Image Classification Using Vision Transformer | An Image is Worth 16x16 Words
About this lesson
This video covers the implementation of vision Transformer - VIT in pytorch . This is the third part of Vision transformer - VIT series in which I build entire vision transformer from scratch. I also cover visualizations of the attention map using attention rollout as well as positional embedding visualization to get more intuition on what the model is learning in this video. I have covered Vision Transformer VIT in three parts: 1. Patch Embedding Module - https://youtu.be/lBicvB4iyYU 2. Attention Block - https://www.youtube.com/watch?v=zT_el_cjiJw 3. Building Vision Transformer and visualizing attention map using attention rollout and also visualizing positional embedding of vision transformer (This Video) *Other Good Resources* Yannic Kilcher | An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale (Paper Explained) - https://www.youtube.com/watch?v=TrdevFK_am4 AI Coffee Break with Letitia | An image is worth 16x16 words: ViT | Vision Transformer explained - https://www.youtube.com/watch?v=DVoHvmww2lQ James Briggs | Vision Transformers (ViT) Explained + Fine-tuning in Python - https://www.youtube.com/watch?v=qU7wO02urYU Good Place to understand general transformer further - https://tinyurl.com/exai-vit-transformer Timestamps : 00:00 Intro 00:28 Architecture of Single Vision Transformer Layer 01:19 Architecture of Vision Transformer 02:00 Dataset for Image Classification 02:35 Code for Vision Transformer Layer 03:48 Code & Implementation of VIT - Vision Transformer 04:50 Attention Map Visualization using Attention Rollout 06:09 Attention Maps for trained model 06:18 Position Embedding Visualization 08:32 Outro Paper Link - https://tinyurl.com/exai-vit-paper Implementation - https://tinyurl.com/exai-vit-code Subscribe to Channel - https://tinyurl.com/exai-channel-link Good Place to understand original transformer further - https://tinyurl.com/exai-vit-transformer Background Track - Fruits of Life by Jimena Contreras Email - explainingai.
DeepCamp AI