webui: Server tools (#21237 )

* wip: server_tools * feat: Integrate with `/tools` endpoint * feat: Builtin + MCP + JSON Schema Tools WIP * refactor * displayName -> display_name * snake_case everywhere * rm redundant field * feat: Improvements * chore: update webui build output * refactor: Updates after server updates * chore: update webui build output * change arg to --tools all * feat: UI improvements * chore: update webui build output * add readme mention * llama-gen-docs * chore: update webui build output * chore: update webui build output * chore: update webui build output * feat: Reorganize settings sections * feat: Separate dialogs for MCP Servers Settings and Import/Export * feat: WIP * feat: WIP * feat: WIP * feat: WIP * feat: WIP * feat: WIP * WIP on allozaur/20677-webui-server-tools * feat: UI improvements * chore: Update package lock * chore: Run `npm audit fix` * feat: UI WIP * feat: UI * refactor: Desktop Icon Strip DRY * feat: Cleaner rendering and transition for ChatScreen * feat: UI improvements * feat: UI improvement * feat: Remove MCP Server "enable" switch from Tools submenu * chore: Run `npm audit fix` * feat: WIP * feat: Logic improvements * refactor: Cleanup * refactor: DRY * test: Fix Chat Sidebar UI Tests * chore: Update package lock * refactor: Cleanup * feat: Chat Message Action Card with Continue and Permission flow implementations * feat: Add agentic steering messages, draft messages and improve chat UX * fix: Search results UI * test: Fix unit test * feat: UI/UX improvements * refactor: Simplify `useToolsPanel` access in components * feat: Implement Processing Info Context API * feat: Implement 'Go back to chat' functionality for settings * feat: Enhance MCP Server management in Chat Form Attachments * style: Minor UI and branding adjustments * chore: Update webui static build output * chore: Formatting, linting & type checks * feat: Draft messages logic * feat: UI improvements * feat: Steering Messages improvements * refactor: Cleanup * refactor: Cleanup * feat: Improve UI * refactor: Settings navigation hook * refactor: DRY code * refactor: DRY ChatMessageUser UI components * refactor: Desktop Icon Strip DRY * refactor: Tools & permissions * fix: Navigation condition * refactor: Cleanup * refactor: Cleanup * refactor: Cleanup * fix: preserve reasoning_content in agentic flow --------- Co-authored-by: Xuan Son Nguyen <son@huggingface.co>
vulkan: add barrier after writetimestamp (#21865 )
2026-04-28 13:54:20 +02:00 · 2026-04-28 14:35:49 +03:00 · 2026-04-28 12:28:12 +02:00 · 2026-04-28 12:19:06 +02:00 · 2026-04-28 10:02:32 +03:00 · 2026-04-28 09:56:02 +03:00
694 changed files with 62575 additions and 20670 deletions
--- a/.devops/intel.Dockerfile
+++ b/.devops/intel.Dockerfile
@@ -1,4 +1,4 @@
-ARG ONEAPI_VERSION=2025.3.2-0-devel-ubuntu24.04
+ARG ONEAPI_VERSION=2025.3.3-0-devel-ubuntu24.04

 ## Build Image

--- a/.devops/nix/package.nix
+++ b/.devops/nix/package.nix
@@ -18,6 +18,7 @@
  vulkan-loader,
  openssl,
  shaderc,
+  spirv-headers,
  useBlas ?
    builtins.all (x: !x) [
      useCuda
@@ -145,6 +146,7 @@ effectiveStdenv.mkDerivation (finalAttrs: {
      ninja
      pkg-config
      git
+      spirv-headers
    ]
    ++ optionals useCuda [
      cudaPackages.cuda_nvcc
--- a/.devops/openvino.Dockerfile
+++ b/.devops/openvino.Dockerfile
@@ -2,7 +2,19 @@ ARG OPENVINO_VERSION_MAJOR=2026.0
 ARG OPENVINO_VERSION_FULL=2026.0.0.20965.c6d6a13a886
 ARG UBUNTU_VERSION=24.04

-# Optional proxy build arguments - empty by default
+# Intel GPU driver versions. https://github.com/intel/compute-runtime/releases
+ARG IGC_VERSION=v2.30.1
+ARG IGC_VERSION_FULL=2_2.30.1+20950
+ARG COMPUTE_RUNTIME_VERSION=26.09.37435.1
+ARG COMPUTE_RUNTIME_VERSION_FULL=26.09.37435.1-0
+ARG IGDGMM_VERSION=22.9.0
+
+# Intel NPU driver versions. https://github.com/intel/linux-npu-driver/releases
+ARG NPU_DRIVER_VERSION=v1.32.0
+ARG NPU_DRIVER_FULL=v1.32.0.20260402-23905121947
+ARG LIBZE1_VERSION=1.27.0-1~24.04~ppa2
+
+# Optional proxy build arguments
 ARG http_proxy=
 ARG https_proxy=

@@ -78,13 +90,47 @@ ARG http_proxy
 ARG https_proxy

 RUN apt-get update \
-    && apt-get install -y libgomp1 libtbb12 curl \
+    && apt-get install -y libgomp1 libtbb12 curl wget ocl-icd-libopencl1 \
    && apt autoremove -y \
    && apt clean -y \
    && rm -rf /tmp/* /var/tmp/* \
    && find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete \
    && find /var/cache -type f -delete

+# Install GPU drivers
+ARG IGC_VERSION
+ARG IGC_VERSION_FULL
+ARG COMPUTE_RUNTIME_VERSION
+ARG COMPUTE_RUNTIME_VERSION_FULL
+ARG IGDGMM_VERSION
+RUN mkdir /tmp/neo/ && cd /tmp/neo/ \
+    && wget https://github.com/intel/intel-graphics-compiler/releases/download/${IGC_VERSION}/intel-igc-core-${IGC_VERSION_FULL}_amd64.deb \
+    && wget https://github.com/intel/intel-graphics-compiler/releases/download/${IGC_VERSION}/intel-igc-opencl-${IGC_VERSION_FULL}_amd64.deb \
+    && wget https://github.com/intel/compute-runtime/releases/download/${COMPUTE_RUNTIME_VERSION}/intel-ocloc-dbgsym_${COMPUTE_RUNTIME_VERSION_FULL}_amd64.ddeb \
+    && wget https://github.com/intel/compute-runtime/releases/download/${COMPUTE_RUNTIME_VERSION}/intel-ocloc_${COMPUTE_RUNTIME_VERSION_FULL}_amd64.deb \
+    && wget https://github.com/intel/compute-runtime/releases/download/${COMPUTE_RUNTIME_VERSION}/intel-opencl-icd-dbgsym_${COMPUTE_RUNTIME_VERSION_FULL}_amd64.ddeb \
+    && wget https://github.com/intel/compute-runtime/releases/download/${COMPUTE_RUNTIME_VERSION}/intel-opencl-icd_${COMPUTE_RUNTIME_VERSION_FULL}_amd64.deb \
+    && wget https://github.com/intel/compute-runtime/releases/download/${COMPUTE_RUNTIME_VERSION}/libigdgmm12_${IGDGMM_VERSION}_amd64.deb \
+    && wget https://github.com/intel/compute-runtime/releases/download/${COMPUTE_RUNTIME_VERSION}/libze-intel-gpu1-dbgsym_${COMPUTE_RUNTIME_VERSION_FULL}_amd64.ddeb \
+    && wget https://github.com/intel/compute-runtime/releases/download/${COMPUTE_RUNTIME_VERSION}/libze-intel-gpu1_${COMPUTE_RUNTIME_VERSION_FULL}_amd64.deb \
+    && dpkg --install *.deb \
+    && rm -rf /tmp/neo/
+
+# Install NPU drivers
+ARG NPU_DRIVER_VERSION
+ARG NPU_DRIVER_FULL
+ARG LIBZE1_VERSION
+RUN mkdir /tmp/npu/ && cd /tmp/npu/ \
+    && wget https://github.com/intel/linux-npu-driver/releases/download/${NPU_DRIVER_VERSION}/linux-npu-driver-${NPU_DRIVER_FULL}-ubuntu2404.tar.gz \
+    && tar -xf linux-npu-driver-${NPU_DRIVER_FULL}-ubuntu2404.tar.gz \
+    && dpkg --install *.deb \
+    && rm -rf /tmp/npu/
+
+RUN cd /tmp \
+    && wget https://snapshot.ppa.launchpadcontent.net/kobuk-team/intel-graphics/ubuntu/20260324T100000Z/pool/main/l/level-zero-loader/libze1_${LIBZE1_VERSION}_amd64.deb \
+    && dpkg --install libze1_${LIBZE1_VERSION}_amd64.deb \
+    && rm libze1_${LIBZE1_VERSION}_amd64.deb
+
 COPY --from=build /app/lib/ /app/

 ### Full (all binaries)
--- a/.devops/vulkan.Dockerfile
+++ b/.devops/vulkan.Dockerfile
@@ -7,7 +7,7 @@ RUN apt update && apt install -y git build-essential cmake wget xz-utils

 # Install SSL and Vulkan SDK dependencies
 RUN apt install -y libssl-dev curl \
-    libxcb-xinput0 libxcb-xinerama0 libxcb-cursor-dev libvulkan-dev glslc
+    libxcb-xinput0 libxcb-xinerama0 libxcb-cursor-dev libvulkan-dev glslc spirv-headers

 # Build it
 WORKDIR /app
--- a/.github/pull_request_template.md
+++ b/.github/pull_request_template.md
@@ -6,7 +6,7 @@

 <!-- You can provide more details and link related discussions here. Delete this section if not applicable -->

-# Requirements
+## Requirements

 <!-- IMPORTANT: Please do NOT delete this section, otherwise your PR may be rejected -->

--- a/.github/workflows/build-and-test-snapdragon.yml
+++ b/.github/workflows/build-and-test-snapdragon.yml
@@ -0,0 +1,116 @@
+name: CI (snapdragon)
+
+on:
+  workflow_dispatch:
+  push:
+    branches:
+      - master
+    paths:
+      - '.github/workflows/build-and-test-snapdragon.yml'
+      - 'ggml/include/ggml-hexagon.h'
+      - 'ggml/src/ggml-hexagon/**'
+      - 'docs/backend/snapdragon/**'
+      - 'scripts/snapdragon/**'
+      - 'CMakePresets.json'
+
+  pull_request:
+    types: [opened, synchronize, reopened]
+    paths:
+      - '.github/workflows/build-and-test-snapdragon.yml'
+      - 'ggml/include/ggml-hexagon.h'
+      - 'ggml/src/ggml-hexagon/**'
+      - 'docs/backend/snapdragon/**'
+      - 'scripts/snapdragon/**'
+      - 'CMakePresets.json'
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.head_ref && github.ref || github.run_id }}
+  cancel-in-progress: true
+
+jobs:
+  android-ndk-snapdragon:
+    runs-on: ubuntu-latest
+    container:
+      image: 'ghcr.io/snapdragon-toolchain/arm64-android:v0.3'
+    defaults:
+      run:
+        shell: bash
+
+    steps:
+      - name: Clone
+        uses: actions/checkout@v6
+        with:
+          fetch-depth: 0
+          lfs: false
+
+      - name: Build Llama.CPP for Snapdragon Android
+        id: build_llama_cpp_snapdragon_android
+        run: |
+          cp docs/backend/snapdragon/CMakeUserPresets.json .
+          cmake --preset arm64-android-snapdragon-release -B build
+          cmake --build build
+          cmake --install build --prefix pkg-snapdragon/llama.cpp
+
+      - name: Upload Llama.CPP Snapdragon Android Build Artifact
+        if: ${{ always() && steps.build_llama_cpp_snapdragon_android.outcome == 'success' }}
+        uses: actions/upload-artifact@v6
+        with:
+          name: llama-cpp-android-arm64-snapdragon
+          path: pkg-snapdragon/llama.cpp
+
+  test-snapdragon-qdc:
+    name: Test on QDC Android Device (${{ matrix.device }})
+    needs: [android-ndk-snapdragon]
+    runs-on: ubuntu-slim
+    strategy:
+      fail-fast: false
+      matrix:
+        device: [SM8750, SM8650, SM8850]
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v6
+
+      - name: Download build artifact
+        uses: actions/download-artifact@v7
+        with:
+          name: llama-cpp-android-arm64-snapdragon
+          path: pkg-snapdragon/llama.cpp
+
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: '3.x'
+          cache: pip
+
+      - name: Install system dependencies
+        run: |
+          sudo apt-get update
+          sudo apt-get install -y curl unzip
+
+      - name: Install QDC SDK wheel
+        run: |
+          curl -fSL -o qdc_sdk.zip https://softwarecenter.qualcomm.com/api/download/software/tools/Qualcomm_Device_Cloud_SDK/All/0.2.3/qualcomm_device_cloud_sdk-0.2.3.zip
+          unzip qdc_sdk.zip -d qdc_sdk
+          pip install qdc_sdk/qualcomm_device_cloud_sdk-0.2.3-py3-none-any.whl
+
+      - name: Check QDC API key
+        id: check_secret
+        env:
+          QDC_API_KEY: ${{ secrets.QDC_API_KEY }}
+        run: echo "has-qdc-key=${{ env.QDC_API_KEY != '' }}" >> "$GITHUB_OUTPUT"
+
+      - name: Run QDC tests (${{ matrix.device }})
+        if: steps.check_secret.outputs.has-qdc-key == 'true'
+        run: |
+          python scripts/snapdragon/qdc/run_qdc_jobs.py \
+              --test       all \
+              --pkg-dir    pkg-snapdragon/llama.cpp \
+              --model-url  "https://huggingface.co/bartowski/Llama-3.2-1B-Instruct-GGUF/resolve/main/Llama-3.2-1B-Instruct-Q4_0.gguf" \
+              --device     ${{ matrix.device }}
+        env:
+          QDC_API_KEY: ${{ secrets.QDC_API_KEY }}
+
+      - name: Cleanup
+        if: always()
+        run: rm -rf pkg-snapdragon qdc_sdk qdc_sdk.zip
--- a/.github/workflows/build-android.yml
+++ b/.github/workflows/build-android.yml
@@ -1,26 +1,24 @@
 name: CI (android)

 on:
-  workflow_dispatch: # allows manual triggering
+  workflow_dispatch:
  push:
    branches:
      - master
-    paths: [
-      '.github/workflows/build-android.yml',
-      '**/CMakeLists.txt',
-      '**/.cmake',
-      '**/*.h',
-      '**/*.hpp',
-      '**/*.c',
-      '**/*.cpp'
-    ]
+    paths:
+      - '.github/workflows/build-android.yml'
+      - '**/CMakeLists.txt'
+      - '**/.cmake'
+      - '**/*.h'
+      - '**/*.hpp'
+      - '**/*.c'
+      - '**/*.cpp'

  pull_request:
    types: [opened, synchronize, reopened]
-    paths: [
-      '.github/workflows/build-android.yml',
-      'examples/llama.android/**'
-    ]
+    paths:
+      - '.github/workflows/build-android.yml'
+      - 'examples/llama.android/**'

 concurrency:
  group: ${{ github.workflow }}-${{ github.head_ref && github.ref || github.run_id }}
@@ -51,7 +49,7 @@ jobs:
          distribution: zulu

      - name: Setup Android SDK
-        uses: android-actions/setup-android@9fc6c4e9069bf8d3d10b2204b1fb8f6ef7065407 # v3
+        uses: android-actions/setup-android@40fd30fb8d7440372e1316f5d1809ec01dcd3699 # v4.0.1
        with:
          log-accepted-android-sdk-licenses: false

@@ -67,35 +65,24 @@ jobs:
    defaults:
      run:
        shell: bash
-    strategy:
-      matrix:
-        include:
-          - build: 'arm64-cpu'
-            defines: '-D ANDROID_ABI=arm64-v8a -D ANDROID_PLATFORM=android-31 -D CMAKE_TOOLCHAIN_FILE=${ANDROID_NDK_ROOT}/build/cmake/android.toolchain.cmake -D GGML_NATIVE=OFF -DGGML_CPU_ARM_ARCH=armv8.5-a+fp16+i8mm -G Ninja -D LLAMA_OPENSSL=OFF -D GGML_OPENMP=OFF'
-          - build: 'arm64-snapdragon'
-            defines: '--preset arm64-android-snapdragon-release'

    steps:
      - name: Clone
-        id: checkout
        uses: actions/checkout@v6
        with:
          fetch-depth: 0
          lfs: false

-      - name: Build Llama.CPP for Hexagon Android
-        id: build_llama_cpp_hexagon_android
+      - name: Build
+        id: ndk_build
        run: |
-          if [[ "${{ matrix.build }}" == "arm64-snapdragon" ]]; then
-            cp docs/backend/snapdragon/CMakeUserPresets.json .
-          fi
-          cmake ${{ matrix.defines }} -B build
+          cmake -D ANDROID_ABI=arm64-v8a -D ANDROID_PLATFORM=android-31 -D CMAKE_TOOLCHAIN_FILE=${ANDROID_NDK_ROOT}/build/cmake/android.toolchain.cmake -D GGML_NATIVE=OFF -DGGML_CPU_ARM_ARCH=armv8.5-a+fp16+i8mm -G Ninja -D LLAMA_OPENSSL=OFF -D GGML_OPENMP=OFF -B build
          cmake --build build
          cmake --install build --prefix pkg-adb/llama.cpp

-      - name: Upload Llama.CPP Hexagon Android Build Artifact
-        if: ${{ always() && steps.build_llama_cpp_hexagon_android.outcome == 'success' }}
+      - name: Upload Android Build Artifact
+        if: ${{ always() && steps.ndk_build.outcome == 'success' }}
        uses: actions/upload-artifact@v6
        with:
-          name: llama-cpp-android-${{ matrix.build }}
+          name: llama-cpp-android-arm64-cpu
          path: pkg-adb/llama.cpp
--- a/.github/workflows/build-cross.yml
+++ b/.github/workflows/build-cross.yml
@@ -246,6 +246,7 @@ jobs:
          apt-get install -y --no-install-recommends \
                  build-essential \
                  glslc \
+                  spirv-headers \
                  gcc-14-loongarch64-linux-gnu \
                  g++-14-loongarch64-linux-gnu \
                  libvulkan-dev:loong64
--- a/.github/workflows/build-openvino.yml
+++ b/.github/workflows/build-openvino.yml
@@ -0,0 +1,120 @@
+name: CI (openvino)
+
+on:
+  workflow_dispatch: # allows manual triggering
+  push:
+    branches:
+      - master
+    paths: [
+      '.github/workflows/build-openvino.yml',
+      '**/CMakeLists.txt',
+      '**/.cmake',
+      '**/*.h',
+      '**/*.hpp',
+      '**/*.c',
+      '**/*.cpp',
+    ]
+
+  pull_request:
+    types: [opened, synchronize, reopened]
+    paths: [
+      '.github/workflows/build-openvino.yml',
+      'ggml/src/ggml-openvino/**'
+    ]
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.head_ref && github.ref || github.run_id }}
+  cancel-in-progress: true
+
+env:
+  GGML_NLOOP: 3
+  GGML_N_THREADS: 1
+  LLAMA_LOG_COLORS: 1
+  LLAMA_LOG_PREFIX: 1
+  LLAMA_LOG_TIMESTAMPS: 1
+
+jobs:
+  ubuntu-24-openvino:
+    name: ubuntu-24-openvino-${{ matrix.openvino_device }}
+
+    concurrency:
+      group: openvino-${{ matrix.variant }}-${{ github.head_ref || github.ref }}
+      cancel-in-progress: false
+
+    strategy:
+      matrix:
+        include:
+          - variant: cpu
+            runner: '"ubuntu-24.04"'
+            openvino_device: "CPU"
+          - variant: gpu
+            runner: '["self-hosted","Linux","Intel","OpenVINO"]'
+            openvino_device: "GPU"
+
+    runs-on: ${{ fromJSON(matrix.runner) }}
+
+    env:
+      # Sync versions in build-openvino.yml, build-self-hosted.yml, release.yml, build-cache.yml, .devops/openvino.Dockerfile
+      OPENVINO_VERSION_MAJOR: "2026.0"
+      OPENVINO_VERSION_FULL: "2026.0.0.20965.c6d6a13a886"
+
+    steps:
+      - name: Clone
+        id: checkout
+        uses: actions/checkout@v6
+
+      - name: ccache
+        if: runner.environment == 'github-hosted'
+        uses: ggml-org/ccache-action@v1.2.21
+        with:
+          key: ubuntu-24-openvino-${{ matrix.variant }}-no-preset-v1
+          evict-old-files: 1d
+          save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
+
+      - name: Dependencies
+        id: depends
+        run: |
+          sudo apt-get update
+          sudo apt-get install -y build-essential libssl-dev libtbb12 cmake ninja-build python3-pip
+          sudo apt-get install -y ocl-icd-opencl-dev opencl-headers opencl-clhpp-headers intel-opencl-icd
+
+      - name: Use OpenVINO Toolkit Cache
+        if: runner.environment == 'github-hosted'
+        uses: actions/cache@v5
+        id: cache-openvino
+        with:
+          path: ./openvino_toolkit
+          key: openvino-toolkit-v${{ env.OPENVINO_VERSION_FULL }}-${{ runner.os }}
+
+      - name: Setup OpenVINO Toolkit
+        if: steps.cache-openvino.outputs.cache-hit != 'true'
+        uses: ./.github/actions/linux-setup-openvino
+        with:
+          path: ./openvino_toolkit
+          version_major: ${{ env.OPENVINO_VERSION_MAJOR }}
+          version_full: ${{ env.OPENVINO_VERSION_FULL }}
+
+      - name: Install OpenVINO dependencies
+        run: |
+          cd ./openvino_toolkit
+          chmod +x ./install_dependencies/install_openvino_dependencies.sh
+          echo "Y" | sudo -E ./install_dependencies/install_openvino_dependencies.sh
+
+      - name: Build
+        id: cmake_build
+        run: |
+          source ./openvino_toolkit/setupvars.sh
+          cmake -B build/ReleaseOV -G Ninja \
+            -DCMAKE_BUILD_TYPE=Release \
+            -DGGML_OPENVINO=ON
+          time cmake --build build/ReleaseOV --config Release -j $(nproc)
+
+      - name: Test
+        id: cmake_test
+        # TODO: fix and re-enable the `test-llama-archs` test below
+        run: |
+          cd ${{ github.workspace }}
+          if [ "${{ matrix.openvino_device }}" = "GPU" ]; then
+            export GGML_OPENVINO_DEVICE=GPU
+          fi
+          ctest --test-dir build/ReleaseOV -L main -E "test-llama-archs" --verbose --timeout 2000
--- a/.github/workflows/build-riscv.yml
+++ b/.github/workflows/build-riscv.yml
@@ -47,22 +47,10 @@ jobs:
    steps:
      - name: Install dependencies
        run: |
-          sudo apt-get update
-
-          # Install necessary packages
-          sudo apt-get install -y libatomic1 libtsan2 gcc-14 g++-14 cmake build-essential wget git-lfs
-
          # Set gcc-14 and g++-14 as the default compilers
          sudo update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-14 100
          sudo update-alternatives --install /usr/bin/g++ g++ /usr/bin/g++-14 100

-          if ! which rustc; then
-            # Install Rust stable version
-            sudo apt-get install -y rustup
-            rustup install stable
-            rustup default stable
-          fi
-
          git lfs install

      - name: GCC version check
@@ -74,12 +62,12 @@ jobs:
        id: checkout
        uses: actions/checkout@v6

-      # FIXME: Enable when ggml-org/ccache-action works on riscv64
-      # - name: ccache
-      #   uses: ggml-org/ccache-action@v1.2.21
-      #   with:
-      #     key: ubuntu-riscv64-native-sanitizer-${{ matrix.sanytizer }}-${{ matrix.build_type }}
-      #     save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
+      - name: ccache
+        uses: ggml-org/ccache-action@afde29e5b5422e5da23cb1f639e8baecadeadfc3 # https://github.com/ggml-org/ccache-action/pull/1
+        with:
+          key: ubuntu-riscv64-native-sanitizer-${{ matrix.sanitizer }}-${{ matrix.build_type }}
+          evict-old-files: 1d
+          save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}

      - name: Build
        id: cmake_build
--- a/.github/workflows/build-self-hosted.yml
+++ b/.github/workflows/build-self-hosted.yml
@@ -97,6 +97,36 @@ jobs:
          vulkaninfo --summary
          GG_BUILD_VULKAN=1 bash ./ci/run.sh ~/results/llama.cpp /mnt/llama.cpp

