Package 

Interface LlamaRuntime

  • All Implemented Interfaces:
    java.lang.AutoCloseable 
    
public interface LlamaRuntime
 implements AutoCloseable

    Thin boundary over the native llama.cpp runtime.

    The JNI/native binding is wired during the GGUF spike (plan Phase 2 — route A: vendored NDK/CMake build of llama.cpp, or route B: a maintained community AAR). Until the native library is bundled, isAvailable is false and GgufEngine does not register the engine, so GGUF models run via the orchestrator's cloud fallback instead of being accepted and failing.

    • Method Summary

      Modifier and Type Method Description
      abstract String generate(String prompt, Integer maxNewTokens)

      • Methods inherited from class ai.zerogpu.android.sdk.adapter.gguf.LlamaRuntime

        close

      • Methods inherited from class java.lang.Object

        clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait