@vintcessun: 3 courses updated in 3 days: Day00 Learning Map, Day01 Transformer Overview, Day02 Input Layer. One course each morning, with illustrations, tables, formulas, and checklists. The greatest value of this path is letting beginners build understanding starting from the input layer, rather than diving directly into papers. What you learn is—understanding tok…
Summary
Published a GitHub repository with an LLM learning path for beginners, covering course content from Transformer basics to LLM techniques, with one course updated daily.
View Cached Full Text
Cached at: 07/04/26, 02:47 PM
3 lessons updated in 3 days: Day00 Learning Map, Day01 Transformer Big Picture, Day02 Input Layer. One lesson each morning, with illustrations, tables, formulas, and checklists. The greatest value of this path is that it lets beginners build cognition starting from the input layer, rather than directly tackling research papers. What you learn: understanding the three-step transformation of token, embedding, and positional encoding enables you to grasp all subsequent LLM structures.
The repo is here 👇
terrense/LLM_path_for_begginers
Source: https://github.com/terrense/LLM_path_for_begginers
LLM Path for Beginners
A large language model technical learning path for beginners. Starting from Transformer fundamentals, it gradually covers LLM, VLM, RAG, SFT, RLHF, quantization, Ollama, vLLM, SGLang, and local deployment experiments.
Course Outline
| No. | Topic | Link |
|---|---|---|
| Day 00 | LLM Learning Path: From Seeing the Forest to Setting Up Your Lab | posts/day-00-learning-map |
| Day 01 | Transformer Big Picture: Why It Changed NLP | posts/day-01-transformer-big-picture |
| Day 02 | Tokens, Embeddings, and Positional Encoding: The Three-Step Transformation of the LLM Input Layer | posts/day-02-tokens-embeddings-position |
Update Schedule
One lesson will be published every morning at 7:00 AM. Each article includes the main text, illustrations, tables, flowcharts, mathematical formulas, practical checklists, and a quiz. Images prioritize report-style technical diagrams, architecture diagrams, flowcharts, and comparison charts, emphasizing readability and learning value.
Similar Articles
@kuririrn: Studying to grasp the recent trends in scratch development of LLMs Stanford CS336 Lang. Modeling from Scratch | Spring …
This course note summarizes the architectural evolution from the original Transformer to modern LLMs, focusing on convergent developments such as pre-normalization, RMS normalization, and RoPE, and provides hyperparameter selection recommendations.
@techNmak: This is the best way to learn how LLMs work. Interactive. 3D. Step-by-step. Covers: → Embedding → Layer Norm → Self-Att…
An interactive 3D step-by-step guide to learning how LLMs work, covering key transformer concepts like embedding, self-attention, and softmax. It recommends a visual approach over reading papers.
@QingQ77: 'Dive into Deep Learning' is an excellent introductory book, but its update speed struggles to keep pace with the field's development. Since the Transformer, content like CLIP, Diffusion, vLLM, and more has proliferated. While online resources are abundant, they are highly fragmented—today you study Attention, tomorrow LoRA, the day after...
This project is a systematic deep learning notes repository covering PyTorch, Transformers, generative models, and more. It aims to address the fragmentation of learning materials and provides code implementations along with practical guides.
@GitHub_Daily: Want to understand the underlying principles of large language models? Most resources only cover theory or provide source code, leaving you still confused. Stumbled upon this open-source tutorial, EveryonesLLM, which guides us step by step to build a complete large language model from scratch on Google Colab, writing code throughout. The whole tutorial is divided into...
EveryonesLLM is an open-source tutorial that provides 29 chapters of Colab notebooks. It teaches users step by step to build a complete large language model from scratch on Google Colab, including pre-training and instruction fine-tuning, and supports Chinese.
@FinanceYF5: Setting aside the story of a friend landing a 750k salary with this, the video itself is pure substance—a 3-hour hands-on walkthrough of building an LLM from scratch. It covers five stages: LLM basics, Transformer architecture, training process, model modernization, and scaling. Code is fully visible, typed out as explained. If you want to understand how ChatGP…
A 3-hour hands-on walkthrough of building an LLM from scratch, covering five stages: LLM basics, Transformer architecture, training, model modernization, and scaling, with code visible throughout.