daisytuner / docc / 20738692415

67%

push

github

Add minimal Torch-MLIR integration for inference with Python bindings and tests (#425)

* Initial plan

* Add PyTorch/ONNX/Torch-MLIR integration stubs with nn, optim, and training modules

Co-authored-by: lukastruemper <31564409+lukastruemper@users.noreply.github.com>

* Add documentation and example for PyTorch integration

Co-authored-by: lukastruemper <31564409+lukastruemper@users.noreply.github.com>

* Add implementation summary documenting all PyTorch integration changes

Co-authored-by: lukastruemper <31564409+lukastruemper@users.noreply.github.com>

* Simplify to minimal inference-only integration (120 LOC vs 1547)

Co-authored-by: lukastruemper <31564409+lukastruemper@users.noreply.github.com>

* Add MLIR to SDFG conversion library with torch-mlir lowering passes

Co-authored-by: lukastruemper <31564409+lukastruemper@users.noreply.github.com>

* Fix MLIR CMake to use canonical MLIR project patterns

Co-authored-by: lukastruemper <31564409+lukastruemper@users.noreply.github.com>

* adds mlir workflows

* Implement minimal PyTorch-MLIR integration with tests

Co-authored-by: lukastruemper <31564409+lukastruemper@users.noreply.github.com>

* removes duplicate cmake command

---------

Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: lukastruemper <31564409+lukastruemper@users.noreply.github.com>
Co-authored-by: Lukas Truemper <lukas.truemper@outlook.de>

2424 of 16890 branches covered (14.35%)

Branch coverage included in aggregate %.

1868 of 7702 relevant lines covered (24.25%)

8.19 hits per line