ML-Manim Animations

This repository aims to explain and animate complex Machine Learning research papers using Manim Community Edition. The goal is to make advanced ML concepts more accessible and understandable through high-quality visual explanations.

Papers & Animations

Research Paper Name	Paper PDF Link	Animation Link	Additional Resources
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding	PDF	BERT Breakthrough	- BERT Explained: A Complete Guide to Understanding BERT - The Illustrated BERT, ELMo, and co. (How NLP Cracked Transfer Learning) - Hugging Face Transformers Library (BERT)
Improving Language Understanding by Generative Pre-training	PDF	GPT Animation	- The Illustrated GPT-2 (Visualizing Transformer Language Models) - Hugging Face Transformers Library (GPT)
The Llama 3 Herd of Models	PDF	Llama 3 Animation	-
DeepSeek-R1	PDF	DeepSeek R1 Animation	- Understanding the DeepSeek R1 Paper - Hugging Face LLM Course - deepseek-ai/DeepSeek-R1 - Hugging Face - How DeepSeek-R1 Leverages Reinforcement Learning to Master ... - Hugging Face Blog
Reasoning Models Don’t Always Say What They Think	PDF	Reasoning Models Animation	- Plain English Summary: AI Models Often Fake Their Step-by-Step Reasoning (DEV.to) - arXiv: Chain-of-Thought Reasoning In The Wild Is Not Always Faithful - Anthropic Blog Announcement

How to Use This Repo

Each animation is typically contained within its own directory (e.g., BERT/).

1. Clone the Repository:

   git clone https://github.com/your-username/ML-Manim_Animations.git
   cd ML-Manim_Animations

2. Install ManimCE: If you haven't already, install Manim Community Edition. It's highly recommended to use a virtual environment.

   pip install manim

   For more detailed installation instructions, refer to the ManimCE Documentation.

3. Run an Animation: Navigate to the directory of the animation you want to render and run the main.py script. For example, to render the BERT animation:

   cd BERT
   manim -pql main.py BERTBreakthrough

   Or to render the GPT animation:

   cd GPT
   manim -pql main.py GPTPaperAnimation

   • -p: Plays the animation after rendering.
   • -q l: Renders in low quality (for quick previews). Use -q h for high quality (1080p) or -q k for 4K.
   • -l: Leaves the progress bars (if configured in main.py).

   The rendered video will be saved in BERT/media/videos/1080p60/ (or similar, depending on quality settings).

Contributing

Contributions are welcome! If you'd like to contribute an animation for a research paper:

1. Fork the repository.
2. Create a new directory for your paper (e.g., YourPaperName/).
3. Add your Manim Python script (main.py) and any necessary media files.
4. Update the README.md with an entry for your new paper and animation.
5. Submit a Pull Request.

Please ensure your animations are clear, accurate, and well-commented.

License

This project is open-sourced under the MIT license. See the LICENSE file for details.