Change8

b8164

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

Summary

This release introduces support for merging gate and exp weights in llama models and adds necessary components for all MoE models, alongside various pre-compiled binaries.

✨ New Features

  • Added option to merge gate and exp weights for llama models.
  • Added support for gate_up for all Mixture of Experts (MoE) models.

🐛 Bug Fixes

  • Simplified the condition for merging tensors in the conversion process.

Affected Symbols