Document supported ops

MLIR.cmake

@@ -73,6 +73,22 @@ function(add_onnx_mlir_dialect_doc dialect dialect_tablegen_file)
 endfunction()
 add_custom_target(onnx-mlir-docs)
 
+# Create the list of supported ops. Pass the input file to scan, and the target architecture.
+# Target will create a docs/SupportedONNXOps-<arch>.md file listed
+# Useful options are "--todo", "--unsupported". Check python documentOps.py -h for more info.
+function(add_onnx_mlir_supported_ops input_file arch)
+  set(GEN_DOC_FILE ${ONNX_MLIR_SRC_ROOT}/docs/SupportedONNXOps-${arch}.md)
+  set(supported_ops_cmd ${Python3_EXECUTABLE} ${ONNX_MLIR_SRC_ROOT}/utils/documentOps.py --arch ${arch} -i ${input_file} -p ${ONNX_MLIR_SRC_ROOT}/utils)
+  add_custom_command(
+    OUTPUT ${GEN_DOC_FILE} 
+    COMMAND ${supported_ops_cmd}  --todo --unsupported > ${GEN_DOC_FILE}
+    DEPENDS ${input_file})
+  add_custom_target(onnx_mlir_supported_ops_${arch} DEPENDS ${GEN_DOC_FILE})
+  add_dependencies(onnx_mlir_supported_ops onnx_mlir_supported_ops_${arch})
+endfunction()
+add_custom_target(onnx_mlir_supported_ops)
+set_target_properties(onnx_mlir_supported_ops PROPERTIES EXCLUDE_FROM_DEFAULT_BUILD ON)
+
 # If an extra parameter, the dialect name, is provided,
 # this function will generate dialect and type from the td file
 function(add_onnx_mlir_dialect dialect dialect_name)

docs/DocumentList.md

@@ -3,6 +3,9 @@
 # Index of documents
 This document serves as an index for onnx-mlir documents.
 
+# Supported ONNX Ops
+* CPU support is covered [here](SupportedONNXOps-cpu.md).
+
 # Working environment
 * Installation is covered by [README.md](../README.md).
 * [Workflow.md](Workflow.md) describes how to contribute in github environment.

