Hello, from the https://github.com/para-lost/AutoPresent/tree/main/slidesbench , both "python code" and "prompt" needs to be created by using running a python script or GPT models.
Is the prompt 100% reproducible to be exactly the same as what is used in the paper?
Can you release the "prompt, python code" dataset for the ease to reproduce the model? Because "prompt, python code" data are no longer in pptx format, there shouldn't be any permission issue from the original slide owner.
And from the slideshare website, as per slide's agreement terms and condition, if a slide owner uploads a slide, by default, the permission is given to others to use the slides.
Hello, from the https://github.com/para-lost/AutoPresent/tree/main/slidesbench , both "python code" and "prompt" needs to be created by using running a python script or GPT models.
Is the prompt 100% reproducible to be exactly the same as what is used in the paper?
Can you release the "prompt, python code" dataset for the ease to reproduce the model? Because "prompt, python code" data are no longer in pptx format, there shouldn't be any permission issue from the original slide owner.
And from the slideshare website, as per slide's agreement terms and condition, if a slide owner uploads a slide, by default, the permission is given to others to use the slides.