LLM

DeepFusion

What’s Inside

• 📖 Comprehensive Book Gain deep insights into a wide array of deep learning and reinforcement learning algorithms. This book blends theory with practical applications, offering clear explanations and real-world relevance.
• 🚀 Platform-Specific Inference Code Ready-to-use implementations for various platforms make deploying AI solutions seamless and straightforward.
• 🔄 Reimplementation of Key Algorithms Study and experiment with re-implemented versions of popular algorithms to deepen your understanding or customize them for your needs.
• ⚡ Model Optimization Learn and apply techniques to optimize your models for better performance, reduced latency, and efficient deployment across devices.
• 🛠️ Additional Tools and Utilities Access a suite of tools for model evaluation, visualization, and debugging, enhancing your workflow and productivity.

📌 Recommendation

For easier navigation of this repository, consider using the Google Octotree browser extension. Octotree provides a sidebar with a file tree structure, making it simple to explore and manage large GitHub projects efficiently.

📚 Repository Structure

[CV]

• Comprehension of Data Labeling, OpenCV, Object Detection, ReID, Object Tracking, Super-Resolution, Camera Calibration, Large Vision Models, Feature Mapping, Stereo Vision.

[NLP]

• Usage introduction of jieba & hanlp, comprehension of Transformers.

[RL]

• Classic Reinforcement Learning code reimplementation.

[Inference Code with Model Optimization]

• Inference code for different platforms, including NVIDIA, Intel, RK, DL, etc.
• Model pruning and quantization.

[Training Related]

• Explanation of various training tricks.

[Frameworks]

• Useful usage examples of PyTorch, NumPy, pandas, Matplotlib, TensorBoard, and PyTorch Quantization.

[Scripts]

• A collection of useful scripts.

[Python]

• Examples of multiprocessing, multi-threading, and other Python utilities.

[C++]

• Frequently used C++ code and C++ multi-threading tutorials.

[Linux]

• Useful Linux operations and commands.

[Docker]

• Basic usage and examples of Docker.

[FFmpeg]

• Basic usage examples of FFmpeg.

[Git]

• Basic usage of Git and version control.

[IDE]

• Tips and tricks for using VSCode and PyCharm.

[Web Related]

• Basic usage of Flask for web development.

[Machine Learning]

• Understanding the fundamentals of mathematics in machine learning.

[Frp]

• Introduction to using frp (Fast Reverse Proxy).

[Interview Questions]

• A collection of common technical interview questions.

[Algorithms Reimplementation]

• Reimplementations of key algorithms such as Canny, ResNet, ROIPooling, RANSAC, CBAM, and Generalized Hough Transform.

[HuggingFace Related]

• Usage and integration with HuggingFace libraries.

[Latex]

• Tips and tricks for using LaTeX.

[Others]

• Miscellaneous resources and utilities not covered by the above categories.

⚠️ Declaration of Original Contributions

This repository includes original algorithms and methods developed exclusively by the author. These contributions represent novel ideas and unique implementations not derived or copied from existing work. If you use these original elements in your research or projects, please cite this repository to acknowledge the effort and innovation behind them.

• [x] YOLOv8-Pruning
• [x] YOLOv11-Pruning
• [x] Classic Reinforcement Learning code reimplementation.

TODO LIST

• [x] YOlOv11-pruning
• [x] C++ code implementation of SGBM
• [ ] Further study of Reinforcement Learning
• [x] C++ version of YOLO-pose inference code uploading
• [x] RTMPose inference code uploading
• [ ] Further study of LLM