Keras Convolutional Neural Neural Networks for MNIST and Fashion MNIST (6.2)

[@miladesn]

you have the best CV class and content ever. If only 1% of Profs had good as course material as yours. Your students don't know how good they have it.

[@HeatonResearch]

Thanks! Kind of you to say

[@yaminadjoudi4357]

@@HeatonResearch i'm agree

[@cchristoff]

Thank you for publishing this. It is explained simply yet in a detailed way. it helped me with my assignment on CNNs.

[@maximinmaster7511]

Thank you for theses courses !

[@aidanmacdougall5530]

I guess more people use the Colab gpus these days (or google changed their free access capabilities) compared to when this was recorded, first epoch took 144s online and 56s on the old jetson xavier... back to the local gpu! Thanks a heap for the content Jeff! This course is a really nice walk through and the assignments are great application exercises for the material

[@rohitdhankar360]

@10:20 - The input "A" and the final Output - 10 Classes :)

[@HeatonResearch]

That is a great point, I never thought of that. Going to have to ask Dr. LeCun about that one!

[@12_vermouth_12]

Actually there is a Conv3D layer implemented within PyTorch… Great video as always, thank you!

[@neelamadhabkhaya3868]

Where are these contents plz

[@nguyenanhnguyen7658]

Sir, ResNet is deep but huge in size, slow etc. Why do people like it that much especially researchers? Can you please explain, thank you.

[@dr.manishaprakashbharati3359]

Sir cant we apply CNN on KDD99 and CSE IDS2018

[@christopheryoung7880]

Has there been a recent update? Section 6 video links in Jupyter Notebook point to Backprop (2017) not the current CNN content.

[@HeatonResearch]

Yes, there have been considerable updates for the fall 2019 semester. I will be double checking all the links soon, some might be broken. If you can't find something, just let me know

[@soumyasrm]

Can you please make videos on unsupervised learning... i am big fan of ur videos

[@HeatonResearch]

Yes, I probably will! Thanks

[@RR-gc4zw]

You have a typo. MNIST is correct.

[@HeatonResearch]

Thanks! Updated title/description/notebooks. Will get the video next cycle.