llama.cpp Releases b9878, Fixes Tensor-Split Parameters

Core summary

One sentence to understand this update

The llama.cpp project released version b9878, primarily addressing stale tensor-split parameters for draft models and optimizing metadata for GQA attention.

Impact & opportunity

What this could mean

Developers relying on llama.cpp should update to this version to ensure correct functionality and performance for draft models and GQA attention.