Adam Optimizer for Neural Network || Lesson 15 || Deep Learning || Learning Monkey ||

[@saswatam]

Sir in the final Equation: Wt = Wt-1 - Alpha * m^t / sqrt(V^t + ε)), here we are representing gt which was dl/dw earlier as m^t / sqrt(V^t + ε))?, because in the same previous Equation: Wt = Wt-1 - Alpha *gt. Where Alpha = Learning Rate

[@babakdianati6515]

Thanks for the nice video

[@muhammadumarbello6574]

Very nice explanation,but how can I get a detail pdf on Adam

[@OnTastySpots]

Wow! Thank you for the detailed explanation! I'm wondering do we have m_t, v_t for each parameter? For example, f(w1,w2,w3) will need 3 m_t's and 3 v_t's in each timestep.

[@ayushx4831]

Send link of that T shirt 🙂

[@AkshayRakate]

What is alpha in the Adam equation? else a brilliant explanation !

[@LearningMonkey]

Please watch our Ada delta video. Alpha is learning rate

[@bangarrajumuppidu8354]

super sir

[@mohammedmunavarbsa573]

super bro