How higiangcoder/higiangcoder works
Deep Learning / AI Projects
• The Final_Project folder holds Jupyter notebooks like “Lenet5_MNIST.ipynb” and several PDFs (e.g., CNN-model-tutorial-part-1.pdf, CNN.pdf, Lenet_Build_From_Scratch.pdf) indicating deep learning tutorials and experiments.
• A “data” subdirectory under Final_Project contains the MNIST dataset files, which are clearly inputs for the training or experimentation processes.
• The Report-FinalProject_2 folder includes additional deep learning components (e.g., UNet.ipynb, UNet_medical_segmentation.pdf) that hint at segmentation tasks, likely involving U-Net architectures.
Competitive Programming / C++ Code
• A mix of individual C++ files (e.g., check2code.cpp, gene.cpp, gene_linux.cpp) as well as a dedicated “old_code” directory containing various algorithm implementations and templates (e.g., MODSORT.cpp, Trie.cpp, mastermind.cpp, etc.).
• There is also a “ymd240925” folder containing versions of a Project Euler problem solution, denoting the iterative nature of problem solving.
UI/Guide Documentation
• The folder “huong_dan_giao_dien” includes a GitHub markdown file that likely provides instructions for a UI or front-end interface, separate from the algorithmic and deep learning content.
Supporting / Infrastructure Files
• The .github/workflows folder with a c-cpp.yml file indicates that there is at least one continuous integration (CI) workflow set up for the C++ parts.
• Additional directories like .jupyter (workspace info) and temporary folders (tmpdir) serve supportive roles for configuration and testing.
A. Component Identification:
1. Deep Learning Module:
• Components: Jupyter Notebooks (e.g., Lenet5_MNIST.ipynb, UNet.ipynb), PDF tutorials/documentation, and the MNIST dataset.
• Role: Provide training, experimentation, and documentation for various CNN/UNet models.
• Dependencies: Likely dependent on Python libraries such as TensorFlow or PyTorch.
2. Competitive Programming Module:
• Components: Various C++ files scattered across the root and in “old_code” and “ymd240925.”
• Role: Contain algorithm implementations, templates, and solved problems.
• Dependencies: Standard C++ toolchain (possibly integrated with CI tools defined in the .github workflows).
3. UI/Guide Module:
• Components: Files under “huong_dan_giao_dien.”
• Role: Provide instructions or guidances for user interfaces or workflow documentation.
4. CI/CD and Project Infrastructure:
• Components: The .github/workflows directory, potentially configuration files and temporary files.
• Role: Facilitate automated builds or tests, especially for the C++ code.
B. Mapping Relationships and Data Flow:
1. Within the Deep Learning Module, illustrate the flow from the “data” directory (MNIST raw data) → preprocessing (possibly embedded in the notebooks) → model training and experiments → generation of PDF reports/documentation.
2. In the Competitive Programming Module, each C++ file or set of related files could be depicted as standalone components or libraries that can be compiled/tested via the CI pipeline (represented by the workflow in .github).
3. The UI/Guide Module is separate yet complementary: these files might be referenced within the README or by users as supplementary material.
C. Architectural Patterns & Relevant Technologies:
• The Deep Learning section follows a typical data-science project architecture: Data → Notebook-based Processing/Training → Reports and Documentation.
• The Competitive Programming part is structured as independent code submissions or modules, potentially developed under agile or iterative problem-solving approaches.
• The presence of CI (via .github/workflows) points to a continuous integration design, ensuring code quality for the C++ modules.
• Technologies/libraries explicitly: Python (with TensorFlow/PyTorch hinted by “Test_Tensorflow-checkpoint” in a Jupyter checkpoint), standard C++ compilers, GitHub Actions for CI, and documentation in PDF/Markdown formats.
1. Use clear labels:
• In the deep learning branch, indicate the flow from raw data → processing (notebooks) → output reports.
• How the CI pipeline (GitHub Actions) is triggered by changes in the C++ modules.
3. Visual Differentiation:
E. Tailoring to the Specific Project:
• For the deep learning part: Emphasize a “notebook-driven” workflow where raw data is ingested, processed, and used to generate experimental outputs.
• For the C++ parts: Showcase them as isolated module implementations that might have a common CI process or template but solve independent problems.
• Additionally, include an “External Resources” or “Guidelines” node that contains UI documentation or instructional markdowns.
Final Note: