About the reproduce results of deepex

[EternalEep]

Hi, thank you for your good work in OIE2016, but I use your default parameters with 1 V100 GPU, but I get the following results:

Did I have some mistakes that I cannot get the 0.72 F1?

Also, I find that the constractive pretrain deep-ranking-model has not been released and the code is inference code now. It needs about 3 hours to test. Am I right?

Hope for your reply!

[HaoyunHong]

Hi @EternalEep thanks very much for your interest! We are taking a look at the reproduce mismatch issue.

The deep-ranking-model is released at https://huggingface.co/Magolor/deepex-ranking-model, which is shown in https://github.com/cgraywang/deepex/blob/a4a4cf60c96e1bfe3ddc8007498bf5ed783af730/scripts/ranking.py#L33 and you can directly use it in our code (with transformers package of huggingface).

The pretrain code of deep-ranking-model will be released later. Yes, it needs 3 hours to test with 1 V100 GPU.

[EternalEep]

Thank you very much, I guess the problem of mismatch may be the environment, such as version of some packages.

[cgraywang]

@EternalEep Thanks, I will close the issue for now. If there are additional ones, please free feel to reach out.