Deep Learning — Study Companion
A study companion and working knowledge base for the Deep Learning textbook by Goodfellow, Bengio & Courville (2016). It indexes all 20 chapters, provides a 2016-to-2026 delta layer, and includes four deterministic tools for reading-path planning, training-failure diagnosis, capacity/regularization planning, and model arithmetic calculations.
✨ What it does
- Indexes all 20 chapters with key content summaries and links to official chapters.
- Provides a 2016-to-2026 delta layer naming what the book got right, what was superseded, and what still holds.
- Includes a topic index mapping concepts to chapters.
- Ships four deterministic tools: reading-path planner, training-failure diagnostic, capacity-and-regularization planner,
- Offers core frameworks and mental models for understanding deep learning concepts.
- Includes supporting files: glossary, patterns, cheatsheet, and references.
🎯 When to use it
- When studying or teaching the Deep Learning book and needing a structured route through its chapters.
- When deciding whether a chapter's advice is still current in 2026 practice.
- When translating the book's mathematical concepts into concrete training decisions.
- When troubleshooting training failures or planning model capacity and regularization.
- When estimating parameter counts, FLOPs, or activation memory for a model.
🚀 How to use
Trigger by asking about the Deep Learning book, a specific topic (e.g., 'regularization'), a chapter (e.g., 'ch07'), or 'is this still true?'. For reading paths, run the reading_path_planner.py script. The skill expects inputs like a topic, chapter number, or a question about the book's content. Example prompts:
What does the book say about regularization?
Is the advice on dropout still valid in 2026?
For tools, use command-line invocations as shown in the SKILL.md.
📄 Output: The skill provides chapter summaries, delta analysis, and tool outputs (e.g., reading paths, diagnostics, plans, calculations).
📦 Add this skill to Claude Code
# 1. Get the skills repo
git clone --depth 1 https://github.com/alirezarezvani/claude-skills /tmp/claude-skills
# 2. Copy this skill into your project (or ~/.claude/skills for all projects)
mkdir -p .claude/skills
cp -r /tmp/claude-skills/engineering/deep-learning-book/skills/deep-learning-book .claude/skills/deep-learning-bookSkill source: engineering/deep-learning-book/skills/deep-learning-book/SKILL.md
⚠️ Good to know
The skill does not cover topics outside the 2016 book, such as RLHF, LLM infrastructure, or MLOps; it only points to official chapters and does not reproduce copyrighted text.
❓ FAQ
Does the skill reproduce the book's content?
No, it is a companion that provides original synthesis and links to official chapters, respecting copyright.
How do I get a reading path?
Run the reading_path_planner.py script with your goal, background, and hours per week.
What does the delta layer cover?
It covers what has changed from 2016 to 2026, such as transformers, AdamW, diffusion, and double descent.
🤖 Overview, features, install steps and FAQ were generated from the project's SKILL.md on Sep 4, 2026. Always check the original source before running commands.