+  # TODO: investigate slight precision issues in some operations for test-backend-ops on the WebGPU backend.
+  #ggml-ci-nvidia-webgpu:
+  #  runs-on: [self-hosted, Linux, NVIDIA]
+
+  #  steps:
+  #    - name: Clone
+  #      id: checkout
+  #      uses: actions/checkout@v6
+
+  #    - name: Dawn Dependency
+  #      id: dawn-depends
+  #      run: |
+  #        DAWN_VERSION="v20260317.182325"
+  #        DAWN_OWNER="google"
+  #        DAWN_REPO="dawn"
+  #        DAWN_ASSET_NAME="Dawn-18eb229ef5f707c1464cc581252e7603c73a3ef0-ubuntu-latest-Release"
+  #        echo "Fetching release asset from https://github.com/google/dawn/releases/download/${DAWN_VERSION}/${DAWN_ASSET_NAME}.tar.gz"
+  #        curl -L -o artifact.tar.gz \
+  #          "https://github.com/google/dawn/releases/download/${DAWN_VERSION}/${DAWN_ASSET_NAME}.tar.gz"
+  #        mkdir dawn
+  #        tar -xvf artifact.tar.gz -C dawn --strip-components=1
+
+  #    - name: Test
+  #      id: ggml-ci
+  #      run: |
+  #        GG_BUILD_WEBGPU=1 \
+  #        GG_BUILD_WEBGPU_DAWN_PREFIX="$GITHUB_WORKSPACE/dawn" \
+  #        GG_BUILD_WEBGPU_DAWN_DIR="$GITHUB_WORKSPACE/dawn/lib64/cmake/Dawn" \
+  #          bash ./ci/run.sh ~/results/llama.cpp /mnt/llama.cpp
+
  # TODO: provision AMX-compatible machine
  #ggml-ci-cpu-amx:
  #  runs-on: [self-hosted, Linux, CPU, AMX]
@@ -141,61 +171,59 @@ jobs:
  #         amd-smi static
  #         GG_BUILD_ROCM=1 GG_BUILD_AMDGPU_TARGETS="gfx1101" bash ./ci/run.sh ~/results/llama.cpp /mnt/llama.cpp

-  # TODO: sandbox Mac runners
-  #  ggml-ci-mac-metal:
-  #    runs-on: [self-hosted, macOS, ARM64]
-  #
-  #    steps:
-  #      - name: Clone
-  #        id: checkout
-  #        uses: actions/checkout@v6
-  #
-  #      - name: Test
-  #        id: ggml-ci
-  #        run: |
-  #          GG_BUILD_METAL=1 bash ./ci/run.sh ~/results/llama.cpp ~/mnt/llama.cpp
-  #
-  #  ggml-ci-mac-webgpu:
-  #    runs-on: [self-hosted, macOS, ARM64]
-  #
-  #    steps:
-  #      - name: Clone
-  #        id: checkout
-  #        uses: actions/checkout@v6
-  #
-  #      - name: Dawn Dependency
-  #        id: dawn-depends
-  #        run: |
-  #          DAWN_VERSION="v2.0.0"
-  #          DAWN_OWNER="reeselevine"
-  #          DAWN_REPO="dawn"
-  #          DAWN_ASSET_NAME="Dawn-5e9a4865b1635796ccc77dd30057f2b4002a1355-macos-latest-Release"
-  #          echo "Fetching release asset from https://github.com/${DAWN_OWNER}/${DAWN_REPO}/releases/download/${DAWN_VERSION}/${DAWN_ASSET_NAME}.zip"
-  #          curl -L -o artifact.zip \
-  #            "https://github.com/${DAWN_OWNER}/${DAWN_REPO}/releases/download/${DAWN_VERSION}/${DAWN_ASSET_NAME}.zip"
-  #          mkdir dawn
-  #          unzip artifact.zip
-  #          tar -xvf ${DAWN_ASSET_NAME}.tar.gz -C dawn --strip-components=1
-  #
-  #      - name: Test
-  #        id: ggml-ci
-  #        run: |
-  #          GG_BUILD_WEBGPU=1 GG_BUILD_WEBGPU_DAWN_PREFIX="$GITHUB_WORKSPACE/dawn" \
-  #            bash ./ci/run.sh ~/results/llama.cpp ~/mnt/llama.cpp
-  #
-  #  ggml-ci-mac-vulkan:
-  #    runs-on: [self-hosted, macOS, ARM64]
-  #
-  #    steps:
-  #      - name: Clone
-  #        id: checkout
-  #        uses: actions/checkout@v6
-  #
-  #      - name: Test
-  #        id: ggml-ci
-  #        run: |
-  #          vulkaninfo --summary
-  #          GG_BUILD_VULKAN=1 bash ./ci/run.sh ~/results/llama.cpp ~/mnt/llama.cpp
+  ggml-ci-mac-metal:
+    runs-on: [self-hosted, macOS, ARM64]
+
+    steps:
+      - name: Clone
+        id: checkout
+        uses: actions/checkout@v6
+
+      - name: Test
+        id: ggml-ci
+        run: |
+          GG_BUILD_METAL=1 bash ./ci/run.sh ~/results/llama.cpp ~/mnt/llama.cpp
+
+  ggml-ci-mac-webgpu:
+    runs-on: [self-hosted, macOS, ARM64]
+
+    steps:
+      - name: Clone
+        id: checkout
+        uses: actions/checkout@v6
+
+      - name: Dawn Dependency
+        id: dawn-depends
+        run: |
+          DAWN_VERSION="v20260317.182325"
+          DAWN_OWNER="google"
+          DAWN_REPO="dawn"
+          DAWN_ASSET_NAME="Dawn-18eb229ef5f707c1464cc581252e7603c73a3ef0-macos-latest-Release"
+          echo "Fetching release asset from https://github.com/google/dawn/releases/download/${DAWN_VERSION}/${DAWN_ASSET_NAME}.tar.gz"
+          curl -L -o artifact.tar.gz \
+            "https://github.com/google/dawn/releases/download/${DAWN_VERSION}/${DAWN_ASSET_NAME}.tar.gz"
+          mkdir dawn
+          tar -xvf artifact.tar.gz -C dawn --strip-components=1
+
+      - name: Test
+        id: ggml-ci
+        run: |
+          GG_BUILD_WEBGPU=1 GG_BUILD_WEBGPU_DAWN_PREFIX="$GITHUB_WORKSPACE/dawn" \
+            bash ./ci/run.sh ~/results/llama.cpp ~/mnt/llama.cpp
+
+  ggml-ci-mac-vulkan:
+    runs-on: [self-hosted, macOS, ARM64]
+
+    steps:
+      - name: Clone
+        id: checkout
+        uses: actions/checkout@v6
+
+      - name: Test
+        id: ggml-ci
+        run: |
+          vulkaninfo --summary
+          GG_BUILD_VULKAN=1 bash ./ci/run.sh ~/results/llama.cpp ~/mnt/llama.cpp

  ggml-ci-linux-intel-vulkan:
    runs-on: [self-hosted, Linux, Intel]
@@ -237,6 +265,10 @@ jobs:
  ggml-ci-intel-openvino-gpu-low-perf:
    runs-on: [self-hosted, Linux, Intel, OpenVINO]

+    concurrency:
+      group: openvino-gpu-${{ github.head_ref || github.ref }}
+      cancel-in-progress: false
+
    env:
      # Sync versions in build.yml, build-self-hosted.yml, release.yml, build-cache.yml, .devops/openvino.Dockerfile
      OPENVINO_VERSION_MAJOR: "2026.0"
--- a/.github/workflows/build-sycl.yml
+++ b/.github/workflows/build-sycl.yml
@@ -0,0 +1,142 @@
+name: CI (sycl)
+
+on:
+  workflow_dispatch: # allows manual triggering
+  push:
+    branches:
+      - master
+    paths: [
+      '.github/workflows/build-sycl.yml',
+      '**/CMakeLists.txt',
+      '**/.cmake',
+      '**/*.h',
+      '**/*.hpp',
+      '**/*.c',
+      '**/*.cpp'
+    ]
+
+  pull_request:
+    types: [opened, synchronize, reopened]
+    paths: [
+      '.github/workflows/build-sycl.yml',
+      'ggml/src/ggml-sycl/**'
+    ]
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.head_ref && github.ref || github.run_id }}
+  cancel-in-progress: true
+
+env:
+  GGML_NLOOP: 3
+  GGML_N_THREADS: 1
+  LLAMA_LOG_COLORS: 1
+  LLAMA_LOG_PREFIX: 1
+  LLAMA_LOG_TIMESTAMPS: 1
+
+jobs:
+
+  ubuntu-24-sycl:
+    strategy:
+      matrix:
+        build: [fp32, fp16]
+        include:
+          - build: fp32
+            fp16: OFF
+          - build: fp16
+            fp16: ON
+
+    runs-on: ubuntu-24.04
+
+    env:
+      ONEAPI_ROOT: /opt/intel/oneapi/
+      ONEAPI_INSTALLER_VERSION: "2025.3.3"
+
+    continue-on-error: true
+
+    steps:
+      - uses: actions/checkout@v6
+
+      - name: Use oneAPI Installation Cache
+        uses: actions/cache@v5
+        id: cache-sycl
+        with:
+          path: ${{ env.ONEAPI_ROOT }}
+          key: oneAPI-${{ env.ONEAPI_INSTALLER_VERSION }}-${{ runner.os }}
+
+      - name: Download & Install oneAPI
+        shell: bash
+        if: steps.cache-sycl.outputs.cache-hit != 'true'
+        run: |
+          cd /tmp
+          wget https://registrationcenter-download.intel.com/akdlm/IRC_NAS/56f7923a-adb8-43f3-8b02-2b60fcac8cab/intel-deep-learning-essentials-2025.3.3.16_offline.sh -O intel-deep-learning-essentials_offline.sh
+          sudo bash intel-deep-learning-essentials_offline.sh -s -a --silent --eula accept
+
+      - name: Clone
+        id: checkout
+        uses: actions/checkout@v6
+
+      - name: ccache
+        uses: ggml-org/ccache-action@v1.2.21
+        with:
+          key: ubuntu-24-sycl-${{ matrix.build }}
+          evict-old-files: 1d
+          save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
+
+      - name: Build
+        id: cmake_build
+        run: |
+          source /opt/intel/oneapi/setvars.sh
+          cmake -B build \
+            -G "Ninja" \
+            -DCMAKE_BUILD_TYPE=Release \
+            -DGGML_SYCL=ON \
+            -DCMAKE_C_COMPILER=icx \
+            -DCMAKE_CXX_COMPILER=icpx \
+            -DLLAMA_OPENSSL=OFF \
+            -DGGML_NATIVE=OFF \
+            -DGGML_SYCL_F16=${{ matrix.fp16 }}
+          time cmake --build build --config Release -j $(nproc)
+
+  windows-latest-sycl:
+    runs-on: windows-2022
+
+    defaults:
+      run:
+        shell: bash
+
+    env:
+      WINDOWS_BASEKIT_URL: https://registrationcenter-download.intel.com/akdlm/IRC_NAS/b60765d1-2b85-4e85-86b6-cb0e9563a699/intel-deep-learning-essentials-2025.3.3.18_offline.exe
+      WINDOWS_DPCPP_MKL: intel.oneapi.win.cpp-dpcpp-common:intel.oneapi.win.mkl.devel:intel.oneapi.win.dnnl:intel.oneapi.win.tbb.devel
+      ONEAPI_ROOT: "C:/Program Files (x86)/Intel/oneAPI"
+      ONEAPI_INSTALLER_VERSION: "2025.3.3"
+    steps:
+      - name: Clone
+        id: checkout
+        uses: actions/checkout@v6
+
+      - name: Use oneAPI Installation Cache
+        uses: actions/cache@v5
+        id: cache-sycl
+        with:
+          path: ${{ env.ONEAPI_ROOT }}
+          key: oneAPI-${{ env.ONEAPI_INSTALLER_VERSION }}-${{ runner.os }}
+
+      - name: Download & Install oneAPI
+        shell: bash
+        if: steps.cache-sycl.outputs.cache-hit != 'true'
+        run: |
+          scripts/install-oneapi.bat $WINDOWS_BASEKIT_URL $WINDOWS_DPCPP_MKL
+
+      - name: ccache
+        uses: ggml-org/ccache-action@v1.2.21
+        with:
+          key: windows-latest-sycl
+          variant: ccache
+          evict-old-files: 1d
+          save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
+
+      # TODO: add ssl support ; we will also need to modify win-build-sycl.bat to accept user-specified args
+
+      - name: Build
+        id: cmake_build
+        run:  examples/sycl/win-build-sycl.bat
--- a/.github/workflows/build-vulkan.yml
+++ b/.github/workflows/build-vulkan.yml
@@ -93,4 +93,5 @@ jobs:
          export GGML_VK_DISABLE_F16=1
          export GGML_VK_DISABLE_COOPMAT=1
          # This is using llvmpipe and runs slower than other backends
-          ctest -L main --verbose --timeout 4800
+          # test-backend-ops is too slow on llvmpipe, skip it
+          ctest -L main -E test-backend-ops --verbose --timeout 900
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -267,6 +267,56 @@ jobs:
          wget https://huggingface.co/ggml-org/models/resolve/main/tinyllamas/stories260K-be.gguf
          ./bin/llama-completion -m stories260K-be.gguf -p "One day, Lily met a Shoggoth" -n 500 -c 256

+  android-arm64:
+    runs-on: ubuntu-latest
+
+    env:
+      NDK_VERSION: "29.0.14206865"
+
+    steps:
+      - name: Clone
+        id: checkout
+        uses: actions/checkout@v6
+
+      - name: ccache
+        uses: ggml-org/ccache-action@v1.2.21
+        with:
+          key: android-arm64
+          evict-old-files: 1d
+          save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
+
+      - name: Set up JDK
+        uses: actions/setup-java@v5
+        with:
+          java-version: 17
+          distribution: temurin
+
+      - name: Setup Android SDK
+        uses: android-actions/setup-android@40fd30fb8d7440372e1316f5d1809ec01dcd3699 # v4.0.1
+        with:
+          log-accepted-android-sdk-licenses: false
+
+      - name: Install NDK
+        run: |
+          sdkmanager "ndk;${{ env.NDK_VERSION }}"
+          echo "ANDROID_NDK=${ANDROID_SDK_ROOT}/ndk/${{ env.NDK_VERSION }}" >> $GITHUB_ENV
+
+      - name: Build
+        id: cmake_build
+        run: |
+          cmake -B build \
+            -DCMAKE_TOOLCHAIN_FILE=${ANDROID_NDK}/build/cmake/android.toolchain.cmake \
+            -DANDROID_ABI=arm64-v8a \
+            -DANDROID_PLATFORM=android-28 \
+            -DLLAMA_FATAL_WARNINGS=ON \
+            -DGGML_BACKEND_DL=ON \
+            -DGGML_NATIVE=OFF \
+            -DGGML_CPU_ALL_VARIANTS=ON \
+            -DGGML_OPENMP=OFF \
+            -DLLAMA_BUILD_BORINGSSL=ON \
+            -DGGML_RPC=ON
+          time cmake --build build --config Release -j $(nproc)
+
  ubuntu-latest-rpc:
    runs-on: ubuntu-latest

@@ -318,7 +368,7 @@ jobs:
        id: depends
        run: |
          sudo apt-get update
-          sudo apt-get install -y gcc-14 g++-14 build-essential glslc libvulkan-dev libssl-dev ninja-build
+          sudo apt-get install -y gcc-14 g++-14 build-essential glslc libvulkan-dev spirv-headers libssl-dev ninja-build
          echo "CC=gcc-14" >> "$GITHUB_ENV"
          echo "CXX=g++-14" >> "$GITHUB_ENV"

@@ -505,186 +555,6 @@ jobs:
            -DGGML_MUSA=ON
          time cmake --build build --config Release -j $(nproc)

-  ubuntu-22-sycl:
-    runs-on: ubuntu-22.04
-
-    continue-on-error: true
-
-    steps:
-      - uses: actions/checkout@v6
-
-      - name: add oneAPI to apt
-        shell: bash
-        run: |
-          cd /tmp
-          wget https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB
-          sudo apt-key add GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB
-          rm GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB
-          sudo add-apt-repository "deb https://apt.repos.intel.com/oneapi all main"
-
-      - name: install oneAPI dpcpp compiler
-        shell: bash
-        run: |
-          sudo apt update
-          sudo apt install intel-oneapi-compiler-dpcpp-cpp libssl-dev
-
-      - name: install oneAPI MKL library
-        shell: bash
-        run: |
-          sudo apt install intel-oneapi-mkl-devel
-
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v6
-
-      - name: ccache
-        uses: ggml-org/ccache-action@v1.2.21
-        with:
-          key: ubuntu-22-sycl
-          evict-old-files: 1d
-          save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
-
-      - name: Build
-        id: cmake_build
-        run: |
-          source /opt/intel/oneapi/setvars.sh
-          cmake -B build \
-            -DGGML_SYCL=ON \
-            -DCMAKE_C_COMPILER=icx \
-            -DCMAKE_CXX_COMPILER=icpx
-          time cmake --build build --config Release -j $(nproc)
-
-  ubuntu-22-sycl-fp16:
-    runs-on: ubuntu-22.04
-
-    continue-on-error: true
-
-    steps:
-      - uses: actions/checkout@v6
-
-      - name: add oneAPI to apt
-        shell: bash
-        run: |
-          cd /tmp
-          wget https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB
-          sudo apt-key add GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB
-          rm GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB
-          sudo add-apt-repository "deb https://apt.repos.intel.com/oneapi all main"
-
-      - name: install oneAPI dpcpp compiler
-        shell: bash
-        run: |
-          sudo apt update
-          sudo apt install intel-oneapi-compiler-dpcpp-cpp libssl-dev ninja-build
-
-      - name: install oneAPI MKL library
-        shell: bash
-        run: |
-          sudo apt install intel-oneapi-mkl-devel
-
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v6
-
-      - name: ccache
-        uses: ggml-org/ccache-action@v1.2.21
-        with:
-          key: ubuntu-22-sycl-fp16
-          evict-old-files: 1d
-          save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
-
-      - name: Build
-        id: cmake_build
-        run: |
-          source /opt/intel/oneapi/setvars.sh
-          cmake -B build \
-            -G "Ninja" \
-            -DCMAKE_BUILD_TYPE=Release \
-            -DGGML_SYCL=ON \
-            -DCMAKE_C_COMPILER=icx \
-            -DCMAKE_CXX_COMPILER=icpx \
-            -DGGML_SYCL_F16=ON
-          time cmake --build build --config Release -j $(nproc)
-
-  ubuntu-24-openvino:
-      name: ubuntu-24-openvino-${{ matrix.openvino_device }}
-      strategy:
-        matrix:
-          include:
-            - variant: cpu
-              runner: '"ubuntu-24.04"'
-              openvino_device: "CPU"
-            - variant: gpu
-              runner: '["self-hosted","Linux","X64","Intel"]'
-              openvino_device: "GPU"
-
-      runs-on: ${{ fromJSON(matrix.runner) }}
-
-      env:
-        # Sync versions in build.yml, build-self-hosted.yml, release.yml, build-cache.yml, .devops/openvino.Dockerfile
-        OPENVINO_VERSION_MAJOR: "2026.0"
-        OPENVINO_VERSION_FULL: "2026.0.0.20965.c6d6a13a886"
-
-      steps:
-        - name: Clone
-          id: checkout
-          uses: actions/checkout@v6
-
-        - name: ccache
-          if: runner.environment == 'github-hosted'
-          uses: ggml-org/ccache-action@v1.2.21
-          with:
-            key: ubuntu-24-openvino-${{ matrix.variant }}-no-preset-v1
-            evict-old-files: 1d
-            save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
-
-        - name: Dependencies
-          id: depends
-          run: |
-            sudo apt-get update
-            sudo apt-get install -y build-essential libssl-dev libtbb12 cmake ninja-build python3-pip
-            sudo apt-get install -y ocl-icd-opencl-dev opencl-headers opencl-clhpp-headers intel-opencl-icd
-
-        - name: Use OpenVINO Toolkit Cache
-          if: runner.environment == 'github-hosted'
-          uses: actions/cache@v5
-          id: cache-openvino
-          with:
-            path: ./openvino_toolkit
-            key: openvino-toolkit-v${{ env.OPENVINO_VERSION_FULL }}-${{ runner.os }}
-
-        - name: Setup OpenVINO Toolkit
-          if: steps.cache-openvino.outputs.cache-hit != 'true'
-          uses: ./.github/actions/linux-setup-openvino
-          with:
-            path: ./openvino_toolkit
-            version_major: ${{ env.OPENVINO_VERSION_MAJOR }}
-            version_full: ${{ env.OPENVINO_VERSION_FULL }}
-
-        - name: Install OpenVINO dependencies
-          run: |
-            cd ./openvino_toolkit
-            chmod +x ./install_dependencies/install_openvino_dependencies.sh
-            echo "Y" | sudo -E ./install_dependencies/install_openvino_dependencies.sh
-
-        - name: Build
-          id: cmake_build
-          run: |
-            source ./openvino_toolkit/setupvars.sh
-            cmake -B build/ReleaseOV -G Ninja \
-              -DCMAKE_BUILD_TYPE=Release \
-              -DGGML_OPENVINO=ON
-            time cmake --build build/ReleaseOV --config Release -j $(nproc)
-
-        - name: Test
-          id: cmake_test
-          # TODO: fix and re-enable the `test-llama-archs` test below
-          run: |
-            cd ${{ github.workspace }}
-            if [ "${{ matrix.openvino_device }}" = "GPU" ]; then
-              export GGML_OPENVINO_DEVICE=GPU
-            fi
-            ctest --test-dir build/ReleaseOV -L main -E "test-llama-archs" --verbose --timeout 2000

  windows-latest:
    runs-on: windows-2025
@@ -893,39 +763,6 @@ jobs:
          cmake --build build --config Release -j %NINJA_JOBS% -t ggml
          cmake --build build --config Release

-  windows-latest-sycl:
-    runs-on: windows-2022
-
-    defaults:
-      run:
-        shell: bash
-
-    env:
-      WINDOWS_BASEKIT_URL: https://registrationcenter-download.intel.com/akdlm/IRC_NAS/24751ead-ddc5-4479-b9e6-f9fe2ff8b9f2/intel-deep-learning-essentials-2025.2.1.25_offline.exe
-      WINDOWS_DPCPP_MKL: intel.oneapi.win.cpp-dpcpp-common:intel.oneapi.win.mkl.devel:intel.oneapi.win.dnnl:intel.oneapi.win.tbb.devel
-      ONEAPI_ROOT: "C:/Program Files (x86)/Intel/oneAPI"
-    steps:
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v6
-
-      - name: ccache
-        uses: ggml-org/ccache-action@v1.2.21
-        with:
-          key: windows-latest-sycl
-          variant: ccache
-          evict-old-files: 1d
-          save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
-
-      - name: Install
-        run:  |
-          scripts/install-oneapi.bat $WINDOWS_BASEKIT_URL $WINDOWS_DPCPP_MKL
-
-      # TODO: add ssl support ; we will also need to modify win-build-sycl.bat to accept user-specified args
-
-      - name: Build
-        id: cmake_build
-        run:  examples/sycl/win-build-sycl.bat

  windows-latest-hip:
    runs-on: windows-2022
@@ -1001,22 +838,14 @@ jobs:
    steps:
      - name: Install dependencies
        run: |
-          sudo apt-get update
-
          # Install necessary packages
-          sudo apt-get install -y libatomic1 libtsan2 gcc-14 g++-14 cmake build-essential libssl-dev wget git-lfs
+          sudo apt-get update
+          sudo apt-get install -y libssl-dev

          # Set gcc-14 and g++-14 as the default compilers
          sudo update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-14 100
          sudo update-alternatives --install /usr/bin/g++ g++ /usr/bin/g++-14 100

-          if ! which rustc; then
-            # Install Rust stable version
-            sudo apt-get install -y rustup
-            rustup install stable
-            rustup default stable
-          fi
-
          git lfs install

      - name: Check environment
@@ -1032,13 +861,12 @@ jobs:
        id: checkout
        uses: actions/checkout@v6

-      # FIXME: Enable when ggml-org/ccache-action works on riscv64
-      # - name: ccache
-      #   uses: ggml-org/ccache-action@v1.2.21
-      #   with:
-      #     key: ubuntu-cpu-riscv64-native
-      #     evict-old-files: 1d
-      #     save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
+      - name: ccache
+        uses: ggml-org/ccache-action@afde29e5b5422e5da23cb1f639e8baecadeadfc3 # https://github.com/ggml-org/ccache-action/pull/1
+        with:
+          key: ubuntu-cpu-riscv64-native
+          evict-old-files: 1d
+          save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}

