llama.cpp Release b10907 Fixes MTP Context KV Cache

Core summary

One sentence to understand this update

llama.cpp version b10907 addresses MTP context KV cache allocation issues for model architectures like deepseek2, glm4moe, and cohere2moe.

Impact & opportunity

What this could mean

This update improves llama.cpp's compatibility and performance with specific MoE model architectures, benefiting developers using these models.