recheck models features and output for its high accurecy and aaceptance

[tahmeedtoqi]

today I have made PR for my model (ToqiNet) but it was rejected and closed (#347) declaring it as a toy model which was in (#345). I have made some modification and updates so far though it inspired by the AlexNet , but compare to AlexNet it has larger parameters and it has higher accuracy compared to other models.

I hereby, request to consider rechecking and test for its actual performance (If necessary)

my account: https://github.com/tahmeedtoqi

repository: https://github.com/tahmeedtoqi/ToqiNet---v1.1.2.git

[NicolasHug]

@tahmeedtoqi how is ToqiNet different from a vanilla multi-layer perceptron? What are the modifications you made to AlexNet? How much more accurate is it? On what datasets? Has this work been peer-reviewed and validated in any shape or form? What do you think would be the benefit of the community from using ToqiNet instead of other baseline models?

All of these are questions that need to be fairly clear before we can consider it for inclusion on the website. Note that nothing prevents you from making toqinet available through torch.hub.load(), you would just have to create a hubconf.py file in your own repo for that (refer to our docs and look at how other repos do it).