      - name: Build
        id: cmake_build
--- a/.github/workflows/close-issue.yml
+++ b/.github/workflows/close-issue.yml
@@ -17,7 +17,7 @@ jobs:
    steps:
      - uses: actions/stale@v10
        with:
-          exempt-issue-labels: "refactoring,help wanted,good first issue,research 🔬,bug,roadmap"
+          exempt-issue-labels: "refactoring,help wanted,good first issue,research 🔬,bug,roadmap,security"
          days-before-issue-stale: 30
          days-before-issue-close: 14
          stale-issue-label: "stale"
--- a/.github/workflows/release.yml
+++ b/.github/workflows/release.yml
@@ -202,7 +202,7 @@ jobs:
            sudo apt-get install -y build-essential mesa-vulkan-drivers vulkan-sdk libssl-dev
          else
            sudo apt-get update -y
-            sudo apt-get install -y gcc-14 g++-14 build-essential glslc libvulkan-dev libssl-dev ninja-build
+            sudo apt-get install -y gcc-14 g++-14 build-essential glslc libvulkan-dev spirv-headers libssl-dev ninja-build
            echo "CC=gcc-14" >> "$GITHUB_ENV"
            echo "CXX=g++-14" >> "$GITHUB_ENV"
          fi
@@ -236,6 +236,75 @@ jobs:
          path: llama-${{ steps.tag.outputs.name }}-bin-ubuntu-vulkan-${{ matrix.build }}.tar.gz
          name: llama-bin-ubuntu-vulkan-${{ matrix.build }}.tar.gz

+  android-arm64:
+    runs-on: ubuntu-latest
+
+    env:
+      NDK_VERSION: "29.0.14206865"
+
+    steps:
+      - name: Clone
+        id: checkout
+        uses: actions/checkout@v6
+        with:
+          fetch-depth: 0
+
+      - name: ccache
+        uses: ggml-org/ccache-action@v1.2.21
+        with:
+          key: android-arm64
+          evict-old-files: 1d
+
+      - name: Set up JDK
+        uses: actions/setup-java@v5
+        with:
+          java-version: 17
+          distribution: temurin
+
+      - name: Setup Android SDK
+        uses: android-actions/setup-android@40fd30fb8d7440372e1316f5d1809ec01dcd3699 # v4.0.1
+        with:
+          log-accepted-android-sdk-licenses: false
+
+      - name: Install NDK
+        run: |
+          sdkmanager "ndk;${{ env.NDK_VERSION }}"
+          echo "ANDROID_NDK=${ANDROID_SDK_ROOT}/ndk/${{ env.NDK_VERSION }}" >> $GITHUB_ENV
+
+      - name: Build
+        id: cmake_build
+        run: |
+          cmake -B build \
+            -DCMAKE_TOOLCHAIN_FILE=${ANDROID_NDK}/build/cmake/android.toolchain.cmake \
+            -DANDROID_ABI=arm64-v8a \
+            -DANDROID_PLATFORM=android-28 \
+            -DCMAKE_INSTALL_RPATH='$ORIGIN' \
+            -DCMAKE_BUILD_WITH_INSTALL_RPATH=ON \
+            -DGGML_BACKEND_DL=ON \
+            -DGGML_NATIVE=OFF \
+            -DGGML_CPU_ALL_VARIANTS=ON \
+            -DLLAMA_FATAL_WARNINGS=ON \
+            -DGGML_OPENMP=OFF \
+            -DLLAMA_BUILD_BORINGSSL=ON \
+            ${{ env.CMAKE_ARGS }}
+          cmake --build build --config Release -j $(nproc)
+
+      - name: Determine tag name
+        id: tag
+        uses: ./.github/actions/get-tag-name
+
+      - name: Pack artifacts
+        id: pack_artifacts
+        run: |
+          cp LICENSE ./build/bin/
+          tar -czvf llama-${{ steps.tag.outputs.name }}-bin-android-arm64.tar.gz --transform "s,./,llama-${{ steps.tag.outputs.name }}/," -C ./build/bin .
+
+      - name: Upload artifacts
+        uses: actions/upload-artifact@v6
+        with:
+          path: llama-${{ steps.tag.outputs.name }}-bin-android-arm64.tar.gz
+          name: llama-bin-android-arm64.tar.gz
+
  ubuntu-24-openvino:
    runs-on: ubuntu-24.04

@@ -529,15 +598,29 @@ jobs:
        shell: bash

    env:
-      WINDOWS_BASEKIT_URL: https://registrationcenter-download.intel.com/akdlm/IRC_NAS/24751ead-ddc5-4479-b9e6-f9fe2ff8b9f2/intel-deep-learning-essentials-2025.2.1.25_offline.exe
+      WINDOWS_BASEKIT_URL: https://registrationcenter-download.intel.com/akdlm/IRC_NAS/b60765d1-2b85-4e85-86b6-cb0e9563a699/intel-deep-learning-essentials-2025.3.3.18_offline.exe
      WINDOWS_DPCPP_MKL: intel.oneapi.win.cpp-dpcpp-common:intel.oneapi.win.mkl.devel:intel.oneapi.win.dnnl:intel.oneapi.win.tbb.devel
      ONEAPI_ROOT: "C:/Program Files (x86)/Intel/oneAPI"
+      ONEAPI_INSTALLER_VERSION: "2025.3.3"

    steps:
      - name: Clone
        id: checkout
        uses: actions/checkout@v6

+      - name: Use oneAPI Installation Cache
+        uses: actions/cache@v5
+        id: cache-sycl
+        with:
+          path: ${{ env.ONEAPI_ROOT }}
+          key: oneAPI-${{ env.ONEAPI_INSTALLER_VERSION }}-${{ runner.os }}
+
+      - name: Download & Install oneAPI
+        shell: bash
+        if: steps.cache-sycl.outputs.cache-hit != 'true'
+        run: |
+          scripts/install-oneapi.bat $WINDOWS_BASEKIT_URL $WINDOWS_DPCPP_MKL
+
      - name: ccache
        uses: ggml-org/ccache-action@v1.2.21
        with:
@@ -545,10 +628,6 @@ jobs:
          variant: ccache
          evict-old-files: 1d

-      - name: Install
-        run:  |
-          scripts/install-oneapi.bat $WINDOWS_BASEKIT_URL $WINDOWS_DPCPP_MKL
-
      - name: Build
        id: cmake_build
        shell: cmd
@@ -601,6 +680,82 @@ jobs:
          path: llama-bin-win-sycl-x64.zip
          name: llama-bin-win-sycl-x64.zip

+  ubuntu-24-sycl:
+    strategy:
+      matrix:
+        build: [fp32, fp16]
+        include:
+          - build: fp32
+            fp16: OFF
+          - build: fp16
+            fp16: ON
+
+    runs-on: ubuntu-24.04
+
+    env:
+      ONEAPI_ROOT: /opt/intel/oneapi/
+      ONEAPI_INSTALLER_VERSION: "2025.3.3"
+
+    steps:
+      - name: Clone
+        id: checkout
+        uses: actions/checkout@v6
+        with:
+          fetch-depth: 0
+
+      - name: Use oneAPI Installation Cache
+        uses: actions/cache@v5
+        id: cache-sycl
+        with:
+          path: ${{ env.ONEAPI_ROOT }}
+          key: oneAPI-${{ env.ONEAPI_INSTALLER_VERSION }}-${{ runner.os }}
+
+      - name: Download & Install oneAPI
+        shell: bash
+        if: steps.cache-sycl.outputs.cache-hit != 'true'
+        run: |
+          cd /tmp
+          wget https://registrationcenter-download.intel.com/akdlm/IRC_NAS/56f7923a-adb8-43f3-8b02-2b60fcac8cab/intel-deep-learning-essentials-2025.3.3.16_offline.sh -O intel-deep-learning-essentials_offline.sh
+          sudo bash intel-deep-learning-essentials_offline.sh -s -a --silent --eula accept
+
+      - name: ccache
+        uses: ggml-org/ccache-action@v1.2.21
+        with:
+          key: ubuntu-24-sycl-${{ matrix.build }}
+          evict-old-files: 1d
+          save: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' }}
+
+      - name: Build
+        id: cmake_build
+        run: |
+          source /opt/intel/oneapi/setvars.sh
+          cmake -B build \
+            -G "Ninja" \
+            -DCMAKE_BUILD_TYPE=Release \
+            -DGGML_SYCL=ON \
+            -DCMAKE_C_COMPILER=icx \
+            -DCMAKE_CXX_COMPILER=icpx \
+            -DLLAMA_OPENSSL=OFF \
+            -DGGML_NATIVE=OFF \
+            -DGGML_SYCL_F16=${{ matrix.fp16 }}
+          time cmake --build build --config Release -j $(nproc)
+
+      - name: Determine tag name
+        id: tag
+        uses: ./.github/actions/get-tag-name
+
+      - name: Pack artifacts
+        id: pack_artifacts
+        run: |
+          cp LICENSE ./build/bin/
+          tar -czvf llama-${{ steps.tag.outputs.name }}-bin-ubuntu-sycl-${{ matrix.build }}-x64.tar.gz --transform "s,./,llama-${{ steps.tag.outputs.name }}/," -C ./build/bin .
+
+      - name: Upload artifacts
+        uses: actions/upload-artifact@v6
+        with:
+          path: llama-${{ steps.tag.outputs.name }}-bin-ubuntu-sycl-${{ matrix.build }}-x64.tar.gz
+          name: llama-bin-ubuntu-sycl-${{ matrix.build }}-x64.tar.gz
+
  ubuntu-22-rocm:
    runs-on: ubuntu-22.04

@@ -618,6 +773,11 @@ jobs:
        with:
          fetch-depth: 0

+      - name: Free up disk space
+        uses: ggml-org/free-disk-space@v1.3.1
+        with:
+          tool-cache: true
+
      - name: ccache
        uses: ggml-org/ccache-action@v1.2.21
        with:
@@ -971,6 +1131,8 @@ jobs:
      - ubuntu-cpu
      - ubuntu-vulkan
      - ubuntu-24-openvino
+      - ubuntu-24-sycl
+      - android-arm64
      - macOS-cpu
      - ios-xcode-build
      - openEuler-cann
@@ -1058,6 +1220,11 @@ jobs:
            - [Ubuntu arm64 (Vulkan)](https://github.com/ggml-org/llama.cpp/releases/download/${{ steps.tag.outputs.name }}/llama-${{ steps.tag.outputs.name }}-bin-ubuntu-vulkan-arm64.tar.gz)
            - [Ubuntu x64 (ROCm 7.2)](https://github.com/ggml-org/llama.cpp/releases/download/${{ steps.tag.outputs.name }}/llama-${{ steps.tag.outputs.name }}-bin-ubuntu-rocm-7.2-x64.tar.gz)
            - [Ubuntu x64 (OpenVINO)](https://github.com/ggml-org/llama.cpp/releases/download/${{ steps.tag.outputs.name }}/llama-${{ steps.tag.outputs.name }}-bin-ubuntu-openvino-${{ needs.ubuntu-24-openvino.outputs.openvino_version }}-x64.tar.gz)
+            - [Ubuntu x64 (SYCL FP32)](https://github.com/ggml-org/llama.cpp/releases/download/${{ steps.tag.outputs.name }}/llama-${{ steps.tag.outputs.name }}-bin-ubuntu-sycl-fp32-x64.tar.gz)
+            - [Ubuntu x64 (SYCL FP16)](https://github.com/ggml-org/llama.cpp/releases/download/${{ steps.tag.outputs.name }}/llama-${{ steps.tag.outputs.name }}-bin-ubuntu-sycl-fp16-x64.tar.gz)
+
+            **Android:**
+            - [Android arm64 (CPU)](https://github.com/ggml-org/llama.cpp/releases/download/${{ steps.tag.outputs.name }}/llama-${{ steps.tag.outputs.name }}-bin-android-arm64.tar.gz)

            **Windows:**
            - [Windows x64 (CPU)](https://github.com/ggml-org/llama.cpp/releases/download/${{ steps.tag.outputs.name }}/llama-${{ steps.tag.outputs.name }}-bin-win-cpu-x64.zip)
--- a/.github/workflows/server-self-hosted.yml
+++ b/.github/workflows/server-self-hosted.yml
@@ -84,41 +84,42 @@ jobs:
          export ${{ matrix.extra_args }}
          pytest -v -x -m "not slow"

-  server-cuda:
-    runs-on: [self-hosted, llama-server, Linux, NVIDIA]
-
-    name: server-cuda (${{ matrix.wf_name }})
-    strategy:
-      matrix:
-        build_type: [Release]
-        wf_name: ["GPUx1"]
-        include:
-          - build_type: Release
-            extra_args: "LLAMA_ARG_BACKEND_SAMPLING=1"
-            wf_name:    "GPUx1, backend-sampling"
-      fail-fast: false
-
-    steps:
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v6
-        with:
-          fetch-depth: 0
-          ref: ${{ github.event.inputs.sha || github.event.pull_request.head.sha || github.sha || github.head_ref || github.ref_name }}
-
-      - name: Build
-        id: cmake_build
-        run: |
-          cmake -B build -DGGML_SCHED_NO_REALLOC=ON
-          cmake --build build --config ${{ matrix.build_type }} -j $(sysctl -n hw.logicalcpu) --target llama-server
-
-      - name: Tests
-        id: server_integration_tests
-        if: ${{ (!matrix.disabled_on_pr || !github.event.pull_request) }}
-        run: |
-          cd tools/server/tests
-          python3 -m venv venv
-          source venv/bin/activate
-          pip install -r requirements.txt
-          export ${{ matrix.extra_args }}
-          pytest -v -x -m "not slow"
+  # TODO: provision CUDA runner
+  #  server-cuda:
+  #    runs-on: [self-hosted, llama-server, Linux, NVIDIA]
+  #
+  #    name: server-cuda (${{ matrix.wf_name }})
+  #    strategy:
+  #      matrix:
+  #        build_type: [Release]
+  #        wf_name: ["GPUx1"]
+  #        include:
+  #          - build_type: Release
+  #            extra_args: "LLAMA_ARG_BACKEND_SAMPLING=1"
+  #            wf_name:    "GPUx1, backend-sampling"
+  #      fail-fast: false
+  #
+  #    steps:
+  #      - name: Clone
+  #        id: checkout
+  #        uses: actions/checkout@v6
+  #        with:
+  #          fetch-depth: 0
+  #          ref: ${{ github.event.inputs.sha || github.event.pull_request.head.sha || github.sha || github.head_ref || github.ref_name }}
+  #
+  #      - name: Build
+  #        id: cmake_build
+  #        run: |
+  #          cmake -B build -DGGML_SCHED_NO_REALLOC=ON
+  #          cmake --build build --config ${{ matrix.build_type }} -j $(sysctl -n hw.logicalcpu) --target llama-server
+  #
+  #      - name: Tests
+  #        id: server_integration_tests
+  #        if: ${{ (!matrix.disabled_on_pr || !github.event.pull_request) }}
+  #        run: |
+  #          cd tools/server/tests
+  #          python3 -m venv venv
+  #          source venv/bin/activate
+  #          pip install -r requirements.txt
+  #          export ${{ matrix.extra_args }}
+  #          pytest -v -x -m "not slow"
--- a/.gitignore
+++ b/.gitignore
@@ -34,7 +34,6 @@
 /.vscode/
 /nppBackup

-
 # Coverage

 /gcovr-report/
@@ -74,6 +73,7 @@
 !/models/templates

 # Zig
+
 /zig-out/
 /zig-cache/

@@ -93,6 +93,7 @@
 !/examples/sycl/*.sh

 # Server Web UI temporary files
+
 /tools/server/webui/node_modules
 /tools/server/webui/dist
 # we no longer use gz for index.html
@@ -106,9 +107,11 @@ __pycache__/
 poetry.toml

 # Nix
+
 /result

 # Test binaries
+
 /tests/test-backend-ops
 /tests/test-double-float
 /tests/test-grad0
@@ -124,6 +127,7 @@ poetry.toml
 /tests/test-tokenizer-1-spm

 # Scripts
+
 !/scripts/install-oneapi.bat

 # Generated by scripts
@@ -132,16 +136,24 @@ poetry.toml
 /wikitext-2-raw/

 # Test models for lora adapters
+
 /lora-tests

 # Local scripts
+
 /run-vim.sh
 /run-chat.sh
 /run-spec.sh
 /.ccache/

 # IDE
+
 /*.code-workspace
 /.windsurf/
 # emscripten
 a.out.*
+
+# AGENTS
+
+AGENTS.local.md
+.pi/SYSTEM.md
--- a/.pi/gg/SYSTEM.md
+++ b/.pi/gg/SYSTEM.md
@@ -0,0 +1,33 @@
+You are a coding agent. Here are some very important rules that you must follow:
+
+General:
+- By very precise and concise when writing code, comments, explanations, etc.
+- PR and commit titles format: `<module> : <title>`. Lookup recents for examples
+- Don't try to build or run the code unless you are explicitly asked to do so
+
+Coding:
+- When in doubt, always refer to the CONTRIBUTING.md file of the project
+- When referencing issues or PRs in comments, use the format:
+  - C/C++ code: `// ref: <url>`
+  - Other (CMake, etc.): `# ref: <url>`
+
+Pull requests (PRs):
+- New branch names are prefixed with "gg/"
+- Before opening a pull request, ask the user to confirm the description
+- When creating a pull request, look for the repository's PR template and follow it
+- For the AI usage disclosure section, write "YES. llama.cpp + pi"
+- Always create the pull requests in draft mode
+
+Commits:
+- On every commit that you make, include a "Assisted-by: llama.cpp:local pi" tag
+- Do not explicitly set the git author in commits - rely on the default git config
+
+Resources (read on demand):
+- [CONTRIBUTING.md](CONTRIBUTING.md)
+- [Build documentation](docs/build.md)
+- [Server usage documentation](tools/server/README.md)
+- [Server development documentation](tools/server/README-dev.md)
+- [PEG parser](docs/development/parsing.md)
+- [Auto parser](docs/autoparser.md)
+- [Jinja engine](common/jinja/README.md)
+- [PR template](.github/pull_request_template.md)
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -225,7 +225,7 @@ foreach(FILE_PATH ${EXTRA_LICENSES})
 endforeach()

 if (LLAMA_BUILD_COMMON)
-    license_generate(common)
+    license_generate(llama-common)
 endif()

 #
@@ -249,6 +249,10 @@ set_target_properties(llama

 install(TARGETS llama LIBRARY PUBLIC_HEADER)

+if (LLAMA_BUILD_COMMON)
+    install(TARGETS llama-common LIBRARY)
+endif()
+
 configure_package_config_file(
        ${CMAKE_CURRENT_SOURCE_DIR}/cmake/llama-config.cmake.in
        ${CMAKE_CURRENT_BINARY_DIR}/llama-config.cmake
--- a/30
+++ b/30
@@ -1,5 +1,21 @@
 # collaborators can optionally add themselves here to indicate their availability for reviewing related PRs
-# multiplie collaborators per item can be specified
+# multiple collaborators per item can be specified
+#
+# ggml-org/ci               : CISC, danbev, ggerganov, netrunnereve, ngxson, taronaeo
+# ggml-org/ggml-cann        : hipudding
+# ggml-org/ggml-cuda        : JohannesGaessler, am17an, IMbackK, ORippler
+# ggml-org/ggml-hexagon     : lhez, max-krasnyansky
+# ggml-org/ggml-metal       : ggerganov
+# ggml-org/ggml-opencl      : lhez, max-krasnyansky
+# ggml-org/ggml-rpc         : rgerganov
+# ggml-org/ggml-sycl        : arthw
+# ggml-org/ggml-vulkan      : 0cc4m, jeffbolznv
+# ggml-org/ggml-webgpu      : reeselevine
+# ggml-org/ggml-zdnn        : taronaeo
+# ggml-org/llama-common     : ggerganov, aldehir, angt, danbev, ngxson, pwilkin
+# ggml-org/llama-mtmd       : ngxson
+# ggml-org/llama-server     : ggerganov, ngxson, allozaur, angt, ServeurpersoCom
+# ggml-org/llama-webui      : allozaur

 /.devops/*.Dockerfile                   @ngxson
 /.github/actions/                       @ggml-org/ci
@@ -7,6 +23,7 @@
 /ci/                                    @ggerganov
 /cmake/                                 @ggerganov
 /common/                                @ggml-org/llama-common
+/common/fit.*                           @JohannesGaessler
 /common/jinja/                          @CISC
 /common/ngram-map.*                     @srogmann
 /convert_*.py                           @CISC
@@ -36,28 +53,29 @@
 /examples/speculative/                  @ggerganov
 /ggml/cmake/                            @ggerganov
 /ggml/include/                          @ggerganov
+/ggml/src/ggml-backend-meta.cpp         @JohannesGaessler
 /ggml/src/ggml-cann/                    @ggml-org/ggml-cann
 /ggml/src/ggml-common.h                 @ggerganov
 /ggml/src/ggml-cpu/                     @ggerganov
 /ggml/src/ggml-cpu/spacemit/            @alex-spacemit
 /ggml/src/ggml-cuda/                    @ggml-org/ggml-cuda
-/ggml/src/ggml-cuda/fattn-wmma*         @IMbackK
-/ggml/src/ggml-hip/                     @IMbackK
 /ggml/src/ggml-cuda/vendors/hip.h       @IMbackK
+/ggml/src/ggml-cuda/fattn-wmma*         @IMbackK
+/ggml/src/ggml-hexagon/                 @ggml-org/ggml-hexagon
+/ggml/src/ggml-hip/                     @IMbackK
 /ggml/src/ggml-impl.h                   @ggerganov
 /ggml/src/ggml-metal/                   @ggml-org/ggml-metal
 /ggml/src/ggml-opencl/                  @ggml-org/ggml-opencl
-/ggml/src/ggml-hexagon/                 @ggml-org/ggml-hexagon
+/ggml/src/ggml-openvino/                @cavusmustafa @wine99
 /ggml/src/ggml-opt.cpp                  @JohannesGaessler
 /ggml/src/ggml-quants.*                 @ggerganov
 /ggml/src/ggml-rpc/                     @ggml-org/ggml-rpc
 /ggml/src/ggml-sycl/                    @ggml-org/ggml-sycl
 /ggml/src/ggml-threading.*              @ggerganov
-/ggml/src/ggml-vulkan/                  @ggml-org/ggml-vulkan
 /ggml/src/ggml-virtgpu/                 @kpouget
+/ggml/src/ggml-vulkan/                  @ggml-org/ggml-vulkan
 /ggml/src/ggml-webgpu/                  @ggml-org/ggml-webgpu
 /ggml/src/ggml-zdnn/                    @ggml-org/ggml-zdnn @Andreas-Krebbel @AlekseiNikiforovIBM
-/ggml/src/ggml-openvino/                @cavusmustafa @wine99
 /ggml/src/ggml.c                        @ggerganov
 /ggml/src/ggml.cpp                      @ggerganov
 /ggml/src/gguf.cpp                      @JohannesGaessler @Green-Sky
--- a/cmake/arm64-linux-clang.cmake
+++ b/cmake/arm64-linux-clang.cmake
@@ -0,0 +1,17 @@
+set( CMAKE_SYSTEM_NAME Linux )
+set( CMAKE_SYSTEM_PROCESSOR arm64 )
+
+set( target aarch64-linux-gnu )
+
+set( CMAKE_C_COMPILER    clang )
+set( CMAKE_CXX_COMPILER  clang++ )
+
+set( CMAKE_C_COMPILER_TARGET   ${target} )
+set( CMAKE_CXX_COMPILER_TARGET ${target} )
+
+set( arch_c_flags "-march=armv8.7-a -fvectorize -ffp-model=fast -fno-finite-math-only" )
+set( warn_c_flags "-Wno-format -Wno-unused-variable -Wno-unused-function -Wno-gnu-zero-variadic-macro-arguments" )
+
+set( CMAKE_C_FLAGS_INIT   "${arch_c_flags} ${warn_c_flags}" )
+set( CMAKE_CXX_FLAGS_INIT "${arch_c_flags} ${warn_c_flags}" )
+
--- a/common/CMakeLists.txt
+++ b/common/CMakeLists.txt
@@ -1,9 +1,11 @@
-# common
-
 find_package(Threads REQUIRED)

 llama_add_compile_flags()

+#
+# llama-common-base
+#
+
 # Build info header

 if(EXISTS "${PROJECT_SOURCE_DIR}/.git")
@@ -33,17 +35,25 @@ endif()

 set(TEMPLATE_FILE "${CMAKE_CURRENT_SOURCE_DIR}/build-info.cpp.in")
 set(OUTPUT_FILE   "${CMAKE_CURRENT_BINARY_DIR}/build-info.cpp")
+
 configure_file(${TEMPLATE_FILE} ${OUTPUT_FILE})

-set(TARGET build_info)
-add_library(${TARGET} OBJECT ${OUTPUT_FILE})
+set(TARGET llama-common-base)
+add_library(${TARGET} STATIC ${OUTPUT_FILE})
+
+target_include_directories(${TARGET} PUBLIC .)
+
 if (BUILD_SHARED_LIBS)
    set_target_properties(${TARGET} PROPERTIES POSITION_INDEPENDENT_CODE ON)
 endif()

-set(TARGET common)
+#
+# llama-common
+#

-add_library(${TARGET} STATIC
+set(TARGET llama-common)
+
+add_library(${TARGET}
    arg.cpp
    arg.h
    base64.hpp
@@ -63,6 +73,8 @@ add_library(${TARGET} STATIC
    debug.h
    download.cpp
    download.h
+    fit.cpp
+    fit.h
    hf-cache.cpp
    hf-cache.h
    http.h
@@ -106,17 +118,24 @@ add_library(${TARGET} STATIC
    jinja/caps.h
    )

+set_target_properties(${TARGET} PROPERTIES
+    VERSION ${LLAMA_INSTALL_VERSION}
+    SOVERSION 0
+    MACHO_CURRENT_VERSION 0 # keep macOS linker from seeing oversized version number
+)
+
 target_include_directories(${TARGET} PUBLIC . ../vendor)
 target_compile_features   (${TARGET} PUBLIC cxx_std_17)

 if (BUILD_SHARED_LIBS)
    set_target_properties(${TARGET} PROPERTIES POSITION_INDEPENDENT_CODE ON)
+
+    # TODO: make fine-grained exports in the future
+    set_target_properties(${TARGET} PROPERTIES WINDOWS_EXPORT_ALL_SYMBOLS ON)
 endif()

-target_link_libraries(${TARGET} PRIVATE
-    build_info
-    cpp-httplib
-)
+target_link_libraries(${TARGET} PUBLIC  llama-common-base)
+target_link_libraries(${TARGET} PRIVATE cpp-httplib)

 if (LLAMA_LLGUIDANCE)
    include(ExternalProject)
--- a/common/arg.cpp
+++ b/common/arg.cpp
--- a/common/arg.h
+++ b/common/arg.h
@@ -25,7 +25,8 @@ struct common_arg {
    const char * value_hint_2 = nullptr; // for second arg value
    const char * env          = nullptr;
    std::string help;
-    bool is_sparam = false; // is current arg a sampling param?
+    bool is_sampling = false; // is current arg a sampling param?
+    bool is_spec = false; // is current arg a speculative decoding param?
    bool is_preset_only = false; // is current arg preset-only (not treated as CLI arg)
    void (*handler_void)   (common_params & params) = nullptr;
    void (*handler_string) (common_params & params, const std::string &) = nullptr;
@@ -74,7 +75,8 @@ struct common_arg {
    common_arg & set_examples(std::initializer_list<enum llama_example> examples);
    common_arg & set_excludes(std::initializer_list<enum llama_example> excludes);
    common_arg & set_env(const char * env);
-    common_arg & set_sparam();
+    common_arg & set_sampling();
+    common_arg & set_spec();
    common_arg & set_preset_only();
    bool in_example(enum llama_example ex);
    bool is_exclude(enum llama_example ex);
--- a/common/build-info.cpp.in
+++ b/common/build-info.cpp.in
@@ -1,4 +1,35 @@
+#include "build-info.h"
+
+#include <cstdio>
+#include <string>
+
 int LLAMA_BUILD_NUMBER = @LLAMA_BUILD_NUMBER@;
-char const *LLAMA_COMMIT = "@LLAMA_BUILD_COMMIT@";
-char const *LLAMA_COMPILER = "@BUILD_COMPILER@";
-char const *LLAMA_BUILD_TARGET = "@BUILD_TARGET@";
+char const * LLAMA_COMMIT = "@LLAMA_BUILD_COMMIT@";
+char const * LLAMA_COMPILER = "@BUILD_COMPILER@";
+char const * LLAMA_BUILD_TARGET = "@BUILD_TARGET@";
+
+int llama_build_number(void) {
+    return LLAMA_BUILD_NUMBER;
+}
+
+const char * llama_commit(void) {
+    return LLAMA_COMMIT;
+}
+
+const char * llama_compiler(void) {
+    return LLAMA_COMPILER;
+}
+
+const char * llama_build_target(void) {
+    return LLAMA_BUILD_TARGET;
+}
+
+const char * llama_build_info(void) {
+    static std::string s = "b" + std::to_string(LLAMA_BUILD_NUMBER) + "-" + LLAMA_COMMIT;
+    return s.c_str();
+}
+
+void llama_print_build_info(void) {
+    fprintf(stderr, "%s: build = %d (%s)\n",      __func__, llama_build_number(), llama_commit());
+    fprintf(stderr, "%s: built with %s for %s\n", __func__, llama_compiler(), llama_build_target());
+}
--- a/common/build-info.h
+++ b/common/build-info.h
@@ -0,0 +1,11 @@
+#pragma once
+
+int llama_build_number(void);
+
+const char * llama_commit(void);
+const char * llama_compiler(void);
+
+const char * llama_build_target(void);
+const char * llama_build_info(void);
+
+void llama_print_build_info(void);
--- a/common/chat-auto-parser-generator.cpp
+++ b/common/chat-auto-parser-generator.cpp
@@ -69,6 +69,10 @@ common_chat_params peg_generator::generate_parser(const common_chat_template &
                auto         schema   = function.contains("parameters") ? function.at("parameters") : json::object();
                builder.resolve_refs(schema);
            });
+            if (has_response_format) {
+                auto schema = inputs.json_schema;
+                builder.resolve_refs(schema);
+            }
            parser.build_grammar(builder, data.grammar_lazy);
        });

@@ -194,10 +198,19 @@ common_peg_parser analyze_tools::build_tool_parser_json_native(parser_build_cont
        args_field = format.function_field + "." + args_field;
    }

-    auto tools_parser = p.standard_json_tools(
-        format.section_start, format.section_end, inputs.tools, inputs.parallel_tool_calls,
-        inputs.tool_choice == COMMON_CHAT_TOOL_CHOICE_REQUIRED, name_field, args_field, format.tools_array_wrapped,
-        format.fun_name_is_key, format.id_field, format.gen_id_field, format.parameter_order);
+    auto tools_parser = p.eps();
+    if (format.section_start.empty() && !format.per_call_start.empty()) {
+        auto single_tool_parser = p.standard_json_tools(
+            format.per_call_start, format.per_call_end, inputs.tools, inputs.parallel_tool_calls,
+            inputs.tool_choice == COMMON_CHAT_TOOL_CHOICE_REQUIRED, name_field, args_field, format.tools_array_wrapped,
+            format.fun_name_is_key, format.id_field, format.gen_id_field, format.parameter_order);
+        tools_parser = p.trigger_rule("tool-calls", p.one_or_more(single_tool_parser + p.space()));
+    } else {
+        tools_parser = p.standard_json_tools(
+            format.section_start, format.section_end, inputs.tools, inputs.parallel_tool_calls,
+            inputs.tool_choice == COMMON_CHAT_TOOL_CHOICE_REQUIRED, name_field, args_field, format.tools_array_wrapped,
+            format.fun_name_is_key, format.id_field, format.gen_id_field, format.parameter_order);
+    }

    // Handle content wrappers if present
    if (ctx.content && ctx.content->is_always_wrapped()) {
@@ -332,58 +345,36 @@ common_peg_parser analyze_tools::build_tool_parser_tag_tagged(parser_build_conte
    const auto & inputs      = ctx.inputs;
    bool         force_tools = inputs.tool_choice == COMMON_CHAT_TOOL_CHOICE_REQUIRED;

+    auto until_suffix = p.rule("until-suffix", p.until(arguments.value_suffix));
+
    common_peg_parser tool_choice = p.choice();

    foreach_function(inputs.tools, [&](const json & tool) {
        const auto &          func       = tool.at("function");
        std::string           name       = func.at("name");
-        const auto &          params     = func.contains("parameters") ? func.at("parameters") : json::object();
+        auto                  params     = func.contains("parameters") ? func.at("parameters") : json::object();
        const auto &          properties = params.contains("properties") ? params.at("properties") : json::object();
+
        std::set<std::string> required;
+        if (params.contains("required")) {
+            params.at("required").get_to(required);
+        }
+
+        auto schema_info = common_schema_info();
+        schema_info.resolve_refs(params);

        // Build parser for each argument, separating required and optional
        std::vector<common_peg_parser> required_parsers;
        std::vector<common_peg_parser> optional_parsers;
        for (const auto & [param_name, param_schema] : properties.items()) {
-            bool        is_required = required.find(param_name) != required.end();
-            std::string type        = "object";
-            if (param_schema.contains("type")) {
-                const auto & type_obj = param_schema.at("type");
-                if (type_obj.is_string()) {
-                    type_obj.get_to(type);
-                } else if (type_obj.is_array()) {
-                    // Handle nullable types like ["string", "null"]
-                    for (const auto & t : type_obj) {
-                        if (t.is_string() && t.get<std::string>() != "null") {
-                            type = t.get<std::string>();
-                            break;
-                        }
-                    }
-                } else if (type_obj.is_object()) {
-                    if (type_obj.contains("type") && type_obj.at("type").is_string()) {
-                        type_obj.at("type").get_to(type);
-                    }
-                }
-            }
-            // Infer string type from enum values when type is unspecified
-            if (type == "object" && param_schema.contains("enum")) {
-                const auto & enum_vals = param_schema.at("enum");
-                if (enum_vals.is_array()) {
-                    for (const auto & v : enum_vals) {
-                        if (v.is_string()) {
-                            type = "string";
-                            break;
-                        }
-                    }
-                }
-            }
+            bool is_required = required.find(param_name) != required.end();

            auto arg =
                p.tool_arg(p.tool_arg_open(arguments.name_prefix + p.tool_arg_name(p.literal(param_name)) +
                                           arguments.name_suffix) +
                           arguments.value_prefix +
-                           (type == "string" ?
-                                p.tool_arg_string_value(p.schema(p.until(arguments.value_suffix),
+                           (schema_info.resolves_to_string(param_schema) ?
+                                p.tool_arg_string_value(p.schema(until_suffix,
                                                                 "tool-" + name + "-arg-" + param_name + "-schema",
                                                                 param_schema, true)) :
                                p.tool_arg_json_value(p.schema(
@@ -414,7 +405,7 @@ common_peg_parser analyze_tools::build_tool_parser_tag_tagged(parser_build_conte
            for (const auto & opt : optional_parsers) {
                any_opt |= opt;
            }
-            args_seq = args_seq + p.repeat(p.space() + any_opt, 0, (int) optional_parsers.size());
+            args_seq = args_seq + p.repeat(p.space() + any_opt, 0, -1);
        }

        if (!arguments.start.empty()) {
@@ -452,14 +443,14 @@ common_peg_parser analyze_tools::build_tool_parser_tag_tagged(parser_build_conte
    if (!format.per_call_start.empty()) {
        auto wrapped_call = format.per_call_start + p.space() + tool_choice + p.space() + format.per_call_end;
        if (inputs.parallel_tool_calls) {
-            tool_calls = p.trigger_rule("tool-call", wrapped_call + p.zero_or_more(p.space() + wrapped_call));
+            tool_calls = p.trigger_rule("tool-call", wrapped_call + p.zero_or_more(p.space() + wrapped_call) + p.space());
        } else {
-            tool_calls = p.trigger_rule("tool-call", wrapped_call);
+            tool_calls = p.trigger_rule("tool-call", wrapped_call + p.space());
        }
        if (!format.section_start.empty()) {
            tool_calls = p.trigger_rule("tool-calls",
                                        p.literal(format.section_start) + p.space() + tool_calls + p.space() +
-                                            (format.section_end.empty() ? p.end() : p.literal(format.section_end)));
+                                            (format.section_end.empty() ? p.end() : p.literal(format.section_end) + p.space()));
        }
    } else {
        std::string separator = ", ";  // Default
--- a/common/chat-auto-parser.h
+++ b/common/chat-auto-parser.h
@@ -308,19 +308,23 @@ struct analyze_tools : analyze_base {

  private:
    // Extract tool calling 'haystack' for further analysis and delegate further analysis based on format
-    void analyze_tool_calls(const analyze_reasoning & reasoning);
+    void analyze_tool_calls(const analyze_reasoning & reasoning, bool supports_parallel_tool_calls);

    // Analyze format based on position of function and argument name in needle
    void analyze_tool_call_format(const std::string &       haystack,
                                  const std::string &       fun_name_needle,
                                  const std::string &       arg_name_needle,
-                                  const analyze_reasoning & reasoning);
+                                  const analyze_reasoning & reasoning,
+                                  bool                      supports_parallel_tool_calls);

    // Analyze specifics of JSON native format (entire tool call is a JSON object)
    void analyze_tool_call_format_json_native(const std::string & clean_haystack,
                                              const std::string & fun_name_needle,
                                              const std::string & arg_name_needle);

+    // Check if parallel calls in JSON native format array wrapped or tag wrapped
+    void analyze_json_native_parallel_calls();
+
    // Analyze specifics of non-JSON native format (tags for function name or for function name and arguments)
    void analyze_tool_call_format_non_json(const std::string & clean_haystack,
                                           const std::string & fun_name_needle);
--- a/common/chat-diff-analyzer.cpp
+++ b/common/chat-diff-analyzer.cpp
@@ -296,7 +296,7 @@ void analyze_reasoning::compare_reasoning_presence() {
            return p.literal(reasoning_content) + p.space() + p.optional(p.tag("post", (p.marker() + p.space())) + p.rest());
        });
        auto parser_wrapped = build_tagged_peg_parser([&](common_peg_parser_builder &p) {
-            return p.tag("pre", p.marker() + p.space()) + p.literal(reasoning_content) + p.space() + p.tag("post", (p.marker() + p.space())) + p.rest();
+            return p.tag("pre", p.marker() + p.space()) + p.literal(reasoning_content) + p.tag("post", (p.space() + p.marker() + p.space())) + p.rest();
        });
        // try the more aggressive parse first, if it fails, fall back to the delimiter one
        auto result = parser_wrapped.parse_anywhere_and_extract(comparison->output_B);
@@ -306,11 +306,11 @@ void analyze_reasoning::compare_reasoning_presence() {
        if (result.result.success()) {
            if (!result.tags["pre"].empty() && !result.tags["post"].empty()) {
                mode = reasoning_mode::TAG_BASED;
-                start = trim_leading_whitespace(result.tags["pre"]);
-                end   = trim_trailing_whitespace(result.tags["post"]);
+                start = result.tags["pre"];
+                end   = result.tags["post"];
            } else if (!result.tags["post"].empty()) {
                mode = reasoning_mode::TAG_BASED;
-                end = trim_trailing_whitespace(result.tags["post"]);
+                end = result.tags["post"];
            }
        }
    }
@@ -558,7 +558,7 @@ analyze_tools::analyze_tools(const common_chat_template & tmpl,
    : analyze_base(tmpl) {
    LOG_DBG(ANSI_ORANGE "Phase 3: Tool call analysis\n" ANSI_RESET);

-    analyze_tool_calls(reasoning);
+    analyze_tool_calls(reasoning, caps.supports_parallel_tool_calls);

    if (format.mode != tool_format::NONE && format.mode != tool_format::JSON_NATIVE) {
        if (caps.supports_parallel_tool_calls) {
@@ -577,7 +577,7 @@ analyze_tools::analyze_tools(const common_chat_template & tmpl,
    }
 }

-void analyze_tools::analyze_tool_calls(const analyze_reasoning & reasoning) {
+void analyze_tools::analyze_tool_calls(const analyze_reasoning & reasoning, bool supports_parallel_tool_calls) {
    json assistant_no_tools = json{
        { "role",    "assistant"   },
        { "content", ASSISTANT_MSG }
@@ -611,13 +611,14 @@ void analyze_tools::analyze_tool_calls(const analyze_reasoning & reasoning) {
        return;
    }

-    analyze_tool_call_format(tool_section, FUN_FIRST, ARG_FIRST, reasoning);
+    analyze_tool_call_format(tool_section, FUN_FIRST, ARG_FIRST, reasoning, supports_parallel_tool_calls);
 }

 void analyze_tools::analyze_tool_call_format(const std::string &       haystack,
                                             const std::string &       fun_name_needle,
                                             const std::string &       arg_name_needle,
-                                             const analyze_reasoning & reasoning) {
+                                             const analyze_reasoning & reasoning,
+                                             bool                      supports_parallel_tool_calls) {
    if (fun_name_needle.empty() || arg_name_needle.empty() || haystack.empty()) {
        return;
    }
@@ -660,6 +661,9 @@ void analyze_tools::analyze_tool_call_format(const std::string &       haystack,

    if (format.mode == tool_format::JSON_NATIVE) {
        analyze_tool_call_format_json_native(clean_haystack, fun_name_needle, arg_name_needle);
+        if (supports_parallel_tool_calls) {
+            analyze_json_native_parallel_calls();
+        }
    } else {
        analyze_tool_call_format_non_json(clean_haystack, fun_name_needle);
    }
@@ -668,6 +672,42 @@ void analyze_tools::analyze_tool_call_format(const std::string &       haystack,
    format.per_call_end = trim_whitespace(format.per_call_end);
 }

+void analyze_tools::analyze_json_native_parallel_calls() {
+    json assistant_one_tool = json{
+        { "role",       "assistant" },
+        { "content",    ""          },
+        { "tool_calls", json::array({ first_tool_call }) }
+    };
+
+    json assistant_two_tools = json{
+        { "role",       "assistant" },
+        { "content",    ""          },
+        { "tool_calls", json::array({ first_tool_call, second_tool_call }) }
+    };
+
+    template_params params;
+    params.messages              = json::array({ user_msg, assistant_one_tool });
+    params.tools                 = tools;
+    params.add_generation_prompt = false;
+    params.enable_thinking       = true;
+
+    auto comparison = compare_variants(
+        *tmpl, params, [&](template_params & p) { p.messages = json::array({ user_msg, assistant_two_tools }); });
+
+    if (!comparison) {
+        LOG_DBG(ANSI_ORANGE "%s: Template application failed\n" ANSI_RESET, __func__);
+        return;
+    }
+
+    std::string & second_call = comparison->diff.right;
+    if (!format.section_start.empty() && second_call.find(format.section_start) != std::string::npos) {
+        format.per_call_start = format.section_start;
+        format.per_call_end = format.section_end;
+        format.section_start.clear();
+        format.section_end.clear();
+    }
+}
+
 void analyze_tools::analyze_tool_call_format_json_native(const std::string & clean_haystack,
                                                         const std::string & fun_name_needle,
                                                         const std::string & arg_name_needle) {
--- a/common/chat-peg-parser.cpp
+++ b/common/chat-peg-parser.cpp
@@ -676,7 +676,7 @@ common_peg_parser common_chat_peg_builder::build_json_tools_nested_keys(
        ordered_json   params   = function.contains("parameters") ? function.at("parameters") : ordered_json::object();

        auto nested_name = literal("\"" + nested_name_field + "\"") + space() + literal(":") + space() +
-                          literal("\"") + tool_name(literal(name)) + literal("\"");
+                          atomic(literal("\"") + tool_name(literal(name)) + literal("\""));
        auto nested_args = literal("\"" + nested_args_field + "\"") + space() + literal(":") + space() +
                          tool_args(schema(json(), "tool-" + name + "-schema", params));

@@ -744,7 +744,7 @@ common_peg_parser common_chat_peg_builder::build_json_tools_flat_keys(
        ordered_json   params   = function.contains("parameters") ? function.at("parameters") : ordered_json::object();

        auto tool_name_ = name_key_parser + space() + literal(":") + space() +
-                         literal("\"") + tool_name(literal(name)) + literal("\"");
+                         atomic(literal("\"") + tool_name(literal(name)) + literal("\""));
        auto tool_args_ = args_key_parser + space() + literal(":") + space() +
                         tool_args(schema(json(), "tool-" + name + "-schema", params));

--- a/common/chat.cpp
+++ b/common/chat.cpp
@@ -397,6 +397,25 @@ json common_chat_msgs_to_json_oaicompat(const std::vector<common_chat_msg> & msg
    return render_message_to_json(msgs, c);
 }

+json common_chat_tools_to_json_oaicompat(const std::vector<common_chat_tool> & tools) {
+    if (tools.empty()) {
+        return json();
+    }
+
+    auto result = json::array();
+    for (const auto & tool : tools) {
+        result.push_back({
+            { "type",     "function" },
+            { "function", {
+                { "name", tool.name },
+                { "description", tool.description },
+                { "parameters", json::parse(tool.parameters) },
+            }},
+        });
+    }
+    return result;
+}
+
 std::vector<common_chat_tool> common_chat_tools_parse_oaicompat(const json & tools) {
    std::vector<common_chat_tool> result;

@@ -432,56 +451,6 @@ std::vector<common_chat_tool> common_chat_tools_parse_oaicompat(const json & too
    return result;
 }

-json common_chat_tools_to_json_oaicompat(const std::vector<common_chat_tool> & tools) {
-    if (tools.empty()) {
-        return json();
-    }
-
-    auto result = json::array();
-    for (const auto & tool : tools) {
-        result.push_back({
-            { "type",     "function" },
-            { "function",
-             {
-                  { "name", tool.name },
-                  { "description", tool.description },
-                  { "parameters", json::parse(tool.parameters) },
-              }                      },
-        });
-    }
-    return result;
-}
-
-json common_chat_msg_diff_to_json_oaicompat(const common_chat_msg_diff & diff) {
-    json delta = json::object();
-    if (!diff.reasoning_content_delta.empty()) {
-        delta["reasoning_content"] = diff.reasoning_content_delta;
-    }
-    if (!diff.content_delta.empty()) {
-        delta["content"] = diff.content_delta;
-    }
-    if (diff.tool_call_index != std::string::npos) {
-        json tool_call;
-        tool_call["index"] = diff.tool_call_index;
-        if (!diff.tool_call_delta.id.empty()) {
-            tool_call["id"]   = diff.tool_call_delta.id;
-            tool_call["type"] = "function";
-        }
-        if (!diff.tool_call_delta.name.empty() || !diff.tool_call_delta.arguments.empty()) {
-            json function = json::object();
-            if (!diff.tool_call_delta.name.empty()) {
-                function["name"] = diff.tool_call_delta.name;
-            }
-            if (!diff.tool_call_delta.arguments.empty()) {
-                function["arguments"] = diff.tool_call_delta.arguments;
-            }
-            tool_call["function"] = function;
-        }
-        delta["tool_calls"] = json::array({ tool_call });
-    }
-    return delta;
-}
-
 bool common_chat_verify_template(const std::string & tmpl, bool use_jinja) {
    if (use_jinja) {
        try {
@@ -575,6 +544,26 @@ bool common_chat_templates_was_explicit(const struct common_chat_templates * tmp
    return tmpls->has_explicit_template;
 }

+// LFM2 format detection: template uses <|tool_list_start|>[...]<|tool_list_end|> around the tool list
+// and <|tool_call_start|>[...]<|tool_call_end|> around each tool call
+static bool is_lfm2_template(const std::string & src) {
+    return src.find("<|tool_list_start|>") != std::string::npos &&
+           src.find("<|tool_list_end|>")   != std::string::npos;
+}
+
+common_chat_prompt_preset common_chat_get_asr_prompt(const common_chat_templates * chat_templates) {
+    common_chat_prompt_preset asr_preset;
+    asr_preset.system = "";
+    asr_preset.user   = "Transcribe audio to text";
+
+    if (chat_templates && chat_templates->template_default && is_lfm2_template(chat_templates->template_default->source())) {
+        asr_preset.system = "Perform ASR.";
+        asr_preset.user   = "";
+    }
+
+    return asr_preset;
+}
+
 std::string common_chat_templates_source(const struct common_chat_templates * tmpls, const std::string & variant) {
    if (!variant.empty()) {
        if (variant == "tool_use") {
@@ -865,9 +854,10 @@ static common_chat_params common_chat_params_init_ministral_3(const common_chat_
        adjusted_messages.push_back(adjusted);
    }

-    auto has_tools         = inputs.tools.is_array() && !inputs.tools.empty();
-    auto extract_reasoning = inputs.reasoning_format != COMMON_REASONING_FORMAT_NONE;
-    auto include_grammar   = true;
+    auto has_tools            = inputs.tools.is_array() && !inputs.tools.empty();
+    auto has_response_format  = inputs.json_schema.is_object() && !inputs.json_schema.empty();
+    auto extract_reasoning    = inputs.reasoning_format != COMMON_REASONING_FORMAT_NONE;
+    auto include_grammar      = true;

    data.supports_thinking  = true;
    data.thinking_start_tag = "[THINK]";
@@ -887,7 +877,7 @@ static common_chat_params common_chat_params_init_ministral_3(const common_chat_
            extract_reasoning ? p.optional("[THINK]" + p.reasoning(p.until("[/THINK]")) + "[/THINK]") : p.eps();

        // Response format parser
-        if (inputs.json_schema.is_object() && !inputs.json_schema.empty()) {
+        if (has_response_format) {
            // Ministral wants to emit json surrounded by code fences
            return generation_prompt + (reasoning << "```json" << p.content(p.schema(p.json(), "response-format", inputs.json_schema)) << "```");
        }
@@ -928,6 +918,10 @@ static common_chat_params common_chat_params_init_ministral_3(const common_chat_
                auto         schema   = function.at("parameters");
                builder.resolve_refs(schema);
            });
+            if (has_response_format) {
+                auto schema = inputs.json_schema;
+                builder.resolve_refs(schema);
+            }
            parser.build_grammar(builder, data.grammar_lazy);
        });

@@ -1063,6 +1057,10 @@ static common_chat_params common_chat_params_init_gpt_oss(const common_chat_temp
                auto         schema   = function.at("parameters");
                builder.resolve_refs(schema);
            });
+            if (has_response_format) {
+                auto schema = inputs.json_schema;
+                builder.resolve_refs(schema);
+            }
            parser.build_grammar(builder, data.grammar_lazy);
        });

@@ -1082,8 +1080,18 @@ static common_chat_params common_chat_params_init_gemma4(const common_chat_templ
    common_chat_params data;

    data.prompt            = common_chat_template_direct_apply_impl(tmpl, inputs);
+
+    if (inputs.add_generation_prompt && string_ends_with(data.prompt, "<turn|>\n")) {
+        // This may happen if the model generates content + tool_call, the
+        // template does not add the model's next turn and confuses the model
+        // from emitting its proper reasoning token sequence.
+        data.prompt += "<|turn>model\n";
+    }
+
    data.format            = COMMON_CHAT_FORMAT_PEG_GEMMA4;
-    data.supports_thinking = true;
+    data.supports_thinking  = true;
+    data.thinking_start_tag = "<|channel>thought";
+    data.thinking_end_tag   = "<channel|>";

    data.preserved_tokens = {
        "<|channel>",
@@ -1102,12 +1110,13 @@ static common_chat_params common_chat_params_init_gemma4(const common_chat_templ
        auto start = p.rule("start", p.prefix(inputs.generation_prompt, "<|channel>"));

        if (extract_reasoning) {
-            p.rule("thought", p.literal("<|channel>thought\n") + p.reasoning(p.until("<channel|>")) + p.literal("<channel|>"));
+            p.rule("thought", p.literal("<|channel>thought") + p.space() + p.reasoning(p.until("<channel|>")) + p.literal("<channel|>"));
        } else {
-            p.rule("thought", p.content(p.literal("<|channel>thought\n") + p.until("<channel|>") + p.literal("<channel|>")));
+            p.rule("thought", p.content(p.literal("<|channel>thought") + p.space() + p.until("<channel|>") + p.literal("<channel|>")));
        }

-        auto thought = (p.peek(p.literal("<|channel>")) + p.ref("thought")) | p.negate(p.literal("<|channel>"));
+        auto consume_empty_channels = p.gbnf(p.zero_or_more(p.literal("<|channel>") + p.negate(p.literal("thought"))), "");
+        auto thought = (p.peek(p.literal("<|channel>")) + consume_empty_channels + p.ref("thought")) | p.negate(p.literal("<|channel>"));

        if (has_response_format) {
            auto response_format = p.literal("```json") <<
@@ -1124,7 +1133,7 @@ static common_chat_params common_chat_params_init_gemma4(const common_chat_templ
            p.rule("gemma4-bool", p.json_bool());
            p.rule("gemma4-null", p.json_null());
            p.rule("gemma4-number", p.json_number());
-            p.rule("gemma4-dict-key", p.rule("gemma4-dict-key-name", p.until(":")) + p.literal(":"));
+            p.rule("gemma4-dict-key", p.rule("gemma4-dict-key-name", p.chars("[^:}]", 1, -1)) + p.literal(":"));
            p.rule("gemma4-dict-kv", p.ref("gemma4-dict-key") + p.space() + p.ref("gemma4-value"));
            p.rule("gemma4-dict", [&]() {
                auto ws = p.space();
@@ -1171,12 +1180,16 @@ static common_chat_params common_chat_params_init_gemma4(const common_chat_templ
                /* max = */ inputs.parallel_tool_calls ? -1 : 1
            ));

-            auto content = p.rule("content", p.content(p.until_one_of({"<|channel>", "<|tool_call>"})));
+            auto scan_to_toolcall = p.rule("scan-to-toolcall", p.until("<|tool_call>"));
+            auto content = p.rule("content", p.content(p.until_one_of({"<|channel>", "<channel|>", "<|tool_call>"})));
            auto message = p.rule("message", thought + content);
-            return start + p.zero_or_more(message) + tool_call;
+            return start + p.zero_or_more(message) + scan_to_toolcall + tool_call;
        }

-        auto content = p.rule("content", p.content(p.until("<|channel>")));
+        // Gemma 4 may emit an extra <|channel>thought\n<channel|> at the end of the content. It may
+        // also emit a single trailing <channel|> token. Consume all complete reasoning blocks and
+        // then stop at the first unmatched <channel|> token.
+        auto content = p.rule("content", p.content(p.until_one_of({"<|channel>", "<channel|>"})));
        auto message = p.rule("message", thought + content);
        return start + p.one_or_more(message);
    });
@@ -1191,6 +1204,10 @@ static common_chat_params common_chat_params_init_gemma4(const common_chat_templ
                auto         schema   = function.at("parameters");
                builder.resolve_refs(schema);
            });
+            if (has_response_format) {
+                auto schema = inputs.json_schema;
+                builder.resolve_refs(schema);
+            }
            parser.build_grammar(builder, data.grammar_lazy);
        });

@@ -1641,6 +1658,173 @@ static common_chat_params common_chat_params_init_gigachat_v3(
    return data;
 }

+static common_chat_params common_chat_params_init_deepseek_v3_2(const common_chat_template &    tmpl,
+                                                                 const autoparser::generation_params & inputs) {
+    common_chat_params data;
+
+    data.prompt            = common_chat_template_direct_apply_impl(tmpl, inputs);
+    data.format            = COMMON_CHAT_FORMAT_PEG_NATIVE;
+    data.supports_thinking = true;
+    data.thinking_start_tag = "<think>";
+    data.thinking_end_tag   = "</think>";
+    data.preserved_tokens  = {
+        "｜DSML｜",
+        "<think>",
+        "</think>",
+    };
+
+    auto has_tools           = inputs.tools.is_array() && !inputs.tools.empty();
+    auto has_response_format = !inputs.json_schema.is_null() && inputs.json_schema.is_object();
+    auto extract_reasoning   = inputs.reasoning_format != COMMON_REASONING_FORMAT_NONE;
+    auto include_grammar     = has_response_format || (has_tools && inputs.tool_choice != COMMON_CHAT_TOOL_CHOICE_NONE);
+
+    const std::string DSML         = "｜DSML｜";
+    const std::string THINK_START  = "<think>";
+    const std::string THINK_END    = "</think>";
+    const std::string FC_START     = "<" + DSML + "function_calls>";
+    const std::string FC_END       = "</" + DSML + "function_calls>";
+    const std::string INVOKE_START = "<" + DSML + "invoke";
+    const std::string INVOKE_END   = "</" + DSML + "invoke>";
+    const std::string PARAM_START  = "<" + DSML + "parameter";
+    const std::string PARAM_END    = "</" + DSML + "parameter>";
+
+    auto parser = build_chat_peg_parser([&](common_chat_peg_builder & p) {
+        auto generation_prompt = p.prefix(inputs.generation_prompt, THINK_START);
+        auto end = p.end();
+
+        auto reasoning = p.eps();
+        if (extract_reasoning && inputs.enable_thinking) {
+            reasoning = p.optional(THINK_START + p.reasoning(p.until(THINK_END)) + THINK_END);
+        } else if (extract_reasoning) {
+            // Thinking disabled but reasoning extraction requested: the generation prompt
+            // contains an empty <think></think> pair that must still be consumed.
+            reasoning = p.optional(p.literal(THINK_START) + p.until(THINK_END) + p.literal(THINK_END));
+        }
+
+        if (has_response_format) {
+            auto response_format = p.rule("response-format",
+                p.literal("```json") + p.space() +
+                p.content(p.schema(p.json(), "response-format-schema", inputs.json_schema)) +
+                p.space() + p.literal("```"));
+            return generation_prompt + reasoning + response_format + end;
+        }
+
+        if (!has_tools || inputs.tool_choice == COMMON_CHAT_TOOL_CHOICE_NONE) {
+            return generation_prompt + reasoning + p.content(p.rest()) + end;
+        }
+
+        auto tool_choice = p.choice();
+        foreach_function(inputs.tools, [&](const json & tool) {
+            const auto & function = tool.at("function");
+            std::string  name     = function.at("name");
+            auto params   = function.contains("parameters") ? function.at("parameters") : json::object();
+            const auto & props    = params.contains("properties") ? params.at("properties") : json::object();
+
+            std::set<std::string> required;
+            if (params.contains("required")) {
+                params.at("required").get_to(required);
+            }
+
+            auto schema_info = common_schema_info();
+            schema_info.resolve_refs(params);
+
+            std::vector<common_peg_parser> required_parsers;
+            std::vector<common_peg_parser> optional_parsers;
+            for (const auto & [param_name, param_schema] : props.items()) {
+                bool is_required = required.find(param_name) != required.end();
+                bool is_string   = schema_info.resolves_to_string(param_schema);
+
+                auto arg = p.tool_arg(
+                    p.tool_arg_open(
+                        p.literal(PARAM_START + " name=\"") +
+                        p.tool_arg_name(p.literal(param_name)) +
+                        p.literal("\" string=\"" + std::string(is_string ? "true" : "false") + "\">")) +
+                    (is_string
+                         ? p.tool_arg_string_value(p.until(PARAM_END))
+                         : p.tool_arg_json_value(p.schema(p.json(),
+                                                          "tool-" + name + "-arg-" + param_name + "-schema",
+                                                          param_schema, false))) +
+                    p.tool_arg_close(p.literal(PARAM_END)));
+
+                auto named_arg = p.rule("tool-" + name + "-arg-" + param_name, arg);
+                if (is_required) {
+                    required_parsers.push_back(named_arg);
+                } else {
+                    optional_parsers.push_back(named_arg);
+                }
+            }
+
+            common_peg_parser args_seq = p.eps();
+            for (size_t i = 0; i < required_parsers.size(); i++) {
+                if (i > 0) {
+                    args_seq = args_seq + p.space();
+                }
+                args_seq = args_seq + required_parsers[i];
+            }
+
+            if (!optional_parsers.empty()) {
+                common_peg_parser any_opt = p.choice();
+                for (const auto & opt : optional_parsers) {
+                    any_opt |= opt;
+                }
+                args_seq = args_seq + p.repeat(p.space() + any_opt, 0, -1);
+            }
+
+            common_peg_parser invoke_body = args_seq;
+            auto func_parser = p.tool(
+                p.tool_open(p.literal(INVOKE_START + " name=\"") +
+                            p.tool_name(p.literal(name)) + p.literal("\">\n")) +
+                invoke_body + p.space() +
+                p.tool_close(p.literal(INVOKE_END)));
+
+            tool_choice |= p.rule("tool-" + name, func_parser);
+        });
+
+        auto require_tools = inputs.tool_choice == COMMON_CHAT_TOOL_CHOICE_REQUIRED;
+
+        common_peg_parser tool_calls = p.eps();
+        if (inputs.parallel_tool_calls) {
+            tool_calls = p.trigger_rule("tool-call",
+                p.literal(FC_START) + p.space() + tool_choice +
+                p.zero_or_more(p.space() + tool_choice) + p.space() + p.literal(FC_END));
+        } else {
+            tool_calls = p.trigger_rule("tool-call",
+                p.literal(FC_START) + p.space() + tool_choice + p.space() + p.literal(FC_END));
+        }
+
+        if (!require_tools) {
+            tool_calls = p.optional(tool_calls);
+        }
+
+        auto content_before_tools = p.content(p.until(FC_START));
+        return generation_prompt + reasoning + content_before_tools + tool_calls + end;
+    });
+
+    data.parser = parser.save();
+
+    if (include_grammar) {
+        data.grammar_lazy = !(has_response_format || (has_tools && inputs.tool_choice == COMMON_CHAT_TOOL_CHOICE_REQUIRED));
+        data.grammar      = build_grammar([&](const common_grammar_builder & builder) {
+            foreach_function(inputs.tools, [&](const json & tool) {
+                const auto & function = tool.at("function");
+                auto         schema   = function.contains("parameters") ? function.at("parameters") : json::object();
+                builder.resolve_refs(schema);
+            });
+            if (has_response_format) {
+                auto schema = inputs.json_schema;
+                builder.resolve_refs(schema);
+            }
+            parser.build_grammar(builder, data.grammar_lazy);
+        });
+
+        data.grammar_triggers = {
+            { COMMON_GRAMMAR_TRIGGER_TYPE_WORD, FC_START },
+        };
+    }
+
+    return data;
+}
+
 namespace workaround {

 static void map_developer_role_to_system(json & messages) {
@@ -1889,10 +2073,7 @@ std::optional<common_chat_params> common_chat_try_specialized_template(
        return common_chat_params_init_kimi_k2(tmpl, params);
    }

-    // LFM2 format detection: template uses <|tool_list_start|>[...]<|tool_list_end|> around the tool list
-    // and <|tool_call_start|>[...]<|tool_call_end|> around each tool call
-    if (src.find("<|tool_list_start|>") != std::string::npos &&
-        src.find("<|tool_list_end|>") != std::string::npos) {
+    if (is_lfm2_template(src)) {
        LOG_DBG("Using specialized template: LFM2\n");
        return common_chat_params_init_lfm2(tmpl, params);
    }
@@ -1912,9 +2093,23 @@ std::optional<common_chat_params> common_chat_try_specialized_template(
        return common_chat_params_init_gigachat_v3(tmpl, params);
    }

+    // DeepSeek V3.2 format detection: template defines dsml_token and uses it for tool calls.
+    // The template source contains the token as a variable assignment, not as a literal in markup.
+    if (src.find("dsml_token") != std::string::npos &&
+        src.find("function_calls") != std::string::npos &&
+        src.find("DSML") != std::string::npos) {
+        LOG_DBG("Using specialized template: DeepSeek V3.2\n");
+        return common_chat_params_init_deepseek_v3_2(tmpl, params);
+    }
+
    // Gemma4 format detection
    if (src.find("'<|tool_call>call:'") != std::string::npos) {
-        workaround::convert_tool_responses_gemma4(params.messages);
+        if (src.find("{#- OpenAI Chat Completions:") == std::string::npos) {
+            // apply workarounds if using the older gemma4 templates
+            LOG_WRN("%s: detected an outdated gemma4 chat template, applying compatibility workarounds. "
+                    "Consider updating to the official template.\n", __func__);
+            workaround::convert_tool_responses_gemma4(params.messages);
+        }
        return common_chat_params_init_gemma4(tmpl, params);
    }

@@ -2125,7 +2320,7 @@ common_chat_msg common_chat_peg_parse(const common_peg_arena &          src_pars
        ? input
        : params.generation_prompt + input;

-    LOG_DBG("Parsing PEG input with format %s: %s\n", common_chat_format_name(params.format), effective_input.c_str());
+    //LOG_DBG("Parsing PEG input with format %s: %s\n", common_chat_format_name(params.format), effective_input.c_str());

    common_peg_parse_flags flags = COMMON_PEG_PARSE_FLAG_LENIENT;
    if (params.debug) {
@@ -2187,4 +2382,3 @@ std::map<std::string, bool> common_chat_templates_get_caps(const common_chat_tem
    GGML_ASSERT(chat_templates->template_default != nullptr);
    return chat_templates->template_default->caps.to_map();
 }
-
--- a/common/chat.h
+++ b/common/chat.h
@@ -256,14 +256,13 @@ bool common_chat_templates_support_enable_thinking(const common_chat_templates *
 // Parses a JSON array of messages in OpenAI's chat completion API format.
 std::vector<common_chat_msg> common_chat_msgs_parse_oaicompat(const nlohmann::ordered_json & messages);

+std::vector<common_chat_tool> common_chat_tools_parse_oaicompat(const nlohmann::ordered_json & tools);
+
 // DEPRECATED: only used in tests
 nlohmann::ordered_json common_chat_msgs_to_json_oaicompat(const std::vector<common_chat_msg> & msgs, bool concat_typed_text = false);

-std::vector<common_chat_tool> common_chat_tools_parse_oaicompat(const nlohmann::ordered_json & tools);
 nlohmann::ordered_json common_chat_tools_to_json_oaicompat(const std::vector<common_chat_tool> & tools);

-nlohmann::ordered_json common_chat_msg_diff_to_json_oaicompat(const common_chat_msg_diff & diff);
-
 // get template caps, useful for reporting to server /props endpoint
 std::map<std::string, bool> common_chat_templates_get_caps(const common_chat_templates * chat_templates);

@@ -275,3 +274,11 @@ std::optional<common_chat_params> common_chat_try_specialized_template(
        const common_chat_template &          tmpl,
        const std::string &                   src,
        autoparser::generation_params & params);
+
+// specialized per-task preset
+struct common_chat_prompt_preset {
+    std::string system;
+    std::string user;
+};
+
+common_chat_prompt_preset common_chat_get_asr_prompt(const common_chat_templates * chat_templates);
--- a/common/common.cpp
+++ b/common/common.cpp
@@ -1,7 +1,9 @@
 #include "ggml.h"
 #include "gguf.h"

+#include "build-info.h"
 #include "common.h"
+#include "fit.h"
 #include "log.h"
 #include "llama.h"
 #include "sampling.h"
@@ -68,7 +70,7 @@ common_time_meas::~common_time_meas() {
 // CPU utils
 //

-int32_t cpu_get_num_physical_cores() {
+int32_t common_cpu_get_num_physical_cores() {
 #ifdef __linux__
    // enumerate the set of thread siblings, num entries is num cores
    std::unordered_set<std::string> siblings;
@@ -183,11 +185,11 @@ static int cpu_count_math_cpus(int n_cpu) {
 /**
 * Returns number of CPUs on system that are useful for math.
 */
-int32_t cpu_get_num_math() {
+int32_t common_cpu_get_num_math() {
 #if defined(__x86_64__) && defined(__linux__) && !defined(__ANDROID__)
    int n_cpu = sysconf(_SC_NPROCESSORS_ONLN);
    if (n_cpu < 1) {
-        return cpu_get_num_physical_cores();
+        return common_cpu_get_num_physical_cores();
    }
    if (is_hybrid_cpu()) {
        cpu_set_t affinity;
@@ -200,7 +202,7 @@ int32_t cpu_get_num_math() {
        }
    }
 #endif
-    return cpu_get_num_physical_cores();
+    return common_cpu_get_num_physical_cores();
 }

 // Helper for setting process priority
@@ -261,7 +263,7 @@ bool set_process_priority(enum ggml_sched_priority prio) {
 //


-void postprocess_cpu_params(cpu_params& cpuparams, const cpu_params* role_model) {
+void postprocess_cpu_params(common_cpu_params & cpuparams, const common_cpu_params * role_model) {
    int32_t n_set = 0;

    if (cpuparams.n_threads < 0) {
@@ -269,7 +271,7 @@ void postprocess_cpu_params(cpu_params& cpuparams, const cpu_params* role_model)
        if (role_model != nullptr) {
            cpuparams = *role_model;
        } else {
-            cpuparams.n_threads = cpu_get_num_math();
+            cpuparams.n_threads = common_cpu_get_num_math();
        }
    }

@@ -372,7 +374,7 @@ void common_init() {
    const char * build_type = " (debug)";
 #endif

-    LOG_DBG("build: %d (%s) with %s for %s%s\n", LLAMA_BUILD_NUMBER, LLAMA_COMMIT, LLAMA_COMPILER, LLAMA_BUILD_TARGET, build_type);
+    LOG_DBG("build: %d (%s) with %s for %s%s\n", llama_build_number(), llama_commit(), llama_compiler(), llama_build_target(), build_type);
 }

 std::string common_params_get_system_info(const common_params & params) {
@@ -1146,7 +1148,7 @@ common_init_result::common_init_result(common_params & params) :

    if (params.fit_params) {
        LOG_INF("%s: fitting params to device memory, for bugs during this step try to reproduce them with -fit off, or provide --verbose logs if the bug only occurs with -fit on\n", __func__);
-        llama_params_fit(params.model.path.c_str(), &mparams, &cparams,
+        common_fit_params(params.model.path.c_str(), &mparams, &cparams,
            params.tensor_split,
            params.tensor_buft_overrides.data(),
            params.fit_params_target.data(),
@@ -1381,7 +1383,7 @@ common_init_result_ptr common_init_from_params(common_params & params) {

 common_init_result::~common_init_result() = default;

-std::string get_model_endpoint() {
+std::string common_get_model_endpoint() {
    const char * model_endpoint_env = getenv("MODEL_ENDPOINT");
    // We still respect the use of environment-variable "HF_ENDPOINT" for backward-compatibility.
    const char * hf_endpoint_env = getenv("HF_ENDPOINT");
@@ -1396,6 +1398,42 @@ std::string get_model_endpoint() {
    return model_endpoint;
 }

+common_context_seq_rm_type common_context_can_seq_rm(llama_context * ctx) {
+    auto * mem = llama_get_memory(ctx);
+    if (mem == nullptr) {
+        return COMMON_CONTEXT_SEQ_RM_TYPE_NO;
+    }
+
+    common_context_seq_rm_type res = COMMON_CONTEXT_SEQ_RM_TYPE_PART;
+
+    llama_memory_clear(mem, true);
+
+    // eval 2 tokens to check if the context is compatible
+    std::vector<llama_token> tmp;
+    tmp.push_back(0);
+    tmp.push_back(0);
+
+    int ret = llama_decode(ctx, llama_batch_get_one(tmp.data(), tmp.size()));
+    if (ret != 0) {
+        LOG_ERR("%s: llama_decode() failed: %d\n", __func__, ret);
+        res = COMMON_CONTEXT_SEQ_RM_TYPE_NO;
+        goto done;
+    }
+
+    // try to remove the last tokens
+    if (!llama_memory_seq_rm(mem, 0, 1, -1)) {
+        LOG_WRN("%s: the target context does not support partial sequence removal\n", __func__);
+        res = COMMON_CONTEXT_SEQ_RM_TYPE_FULL;
+        goto done;
+    }
+
+done:
+    llama_memory_clear(mem, true);
+    llama_synchronize(ctx);
+
+    return res;
+}
+
 void common_set_adapter_lora(struct llama_context * ctx, std::vector<common_adapter_lora_info> & lora) {
    std::vector<llama_adapter_lora *> loras;
    std::vector<float> scales;
@@ -1483,7 +1521,7 @@ struct llama_context_params common_context_params_to_llama(const common_params &
    return cparams;
 }

-struct ggml_threadpool_params ggml_threadpool_params_from_cpu_params(const cpu_params & params) {
+struct ggml_threadpool_params ggml_threadpool_params_from_cpu_params(const common_cpu_params & params) {
    struct ggml_threadpool_params tpp;

    ggml_threadpool_params_init(&tpp, params.n_threads); // setup the defaults
--- a/common/common.h
+++ b/common/common.h
@@ -2,15 +2,15 @@

 #pragma once

+#include "llama-cpp.h"
+
 #include "ggml-opt.h"
 #include "ggml.h"
-#include "llama-cpp.h"

 #include <set>
 #include <sstream>
 #include <string>
 #include <string_view>
-#include <variant>
 #include <vector>
 #include <map>

@@ -27,11 +27,6 @@
 #define die(msg)          do { fputs("error: " msg "\n", stderr);                exit(1); } while (0)
 #define die_fmt(fmt, ...) do { fprintf(stderr, "error: " fmt "\n", __VA_ARGS__); exit(1); } while (0)

-#define print_build_info() do {                                                                     \
-    fprintf(stderr, "%s: build = %d (%s)\n",      __func__, LLAMA_BUILD_NUMBER, LLAMA_COMMIT);      \
-    fprintf(stderr, "%s: built with %s for %s\n", __func__, LLAMA_COMPILER, LLAMA_BUILD_TARGET);    \
-} while(0)
-
 struct common_time_meas {
    common_time_meas(int64_t & t_acc, bool disable = false);
    ~common_time_meas();
@@ -53,21 +48,13 @@ struct common_adapter_lora_info {

 using llama_tokens = std::vector<llama_token>;

-// build info
-extern int LLAMA_BUILD_NUMBER;
-extern const char * LLAMA_COMMIT;
-extern const char * LLAMA_COMPILER;
-extern const char * LLAMA_BUILD_TARGET;
-
-const static std::string build_info("b" + std::to_string(LLAMA_BUILD_NUMBER) + "-" + LLAMA_COMMIT);
-
 struct common_control_vector_load_info;

 //
 // CPU utils
 //

-struct cpu_params {
+struct common_cpu_params {
    int      n_threads                   = -1;
    bool     cpumask[GGML_MAX_N_THREADS] = {false}; // CPU affinity mask.
    bool     mask_valid                  = false;   // Default: any CPU
@@ -76,8 +63,8 @@ struct cpu_params {
    uint32_t poll                        = 50;      // Polling (busywait) level (0 - no polling, 100 - mostly polling)
 };

