Skip to content

ScratchMind/ScratchSeq

Repository files navigation

📚 ScratchSeq

ScratchSeq is a from-scratch learning project for sequence modeling and language understanding, implemented with a PyTorch-first mindset.
The goal is not benchmark performance, but mechanistic understanding — how sequence models evolved, why each innovation mattered, and how to implement them cleanly.


🎯 Philosophy

  • Implement core models before using abstractions
  • Read original papers alongside code
  • Prefer minimal, inspectable implementations
  • Focus on learning signals, gradients, and inductive biases

This repository is designed as a learning timeline, not a model zoo.


ROADMAP available at TIMELINE


🔍 Non-Goals

  • ❌ No large-scale pretraining
  • ❌ No SOTA chasing
  • ❌ No heavy frameworks or wrappers
  • ❌ No “black box” usage

📌 Outcome

By completing ScratchSeq, one should be able to:

  • Derive sequence models from first principles
  • Understand why transformers replaced recurrence
  • Reason about attention, memory, and scaling limits
  • Read modern LLM papers without hand-waving gaps

🧱 Status

🚧 Work in progress — built incrementally alongside paper reading and experimentation.


ScratchSeq is about earning intuition, not importing it.

About

ScratchSeq is a from-scratch learning project for sequence modeling and language understanding.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors