* Reorganize main function.
* Follow review comments.
* Emit constants are globals in Krnl and LLVM dialects.
* Replace internal malloc with memory pool and getref instruction.
* Lower krnl.getref to LLVM.
* Fix formatting issues.
* Add tests.
* Add missing dependency.
* Improve LLVM lowering.
* Add test to show getref is generic.
* 1. Add shape inference for the following ops:
- Atan
- Tan
- Sin
- Cast
- ConvTranspose
- Flatten
- DynamicQuantizeLinear
- QuantizeLinear
- DequantizeLinear
- ConvInteger
2. Import attributes for generic nodes
3. Fixes for cases where .cast<> should be .isa<> (ONNXConcat::inferShapes)
* Fix foormatting issues
* Address comments:
- SmallVector<> * -> SmallVectorImpl<> &
- switch-case -> helper function
- Inside helper function, preserve signed-ness
- add TODOs
* Can't use signed integers yet in convertONNXTypeToMLIRType, add TODO
Co-authored-by: Tian Jin <tjingrant@gmail.com>
* initial const prop attempt
* added support for broadcast ops
* adde all binary broadcast ops into custom builders with precise type
* added test example
* working
* format
* fixed suggestion by Tung, start woring on unary
* added subtraction and neg the right way, and added elementwise mul too
* formatting changes
* format
* format
* added instructions to add new optimizations
* enable promote attr for pad
* use optional arguments for pad
* shape infereance for pad
* Lowering Pad
* format file
* use DenseTensor for the attribute
* use Pad in ONNXRewrite
* fix the merge conflict
* fix the attr given to constantOp
* handle ONNXConstantOp in attribute promotion
* Fix bug when AttributePromotion is called more than once
* update ONNXOps.td.inc with correct version of onnx
* update onnx.md
* responses to review
* fix the build error
* change the implementation of Pad
* delete commented out code
* clang format
Co-authored-by: Tian Jin <tjingrant@gmail.com>
* Use AffineMap
* Shared AffineMap
* AffineMap for Conv/Pooling
* Create helper files
* Remove changes for Relu
* Remove redundant includes
* Use AffineMap for AveragePool's shape inference
* Add MLIR tests for unknown dimension case
* Extract a method AffineMapIntConstant
* Comment stylist and include path
Co-authored-by: Gheorghe-Teodor Bercea <gt.bercea@gmail.com>
* Specialize the op lowering logic for elementwise operations
* Fix clang-format error.
* Update tests for LSTM since LSTM uses element-wise ops
Co-authored-by: Tian Jin <tjingrant@gmail.com>
* Support dilations and enable e2e tests
* Fix allocating memory for dynamic shape
* Edit comments
* Do dilation by computing an offset from kernel index
* Correct dilation formula, add an example of out-of-bound, and add a test for dilation
* Import optional outputs as NoneType
* Shape inference for ONNXLSTM
* Edit ONNXLSTM::inferShape()
* Shape inference for ONNXLSTMOp
* Create a common function for inferring shape for RNN ops
* CheckInsertDeallocation for a specific result
* Allocate memory for LSTM
* First round of lowering
* Allocate memory for hidden and cell states
* Test with custom Tanh
* Fix an error in Ct's formula
* Add E2E tests
* Return outputs
* Refactor the code
* Enable E2E tests
* Support reverse and bidirectional directions
* Minor revision
* Return all intermediate hidden states
* Call existing activation functions
* Structs for activation functions
* Call existing activations in ONNX
* Minor revision
* Compare strings ignoring case
* Use memreftype of rank 0 for calling activation functions
* Fix getActivationPack()
* Revise the code
* Add one MLIR test
* Add MLIR tests for reverse and bidirectional modes
* Make the order of emiting instructions deterministic
* Use OperandAdaptor instead of directly use an operand index
* Use literal assignments
* Change some variable names
* Use literal assignments
* Use literal assignments
* Format the code
Co-authored-by: Gheorghe-Teodor Bercea <gt.bercea@gmail.com>
* Implement shape inference for SplitOp
* Change spitOpt to SplitAttribute and check the axis range before updating the axis attribute
Co-authored-by: Gheorghe-Teodor Bercea <gt.bercea@gmail.com>
* Reorganize main function.
* Follow review comments.
* Emit constants are globals in Krnl and LLVM dialects.
* Output of non-value constants. Write full source to file.
* Fix e2e tests.
* Output constant free and full code in separate files.
* Emit separate files.
* Move file output management to utils.
* Elide the values of glotbal krnl constants.
* Add dual file output for Basic flag.
* Add tests.
* Add passes to cmake file.