-int32_t cpu_get_num_physical_cores();
-int32_t cpu_get_num_math();
+int32_t common_cpu_get_num_physical_cores();
+int32_t common_cpu_get_num_math();

 //
 // Common params
@@ -287,6 +274,7 @@ struct common_params_sampling {
    std::vector<llama_token> reasoning_budget_start;           // start tag token sequence
    std::vector<llama_token> reasoning_budget_end;             // end tag token sequence
    std::vector<llama_token> reasoning_budget_forced;          // forced sequence (message + end tag)
+    std::string              reasoning_budget_message;         // message injected before end tag when budget exhausted

    bool backend_sampling = false;

@@ -309,34 +297,19 @@ struct common_params_model {

 struct common_ngram_mod;

-struct common_params_speculative {
-    common_speculative_type type = COMMON_SPECULATIVE_TYPE_NONE; // type of speculative decoding
+// draft-model-based speculative decoding parameters
+struct common_params_speculative_draft {
+    int32_t n_max = 16; // maximum number of tokens to draft during speculative decoding
+    int32_t n_min = 0;  // minimum number of draft tokens to use for speculative decoding

-    // general-purpose speculative decoding parameters
+    float p_split = 0.1f;  // speculative decoding split probability
+    float p_min   = 0.75f; // minimum speculative decoding probability (greedy)

-    int32_t n_max   = 16; // maximum number of tokens to draft during speculative decoding
-    int32_t n_min   = 0; // minimum number of draft tokens to use for speculative decoding
-    float   p_split = 0.1f; // speculative decoding split probability
-    float   p_min   = 0.75f; // minimum speculative decoding probability (greedy)
+    common_params_model mparams;

-    // ngram-based speculative decoding
+    llama_model * model = nullptr; // a llama_model that can be shared by multiple speculative contexts

-    uint16_t ngram_size_n     = 12; // ngram size for lookup
-    uint16_t ngram_size_m     = 48; // mgram size for speculative tokens
-    uint16_t ngram_min_hits   =  1; // minimum hits at ngram/mgram lookup for mgram to be proposed
-
-    std::shared_ptr<common_ngram_mod> ngram_mod;
-
-    std::string lookup_cache_static;  // path of static ngram cache file for lookup decoding           // NOLINT
-    std::string lookup_cache_dynamic; // path of dynamic ngram cache file for lookup decoding          // NOLINT
-
-    // draft-model speculative decoding
-
-    struct common_params_model mparams_dft;
-
-    llama_model * model_dft = nullptr; // a llama_model that can be shared by multiple speculative contexts
-
-    llama_context_params cparams_dft; // these are the parameters for the draft llama_context
+    llama_context_params cparams; // these are the parameters for the draft llama_context

    int32_t n_ctx        = 0;  // draft context size
    int32_t n_gpu_layers = -1; // number of layers to store in VRAM for the draft model (-1 - use default)
@@ -344,25 +317,60 @@ struct common_params_speculative {
    ggml_type cache_type_k = GGML_TYPE_F16; // KV cache data type for the K
    ggml_type cache_type_v = GGML_TYPE_F16; // KV cache data type for the V

-    struct cpu_params cpuparams;
-    struct cpu_params cpuparams_batch;
+    common_cpu_params cpuparams;
+    common_cpu_params cpuparams_batch;

    std::vector<ggml_backend_dev_t> devices; // devices to use for offloading

    std::vector<std::pair<std::string, std::string>> replacements; // main to speculative model replacements
    std::vector<llama_model_tensor_buft_override> tensor_buft_overrides;
+};
+
+struct common_params_speculative_ngram_mod {
+    int32_t n_match = 24;
+
+    int32_t n_max = 64;
+    int32_t n_min = 48;
+
+    // shared instance of the ngram container for all speculative decoding contexts
+    std::shared_ptr<common_ngram_mod> obj;
+};
+
+struct common_params_speculative_ngram_map {
+    uint16_t size_n   = 12; // ngram size for lookup
+    uint16_t size_m   = 48; // mgram size for speculative tokens
+    uint16_t min_hits = 1;  // minimum hits at ngram/mgram lookup for mgram to be proposed
+};
+
+struct common_params_speculative_ngram_cache {
+    std::string lookup_cache_static;  // path of static ngram cache file for lookup decoding
+    std::string lookup_cache_dynamic; // path of dynamic ngram cache file for lookup decoding
+};
+
+struct common_params_speculative {
+    // TODO: become a vector in order to support "chains of speculators"
+    common_speculative_type type = COMMON_SPECULATIVE_TYPE_NONE;
+
+    common_params_speculative_draft draft;
+
+    common_params_speculative_ngram_mod ngram_mod;
+    common_params_speculative_ngram_map ngram_simple;
+    common_params_speculative_ngram_map ngram_map_k;
+    common_params_speculative_ngram_map ngram_map_k4v;
+
+    common_params_speculative_ngram_cache ngram_cache;

    bool has_dft() const {
-        return !mparams_dft.path.empty() || !mparams_dft.hf_repo.empty();
+        return !draft.mparams.path.empty() || !draft.mparams.hf_repo.empty();
    }
 };

 struct common_params_vocoder {
    struct common_params_model model;

-    std::string speaker_file = ""; // speaker file path                                      // NOLINT
+    std::string speaker_file; // speaker file path

-    bool use_guide_tokens = false; // enable guide tokens to improve TTS accuracy            // NOLINT
+    bool use_guide_tokens = false; // enable guide tokens to improve TTS accuracy
 };

 struct common_params_diffusion {
@@ -433,19 +441,20 @@ struct common_params {
    // offload params
    std::vector<ggml_backend_dev_t> devices; // devices to use for offloading

-    int32_t n_gpu_layers       = -1;   // number of layers to store in VRAM, -1 is auto, <= -2 is all
-    int32_t main_gpu           = 0;    // the GPU that is used for scratch and small tensors
-    float   tensor_split[128]  = {0};  // how split tensors should be distributed across GPUs
-    bool    fit_params         = true; // whether to fit unset model/context parameters to free device memory
-    int32_t fit_params_min_ctx = 4096; // minimum context size to set when trying to reduce memory use
+    int32_t n_gpu_layers       = -1;    // number of layers to store in VRAM, -1 is auto, <= -2 is all
+    int32_t main_gpu           = 0;     // the GPU that is used for scratch and small tensors
+    float   tensor_split[128]  = {0};   // how split tensors should be distributed across GPUs
+    bool    fit_params         = true;  // whether to fit unset model/context parameters to free device memory
+    bool    fit_params_print   = false; // print the estimated required memory to run the model
+    int32_t fit_params_min_ctx = 4096;  // minimum context size to set when trying to reduce memory use

    // margin per device in bytes for fitting parameters to free memory:
    std::vector<size_t> fit_params_target = std::vector<size_t>(llama_max_devices(), 1024 * 1024*1024);

    enum llama_split_mode split_mode = LLAMA_SPLIT_MODE_LAYER; // how to split the model across GPUs

-    struct cpu_params cpuparams;
-    struct cpu_params cpuparams_batch;
+    common_cpu_params cpuparams;
+    common_cpu_params cpuparams_batch;

    ggml_backend_sched_eval_callback cb_eval = nullptr;
    void * cb_eval_user_data                 = nullptr;
@@ -579,7 +588,7 @@ struct common_params {
    int32_t n_threads_http      = -1;    // number of threads to process HTTP requests (TODO: support threadpool)
    int32_t n_cache_reuse       = 0;     // min chunk size to reuse from the cache via KV shifting
    bool    cache_prompt        = true;  // whether to enable prompt caching
-    bool    clear_idle          = true;  // save and clear idle slots upon starting a new task
+    bool    cache_idle_slots    = true;  // save and clear idle slots upon starting a new task
    int32_t n_ctx_checkpoints   = 32;    // max number of context checkpoints per slot
    int32_t checkpoint_every_nt = 8192;  // make a checkpoint every n tokens during prefill
    int32_t cache_ram_mib       = 8192;  // -1 = no limit, 0 - disable, 1 = 1 MiB, etc.
@@ -593,8 +602,6 @@ struct common_params {
    bool force_pure_content_parser = false;
    common_reasoning_format reasoning_format = COMMON_REASONING_FORMAT_DEEPSEEK;
    int enable_reasoning = -1; // -1 = auto, 0 = disable, 1 = enable
-    int reasoning_budget = -1;
-    std::string reasoning_budget_message; // message injected before end tag when budget exhausted
    bool prefill_assistant = true; // if true, any trailing assistant message will be prefilled into the response
    int sleep_idle_seconds = -1;   // if >0, server will sleep after this many seconds of idle time

@@ -691,7 +698,7 @@ std::string common_params_get_system_info(const common_params & params);

 bool parse_cpu_range(const std::string & range, bool(&boolmask)[GGML_MAX_N_THREADS]);
 bool parse_cpu_mask(const std::string & mask, bool(&boolmask)[GGML_MAX_N_THREADS]);
-void postprocess_cpu_params(cpu_params & cpuparams, const cpu_params * role_model = nullptr);
+void postprocess_cpu_params(common_cpu_params & cpuparams, const common_cpu_params * role_model = nullptr);
 bool set_process_priority(enum ggml_sched_priority prio);

 //
@@ -759,6 +766,11 @@ inline bool string_starts_with(std::string_view str, std::string_view prefix) {
           str.compare(0, prefix.size(), prefix) == 0;
 }

+// remove when moving to c++20
+inline bool string_starts_with(std::string_view str, char prefix) {
+    return !str.empty() && str.front() == prefix;
+}
+
 // remove when moving to c++20
 inline bool string_ends_with(std::string_view str, std::string_view suffix) {
    return str.size() >= suffix.size() &&
@@ -854,12 +866,28 @@ common_init_result_ptr common_init_from_params(common_params & params);

 struct llama_model_params     common_model_params_to_llama  (      common_params & params);
 struct llama_context_params   common_context_params_to_llama(const common_params & params);
-struct ggml_threadpool_params ggml_threadpool_params_from_cpu_params(const cpu_params & params);
+struct ggml_threadpool_params ggml_threadpool_params_from_cpu_params(const common_cpu_params & params);

 // clear LoRA adapters from context, then apply new list of adapters
 void common_set_adapter_lora(struct llama_context * ctx, std::vector<common_adapter_lora_info> & lora);

-std::string                   get_model_endpoint();
+// model endpoint from env
+std::string common_get_model_endpoint();
+
+//
+// Context utils
+//
+
+enum common_context_seq_rm_type {
+    COMMON_CONTEXT_SEQ_RM_TYPE_NO   = 0, // seq_rm not supported (e.g. no memory module)
+    COMMON_CONTEXT_SEQ_RM_TYPE_PART = 1, // can seq_rm partial sequences
+    COMMON_CONTEXT_SEQ_RM_TYPE_FULL = 2, // can seq_rm full sequences only
+};
+
+// check if the llama_context can remove sequences
+// note: clears the memory of the context
+common_context_seq_rm_type common_context_can_seq_rm(llama_context * ctx);
+

 //
 // Batch utils
--- a/common/debug.cpp
+++ b/common/debug.cpp
@@ -1,9 +1,38 @@
 #include "debug.h"

+#include "common.h"
 #include "log.h"

 #include <cmath>
+#include <regex>
 #include <string>
+#include <vector>
+
+struct common_debug_cb_user_data::impl {
+    std::vector<uint8_t>    data;
+    std::vector<std::regex> tensor_filters;
+    bool                    abort_on_nan{false};
+};
+
+common_debug_cb_user_data::common_debug_cb_user_data() : pimpl(std::make_unique<impl>()) {}
+common_debug_cb_user_data::~common_debug_cb_user_data() = default;
+
+common_debug_cb_user_data::common_debug_cb_user_data(common_params & params, const std::vector<std::string> & filter_patterns, bool abort_on_nan)
+    : pimpl(std::make_unique<impl>())
+{
+    for (const auto & pattern : filter_patterns) {
+        try {
+            std::string anchored_pattern = "^" + pattern;
+            pimpl->tensor_filters.emplace_back(anchored_pattern, std::regex::optimize);
+        } catch (const std::regex_error & e) {
+            throw std::runtime_error("Invalid regex pattern '" + pattern + "': " + e.what());
+        }
+    }
+    pimpl->abort_on_nan = abort_on_nan;
+
+    params.cb_eval           = common_debug_cb_eval;
+    params.cb_eval_user_data = this;
+}

 static std::string common_ggml_ne_string(const ggml_tensor * t) {
    std::string str;
@@ -47,8 +76,7 @@ static float common_ggml_get_float_value(const uint8_t * data,

 #define INDENT "    "

-template <bool abort>
-void common_debug_print_tensor(uint8_t * data, ggml_type type, const int64_t * ne, const size_t * nb, int64_t n) {
+static void common_debug_print_tensor(uint8_t * data, ggml_type type, const int64_t * ne, const size_t * nb, int64_t n, bool abort_on_nan) {
    GGML_ASSERT(n > 0);
    float sum = 0;
    for (int64_t i3 = 0; i3 < ne[3]; i3++) {
@@ -94,7 +122,7 @@ void common_debug_print_tensor(uint8_t * data, ggml_type type, const int64_t * n
        LOG(INDENT "sum = %f\n", sum);
    }

-    if constexpr (abort) {
+    if (abort_on_nan) {
        if (std::isnan(sum)) {
            LOG("encountered NaN - aborting\n");
            exit(0);
@@ -112,8 +140,9 @@ void common_debug_print_tensor(uint8_t * data, ggml_type type, const int64_t * n
 * @param user_data user data to pass at each call back
 * @return true to receive data or continue the graph, false otherwise
 */
-template <bool abort_on_nan> bool common_debug_cb_eval(struct ggml_tensor * t, bool ask, void * user_data) {
-    auto * cb_data = (base_callback_data *) user_data;
+bool common_debug_cb_eval(struct ggml_tensor * t, bool ask, void * user_data) {
+    auto * cb_data = (common_debug_cb_user_data *) user_data;
+    auto * pimpl = cb_data->pimpl.get();

    const struct ggml_tensor * src0 = t->src[0];
    const struct ggml_tensor * src1 = t->src[1];
@@ -122,10 +151,10 @@ template <bool abort_on_nan> bool common_debug_cb_eval(struct ggml_tensor * t, b
        return true;  // Always retrieve data
    }

-    bool matches_filter = cb_data->tensor_filters.empty();
+    bool matches_filter = pimpl->tensor_filters.empty();

    if (!matches_filter) {
-        for (const auto & filter : cb_data->tensor_filters) {
+        for (const auto & filter : pimpl->tensor_filters) {
            if (std::regex_search(t->name, filter)) {
                matches_filter = true;
                break;
@@ -148,20 +177,14 @@ template <bool abort_on_nan> bool common_debug_cb_eval(struct ggml_tensor * t, b

    if (!is_host) {
        auto n_bytes = ggml_nbytes(t);
-        cb_data->data.resize(n_bytes);
-        ggml_backend_tensor_get(t, cb_data->data.data(), 0, n_bytes);
+        pimpl->data.resize(n_bytes);
+        ggml_backend_tensor_get(t, pimpl->data.data(), 0, n_bytes);
    }

    if (!ggml_is_quantized(t->type) && matches_filter) {
-        uint8_t * data = is_host ? (uint8_t *) t->data : cb_data->data.data();
-        common_debug_print_tensor<abort_on_nan>(data, t->type, t->ne, t->nb, 3);
+        uint8_t * data = is_host ? (uint8_t *) t->data : pimpl->data.data();
+        common_debug_print_tensor(data, t->type, t->ne, t->nb, 3, pimpl->abort_on_nan);
    }

    return true;
 }
-
-// Explicit template instantiations
-template bool common_debug_cb_eval<false>(ggml_tensor *, bool, void *);
-template bool common_debug_cb_eval<true>(ggml_tensor *, bool, void *);
-template void common_debug_print_tensor<false>(uint8_t *, ggml_type, const int64_t *, const size_t *, int64_t);
-template void common_debug_print_tensor<true>(uint8_t *, ggml_type, const int64_t *, const size_t *, int64_t);
--- a/common/debug.h
+++ b/common/debug.h
@@ -1,43 +1,31 @@
 #pragma once
-#include "common.h"
+
+#include <memory>
 #include <string>
 #include <vector>
-#include <regex>

 // common debug functions and structs

-// Print a tensor's detailed data
-// data - the tensor's data in byte format
-// type - the tensor's quantization type
-// ne   - the tensor dimensions array
-// nb   - the tensor strides array
-// n    - the number of rows/columns to fully print
-template <bool abort_on_nan> void common_debug_print_tensor(uint8_t * data, ggml_type type, const int64_t * ne, const size_t * nb, int64_t n);
+struct common_params;

 // Intended to use as callback for ggml_backend_sched_eval_callback
 // prints tensors that are processed in the computation graph
-// by default prints all tensors, but can be configured by creating a `base_callback_data` instance with
-// non-empty filter_patterns. See examples/debug.ccp for possible usage patterns
-// The template parameter determines whether an error should be thrown whenever a NaN is encountered
+// by default prints all tensors, but can be configured by creating a `common_debug_cb_user_data` instance with
+// non-empty filter_patterns. See examples/debug.cpp for possible usage patterns
+// `common_debug_cb_user_data` contains `abort_on_nan` flag that determines whether an error should be thrown whenever a NaN is encountered
 // in a tensor (useful for stopping debug sessions on first erroneous tensor)
 // The callback data will be passed as the third parameter (user_data)
-template <bool abort_on_nan> bool common_debug_cb_eval(struct ggml_tensor * t, bool ask, void * user_data);
-struct base_callback_data {
-    std::vector<uint8_t>    data;
-    std::vector<std::regex> tensor_filters;
+bool common_debug_cb_eval(struct ggml_tensor * t, bool ask, void * user_data);

-    base_callback_data() = default;
+struct common_debug_cb_user_data {
+    struct impl;
+    std::unique_ptr<impl> pimpl;

-    base_callback_data(common_params & params, const std::vector<std::string> & filter_patterns) {
-        for (const auto & pattern : filter_patterns) {
-            try {
-                std::string anchored_pattern = "^" + pattern;
-                tensor_filters.emplace_back(anchored_pattern, std::regex::optimize);
-            } catch (const std::regex_error & e) {
-                throw std::runtime_error("Invalid regex pattern '" + pattern + "': " + e.what());
-            }
-        }
-        params.cb_eval           = common_debug_cb_eval<false>;
-        params.cb_eval_user_data = this;
-    }
+    common_debug_cb_user_data();
+    ~common_debug_cb_user_data();
+
+    common_debug_cb_user_data(const common_debug_cb_user_data &) = delete;
+    common_debug_cb_user_data & operator=(const common_debug_cb_user_data &) = delete;
+
+    common_debug_cb_user_data(common_params & params, const std::vector<std::string> & filter_patterns, bool abort_on_nan = false);
 };
--- a/common/download.cpp
+++ b/common/download.cpp
@@ -1,5 +1,6 @@
 #include "arg.h"

+#include "build-info.h"
 #include "common.h"
 #include "log.h"
 #include "download.h"
@@ -114,7 +115,7 @@ std::pair<std::string, std::string> common_download_split_repo_tag(const std::st
    return {hf_repo, tag};
 }

-class ProgressBar {
+class ProgressBar : public common_download_callback {
    static inline std::mutex mutex;
    static inline std::map<const ProgressBar *, int> lines;
    static inline int max_line = 0;
@@ -138,7 +139,11 @@ class ProgressBar {
    }

 public:
-    ProgressBar(const std::string & url = "") : filename(url) {
+    ProgressBar() = default;
+
+    void on_start(const common_download_progress & p) override {
+        filename = p.url;
+
        if (auto pos = filename.rfind('/'); pos != std::string::npos) {
            filename = filename.substr(pos + 1);
        }
@@ -156,13 +161,13 @@ public:
        }
    }

-    ~ProgressBar() {
+    void on_done(const common_download_progress &, bool) override {
        std::lock_guard<std::mutex> lock(mutex);
        cleanup(this);
    }

-    void update(size_t current, size_t total) {
-        if (!total || !is_output_a_tty()) {
+    void on_update(const common_download_progress & p) override {
+        if (!p.total || !is_output_a_tty()) {
            return;
        }

@@ -174,17 +179,17 @@ public:
        }
        int lines_up = max_line - lines[this];

-        size_t bar = 55 - len;
-        size_t pct = (100 * current) / total;
-        size_t pos = (bar * current) / total;
+        size_t bar = (55 - len) * 2;
+        size_t pct = (100 * p.downloaded) / p.total;
+        size_t pos = (bar * p.downloaded) / p.total;

        if (lines_up > 0) {
            std::cout << "\033[" << lines_up << "A";
        }
        std::cout << '\r' << "Downloading " << filename << " ";

-        for (size_t i = 0; i < bar; ++i) {
-            std::cout << (i < pos ? "—" : " ");
+        for (size_t i = 0; i < bar; i += 2) {
+            std::cout << (i + 1 < pos ? "─" : (i < pos ? "╴" : " "));
        }
        std::cout << std::setw(4) << pct << "%\033[K";

@@ -193,7 +198,7 @@ public:
        }
        std::cout << '\r' << std::flush;

-        if (current == total) {
+        if (p.downloaded == p.total) {
            cleanup(this);
        }
    }
@@ -206,8 +211,8 @@ static bool common_pull_file(httplib::Client & cli,
                             const std::string & resolve_path,
                             const std::string & path_tmp,
                             bool supports_ranges,
-                             size_t existing_size,
-                             size_t & total_size) {
+                             common_download_progress & p,
+                             common_download_callback * callback) {
    std::ofstream ofs(path_tmp, std::ios::binary | std::ios::app);
    if (!ofs.is_open()) {
        LOG_ERR("%s: error opening local file for writing: %s\n", __func__, path_tmp.c_str());
@@ -215,29 +220,27 @@ static bool common_pull_file(httplib::Client & cli,
    }

    httplib::Headers headers;
-    if (supports_ranges && existing_size > 0) {
-        headers.emplace("Range", "bytes=" + std::to_string(existing_size) + "-");
+    if (supports_ranges && p.downloaded > 0) {
+        headers.emplace("Range", "bytes=" + std::to_string(p.downloaded) + "-");
    }

    const char * func = __func__; // avoid __func__ inside a lambda
-    size_t downloaded = existing_size;
    size_t progress_step = 0;
-    ProgressBar bar(resolve_path);

    auto res = cli.Get(resolve_path, headers,
        [&](const httplib::Response &response) {
-            if (existing_size > 0 && response.status != 206) {
+            if (p.downloaded > 0 && response.status != 206) {
                LOG_WRN("%s: server did not respond with 206 Partial Content for a resume request. Status: %d\n", func, response.status);
                return false;
            }
-            if (existing_size == 0 && response.status != 200) {
+            if (p.downloaded == 0 && response.status != 200) {
                LOG_WRN("%s: download received non-successful status code: %d\n", func, response.status);
                return false;
            }
-            if (total_size == 0 && response.has_header("Content-Length")) {
+            if (p.total == 0 && response.has_header("Content-Length")) {
                try {
                    size_t content_length = std::stoull(response.get_header_value("Content-Length"));
-                    total_size = existing_size + content_length;
+                    p.total = p.downloaded + content_length;
                } catch (const std::exception &e) {
                    LOG_WRN("%s: invalid Content-Length header: %s\n", func, e.what());
                }
@@ -250,11 +253,16 @@ static bool common_pull_file(httplib::Client & cli,
                LOG_ERR("%s: error writing to file: %s\n", func, path_tmp.c_str());
                return false;
            }
-            downloaded += len;
+            p.downloaded += len;
            progress_step += len;

-            if (progress_step >= total_size / 1000 || downloaded == total_size) {
-                bar.update(downloaded, total_size);
+            if (progress_step >= p.total / 1000 || p.downloaded == p.total) {
+                if (callback) {
+                    callback->on_update(p);
+                    if (callback->is_cancelled()) {
+                        return false;
+                    }
+                }
                progress_step = 0;
            }
            return true;
@@ -275,28 +283,13 @@ static bool common_pull_file(httplib::Client & cli,

 // download one single file from remote URL to local path
 // returns status code or -1 on error
-static int common_download_file_single_online(const std::string        & url,
-                                              const std::string        & path,
-                                              const std::string        & bearer_token,
-                                              const common_header_list & custom_headers,
-                                              bool                       skip_etag = false) {
+static int common_download_file_single_online(const std::string & url,
+                                              const std::string & path,
+                                              const common_download_opts & opts,
+                                              bool skip_etag) {
    static const int max_attempts        = 3;
    static const int retry_delay_seconds = 2;

