Change8

b8477

📦 llama-cppView on GitHub →
2 features🐛 3 fixes🔧 2 symbols

Summary

This release introduces dynamic high-resolution image preprocessing for the InternVL model and updates GGUF metadata handling, alongside various internal cleanups and bug fixes.

✨ New Features

  • Added dynamic high-resolution image preprocessing support for the InternVL model (requires Qianfan-OCR).
  • Added min/max dynamic patch support to GGUF metadata.

🐛 Bug Fixes

  • Fixed an issue where 0 values were being written to GGUF files.
  • Removed duplicated resolution candidates using an improved algorithm.
  • Added protection against division by zero errors.

Affected Symbols