Topic: early return

The ggml/llama.cpp project released a fix for a memory leak in early return functionality (commit b10819). This update is crucial as it addresses a potential stability issue, enhancing the reliability of the software.