Python based command-line tool for inspecting AI/ML models.
Testing | |
Meta | |
ai-forensics-python/
├── images/ # 📁 Project-related images
│ └── AI_Forensics.png # Project image/logo
├── ai_forensics/ # 📁 Main source code package
│ ├── __init__.py # Initializes the package
│ ├── __main__.py # Enables `python -m ai_forensics` to run the CLI
│ ├── ascii.py # Contains ASCII art display class using Rich
│ └── cli.py # CLI entry point
├── tests/ # 📁 Test suite
│ └── __init__.py # Marks this directory as a Python package
├── .gitignore # 📄 Specifies intentionally untracked files to ignore in Git
├── .pylintrc # 📄 Pylint configuration file for static code analysis
├── LICENSE # 📄 License file (e.g., MIT) defining terms of use
├── mypy.ini # 📄 Configuration for mypy static type checker
├── poetry.lock # 📄 Locked dependency versions (auto-generated by Poetry)
├── pyproject.toml # 📄 Main project configuration for Poetry and build system
├── README.md # 📄 Project description, usage, setup instructions
# ASCII banner (your existing behavior)
aifx
# Scan a file
aifx scan /path/to/model.gguf --debug
aifx scan /path/to/model.safetensors --json-out report.json
aiforensics
is currently under development. More to come in future versions.
MIT (See LICENSE)
This project is heavily inspired by the following projects:
- [] Step through code.
- [] Include address space for header information.
- [] Support downloading model and metadata from huggingface.
- [] Support S3 interface.
- [] Add Precision to each row. https://huggingface.co/bartowski/Meta-Llama-3.1-8B-Instruct-GGUF/blob/main/Meta-Llama-3.1-8B-Instruct-f32.gguf
- [] Generate table in console report to make more readable.
- [] Create HTML report similar to huggingface link above.
- [] Confirm that all address space was covered and no overlap. Add to report.