-    auto [cli, parts] = common_http_client(url);
-
-    httplib::Headers headers;
-    for (const auto & h : custom_headers) {
-        headers.emplace(h.first, h.second);
-    }
-    if (headers.find("User-Agent") == headers.end()) {
-        headers.emplace("User-Agent", "llama-cpp/" + build_info);
-    }
-    if (!bearer_token.empty()) {
-        headers.emplace("Authorization", "Bearer " + bearer_token);
-    }
-    cli.set_default_headers(headers);
-
    const bool file_exists = std::filesystem::exists(path);

    if (file_exists && skip_etag) {
@@ -304,6 +297,20 @@ static int common_download_file_single_online(const std::string        & url,
        return 304; // 304 Not Modified - fake cached response
    }

+    auto [cli, parts] = common_http_client(url);
+
+    httplib::Headers headers;
+    for (const auto & h : opts.headers) {
+        headers.emplace(h.first, h.second);
+    }
+    if (headers.find("User-Agent") == headers.end()) {
+        headers.emplace("User-Agent", "llama-cpp/" + std::string(llama_build_info()));
+    }
+    if (!opts.bearer_token.empty()) {
+        headers.emplace("Authorization", "Bearer " + opts.bearer_token);
+    }
+    cli.set_default_headers(headers);
+
    std::string last_etag;
    if (file_exists) {
        last_etag = read_etag(path);
@@ -326,10 +333,11 @@ static int common_download_file_single_online(const std::string        & url,
        etag = head->get_header_value("ETag");
    }

-    size_t total_size = 0;
+    common_download_progress p;
+    p.url = url;
    if (head->has_header("Content-Length")) {
        try {
-            total_size = std::stoull(head->get_header_value("Content-Length"));
+            p.total = std::stoull(head->get_header_value("Content-Length"));
        } catch (const std::exception& e) {
            LOG_WRN("%s: invalid Content-Length in HEAD response: %s\n", __func__, e.what());
        }
@@ -357,14 +365,21 @@ static int common_download_file_single_online(const std::string        & url,

    { // silent
        std::error_code ec;
-        std::filesystem::path p(path);
-        std::filesystem::create_directories(p.parent_path(), ec);
+        std::filesystem::create_directories(std::filesystem::path(path).parent_path(), ec);
    }

+    bool success = false;
    const std::string path_temporary = path + ".downloadInProgress";
    int delay = retry_delay_seconds;

+    if (opts.callback) {
+        opts.callback->on_start(p);
+    }
+
    for (int i = 0; i < max_attempts; ++i) {
+        if (opts.callback && opts.callback->is_cancelled()) {
+            break;
+        }
        if (i) {
            LOG_WRN("%s: retrying after %d seconds...\n", __func__, delay);
            std::this_thread::sleep_for(std::chrono::seconds(delay));
@@ -378,28 +393,44 @@ static int common_download_file_single_online(const std::string        & url,
                existing_size = std::filesystem::file_size(path_temporary);
            } else if (remove(path_temporary.c_str()) != 0) {
                LOG_ERR("%s: unable to delete file: %s\n", __func__, path_temporary.c_str());
-                return -1;
+                break;
            }
        }

+        p.downloaded = existing_size;
+
        LOG_DBG("%s: downloading from %s to %s (etag:%s)...\n",
                __func__, common_http_show_masked_url(parts).c_str(),
                path_temporary.c_str(), etag.c_str());

-        if (common_pull_file(cli, parts.path, path_temporary, supports_ranges, existing_size, total_size)) {
+        if (common_pull_file(cli, parts.path, path_temporary, supports_ranges, p, opts.callback)) {
            if (std::rename(path_temporary.c_str(), path.c_str()) != 0) {
                LOG_ERR("%s: unable to rename file: %s to %s\n", __func__, path_temporary.c_str(), path.c_str());
-                return -1;
+                break;
            }
            if (!etag.empty() && !skip_etag) {
                write_etag(path, etag);
            }
-            return head->status;
+            success = true;
+            break;
        }
    }

-    LOG_ERR("%s: download failed after %d attempts\n", __func__, max_attempts);
-    return -1; // max attempts reached
+    if (opts.callback) {
+        opts.callback->on_done(p, success);
+    }
+    if (opts.callback && opts.callback->is_cancelled() &&
+        std::filesystem::exists(path_temporary)) {
+        if (remove(path_temporary.c_str()) != 0) {
+            LOG_ERR("%s: unable to delete temporary file: %s\n", __func__, path_temporary.c_str());
+        }
+    }
+    if (!success) {
+        LOG_ERR("%s: download failed after %d attempts\n", __func__, max_attempts);
+        return -1; // max attempts reached
+    }
+
+    return head->status;
 }

 std::pair<long, std::vector<char>> common_remote_get_content(const std::string          & url,
@@ -411,7 +442,7 @@ std::pair<long, std::vector<char>> common_remote_get_content(const std::string
        headers.emplace(h.first, h.second);
    }
    if (headers.find("User-Agent") == headers.end()) {
-        headers.emplace("User-Agent", "llama-cpp/" + build_info);
+        headers.emplace("User-Agent", "llama-cpp/" + std::string(llama_build_info()));
    }

    if (params.timeout > 0) {
@@ -438,12 +469,15 @@ std::pair<long, std::vector<char>> common_remote_get_content(const std::string

 int common_download_file_single(const std::string & url,
                                const std::string & path,
-                                const std::string & bearer_token,
-                                bool offline,
-                                const common_header_list & headers,
+                                const common_download_opts & opts,
                                bool skip_etag) {
-    if (!offline) {
-        return common_download_file_single_online(url, path, bearer_token, headers, skip_etag);
+    if (!opts.offline) {
+        ProgressBar tty_cb;
+        common_download_opts online_opts = opts;
+        if (!online_opts.callback) {
+            online_opts.callback = &tty_cb;
+        }
+        return common_download_file_single_online(url, path, online_opts, skip_etag);
    }

    if (!std::filesystem::exists(path)) {
@@ -452,6 +486,16 @@ int common_download_file_single(const std::string & url,
    }

    LOG_DBG("%s: using cached file (offline mode): %s\n", __func__, path.c_str());
+
+    // notify the callback that the file was cached
+    if (opts.callback) {
+        common_download_progress p;
+        p.url = url;
+        p.cached = true;
+        opts.callback->on_start(p);
+        opts.callback->on_done(p, true);
+    }
+
    return 304; // Not Modified - fake cached response
 }

@@ -583,7 +627,7 @@ static hf_cache::hf_file find_best_model(const hf_cache::hf_files & files,
    if (!tag.empty()) {
        tags.push_back(tag);
    } else {
-        tags = {"Q4_K_M", "Q4_0"};
+        tags = {"Q4_K_M", "Q8_0"};
    }

    for (const auto & t : tags) {
@@ -631,16 +675,16 @@ struct hf_plan {
    hf_cache::hf_file mmproj;
 };

-static hf_plan get_hf_plan(const common_params_model        & model,
-                           const std::string                & token,
-                           const common_download_model_opts & opts) {
+static hf_plan get_hf_plan(const common_params_model  & model,
+                           const common_download_opts & opts,
+                           bool download_mmproj) {
    hf_plan plan;
    hf_cache::hf_files all;

    auto [repo, tag] = common_download_split_repo_tag(model.hf_repo);

    if (!opts.offline) {
-        all = hf_cache::get_repo_files(repo, token);
+        all = hf_cache::get_repo_files(repo, opts.bearer_token);
    }
    if (all.empty()) {
        all = hf_cache::get_cached_files(repo);
@@ -675,7 +719,7 @@ static hf_plan get_hf_plan(const common_params_model        & model,
    plan.primary = primary;
    plan.model_files = get_split_files(all, primary);

-    if (opts.download_mmproj) {
+    if (download_mmproj) {
        plan.mmproj = find_best_mmproj(all, primary.path);
    }

@@ -710,10 +754,9 @@ static std::vector<download_task> get_url_tasks(const common_params_model & mode
    return tasks;
 }

-common_download_model_result common_download_model(const common_params_model        & model,
-                                                   const std::string                & bearer_token,
-                                                   const common_download_model_opts & opts,
-                                                   const common_header_list         & headers) {
+common_download_model_result common_download_model(const common_params_model  & model,
+                                                   const common_download_opts & opts,
+                                                   bool download_mmproj) {
    common_download_model_result result;
    std::vector<download_task> tasks;
    hf_plan hf;
@@ -721,7 +764,7 @@ common_download_model_result common_download_model(const common_params_model
    bool is_hf = !model.hf_repo.empty();

    if (is_hf) {
-        hf = get_hf_plan(model, bearer_token, opts);
+        hf = get_hf_plan(model, opts, download_mmproj);
        for (const auto & f : hf.model_files) {
            tasks.push_back({f.url, f.local_path});
        }
@@ -742,8 +785,8 @@ common_download_model_result common_download_model(const common_params_model
    std::vector<std::future<bool>> futures;
    for (const auto & task : tasks) {
        futures.push_back(std::async(std::launch::async,
-            [&task, &bearer_token, offline = opts.offline, &headers, is_hf]() {
-                int status = common_download_file_single(task.url, task.path, bearer_token, offline, headers, is_hf);
+            [&task, &opts, is_hf]() {
+                int status = common_download_file_single(task.url, task.path, opts, is_hf);
                return is_http_status_ok(status);
            }
        ));
@@ -879,7 +922,9 @@ std::string common_docker_resolve_model(const std::string & docker) {
        std::string local_path = fs_get_cache_file(model_filename);

        const std::string blob_url = url_prefix + "/blobs/" + gguf_digest;
-        const int http_status = common_download_file_single(blob_url, local_path, token, false, {});
+        common_download_opts opts;
+        opts.bearer_token = token;
+        const int http_status = common_download_file_single(blob_url, local_path, opts);
        if (!is_http_status_ok(http_status)) {
            throw std::runtime_error("Failed to download Docker Model");
        }
--- a/common/download.h
+++ b/common/download.h
@@ -8,6 +8,22 @@ struct common_params_model;
 using common_header      = std::pair<std::string, std::string>;
 using common_header_list = std::vector<common_header>;

+struct common_download_progress {
+    std::string url;
+    size_t downloaded = 0;
+    size_t total      = 0;
+    bool cached       = false;
+};
+
+class common_download_callback {
+public:
+    virtual ~common_download_callback() = default;
+    virtual void on_start(const common_download_progress & p) = 0;
+    virtual void on_update(const common_download_progress & p) = 0;
+    virtual void on_done(const common_download_progress & p, bool ok) = 0;
+    virtual bool is_cancelled() const { return false; }
+};
+
 struct common_remote_params {
    common_header_list headers;
    long timeout  = 0;           // in seconds, 0 means no timeout
@@ -31,10 +47,12 @@ struct common_cached_model_info {
    }
 };

-// Options for common_download_model
-struct common_download_model_opts {
-    bool download_mmproj = false;
-    bool offline         = false;
+// Options for common_download_model and common_download_file_single
+struct common_download_opts {
+    std::string bearer_token;
+    common_header_list headers;
+    bool offline = false;
+    common_download_callback * callback = nullptr;
 };

 // Result of common_download_model
@@ -69,9 +87,8 @@ struct common_download_model_result {
 // returns result with model_path and mmproj_path (empty on failure)
 common_download_model_result common_download_model(
    const common_params_model & model,
-    const std::string & bearer_token,
-    const common_download_model_opts & opts = {},
-    const common_header_list & headers = {}
+    const common_download_opts & opts = {},
+    bool download_mmproj = false
 );

 // returns list of cached models
@@ -82,9 +99,7 @@ std::vector<common_cached_model_info> common_list_cached_models();
 // skip_etag: if true, don't read/write .etag files (for HF cache where filename is the hash)
 int common_download_file_single(const std::string & url,
                                const std::string & path,
-                                const std::string & bearer_token,
-                                bool offline,
-                                const common_header_list & headers = {},
+                                const common_download_opts & opts = {},
                                bool skip_etag = false);

