karpathy / convnetjs

Deep Learning in Javascript. Train Convolutional Neural Networks (or ordinary ones) in your browser.
MIT License
10.86k stars 2.04k forks source link

Refactored the demos #33

Closed pavelpep closed 9 years ago

pavelpep commented 9 years ago

I extracted styles and javascript from all the demo files, explicitly declared several magic constants in mnist and cifar demos, unified the javascript for mnist and cifar10 into one js file, moved js and css files to respective folders and updated references in html added ability to test with own image in cifar10 and mnist demos

still left to do: refactor and/or unify the javascript files for the rest of the demos

karpathy commented 9 years ago

Whoa, nice! This looks like an extensive commit; I assume you verified that the demos work? I skimmed through the changes but I don't think I'll have time for a thorough look.

pavelpep commented 9 years ago

Yes, I tested all the demos and they behave correctly. I also checked the console for errors and unimported js/css files, all clear :)

karpathy commented 9 years ago

looks good, great, thank you!

pavelpep commented 9 years ago

My pleasure! Great job on the whole project :)

cbovar commented 9 years ago

@pavelpep should it be <image_channels rather than <3 here and here?

Thanks for the work! (Should I do a pull request if I spot such error?)

pavelpep commented 9 years ago

@cbovar nice catch! You are correct, it should be <image_channels instead of <3