search

johko / computer-vision-course

This repo is the homebase of a community driven course on Computer Vision with Neural Networks. Feel free to join us on the Hugging Face discord: hf.co/join/discord
MIT License
376 stars 124 forks source link

Initial Draft Outline for VGG19 #57

Closed sezan92 closed 7 months ago

sezan92 commented 8 months ago

This PR is a testing PR for initial VGG19. The writing had been done by albertkao227. The other members reviewed and we have opened our first PR to check if this process is fine.

merveenoyan commented 8 months ago

I'd like @MKhalusova to give a review to get an earlier sense of tone of the course.

sezan92 commented 8 months ago

thank you @lunarflu @MKhalusova @merveenoyan for the review. I will address them within 2-3 days God willing. hopefully that is fine.

sezan92 commented 8 months ago

It assumes that the reader understands what a convolution kernel is, are max-pooling layers, etc. We'll need to double-check that this is actually the case to avoid confusion

@MKhalusova I understand this point . I think the best way is to discuss with the contributors of the previous chapter. I have not seen their outline, I will ping them in discord.

Here are some questions that a reader may have. Why did we choose this model as one of the examples? What is so special about it? What is the ImageNet Large Scale Visual Recognition Challenge (ILSVRC)? What tasks would one use this model for? (And maybe in what cases it wouldn't be the best choice)

these are good questions. But I think they need a separate markdown files for details. Don't you think theses questions deserve a separate PR? if yes, may be we can address them in another PR.

MKhalusova commented 8 months ago

@sezan92 Why do you think that there's a need for a separate file to answer these questions? From my perspective, this file is a great place to address them.

sezan92 commented 8 months ago

Why do you think that there's a need for a separate file to answer these questions? From my perspective, this file is a great place to address them.

@MKhalusova I think the reason is that all of the models mentioned in this chapter are trying to solve the challenge. So the details of this discussion are related to all of the models. That is why I think it deserves another mdx file so that we can refer to the details of the Imagenet challenge from other models' descriptions.

sezan92 commented 7 months ago

@merveenoyan @michaelshekasta sorry for late , i will address the comments this week iA

lunarflu commented 7 months ago

Thanks @sezan92 ! 🤗 ❤️ Always let us know if there is any confusion or if there's a part you get stuck on, happy to help!

I think we are close to merging after a few more fixes which will be great! 🚀

merveenoyan commented 7 months ago

Hey @sezan92 after you change the path of this file we can merge, content wise it's nice!

sezan92 commented 7 months ago

Hey @sezan92 after you change the path of this file we can merge, content wise it's nice!

@merveenoyan sorry for late reply, I had to take some break. I am back. I am updating

sezan92 commented 7 months ago

@merveenoyan @AI-ANK @MKhalusova sorry, i think we can merge this PR when approved.