v0.0.9

* Major internal refactoring to improve code organization and maintainability
* Fixed critical bug where subsequent prompts would fail due to batch seq_id memory management
* Improved position tracking for continuous conversation support
* Enhanced error handling and debugging capabilities
* Added foundation for future chat optimization features
* Breaking change: Internal API restructuring (public API remains stable)

Author: netdur

BUILD.md

@@ -62,7 +62,7 @@ The project includes llama.cpp as a submodule and requires native binaries for v
 
 5. Run the build script:
    ```bash
-   ./build.sh
+   ./darwin/build.sh
    ```
 
 ### Available Platforms

CHANGELOG.md

@@ -1,3 +1,11 @@
+## 0.0.9
+* Major internal refactoring to improve code organization and maintainability
+* Fixed critical bug where subsequent prompts would fail due to batch seq_id memory management
+* Improved position tracking for continuous conversation support
+* Enhanced error handling and debugging capabilities
+* Added foundation for future chat optimization features
+* Breaking change: Internal API restructuring (public API remains stable)
+
 ## 0.0.8
 * disabled llava
 * compatible with llama.cpp 42ae10bb

darwin/run_build.sh

@@ -41,6 +41,7 @@ build_for_platform() {
           -DLLAMA_BUILD_TESTS=OFF \
           -DLLAMA_BUILD_EXAMPLES=OFF \
           -DLLAMA_BUILD_SERVER=OFF \
+          -DLLAMA_BUILD_TOOLS=OFF \
           -DCMAKE_BUILD_TYPE=Release \
           -G Xcode \
           -DCMAKE_TOOLCHAIN_FILE="${script_dir}/ios-arm64.toolchain.cmake" \
@@ -54,7 +55,7 @@ build_for_platform() {
           -DCMAKE_INSTALL_PREFIX="./install" \
           ..
 
-    cmake --build . --config Release
+    cmake --build . --config Release -j
     cmake --install . --config Release
 
     # Copy libraries

example/android/gradle/wrapper/gradle-wrapper.jar

@@ -18,7 +18,7 @@ void main() async {
     samplerParams.penaltyRepeat = 1.1;
 
     Llama.libraryPath = "bin/MAC_ARM64/libllama.dylib";
-    String modelPath = "/Users/adel/Workspace/gguf/gemma-3-12b-it-Q4_K_M.gguf";
+    String modelPath = "/Users/adel/Workspace/gguf/gemma-3-12b-it-q4_0.gguf";
     Llama llama = Llama(modelPath, ModelParams(), contextParams, samplerParams);
 
     llama.setPrompt(