●Change8

b8498

📦 llama-cppView on GitHub →
✨ 2 features🐛 4 fixes🔧 1 symbols

Summary

This release introduces standard Hugging Face cache support for model handling and includes various platform-specific binary builds.

Migration Steps

  1. Users should be aware that manifests are migrated to the Hugging Face cache at startup.

✨ New Features

  • Added standard Hugging Face cache support in the common module.
  • Implemented migration of all manifests to Hugging Face cache at startup.

🐛 Bug Fixes

  • Improved error handling and reporting for API errors.
  • Restored common_cached_model_info and aligned mmproj filtering.
  • Prefer 'main' branch when retrieving cached references.
  • Used cached files when Hugging Face API calls fail.

Affected Symbols

common