docs/SupportedONNXOps-cpu.md

@@ -0,0 +1,164 @@
+<!--- Automatically generated, do not edit. -->
+<!--- python documentOps.py --arch cpu --input /workdir/onnx-mlir/test/backend/inference_backend.py --path /workdir/onnx-mlir/utils --todo --unsupported -->
+
+# Supported ONNX Operation for Target *cpu*.
+
+Onnx-mlir currently support ONNX operations targeting up to opset 16. Limitations are listed when applicable.
+
+| Op |Opset |Limitations |Todo |
+| --- |--- |--- |--- |
+| **Abs** |13 | | |
+| **Acos** |7 | | |
+| **Acosh** |9 | | |
+| **Adagrad** |1 | | |
+| **Adam** |1 | | |
+| **Add** |14 |No support for short integers. | |
+| **And** |7 | | |
+| **ArgMax** |13 | | |
+| **ArgMin** |13 | | |
+| **Asin** |7 | | |
+| **Asinh** |9 | | |
+| **Atan** |7 | | |
+| **Atanh** |9 | | |
+| **AveragePool** |11 | | |
+| **BatchNormalization** |15 |Training not supported. | |
+| **Bernoulli** |15 | | |
+| **Bitshift** |unsupported | | |
+| **Cast** |13 |Cast only between float and double types. | |
+| **CastLike** |15 | | |
+| **Ceil** |13 | | |
+| **Celu** |12 | | |
+| **Clip** |13, 12, 11, 6 |No support for short integers. | |
+| **Compress** |11 | | |
+| **Concat** |13 | | |
+| **Constant** |13 | | |
+| **ConstantOfShape** |9 | | |
+| **Conv** |11 | | |
+| **ConvInteger** |10 | | |
+| **ConvTranspose** |11 | | |
+| **Cos** |7 | | |
+| **Cosh** |9 | | |
+| **CumSum** |14 | | |
+| **DepthToSpace** |13 | | |
+| **DequatizeLinear** |unsupported | | |
+| **Det** |11 | | |
+| **Div** |14 |No support for short integers. | |
+| **Dropout** |13 |Does not support masked and training. | |
+| **DynamicQuantizeLinear** |11 | | |
+| **EinSum** |unsupported | | |
+| **Elu** |6 | | |
+| **Equal** |13 | | |
+| **Erf** |13 | | |
+| **Exp** |13 | | |
+| **Expand** |13 | | |
+| **Eyelike** |unsupported | | |
+| **Flatten** |13 | | |
+| **Floor** |13 | | |
+| **GRU** |14 |Batchwise test is not supported. | |
+| **Gather** |13 | | |
+| **GatherElements** |13 | | |
+| **GatherND** |13 | | |
+| **Gemm** |13 | | |
+| **GlobalAveragePool** |1 | | |
+| **GlobalMaxPool** |1 | | |
+| **Greater** |13 | | |
+| **GreaterOrEqual** |16 | | |
+| **GridSample** |16 | | |
+| **HardSigmoid** |6 | | |
+| **HardSwish** |14 | | |
+| **Hardmax** |13 | | |
+| **Identity** |16 |Sequence identity not supported. | |
+| **If** |16 | | |
+| **InstanceNormalization** |6 | | |
+| **IsInf** |10 | | |
+| **IsNan** |unsupported | | |
+| **LRN** |13 | | |
+| **LSTM** |14 |No support for batchwise examples. | |
+| **LeakyRelu** |16 | | |
+| **Less** |13 | | |
+| **LessOrEqual** |16 | | |
+| **Log** |13 | | |
+| **LogSoftmax** |13 |Axis 0, 1, and default currently disabled due to changes in ONNX 1.8.1/Opset 13. |Temporally removed due to changes in onnx 1.8.1. |
+| **Loop** |16 |No support for opset 13 and 16 at this time. | |
+| **MatMul** |13 | | |
+| **MatMulInteger** |10 | | |
+| **Max** |13 |No support for short floats and unsigned int. | |
+| **MaxPool** |12 |Does not support argmax and short ints. Support single output only. | |
+| **MaxUnpool** |11 | | |
+| **Mean** |13 | | |
+| **MeanVarianceNormalization** |13 | | |
+| **Min** |13 |Does not support short floats and unsigned numbers. | |
+| **Mod** |13 |Support float and double only. | |
+| **Momentum** |1 | | |
+| **Mul** |14 |Does not support short integers. | |
+| **Neg** |13 | | |
+| **NegativeLogLikelihoodLoss** |13 | | |
+| **NonMaxSuppression** |11 | | |
+| **NonZero** |13 | | |
+| **Not** |1 | | |
+| **OneHot** |11 | | |
+| **OptionalGetElement** |15 | | |
+| **OptionalHasElement** |15 | | |
+| **Or** |7 | | |
+| **PRelu** |16 | | |
+| **Pad** |13, 11, 2 | | |
+| **Pow** |15 |No support for power with integer types. | |
+| **QLinearConv** |10 | | |
+| **QLinearMatmul** |unsupported | | |
+| **QuantizeLinear** |13 | | |
+| **RNN** |14 |Batchwise not supported. | |
+| **Range** |11 | | |
+| **Reciprocal** |13 | | |
+| **ReduceL1** |13 | | |
+| **ReduceL2** |13 | | |
+| **ReduceLogSum** |13 | | |
+| **ReduceLogSumExp** |13 | | |
+| **ReduceMax** |13 | | |
+| **ReduceMean** |13 | | |
+| **ReduceMin** |13 | | |
+| **ReduceProd** |13 | | |
+| **ReduceSum** |13, 11 |Default axis and do_not_keep_dim not supported. |Default axis and do_not_keep_dim temporarily removed due to changes in onnx 1.8.1. |
+| **ReduceSumSquare** |13 | | |
+| **Relu** |14 | | |
+| **Reshape** |14 | | |
+| **Resize** |13, 11, 10 |Missing support for linear, cubic, crop, pytorch_half_pixel, and floor. | |
+| **ReverseSequence** |10 | | |
+| **RoiAlign** |16 | | |
+| **Round** |11 | | |
+| **Scan** |16 |Does not support dynamic shapes. |Precision issue with newer opset, maybe just unsupported. Dynamic shape?. |
+| **ScatterElements** |16 |Does not support duplicate indices. | |
+| **ScatterND** |16 |Does not support scatternd add/multiply. | |
+| **Selu** |6 | | |
+| **SequenceInsert** |11 | | |
+| **Shape** |15 | | |
+| **Shrink** |9 | | |
+| **Sigmoid** |13 | | |
+| **Sign** |13 | | |
+| **Sin** |7 | | |
+| **Sinh** |9 | | |
+| **Size** |13 | | |
+| **Slice** |13 |Axis must be a constant argument. |Add tests to slices, currently have none. |
+| **Softmax** |13 | | |
+| **SoftmaxCrossEntropyLoss** |13 | | |
+| **Softplus** |1 | | |
+| **Softsign** |1 | | |
+| **SpaceToDepth** |13 | |Example works, the other is imprecise. To investigate. |
+| **Split** |13, 11 |Does not support static and dynamic shape, zero size splits. |Temporally removed due to changes in onnx 1.8.1. |
+| **Sqrt** |13 | | |
+| **Squeeze** |13, 11 |Does not support static and dynamic shape. |Temporally removed due to changes in onnx 1.8.1. |
+| **StrNormalizer** |unsupported | | |
+| **Sub** |14 |Does not support short integers. | |
+| **Sum** |13 | | |
+| **Tan** |7 | | |
+| **Tanh** |13 | | |
+| **TfdfVectorizer** |unsupported | | |
+| **ThresholdRelu** |unsupported | | |
+| **Tile** |13 | | |
+| **TopK** |11 | | |
+| **Transpose** |13 | | |
+| **Trilu** |14 | | |
+| **Unique** |11 | | |
+| **Unsqueeze** |13, 11 |Does not support static and dynamic shape. |Temporally removed due to changes in onnx 1.8.1. |
+| **Upsample** |10, 9, 7 | | |
+| **Where** |16 | | |
+| **Xor** |7 | | |