INT8	x_zero_point must be zero
Det	N
Div	Y	FP32, FP16, INT32
Dropout	Y	FP32, FP16
DynamicQuantizeLinear	N
Einsum	Y	FP32, FP16	Ellipsis and diagonal operations are not supported. Broadcasting between inputs is not supported
Elu	Y	FP32, FP16, INT8
Equal	Y	FP32, FP16, INT32
Erf	Y	FP32, FP16
Exp	Y	FP32, FP16
Expand	Y	FP32, FP16, INT32, BOOL
EyeLike	Y	FP32, FP16, INT32, BOOL
Flatten	Y	FP32, FP16, INT32, BOOL
Floor	Y	FP32, FP16
Gather	Y	FP32, FP16, INT8, INT32
GatherElements	Y	FP32, FP16, INT8, INT32
GatherND	Y	FP32, FP16, INT8, INT32
Gemm	Y	FP32, FP16, INT8
GlobalAveragePool	Y	FP32, FP16, INT8
GlobalLpPool	Y	FP32, FP16, INT8
GlobalMaxPool	Y	FP32, FP16, INT8
Greater	Y	FP32, FP16, INT32
GreaterOrEqual	Y	FP32, FP16, INT32
GRU	Y	FP32, FP16	For bidirectional GRUs, activation functions must be the same for both the forward and reverse pass
HardSigmoid	Y	FP32, FP16, INT8
Hardmax	N
Identity	Y	FP32, FP16, INT32, INT8, BOOL
If	Y	FP32, FP16, INT32, BOOL	Output tensors of the two conditional branches must have broadcastable shapes, and must have different names
ImageScaler	Y	FP32, FP16
InstanceNormalization	Y	FP32, FP16	Scales scale and biases B must be initializers. Input rank must be >=3 & <=5
IsInf	N
IsNaN	Y	FP32, FP16, INT32
LeakyRelu	Y	FP32, FP16, INT8
Less	Y	FP32, FP16, INT32
LessOrEqual	Y	FP32, FP16, INT32
Log	Y	FP32, FP16
LogSoftmax	Y	FP32, FP16
Loop	Y	FP32, FP16, INT32, BOOL
LRN	Y	FP32, FP16
LSTM	Y	FP32, FP16	For bidirectional LSTMs, activation functions must be the same for both the forward and reverse pass
LpNormalization	Y	FP32, FP16
LpPool	Y	FP32, FP16, INT8
MatMul	Y	FP32, FP16
MatMulInteger	N
Max	Y	FP32, FP16, INT32
MaxPool	Y	FP32, FP16, INT8	2D or 3D pooling only. Indices output tensor unsupported
MaxRoiPool	N
MaxUnpool	N
Mean	Y	FP32, FP16, INT32
MeanVarianceNormalization	N
Min	Y	FP32, FP16, INT32
Mod	N
Mul	Y	FP32, FP16, INT32
Multinomial	N
Neg	Y	FP32, FP16, INT32
NegativeLogLikelihoodLoss	N
NonMaxSuppression	Y [EXPERIMENTAL]	FP32, FP16	Inputs max_output_boxes_per_class, iou_threshold, and score_threshold must be initializers. Output has fixed shape and is padded to [max_output_boxes_per_class, 3].
NonZero	N
Not	Y	BOOL
OneHot	N
Or	Y	BOOL
Pad	Y	FP32, FP16, INT8, INT32
ParametricSoftplus	Y	FP32, FP16, INT8
Pow	Y	FP32, FP16
PRelu	Y	FP32, FP16, INT8
QLinearConv	N
QLinearMatMul	N
QuantizeLinear	Y	FP32, FP16	y_zero_point must be 0
RandomNormal	N
RandomNormalLike	N
RandomUniform	Y	FP32, FP16	seed value is ignored by TensorRT
RandomUniformLike	Y	FP32, FP16	seed value is ignored by TensorRT
Range	Y	FP32, FP16, INT32	Floating point inputs are only supported if start, limit, and delta inputs are initializers
Reciprocal	N