Go to file
Tian Jin b63b6bfd99 Limit build parallelism because CI is running out of memory. 2019-12-30 22:57:51 -05:00
.buildbot Exit Bash Script on Error (#381) 2019-12-21 00:50:30 -05:00
.circleci Limit build parallelism because CI is running out of memory. 2019-12-30 22:57:51 -05:00
src Transition to value-typed Value, rename Value* -> Value, in accordance with upstream MLIR style change. 2019-12-30 22:42:13 -05:00
test revert unnecessary changes 2019-12-24 03:43:47 -05:00
third_party Enable building of first commit. 2019-12-19 14:24:37 -05:00
.clang-format clean up 2019-12-21 02:07:24 -05:00
.gitignore Initial commit 2019-12-18 10:18:14 -05:00
.gitmodules Enable building of first commit. 2019-12-19 14:24:37 -05:00
CMakeLists.txt flatten src directory structure 2019-12-23 00:13:52 -05:00
LICENSE Initial commit 2019-12-18 10:18:14 -05:00
MLIR.cmake Update CMake build script. 2019-12-30 22:41:37 -05:00
README.md nit 2019-12-24 04:01:53 -05:00
SharingWork.md [MLIR] Add broadcasting support for element wise operations (#398) 2019-12-21 02:08:27 -05:00

README.md

ONNF

Open Neural Network Frontend : an ONNX frontend for MLIR.

CircleCI

Installation

We assume an existing installation of MLIR. The LLVM-Project repo commit hash we used to test against is 9b6ad8466bb8b97082b705270603ad7f4559e931 and the MLIR repo commit hash we used is 0710266d0f56cf6ab0f437badbd7416b6cecdf5f.

Two environment variables need to be set:

  • LLVM_SRC should point to the llvm src directory (e.g., llvm-project/llvm).
  • LLVM_BUILD should point to the llvm build directory (e.g., llvm-project/build).

To build ONNF, use the following command:

git clone --recursive git@github.com:clang-ykt/ONNF.git
mkdir build
cd build
cmake ..
cmake --build . --target all

After the above commands succeed, an onnf executable should appear in the bin directory.

Using ONNF

The usage of onnf is as such:

OVERVIEW: ONNF MLIR modular optimizer driver

USAGE: onnf [options] <input file>

OPTIONS:

Generic Options:

  --help        - Display available options (--help-hidden for more)
  --help-list   - Display list of available options (--help-list-hidden for more)
  --version     - Display the version of this program

ONNF Options:
These are frontend options.

  Choose target to emit:
      --EmitONNXIR - Ingest ONNX and emit corresponding ONNX dialect.
      --EmitMLIR   - Lower model to MLIR built-in transformation dialect.
      --EmitLLVMIR - Lower model to LLVM IR (LLVM dialect).
      --EmitLLVMBC - Lower model to LLVM IR and emit (to file) LLVM bitcode for model.

Example

For example, to lower an ONNX model (e.g., add.onnx) to ONNX dialect, use the following command:

./onnf --EmitONNXIR add.onnx

The output should look like:

module {
  func @main_graph(%arg0: tensor<10x10x10xf32>, %arg1: tensor<10x10x10xf32>) -> tensor<10x10x10xf32> {
    %0 = "onnx.Add"(%arg0, %arg1) : (tensor<10x10x10xf32>, tensor<10x10x10xf32>) -> tensor<10x10x10xf32>
    return %0 : tensor<10x10x10xf32>
  }
}