[doc] update supported nnapi ep op md (#12273)

This commit is contained in:
Rachel Guo 2022-07-27 15:49:37 -07:00 committed by GitHub
parent b65d8fe405
commit 57a545e056
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23

View file

@ -102,12 +102,14 @@ Following ops are supported by the NNAPI Execution Provider,
|ai.onnx:Clip||
|ai.onnx:Concat||
|ai.onnx:Conv|Only 2D Conv is supported.<br/>Weights and bias should be constant.|
|ai.onnx:DepthToSpace|Only DCR mode DepthToSpace is supported.|
|ai.onnx:DequantizeLinear|All quantization scales and zero points should be constant.|
|ai.onnx:Div||
|ai.onnx:Elu||
|ai.onnx:Exp||
|ai.onnx:Flatten||
|ai.onnx:Floor||
|ai.onnx:Gather|Input indices should be constant if not int32 type.|
|ai.onnx:Gemm|If input B is not constant, transB should be 1.|
|ai.onnx:GlobalAveragePool|Only 2D Pool is supported.|
|ai.onnx:GlobalMaxPool|Only 2D Pool is supported.|
@ -119,7 +121,9 @@ Following ops are supported by the NNAPI Execution Provider,
|ai.onnx:Min||
|ai.onnx:Mul||
|ai.onnx:Neg||
|ai.onnx:Pad|Only constant mode Pad is supported.<br/>Input pads and constant_value should be constant.<br/>Input pads values should be non-negative.|
|ai.onnx:Pow||
|ai.onnx:PRelu||
|ai.onnx:QLinearConv|Only 2D Conv is supported.<br/>Weights and bias should be constant.<br/>All quantization scales and zero points should be constant.|
|ai.onnx:QLinearMatMul|All quantization scales and zero points should be constant.|
|ai.onnx:QuantizeLinear|All quantization scales and zero points should be constant.|
@ -131,10 +135,11 @@ Following ops are supported by the NNAPI Execution Provider,
|ai.onnx:Slice||
|ai.onnx:Softmax||
|ai.onnx:Sqrt||
|ai.onnx:Squeeze||
|ai.onnx:Squeeze|Input axes should be constant.|
|ai.onnx:Sub||
|ai.onnx:Tanh||
|ai.onnx:Transpose||
|ai.onnx:Unsqueeze|Input axes should be constant.|
|com.microsoft:QLinearAdd|All quantization scales and zero points should be constant.|
|com.microsoft:QLinearAveragePool|Only 2D Pool is supported.<br/>All quantization scales and zero points should be constant.|
|com.microsoft:QLinearSigmoid|All quantization scales and zero points should be constant.|