ATTENTION | An Image is Worth 16x16 Words | Vision Transformers (ViT) Explanation and Implementation

[@Explaining-AI]

*Github Code* - github.com/explainingai-code/VIT-Pytorch
*Patch Embedding* - Vision Transformer (Part One) - ua-cam.com/video/lBicvB4iyYU/v-deo.html
*Attention* in Vision Transformer (Part Two) - ua-cam.com/video/zT_el_cjiJw/v-deo.html
*Implementing Vision Transformer* (Part Three) - ua-cam.com/video/G6_IA5vKXRI/v-deo.html

[@sladewinter]

Best explanation of multi-head attention i have attended to! I already had a reasonable intuition but still gathered so much more, massive respect to your work 🙏

[@Explaining-AI]

Thank you! Really glad that it was of any help.

[@DrAIScience]

Amazing explanation... i did not come accross the beautiful and easy explanation of transformers that seems extremely difficult... this channel deserves millions of subscribers 🎉

[@Explaining-AI]

Thank you for the kind words :)

[@shashankdevraj8633]

Sir Can you explain dual attention vision transformers (Davit)please

[@sladewinter]

Would rearranging by heads before splitting into q, k , v cause any logical difference. Just means fewer lines of code, and operations, but mostly was just curious to verify as it felt same to me.

[@Explaining-AI]

@sladewinter Yes, I agree with you, both seem the same to me as well.

[@sebastiancavada]

Great content! This is helping a lot!! Keep it up :)

[@Explaining-AI]

Thank you:)

[@muhammadawais2173]

Helping, much appreciated. Sir how about self attention in image context

[@Explaining-AI]

Thank you!. I didn't get what exactly you mean by self attention in image context. Could you clarify a bit.