fxmarty

Having `"on"` transformed into `True` is... interesting design choice.

Hello @shivalikasingh95 , great suggestion! If you want to contribute to adding `ORTModelForObjectDetection`, I suggest: * Reading https://github.com/huggingface/optimum/blob/main/CONTRIBUTING.md * Most of the work will be extending `modeling_ort.py`, you can take...

Of course no worries! Don't hesitate to open a draft PR once you are on it if you need any help; if a team member wants to implement the class...

Related issues: https://github.com/huggingface/optimum/issues/524 https://github.com/microsoft/onnxruntime/issues/13808 Looking back, understanding/sharing upfront the implications of this design choice would have helped avoid many issues.

@vgrabovets Thanks for the feedback! Yes, it's very subideal to have this dependency on `torch`. We would need to make `generate()` (or part of) `torch.jit.script`-able to get rid of the...

@TheBloke yeah I'm already spending more time than I should maintaining the repo. Let's wait if @PanQiWei comes back.

@qwopqwop200 @Qubitium thanks a lot for fixing this. I will review & merge hopefully today/in the coming days, and do a release thereafter.

@Qubitium I'll push some modifs if that's fine to you

@Qubitium Do you think we should by default save models in the old `gptq` format?

@Qubitium yes, the issue is maybe underflow in https://github.com/Qubitium/AutoGPTQ/blob/9d88639600d263e46cec73fe5d03cc609bf3705b/auto_gptq/modeling/_utils.py#L192-L201