Self-attention mechanism explained | Self-attention explained | scaled dot product attention

[@ravi8908]

This is far better than other videos on the same topic- finally I understood the topic. I slept off in all other videos and grasped the topic here fully!

[@lavasaigo]

Excellent explanation sir. Thank you!!

[@bakyt_yrysov]

Thank you very much! So far the best channel for Data Science! Please, keep it up!

[@riyazbagban9190]

nice and neat explaination

[@kcbaskar]

You have explained the Simple Attention model clearly with mathematical examples. If you could make a small vocabulary set and show the next predicted word through the calculation, we can understand much better. We will know what these calculations are doing and how they are finding the next word. If possible, use a Python program for the example.

[@danzellbanksgoree23]

This was so great!!!!

[@SanthoshKumar-m2s]

Thank you for your clear explanation

[@mahieerati6505]

Great explanation

[@ajitkulkarni1702]

Best expalinination on self attention !!!

[@dinu9670]

You are a saviour man. Great explanation. Please keep doing these videos 🙏

[@UnfoldDataScience]

Thanks, will do!

[@jayeshsingh116]

well explained thank you for covering these topics

[@cgqqqq]

a true gem

[@AnkitGupta-rj4yy]

Thank you for provide us ❤ in easy way

[@PAadhan]

nice explanation, Thank you how to calculate the vector value x1, x2, x3 f

[@irfanhaider3021]

Kindly make a video on GRU layer as well.

[@manoj1bk]

can be used as self attention mechanism(as an embedding layer) before LSTM in the context of time series analysis?

[@mohammedajaz6034]

Thanks for the video.

[@UnfoldDataScience]

Welcome Ajaz.

[@funwithtechnology6526]

Thank you for the very clear explanation :) . I have a small question here. In self-attention, is there a limit to the dimension of the final attention embedding space?

[@dhirajpatil6776]

Please made video on explanation of transformers architecture

[@RakeshKumarSharma-nc3cj]

awesome video

[@UnfoldDataScience]

Thanks Rakesh

[@ajitkulkarni1702]

Please make viodes on multi head attention...