●Change8

b10244

Breaking Changes
📦 llama-cppView on GitHub →
⚠ 1 breaking🐛 1 fixes🔧 2 symbols

Summary

The MSA logic has been refactored into a new memory implementation, moving it from `llama-kv-cache` to `llama-kv-cache-msa`. This release also includes minor whitespace fixes.

⚠️ Breaking Changes

  • The MSA logic has been moved from `llama-kv-cache` to a new implementation in `llama-kv-cache-msa`. This may require updating any code that directly references the old location of the MSA logic.

Migration Steps

  1. Update any code that directly references the MSA logic to use the new `llama-kv-cache-msa` module instead of `llama-kv-cache`.

🐛 Bug Fixes

  • Minor whitespace fixes.

Affected Symbols

llama-kv-cachellama-kv-cache-msa