Callbacks in Keras

preview_player
Показать описание


Рекомендации по теме
Комментарии
Автор

This video was incredibly helpful! You went into so much detail, and explained way more than I expected.

Aperfull
Автор

This has been incredibly helpful! Many thanks! Please do more Keras content as YouTube currently doesn't have much at the moment.

DanielSchaefer
Автор

Very nice video, was searching for how to use tensorboard with keras, and you nailed it. Thanks

jakobsaadbye
Автор

Another absolutely superb video, Ajay. Thank you. I have a question/comment. At about 4:00 into the video you set activation="relu" when constructing the model using Keras. Isn't "relu" actually piecewise-linear? It is, but I guess it just works as an approximation of a nonlinear activation function. Yes? Anyway, my question is more of a comment since I kinda know the answer, but I'm uncertain. If "relu" is piecewise-linear then a deep neural network using "relu" is actually piecewise linear and therefore amenable to mathematical analysis using Linear System Theory. Yes? I think I've heard of papers that actually do this analysis. I'd love to see a discussion about that someday. Thanks again for all you do. CodeEmporium is fantastic.

vtrandal
Автор

Good video! I noticed however that you used a dropout rate of 0.9. In TensorFlow, this is the KEEP RATE. But in Keras, it's the DROP RATE. I assume you wanted to only shut off 10% and not 90% of the neurons, :-)

LemurDrengene
Автор

may i know where you explained about ModelCheckpoint?

deborahooi
Автор

Is there a playlist for this ? Its very hard to know the sequence of videos

ssanand
Автор

Can I use tensorboard and ModelCheckpoint at the same time? I tried but it does not work.

MaxOdious
Автор

why does the metric parameters need square brackets?

Gavinnnnnnnnnnnnnnn
join shbcf.ru