Implementing GPT Architecture From Scratch: Training and OutputThis is a follow-up of my previous post "Implementing GPT Architecture From Scratch: A Deep Dive into Transformers and Attention" This will be a very short post explaining how i trained the untrained6 min read·ish4n10.hashnode.devMar 8
Implementing GPT Architecture From Scratch: A Deep Dive into Transformers and AttentionA deep dive into building a modern decoder-only transformer from scratch in PyTorch covering Grouped Query Attention, Rotary Positional Embeddings etc13 min read·ish4n10.hashnode.devMar 6
An SMT solver from scratch - Part 2: SMT Theories and LogicsI have covered the basics of a SAT solver and their working mechanism in detail in the first part, in this part I'll try to delve deep-inside SMT theory and how they function being built on-top of a SAT solver. Table of Contents Basic Terminologies ...9 min read·ish4n10.hashnode.devApr 13, 2025
An SMT solver from scratch - Part 1: TheorySAT, CDCL and Theories In the first post of the series, I'll discuss the main idea of an SMT solver and the theories involved to understand the algorithm behind the SMT solver such as Z3. Table of Contents Basic Terminologies What is a SAT Solver? ...6 min read·ish4n10.hashnode.devApr 6, 2025