* Create a template for pooling and add support for AveragePool
* Edit MLIR tests for MaxPool according to the new lowering template for pooling
* Dealloc temporary variables
* Support count_include_pad for AveragePool
* Add MLIR tests for AveragePool lowering
* Make changes according to Tian's comments
* Push AffineMap as upper bound for KrnlIterateOp
* Test AffineMap to use in Pooling
* Replace the old implementaion by a new one using AffineMap
* Fix the computation when dilations are non-unit
* Clean up the old code
* Remove AveragePool from Canonicalization pass
* Fix computing the end indices of a filter window
* Refactor the code for pooling
* Revise pushAffineMapBound
* Add MLIR tests
* Remove unused functions
* Fix check-onnx-backend build on x86 Linux. (#91)
* Add the split marker to test files (#90)
Co-authored-by: Tian Jin <tjingrant@gmail.com>
Co-authored-by: gongsu832 <gong_su@hotmail.com>
Co-authored-by: Tian Jin <tjingrant@gmail.com>
* implement shape inference for concat
* better checking of axis being concatenated: constant values only
* lowering of Concat with lit and backend tests
* fixes
Co-authored-by: Gheorghe-Teodor Bercea <gt.bercea@gmail.com>
* Support dilations and enable e2e tests
* Fix allocating memory for dynamic shape
* Edit comments
* Do dilation by computing an offset from kernel index
* Correct dilation formula, add an example of out-of-bound, and add a test for dilation
Co-authored-by: Gheorghe-Teodor Bercea <gt.bercea@gmail.com>
* Fix reshape when a dynamic shape is given.
* Fix default attributes for ConvNoBias.
* Fix comment.
* Resolve comment.
* Improve checks.
* Handle zero dim case.
* Add helper to fetch constants. Add test for dynamic reshape.
* Add test for zero.
* Use shortcut method for size.
* Support Pads for MaxPoolSingleOut
* Regenerate onnx.md to include the new op
* Edit comments
* Undo redundant parts that were unintentionally changed
* Move declarative rewriting rules into canonicalize to avoid creating a new op
* Reformat the rewriting rule pattern of MaxPoolSingleOut
* Put ONNXPadConstantValuePadOp's build method into a .cpp file instead of a tablegen file
* Use the same helper function as the one in inferShape for the ONNXPadConstantValuePadOp's build method
* Change function names and fix padding for the spatial dimensions
* Call shape-inference again after canonicalization to infer shape for newly added ops during canonicalization.
* Fix typos
* Lower MaxPoolSingleOutOp to Krnl dialect
* Edit comments
* Update changes according to the new folder structure
* Add MLIR tests
* Support ceil_mode
* Merge the first two krnl loops into one krnl loop; remove attribute checks
* Dynamically allocate memory for the result if the result has unknown dimensions
Co-authored-by: Gheorghe-Teodor Bercea <gt.bercea@gmail.com>
* Remove rank constraints in gemm fusion
* Add an MLIR test
Co-authored-by: Tian Jin <tjingrant@gmail.com>
Co-authored-by: Gheorghe-Teodor Bercea <gt.bercea@gmail.com>
* 1. Combine variadicIn/Out with expectedNumOperands/Results to simplify import function arguments.
2. Generic improvements to code readability in gen_doc.py.
* Update ONNX Dialect doc.
* Remove redundant code in ImportNode.
* Prettify op_build_table.inc.
* 1. Remove irrelevant code in gen_doc.py
* Refactor code to be more readable.
* Further refactoring for readability improvements.
* Allow gemm to have an optional operand (bias term), and include an example of declarative optimization pattern targeting gemm with bias term ommitted.
* Make shape inference/lowering of gemm op compatible with optional operand declaration.
* Apply canonicalization again after lowering from onnx -> std dialects.
* Make hasBias compatible with the situation of GemmNoBias op.
* Update doc.
* Add a canonicalization test.
* Remove special handler for importing Gemm op, as it's redundant now.
* Add ONNXBatchNormalizationTestModeOp and its shape inference
* Lower batchnormalization test mode
* re-use scale, bias, mean, and variance
* Add MLIR tests
* Add e2e tests
* fix typos
* Fix a bug in MLIR tests
* Change type from int to int64_t for indices
* Uncomment e2e tests due to segmentation fault
* Uncomment e2e tests due to segmentation fault
* Revise the code
* [Tian] Fix segmentation fault in e2e tests
* Re-generate onnx.md to include BatchNormalizationTestModeOp
* Reverse an unintentional change
* Fix some typos in comments
* Use convertToMemRefType from the master branch
Co-authored-by: Gheorghe-Teodor Bercea <gt.bercea@gmail.com>
Gheorghe-Teodor Bercea
Aman LaChapelle
Alexandre Eichenberger
Tung D. Le
chentong319
Gheorghe-Teodor Bercea
Tian Jin