optimum

2155fbe6 - Enable inference with a merged decoder in `ORTModelForCausalLM` (#647)

Enable inference with a merged decoder in `ORTModelForCausalLM` (#647) * Add save option * Add test for saving * Fix test path * Allow str path for merging * Add Path and remove merged names * Finish adapting ORTModelDecoder * Prepare extra inputs * do not store merged in place * Support I/O binding for merged * Extend to multiple patterns * Add test for inference * Fix test * update test * Remove prints * add back missing method * fix mess * fix post merge * none defaut * fix * fix errors * remove nonsense tests * fix doc * ongoing * debug * fix style * fix post merge * hopefully working! * add tests * rename * add constants * fix test * fix names * ort support * wip * fix * fix * tests for merged * stype * fix * fix merge errors * fix tests * fix test * remove irrelevant test * Update optimum/exporters/onnx/__main__.py Co-authored-by: Michael Benayoun <mickbenayoun@gmail.com> * Update optimum/exporters/onnx/base.py Co-authored-by: Michael Benayoun <mickbenayoun@gmail.com> * Update optimum/exporters/onnx/config.py Co-authored-by: Michael Benayoun <mickbenayoun@gmail.com> * Update optimum/exporters/onnx/config.py Co-authored-by: Michael Benayoun <mickbenayoun@gmail.com> * Update optimum/exporters/onnx/config.py Co-authored-by: Michael Benayoun <mickbenayoun@gmail.com> * Update optimum/exporters/onnx/config.py Co-authored-by: Michael Benayoun <mickbenayoun@gmail.com> * Update optimum/exporters/onnx/config.py Co-authored-by: Michael Benayoun <mickbenayoun@gmail.com> * Update optimum/onnxruntime/modeling_decoder.py Co-authored-by: Michael Benayoun <mickbenayoun@gmail.com> * Update tests/exporters/onnx/test_exporters_onnx_cli.py Co-authored-by: Michael Benayoun <mickbenayoun@gmail.com> * fix on suggestions * fix import of dummyinputgenerators * skip unwanted tests * fix diffusion model * fix tests * Update optimum/commands/export/onnx.py Co-authored-by: Jingya HUANG <44135271+JingyaHuang@users.noreply.github.com> * Update optimum/onnxruntime/modeling_decoder.py Co-authored-by: Jingya HUANG <44135271+JingyaHuang@users.noreply.github.com> * fix last tests * Update optimum/onnx/graph_transformations.py Co-authored-by: Michael Benayoun <mickbenayoun@gmail.com> * Update optimum/onnx/graph_transformations.py Co-authored-by: Michael Benayoun <mickbenayoun@gmail.com> * Update optimum/onnx/graph_transformations.py Co-authored-by: Michael Benayoun <mickbenayoun@gmail.com> * Update optimum/onnx/graph_transformations.py Co-authored-by: Michael Benayoun <mickbenayoun@gmail.com> * Update optimum/onnxruntime/modeling_decoder.py Co-authored-by: Michael Benayoun <mickbenayoun@gmail.com> * fix signature and docstrings * add error message if post process fails * tiny fix * last fixes * typo --------- Co-authored-by: Félix Marty <9808326+fxmarty@users.noreply.github.com> Co-authored-by: Michael Benayoun <mickbenayoun@gmail.com>