Feature Maps Visualization Of CNN | Interpretation Of Output Of Conv2D And Maxpooling Layer

[@ecp5758]

WUuuauu! Very good video!

[@Antz_411]

Great explanation! Is there a link to the Colab Notebook? Please...

[@ManishKumar-bb5ql]

Great Video very informative yet simple

[@PathshalaTutorials]

Thank you..

[@jennilthiyam1261]

what are the dark output in the strip ?..what are they capturing?

[@rohinv2752]

if your guys' strips are too small, you can either drag each layer to the new tab button(on chrome at least), or you can simply double click the strip you want to see

[@udaya349]

sir plz do vedio on feature engineering .ur vedios are too good ,

[@PathshalaTutorials]

Will do.. Thank you for your feedback..

[@wongkahquan-nelson6415]

Great video!!! Really helped me! However, my strip are really small, any idea on that?

[@satya8411]

Should we do it before fitting the model(ie training) or after? Please say

[@Embassador4Peace]

Before training, unless you are using pretrained network, output will be random image. After training visualization makes lots of sense.

[@mikacueto8647]

what versionof tensorflow & Keras are u using?

[@سجىمحسنعباسعلي]

Can one help me The code not executed Error appear Could not broadcast input array from shape (200,200,1) into shap (200,200)

[@techtalk9352]

please provide the link of notebook

[@neomuks]

very good video,

[@PathshalaTutorials]

Thank you..

[@juniorcomsono]

i don't understand how the input image is 3 channel and when checking the shape of the feature maps its 8 actually, really got lost here

[@arhumboyyyy]

The last CNN layer had 8 different filters applied to the output of the previous CNN layer, hence 8 channels or images are the result as was shown in the visualization.

[@juniorcomsono]

@@arhumboyyyy I wish I'd read that 2 weeks ago XD thanks dude

[@satya8411]

Should we do it before fitting the model(ie training) or after? Please say

[@juniorcomsono]

@@satya8411 do what?

[@satya8411]

@@juniorcomsono visualizing

[@rajshreesrivastava7650]

Nice can u share the code

[@PathshalaTutorials]

Thank you so much.. Google Colab link is : colab.research.google.com/drive/1dz4m_aLt_1fkCoPxyNp5cIn5U2Z5e9Vk?usp=sharing

[@emeebritto]

1.3x is the normal speed actually... lol