Embedding Interpret with Eigen Prerequisites 1. Eigenvalue and Eigenvector 2. Embedding Eigen can interpret the transformation space and dimension reduction. Embedding is vector of representati...

Transformer Insights Prerequisites 1. Transformer Transformer structure should be understood by yourself Transformer Process 1. Transformation Space to Query, Key, Value? Space Transformati...

2D CNN Case Study Models 1. AlexNet 2. ZFNet 3. VGGNet 4. GoogLeNet (Inception v1) 5. ResNet AlexNet 1. Structure \begin{table}[h] \centering \begin{tabular}{llll} \hline Layer & ...

BERT Prerequisites 1. Transformer Transformer is very important turning point of model with good performance. But at the moment, the focus of task is just translation. What is BERT 1. What ...

Transformer Prerequisites 1. Multi-layer Perceptron 2. Convolutional Neural Network 3. Recurrent Neural Network 4. Seq2Seq 5. Attention MLP/CNN/RNN learn fixed weights $W$ to map $x \to y$ Atte...

Attention Prerequisites 1. Recurrent Neural Network 2. Seq2Seq Recurrent Neural Network have a problem that have rough 1:1 aligment. But machine translation breaks these assumptions. input/outp...

🎯 Instance Segmentation & Mask R-CNN – Complete Mathematical Notes 1️⃣ Instance Segmentation Problem Instance Segmentation = [\text{Object Detection} + \text{Semantic Segmentation}] Goal:...

🧠 Transformer-based Semantic Segmentation SETR · Segmenter · DPT – Complete Mathematical & Architectural Notes 1️⃣ SETR (SEgmentation TRansformer) 📄 Paper: https://arxiv.org/abs/2012.15840...

🎯 Semantic Segmentation, FCN, Deconvolution & U-Net 1️⃣ Computer Vision Tasks Overview Classification Input: Image Output: Single label No spatial information Object Detection ...

🚀 Detection Transformer (DETR) 🟢 1. Motivation Traditional detectors: R-CNN family → Anchor-based, NMS required YOLO / SSD → Dense prediction + NMS ❗ Problems: Hand-designed anchors ...