Added - Added explicit links to GitHub repository and PyPI package in documentation - Improved model card template to include better credit to Model Quantizer - Enhanced installation instructions with direct links to PyPI
0.3.2
Fixed - Fixed installation issue with run-benchmark command not finding the module - Added py_modules to setup.py to include standalone scripts in the package - Added fix_installation.py script to help users fix existing installations
0.3.1
Fixed - Fixed GPTQ quantization issue by explicitly installing optimum with GPTQ support (`optimum[gptq]`) - Updated installation scripts to ensure proper installation of GPTQ dependencies - Improved requirements-all.txt with explicit GPTQ integration for optimum
Added - Confirmed compatibility with Python 3.11 and 3.12 on macOS Sonoma 14.2 - Added detailed platform-specific documentation for macOS, Windows, and Linux - Enhanced troubleshooting guide with platform-specific information
0.3.0
Fixed - Fixed dependency installation order issue with gptqmodel requiring torch to be installed first - Added installation scripts (install_dependencies.sh and install_dependencies.bat) for reliable dependency installation - Updated requirements-all.txt with clear instructions on installation order - Improved README.md with detailed installation options and troubleshooting guidance
0.2.9
Added - Created `requirements-all.txt` for one-shot installation of all dependencies - Includes all core, quantization, visualization, and data handling packages - Provides a simple way for users to install everything needed with a single command
0.2.8
Changed - Further refined dependencies to absolute minimum required set - Moved optimum from core dependencies to gptq extras - Improved organization of dependencies in requirements.txt - Added detailed comments explaining the purpose of each dependency - Grouped GPTQ dependencies together in the extras