[ORT Training] Some important updates of ONNX Runtime training APIs (#1335)
* update trainer
* update args
* update to main
* update to 4.33
* fix style
* make style
* fix when testing
* Update optimum/onnxruntime/trainer_seq2seq.py
Co-authored-by: fxmarty <9808326+fxmarty@users.noreply.github.com>
* deprecate ort inf
* deprectae ort inf for seq2seq
* update trainer and its args to main
* try CI permission
* update tests
* update examples
* withdraw CI change
---------
Co-authored-by: JingyaHuang <jingya@huggingface.co>
Co-authored-by: fxmarty <9808326+fxmarty@users.noreply.github.com>