 // resolve and download model from Docker registry
--- a/common/fit.cpp
+++ b/common/fit.cpp
@@ -0,0 +1,951 @@
+#include "fit.h"
+
+#include "log.h"
+
+#include "../src/llama-ext.h"
+
+#include <array>
+#include <cassert>
+#include <stdexcept>
+#include <cinttypes>
+#include <set>
+#include <string>
+#include <vector>
+
+// this enum is only used in llama_params_fit_impl but needs to be defined outside of it to fix a Windows compilation issue
+// enum to identify part of a layer for distributing its tensors:
+enum common_layer_fraction_t {
+    LAYER_FRACTION_NONE = 0, // nothing
+    LAYER_FRACTION_ATTN = 1, // attention
+    LAYER_FRACTION_UP   = 2, // attention + up
+    LAYER_FRACTION_GATE = 3, // attention + up + gate
+    LAYER_FRACTION_MOE  = 4, // everything but sparse MoE weights
+};
+
+class common_params_fit_exception : public std::runtime_error {
+    using std::runtime_error::runtime_error;
+};
+
+static std::vector<llama_device_memory_data> common_get_device_memory_data(
+        const char * path_model,
+        const llama_model_params * mparams,
+        const llama_context_params * cparams,
+        std::vector<ggml_backend_dev_t> & devs,
+        uint32_t & hp_ngl,
+        uint32_t & hp_n_ctx_train,
+        uint32_t & hp_n_expert,
+        ggml_log_level log_level) {
+    struct user_data_t {
+        struct {
+            ggml_log_callback callback;
+            void * user_data;
+        } original_logger;
+        ggml_log_level min_level; // prints below this log level go to debug log
+    };
+    user_data_t ud;
+    llama_log_get(&ud.original_logger.callback, &ud.original_logger.user_data);
+    ud.min_level = log_level;
+
+    llama_log_set([](ggml_log_level level, const char * text, void * user_data) {
+        const user_data_t * ud = (const user_data_t *) user_data;
+        const ggml_log_level level_eff = level >= ud->min_level ? level : GGML_LOG_LEVEL_DEBUG;
+        ud->original_logger.callback(level_eff, text, ud->original_logger.user_data);
+    }, &ud);
+
+    llama_model_params mparams_copy = *mparams;
+    mparams_copy.no_alloc  = true;
+    mparams_copy.use_mmap  = false;
+    mparams_copy.use_mlock = false;
+
+    llama_model * model = llama_model_load_from_file(path_model, mparams_copy);
+    if (model == nullptr) {
+        llama_log_set(ud.original_logger.callback, ud.original_logger.user_data);
+        throw std::runtime_error("failed to load model");
+    }
+
+    llama_context * ctx = llama_init_from_model(model, *cparams);
+    if (ctx == nullptr) {
+        llama_model_free(model);
+        llama_log_set(ud.original_logger.callback, ud.original_logger.user_data);
+        throw std::runtime_error("failed to create llama_context from model");
+    }
+
+    const size_t nd = llama_model_n_devices(model);
+    std::vector<llama_device_memory_data> ret(nd + 1);
+
+    llama_memory_breakdown memory_breakdown = llama_get_memory_breakdown(ctx);
+
+    for (const auto & [buft, mb] : memory_breakdown) {
+        if (ggml_backend_buft_is_host(buft)) {
+            ret.back().mb.model   += mb.model;
+            ret.back().mb.context += mb.context;
+            ret.back().mb.compute += mb.compute;
+            continue;
+        }
+
+        ggml_backend_dev_t dev = ggml_backend_buft_get_device(buft);
+        if (!dev) {
+            continue;
+        }
+        for (size_t i = 0; i < nd; i++) {
+            if (dev == llama_model_get_device(model, i)) {
+                ret[i].mb.model   += mb.model;
+                ret[i].mb.context += mb.context;
+                ret[i].mb.compute += mb.compute;
+                break;
+            }
+        }
+    }
+
+    {
+        ggml_backend_dev_t cpu_dev = ggml_backend_dev_by_type(GGML_BACKEND_DEVICE_TYPE_CPU);
+        if (cpu_dev == nullptr) {
+            throw std::runtime_error("no CPU backend found");
+        }
+        size_t free;
+        size_t total;
+        ggml_backend_dev_memory(cpu_dev, &free, &total);
+        ret.back().free  = free;
+        ret.back().total = total;
+    }
+    for (size_t i = 0; i < nd; i++) {
+        size_t free;
+        size_t total;
+        ggml_backend_dev_memory(llama_model_get_device(model, i), &free, &total);
+
+        // devices can return 0 bytes for free and total memory if they do not
+        // have any to report. in this case, we will use the host memory as a fallback
+        // fixes: https://github.com/ggml-org/llama.cpp/issues/18577
+        if (free == 0 && total == 0) {
+            free  = ret.back().free;
+            total = ret.back().total;
+        }
+        ret[i].free  = free;
+        ret[i].total = total;
+    }
+
+    devs.clear();
+    for (int i = 0; i < llama_model_n_devices(model); i++) {
+        devs.push_back(llama_model_get_device(model, i));
+    }
+
+    hp_ngl         = llama_model_n_layer(model);
+    hp_n_ctx_train = llama_model_n_ctx_train(model);
+    hp_n_expert    = llama_model_n_expert(model);
+
+    common_memory_breakdown_print(ctx);
+
+    llama_free(ctx);
+    llama_model_free(model);
+    llama_log_set(ud.original_logger.callback, ud.original_logger.user_data);
+
+    return ret;
+}
+
+static void common_params_fit_impl(
+        const char * path_model, struct llama_model_params * mparams, struct llama_context_params * cparams,
+        float * tensor_split, struct llama_model_tensor_buft_override * tensor_buft_overrides,
+        size_t * margins_s, uint32_t n_ctx_min, enum ggml_log_level log_level) {
+    if (mparams->split_mode == LLAMA_SPLIT_MODE_TENSOR) {
+        throw common_params_fit_exception("llama_params_fit is not implemented for SPLIT_MODE_TENSOR, abort");
+    }
+    constexpr int64_t MiB = 1024*1024;
+    typedef std::vector<llama_device_memory_data> dmds_t;
+    const llama_model_params default_mparams = llama_model_default_params();
+
+    std::vector<ggml_backend_dev_t> devs;
+    uint32_t hp_ngl = 0; // hparams.n_gpu_layers
+    uint32_t hp_nct = 0; // hparams.n_ctx_train
+    uint32_t hp_nex = 0; // hparams.n_expert
+
+    // step 1: get data for default parameters and check whether any changes are necessary in the first place
+
+    LOG_INF("%s: getting device memory data for initial parameters:\n", __func__);
+    const dmds_t dmds_full = common_get_device_memory_data(path_model, mparams, cparams, devs, hp_ngl, hp_nct, hp_nex, log_level);
+    const size_t nd = devs.size(); // number of devices
+
+    std::vector<int64_t> margins; // this function uses int64_t rather than size_t for memory sizes to more conveniently handle deficits
+    margins.reserve(nd);
+    if (nd == 0) {
+        margins.push_back(margins_s[0]);
+    } else {
+        for (size_t id = 0; id < nd; id++) {
+            margins.push_back(margins_s[id]);
+        }
+    }
+
+    std::vector<std::string> dev_names;
+    {
+        dev_names.reserve(nd);
+        size_t max_length = 0;
+        for (const auto & dev : devs) {
+            std::string name = ggml_backend_dev_name(dev);
+            name += " (";
+            name += ggml_backend_dev_description(dev);
+            name += ")";
+            dev_names.push_back(name);
+            max_length = std::max(max_length, name.length());
+        }
+        for (std::string & dn : dev_names) {
+            dn.insert(dn.end(), max_length - dn.length(), ' ');
+        }
+    }
+
+    int64_t sum_free            = 0;
+    int64_t sum_projected_free  = 0;
+    int64_t sum_projected_used  = 0;
+    int64_t sum_projected_model = 0;
+    std::vector<int64_t> projected_free_per_device;
+    projected_free_per_device.reserve(nd);
+
+    if (nd == 0) {
+        sum_projected_used = dmds_full.back().mb.total();
+        sum_free           = dmds_full.back().total;
+        sum_projected_free = sum_free - sum_projected_used;
+        LOG_INF("%s: projected to use %" PRId64 " MiB of host memory vs. %" PRId64 " MiB of total host memory\n",
+            __func__, sum_projected_used/MiB, sum_free/MiB);
+        if (sum_projected_free >= margins[0]) {
+            LOG_INF("%s: will leave %" PRId64 " >= %" PRId64 " MiB of system memory, no changes needed\n",
+                __func__, sum_projected_free/MiB, margins[0]/MiB);
+            return;
+        }
+    } else {
+        if (nd > 1) {
+            LOG_INF("%s: projected memory use with initial parameters [MiB]:\n", __func__);
+        }
+        for (size_t id = 0; id < nd; id++) {
+            const llama_device_memory_data & dmd = dmds_full[id];
+
+            const int64_t projected_used = dmd.mb.total();
+            const int64_t projected_free = dmd.free - projected_used;
+            projected_free_per_device.push_back(projected_free);
+
+            sum_free            += dmd.free;
+            sum_projected_used  += projected_used;
+            sum_projected_free  += projected_free;
+            sum_projected_model += dmd.mb.model;
+
+            if (nd > 1) {
+                LOG_INF("%s:   - %s: %6" PRId64 " total, %6" PRId64 " used, %6" PRId64 " free vs. target of %6" PRId64 "\n",
+                    __func__, dev_names[id].c_str(), dmd.total/MiB, projected_used/MiB, projected_free/MiB, margins[id]/MiB);
+            }
+        }
+        assert(sum_free >= 0 && sum_projected_used >= 0);
+        LOG_INF("%s: projected to use %" PRId64 " MiB of device memory vs. %" PRId64 " MiB of free device memory\n",
+            __func__, sum_projected_used/MiB, sum_free/MiB);
+        if (nd == 1) {
+            if (projected_free_per_device[0] >= margins[0]) {
+                LOG_INF("%s: will leave %" PRId64 " >= %" PRId64 " MiB of free device memory, no changes needed\n",
+                    __func__, projected_free_per_device[0]/MiB, margins[0]/MiB);
+                return;
+            }
+        } else {
+            bool changes_needed = false;
+            for (size_t id = 0; id < nd; id++) {
+                if (projected_free_per_device[id] < margins[id]) {
+                    changes_needed = true;
+                    break;
+                }
+            }
+            if (!changes_needed) {
+                LOG_INF("%s: targets for free memory can be met on all devices, no changes needed\n", __func__);
+                return;
+            }
+        }
+    }
+
+    // step 2: try reducing memory use by reducing the context size
+
+    {
+        int64_t global_surplus = sum_projected_free;
+        if (nd == 0) {
+            global_surplus -= margins[0];
+        } else {
+            for (size_t id = 0; id < nd; id++) {
+                global_surplus -= margins[id];
+            }
+        }
+        if (global_surplus < 0) {
+            if (nd <= 1) {
+                LOG_INF("%s: cannot meet free memory target of %" PRId64 " MiB, need to reduce device memory by %" PRId64 " MiB\n",
+                    __func__, margins[0]/MiB, -global_surplus/MiB);
+            } else {
+                LOG_INF(
+                    "%s: cannot meet free memory targets on all devices, need to use %" PRId64 " MiB less in total\n",
+                    __func__, -global_surplus/MiB);
+            }
+            if (cparams->n_ctx == 0) {
+                if (hp_nct > n_ctx_min) {
+                    int64_t sum_used_target = sum_free;
+                    if (nd == 0) {
+                        sum_used_target -= margins[0];
+                    } else {
+                        for (size_t id = 0; id < nd; id++) {
+                            sum_used_target -= margins[id];
+                        }
+                    }
+                    if (nd > 1) {
+                        // for multiple devices we need to be more conservative in terms of how much context we think can fit:
+                        //   - for dense models only whole layers can be assigned to devices
+                        //   - for MoE models only whole tensors can be assigned to devices, which we estimate to be <= 1/3 of a layer
+                        //   - on average we expect a waste of 0.5 layers/tensors per device
+                        //   - use slightly more than the expected average for nd devices to be safe
+                        const int64_t model_per_layer = sum_projected_model / std::min(uint32_t(mparams->n_gpu_layers), hp_ngl);
+                        sum_used_target -= (nd + 1) * model_per_layer / (hp_nex == 0 ? 2 : 6);
+                    }
+
+                    int64_t sum_projected_used_min_ctx = 0;
+                    cparams->n_ctx = n_ctx_min;
+                    const dmds_t dmds_min_ctx = common_get_device_memory_data(path_model, mparams, cparams, devs, hp_ngl, hp_nct, hp_nex, log_level);
+                    if (nd == 0) {
+                        sum_projected_used_min_ctx = dmds_min_ctx.back().mb.total();
+                    } else {
+                        for (size_t id = 0; id < nd; id++) {
+                            sum_projected_used_min_ctx += dmds_min_ctx[id].mb.total();
+                        }
+                    }
+                    if (sum_used_target > sum_projected_used_min_ctx) {
+                        // linear interpolation between minimum and maximum context size:
+                        cparams->n_ctx += (hp_nct - n_ctx_min) * (sum_used_target - sum_projected_used_min_ctx)
+                            / (sum_projected_used - sum_projected_used_min_ctx);
+                        cparams->n_ctx = std::max(cparams->n_ctx - cparams->n_ctx % 256, n_ctx_min); // round down context for CUDA backend
+
+                        const int64_t bytes_per_ctx = (sum_projected_used - sum_projected_used_min_ctx) / (hp_nct - n_ctx_min);
+                        const int64_t memory_reduction = (hp_nct - cparams->n_ctx) * bytes_per_ctx;
+                        LOG_INF("%s: context size reduced from %" PRIu32 " to %" PRIu32 " -> need %" PRId64 " MiB less memory in total\n",
+                            __func__, hp_nct, cparams->n_ctx, memory_reduction/MiB);
+                        if (nd <= 1) {
+                            LOG_INF("%s: entire model can be fit by reducing context\n", __func__);
+                            return;
+                        }
+                        LOG_INF("%s: entire model should be fit across devices by reducing context\n", __func__);
+                    } else {
+                        const int64_t memory_reduction = sum_projected_used - sum_projected_used_min_ctx;
+                        LOG_INF("%s: context size reduced from %" PRIu32 " to %" PRIu32 " -> need %" PRId64 " MiB less memory in total\n",
+                            __func__, hp_nct, cparams->n_ctx, memory_reduction/MiB);
+                    }
+                } else {
+                    if (n_ctx_min == UINT32_MAX) {
+                        LOG_INF("%s: user has requested full context size of %" PRIu32 " -> no change\n", __func__, hp_nct);
+                    } else {
+                        LOG_INF("%s: default model context size is %" PRIu32 " which is <= the min. context size of %" PRIu32 " -> no change\n",
+                            __func__, hp_nct, n_ctx_min);
+                    }
+                }
+            } else {
+                LOG_INF("%s: context size set by user to %" PRIu32 " -> no change\n", __func__, cparams->n_ctx);
+            }
+        }
+    }
+    if (nd == 0) {
+        throw common_params_fit_exception("was unable to fit model into system memory by reducing context, abort");
+    }
+
+    if (mparams->n_gpu_layers != default_mparams.n_gpu_layers) {
+        throw common_params_fit_exception("n_gpu_layers already set by user to " + std::to_string(mparams->n_gpu_layers) + ", abort");
+    }
+    if (nd > 1) {
+        if (!tensor_split) {
+            throw common_params_fit_exception("did not provide a buffer to write the tensor_split to, abort");
+        }
+        if (mparams->tensor_split) {
+            for (size_t id = 0; id < nd; id++) {
+                if (mparams->tensor_split[id] != 0.0f) {
+                    throw common_params_fit_exception("model_params::tensor_split already set by user, abort");
+                }
+            }
+        }
+        if (mparams->split_mode == LLAMA_SPLIT_MODE_ROW) {
+            throw common_params_fit_exception("changing weight allocation for LLAMA_SPLIT_MODE_ROW not implemented, abort");
+        }
+    }
+    if (!tensor_buft_overrides) {
+        throw common_params_fit_exception("did not provide buffer to set tensor_buft_overrides, abort");
+    }
+    if (mparams->tensor_buft_overrides && (mparams->tensor_buft_overrides->pattern || mparams->tensor_buft_overrides->buft)) {
+        throw common_params_fit_exception("model_params::tensor_buft_overrides already set by user, abort");
+    }
+
+    // step 3: iteratively fill the back to front with "dense" layers
+    //   - for a dense model simply fill full layers, giving each device a contiguous slice of the model
+    //   - for a MoE model, same as dense model but with all MoE tensors in system memory
+
+    // utility function that returns a static C string matching the tensors for a specific layer index and layer fraction:
+    auto get_overflow_pattern = [&](const size_t il, const common_layer_fraction_t lf) -> const char * {
+        constexpr size_t n_strings = 1000;
+        if (il >= n_strings) {
+            throw std::runtime_error("at most " + std::to_string(n_strings) + " model layers are supported");
+        }
+        switch (lf) {
+            case LAYER_FRACTION_ATTN: {
+                static std::array<std::string, n_strings> patterns;
+                if (patterns[il].empty()) {
+                    patterns[il] = "blk\\." + std::to_string(il) + "\\.ffn_(gate|up|gate_up|down).*";
+                }
+                return patterns[il].c_str();
+            }
+            case LAYER_FRACTION_UP: {
+                static std::array<std::string, n_strings> patterns;
+                if (patterns[il].empty()) {
+                    patterns[il] = "blk\\." + std::to_string(il) + "\\.ffn_(gate|gate_up|down).*";
+                }
+                return patterns[il].c_str();
+            }
+            case LAYER_FRACTION_GATE: {
+                static std::array<std::string, n_strings> patterns;
+                if (patterns[il].empty()) {
+                    patterns[il] = "blk\\." + std::to_string(il) + "\\.ffn_down.*";
+                }
+                return patterns[il].c_str();
+            }
+            case LAYER_FRACTION_MOE: {
+                static std::array<std::string, n_strings> patterns;
+                if (patterns[il].empty()) {
+                    patterns[il] = "blk\\." + std::to_string(il) + "\\.ffn_(up|down|gate_up|gate)_(ch|)exps";
+                }
+                return patterns[il].c_str();
+            }
+            default:
+                GGML_ABORT("fatal error");
+        }
+    };
+
+    struct ngl_t {
+        uint32_t n_layer = 0; // number of total layers
+        uint32_t n_part  = 0; // number of partial layers, <= n_layer
+
+        // for the first partial layer varying parts can overflow, all further layers use LAYER_FRACTION_MOE:
+        common_layer_fraction_t overflow_type = LAYER_FRACTION_MOE;
+
+        uint32_t n_full() const {
+            assert(n_layer >= n_part);
+            return n_layer - n_part;
+        }
+    };
+
+    const size_t ntbo = llama_max_tensor_buft_overrides();
+
+    // utility function to set n_gpu_layers and tensor_split
+    auto set_ngl_tensor_split_tbo = [&](
+            const std::vector<ngl_t> & ngl_per_device,
+            const std::vector<ggml_backend_buffer_type_t> & overflow_bufts,
+            llama_model_params & mparams) {
+        mparams.n_gpu_layers = 0;
+        for (size_t id = 0; id < nd; id++) {
+            mparams.n_gpu_layers += ngl_per_device[id].n_layer;
+            if (nd > 1) {
+                tensor_split[id] = ngl_per_device[id].n_layer;
+            }
+        }
+        assert(uint32_t(mparams.n_gpu_layers) <= hp_ngl + 1);
+        uint32_t il0 = hp_ngl + 1 - mparams.n_gpu_layers; // start index for tensor buft overrides
+
+        mparams.tensor_split = tensor_split;
+
+        size_t itbo = 0;
+        for (size_t id = 0; id < nd; id++) {
+            il0 += ngl_per_device[id].n_full();
+            for (uint32_t il = il0; il < il0 + ngl_per_device[id].n_part; il++) {
+                if (itbo + 1 >= ntbo) {
+                    tensor_buft_overrides[itbo].pattern = nullptr;
+                    tensor_buft_overrides[itbo].buft    = nullptr;
+                    itbo++;
+                    mparams.tensor_buft_overrides = tensor_buft_overrides;
+                    throw common_params_fit_exception("llama_max_tensor_buft_overrides() == "
+                        + std::to_string(ntbo) + " is insufficient for model");
+                }
+                tensor_buft_overrides[itbo].pattern = get_overflow_pattern(il, il == il0 ? ngl_per_device[id].overflow_type : LAYER_FRACTION_MOE);
+                tensor_buft_overrides[itbo].buft = il == il0 ? overflow_bufts[id] : ggml_backend_cpu_buffer_type();
+                itbo++;
+            }
+            il0 += ngl_per_device[id].n_part;
+        }
+        tensor_buft_overrides[itbo].pattern = nullptr;
+        tensor_buft_overrides[itbo].buft    = nullptr;
+        itbo++;
+        mparams.tensor_buft_overrides = tensor_buft_overrides;
+    };
+
+    // utility function that returns the memory use per device for given numbers of layers per device
+    auto get_memory_for_layers = [&](
+            const char * func_name,
+            const std::vector<ngl_t> & ngl_per_device,
+            const std::vector<ggml_backend_buffer_type_t> & overflow_bufts) -> std::vector<int64_t> {
+        llama_model_params mparams_copy = *mparams;
+        set_ngl_tensor_split_tbo(ngl_per_device, overflow_bufts, mparams_copy);
+
+        const dmds_t dmd_nl = common_get_device_memory_data(
+            path_model, &mparams_copy, cparams, devs, hp_ngl, hp_nct, hp_nex, log_level);
+
+        LOG_INF("%s: memory for test allocation by device:\n", func_name);
+        for (size_t id = 0; id < nd; id++) {
+            const ngl_t & n = ngl_per_device[id];
+            LOG_INF(
+                "%s: id=%zu, n_layer=%2" PRIu32 ", n_part=%2" PRIu32 ", overflow_type=%d, mem=%6" PRId64 " MiB\n",
+                func_name, id, n.n_layer, n.n_part, int(n.overflow_type), dmd_nl[id].mb.total()/MiB);
+        }
+
+        std::vector<int64_t> ret;
+        ret.reserve(nd);
+        for (size_t id = 0; id < nd; id++) {
+            ret.push_back(dmd_nl[id].mb.total());
+        }
+        return ret;
+    };
+
+    int64_t global_surplus_cpu_moe = 0;
+    if (hp_nex > 0) {
+        const static std::string pattern_moe_all = "blk\\.\\d+\\.ffn_(up|down|gate_up|gate)_(ch|)exps"; // matches all MoE tensors
+        ggml_backend_buffer_type_t cpu_buft = ggml_backend_cpu_buffer_type();
+        tensor_buft_overrides[0] = {pattern_moe_all.c_str(), cpu_buft};
+        tensor_buft_overrides[1] = {nullptr, nullptr};
+        mparams->tensor_buft_overrides = tensor_buft_overrides;
+
+        LOG_INF("%s: getting device memory data with all MoE tensors moved to system memory:\n", __func__);
+        const dmds_t dmds_cpu_moe = common_get_device_memory_data(
+            path_model, mparams, cparams, devs, hp_ngl, hp_nct, hp_nex, log_level);
+
+        for (size_t id = 0; id < nd; id++) {
+            global_surplus_cpu_moe += dmds_cpu_moe[id].free;
+            global_surplus_cpu_moe -= int64_t(dmds_cpu_moe[id].mb.total()) + margins[id];
+        }
+
+        if (global_surplus_cpu_moe > 0) {
+            LOG_INF("%s: with only dense weights in device memory there is a total surplus of %" PRId64 " MiB\n",
+                __func__, global_surplus_cpu_moe/MiB);
+        } else {
+            LOG_INF("%s: with only dense weights in device memory there is still a total deficit of %" PRId64 " MiB\n",
+                __func__, -global_surplus_cpu_moe/MiB);
+        }
+
+        // reset
+        tensor_buft_overrides[0] = {nullptr, nullptr};
+        mparams->tensor_buft_overrides = tensor_buft_overrides;
+    }
+
+    std::vector<int64_t> targets; // maximum acceptable memory use per device
+    targets.reserve(nd);
+    for (size_t id = 0; id < nd; id++) {
+        targets.push_back(dmds_full[id].free - margins[id]);
+        LOG_INF("%s: id=%zu, target=%" PRId64 " MiB\n", __func__, id, targets[id]/MiB);
+    }
+
+    std::vector<ggml_backend_buffer_type_t> overflow_bufts; // which bufts the first partial layer of a device overflows to:
+    overflow_bufts.reserve(nd);
+    for (size_t id = 0; id < nd; id++) {
+        overflow_bufts.push_back(ggml_backend_cpu_buffer_type());
+    }
+
+    std::vector<ngl_t> ngl_per_device(nd);
+    std::vector<int64_t> mem = get_memory_for_layers(__func__, ngl_per_device, overflow_bufts);
+
+    // optimize the number of layers per device using the method of false position:
+    //   - ngl_per_device has 0 layers for each device, lower bound
+    //   - try a "high" configuration where a device is given all unassigned layers
+    //   - interpolate the memory use / layer between low and high linearly to get a guess where it meets our target
+    //   - check memory use of our guess, replace either the low or high bound
+    //   - once we only have a difference of a single layer, stop and return the lower bound that just barely still fits
+    //   - the last device has the output layer, which cannot be a partial layer
+    if (hp_nex == 0) {
+        LOG_INF("%s: filling dense layers back-to-front:\n", __func__);
+    } else {
+        LOG_INF("%s: filling dense-only layers back-to-front:\n", __func__);
+    }
+    for (int id = nd - 1; id >= 0; id--) {
+        uint32_t n_unassigned = hp_ngl + 1;
+        for (size_t jd = id + 1; jd < nd; ++jd) {
+            assert(n_unassigned >= ngl_per_device[jd].n_layer);
+            n_unassigned -= ngl_per_device[jd].n_layer;
+        }
+
+        std::vector<ngl_t> ngl_per_device_high = ngl_per_device;
+        ngl_per_device_high[id].n_layer = n_unassigned;
+        if (hp_nex > 0) {
+            ngl_per_device_high[id].n_part = size_t(id) < nd - 1 ? ngl_per_device_high[id].n_layer : ngl_per_device_high[id].n_layer - 1;
+        }
+        if (ngl_per_device_high[id].n_layer > 0) {
+            std::vector<int64_t> mem_high = get_memory_for_layers(__func__, ngl_per_device_high, overflow_bufts);
+            if (mem_high[id] > targets[id]) {
+                assert(ngl_per_device_high[id].n_layer > ngl_per_device[id].n_layer);
+                uint32_t delta = ngl_per_device_high[id].n_layer - ngl_per_device[id].n_layer;
+                LOG_INF("%s: start filling device %" PRIu32 ", delta=%" PRIu32 "\n", __func__, id, delta);
+                while (delta > 1) {
+                    uint32_t step_size = int64_t(delta) * (targets[id] - mem[id]) / (mem_high[id] - mem[id]);
+                    step_size = std::max(step_size, uint32_t(1));
+                    step_size = std::min(step_size, delta - 1);
+
+                    std::vector<ngl_t> ngl_per_device_test = ngl_per_device;
+                    ngl_per_device_test[id].n_layer += step_size;
+                    if (hp_nex) {
+                        ngl_per_device_test[id].n_part += size_t(id) == nd - 1 && ngl_per_device_test[id].n_part == 0 ?
+                            step_size - 1 : step_size; // the first layer is the output layer which must always be full
+                    }
+                    const std::vector<int64_t> mem_test = get_memory_for_layers(__func__, ngl_per_device_test, overflow_bufts);
+
+                    if (mem_test[id] <= targets[id]) {
+                        ngl_per_device = ngl_per_device_test;
+                        mem            = mem_test;
+                        LOG_INF("%s: set ngl_per_device[%d].n_layer=%" PRIu32 "\n", __func__, id, ngl_per_device[id].n_layer);
+                    } else {
+                        ngl_per_device_high = ngl_per_device_test;
+                        mem_high            = mem_test;
+                        LOG_INF("%s: set ngl_per_device_high[%d].n_layer=%" PRIu32 "\n", __func__, id, ngl_per_device_high[id].n_layer);
+                    }
+                    delta = ngl_per_device_high[id].n_layer - ngl_per_device[id].n_layer;
+                }
+            } else {
+                assert(ngl_per_device_high[id].n_layer == n_unassigned);
+                ngl_per_device = ngl_per_device_high;
+                mem            = mem_high;
+                LOG_INF("%s: set ngl_per_device[%d].n_layer=%" PRIu32 "\n", __func__, id, ngl_per_device[id].n_layer);
+            }
+        }
+
+        const int64_t projected_margin = dmds_full[id].free - mem[id];
+        LOG_INF(
+            "%s:   - %s: %2" PRIu32 " layers, %6" PRId64 " MiB used, %6" PRId64 " MiB free\n",
+            __func__, dev_names[id].c_str(), ngl_per_device[id].n_layer, mem[id]/MiB, projected_margin/MiB);
+    }
+    if (hp_nex == 0 || global_surplus_cpu_moe <= 0) {
+        set_ngl_tensor_split_tbo(ngl_per_device, overflow_bufts, *mparams);
+        return;
+    }
+
+    // step 4: for a MoE model where all dense tensors fit,
+    //     convert the dense-only layers in the back to full layers in the front until all devices are full
+    // essentially the same procedure as for the dense-only layers except front-to-back
+    // also, try fitting at least part of one more layer to reduce waste for "small" GPUs with e.g. 24 GiB VRAM
+
+    size_t id_dense_start = nd;
+    for (int id = nd - 1; id >= 0; id--) {
+        if (ngl_per_device[id].n_layer > 0) {
+            id_dense_start = id;
+            continue;
+        }
+        break;
+    }
+    assert(id_dense_start < nd);
+
+    LOG_INF("%s: converting dense-only layers to full layers and filling them front-to-back with overflow to next device/system memory:\n", __func__);
+    for (size_t id = 0; id <= id_dense_start && id_dense_start < nd; id++) {
+        std::vector<ngl_t> ngl_per_device_high = ngl_per_device;
+        for (size_t jd = id_dense_start; jd < nd; jd++) {
+            const uint32_t n_layer_move = jd < nd - 1 ? ngl_per_device_high[jd].n_layer : ngl_per_device_high[jd].n_layer - 1;
+            ngl_per_device_high[id].n_layer += n_layer_move;
+            ngl_per_device_high[jd].n_layer -= n_layer_move;
+            ngl_per_device_high[jd].n_part = 0;
+        }
+        size_t id_dense_start_high = nd - 1;
+        std::vector<int64_t> mem_high = get_memory_for_layers(__func__, ngl_per_device_high, overflow_bufts);
+
+        if (mem_high[id] > targets[id]) {
+            assert(ngl_per_device_high[id].n_full() >= ngl_per_device[id].n_full());
+            uint32_t delta = ngl_per_device_high[id].n_full() - ngl_per_device[id].n_full();
+            while (delta > 1) {
+                uint32_t step_size = int64_t(delta) * (targets[id] - mem[id]) / (mem_high[id] - mem[id]);
+                step_size = std::max(step_size, uint32_t(1));
+                step_size = std::min(step_size, delta - 1);
+
+                std::vector<ngl_t> ngl_per_device_test = ngl_per_device;
+                size_t id_dense_start_test = id_dense_start;
+                uint32_t n_converted_test = 0;
+                for (;id_dense_start_test < nd; id_dense_start_test++) {
+                    const uint32_t n_convert_jd = std::min(step_size - n_converted_test, ngl_per_device_test[id_dense_start_test].n_part);
+                    ngl_per_device_test[id_dense_start_test].n_layer -= n_convert_jd;
+                    ngl_per_device_test[id_dense_start_test].n_part -= n_convert_jd;
+                    ngl_per_device_test[id].n_layer += n_convert_jd;
+                    n_converted_test += n_convert_jd;
+
+                    if (ngl_per_device_test[id_dense_start_test].n_part > 0) {
+                        break;
+                    }
+                }
+                const std::vector<int64_t> mem_test = get_memory_for_layers(__func__, ngl_per_device_test, overflow_bufts);
+
+                if (mem_test[id] <= targets[id]) {
+                    ngl_per_device = ngl_per_device_test;
+                    mem            = mem_test;
+                    id_dense_start = id_dense_start_test;
+                    LOG_INF("%s: set ngl_per_device[%zu].(n_layer, n_part)=(%" PRIu32 ", %" PRIu32 "), id_dense_start=%zu\n",
+                        __func__, id, ngl_per_device[id].n_layer, ngl_per_device[id].n_part, id_dense_start);
+                } else {
+                    ngl_per_device_high = ngl_per_device_test;
+                    mem_high            = mem_test;
+                    id_dense_start_high = id_dense_start_test;
+                    LOG_INF("%s: set ngl_per_device_high[%zu].(n_layer, n_part)=(%" PRIu32 ", %" PRIu32 "), id_dense_start_high=%zu\n",
+                        __func__, id, ngl_per_device_high[id].n_layer, ngl_per_device_high[id].n_part, id_dense_start_high);
+                }
+                assert(ngl_per_device_high[id].n_full() >= ngl_per_device[id].n_full());
+                delta = ngl_per_device_high[id].n_full() - ngl_per_device[id].n_full();
+            }
+        } else {
+            ngl_per_device = ngl_per_device_high;
+            mem            = mem_high;
+            id_dense_start = id_dense_start_high;
+            LOG_INF("%s: set ngl_per_device[%zu].(n_layer, n_part)=(%" PRIu32 ", %" PRIu32 "), id_dense_start=%zu\n",
+                __func__, id, ngl_per_device[id].n_layer, ngl_per_device[id].n_part, id_dense_start);
+        }
+
+        // try to fit at least part of one more layer
+        if (ngl_per_device[id_dense_start].n_layer > (id < nd - 1 ? 0 : 1)) {
+            std::vector<ngl_t> ngl_per_device_test = ngl_per_device;
+            size_t id_dense_start_test = id_dense_start;
+            ngl_per_device_test[id_dense_start_test].n_layer--;
+            ngl_per_device_test[id_dense_start_test].n_part--;
+            ngl_per_device_test[id].n_layer++;
+            ngl_per_device_test[id].n_part++;
+            if (ngl_per_device_test[id_dense_start_test].n_part == 0) {
+                id_dense_start_test++;
+            }
+            ngl_per_device_test[id].overflow_type = LAYER_FRACTION_UP;
+            std::vector<ggml_backend_buffer_type_t> overflow_bufts_test = overflow_bufts;
+            if (id < nd - 1) {
+                overflow_bufts_test[id] = ggml_backend_dev_buffer_type(devs[id + 1]);
+            }
+            LOG_INF("%s: trying to fit one extra layer with overflow_type=LAYER_FRACTION_UP\n", __func__);
+            std::vector<int64_t> mem_test = get_memory_for_layers(__func__, ngl_per_device_test, overflow_bufts_test);
+            if (mem_test[id] < targets[id] && (id + 1 == nd || mem_test[id + 1] < targets[id + 1])) {
+                ngl_per_device = ngl_per_device_test;
+                overflow_bufts = overflow_bufts_test;
+                mem            = mem_test;
+                id_dense_start = id_dense_start_test;
+                LOG_INF("%s: set ngl_per_device[%zu].(n_layer, n_part, overflow_type)=(%" PRIu32 ", %" PRIu32 ", UP), id_dense_start=%zu\n",
+                    __func__, id, ngl_per_device[id].n_layer, ngl_per_device[id].n_part, id_dense_start);
+
+                ngl_per_device_test[id].overflow_type = LAYER_FRACTION_GATE;
+                LOG_INF("%s: trying to fit one extra layer with overflow_type=LAYER_FRACTION_GATE\n", __func__);
+                mem_test = get_memory_for_layers(__func__, ngl_per_device_test, overflow_bufts_test);
+                if (mem_test[id] < targets[id] && (id + 1 == nd || mem_test[id + 1] < targets[id + 1])) {
+                    ngl_per_device = ngl_per_device_test;
+                    overflow_bufts = overflow_bufts_test;
+                    mem            = mem_test;
+                    id_dense_start = id_dense_start_test;
+                    LOG_INF("%s: set ngl_per_device[%zu].(n_layer, n_part, overflow_type)=(%" PRIu32 ", %" PRIu32 ", GATE), id_dense_start=%zu\n",
+                        __func__, id, ngl_per_device[id].n_layer, ngl_per_device[id].n_part, id_dense_start);
+                }
+            } else {
+                ngl_per_device_test[id].overflow_type = LAYER_FRACTION_ATTN;
+                LOG_INF("%s: trying to fit one extra layer with overflow_type=LAYER_FRACTION_ATTN\n", __func__);
+                mem_test = get_memory_for_layers(__func__, ngl_per_device_test, overflow_bufts_test);
+                if (mem_test[id] < targets[id] && (id + 1 == nd || mem_test[id + 1] < targets[id + 1])) {
+                    ngl_per_device = ngl_per_device_test;
+                    overflow_bufts = overflow_bufts_test;
+                    mem            = mem_test;
+                    id_dense_start = id_dense_start_test;
+                    LOG_INF("%s: set ngl_per_device[%zu].(n_layer, n_part, overflow_type)=(%" PRIu32 ", %" PRIu32 ", ATTN), id_dense_start=%zu\n",
+                        __func__, id, ngl_per_device[id].n_layer, ngl_per_device[id].n_part, id_dense_start);
+                }
+            }
+        }
+
+        const int64_t projected_margin = dmds_full[id].free - mem[id];
+        LOG_INF(
+            "%s:   - %s: %2" PRIu32 " layers (%2" PRIu32 " overflowing), %6" PRId64 " MiB used, %6" PRId64 " MiB free\n",
+            __func__, dev_names[id].c_str(), ngl_per_device[id].n_layer, ngl_per_device[id].n_part, mem[id]/MiB, projected_margin/MiB);
+    }
+
+    // print info for devices that were not changed during the conversion from dense only to full layers:
+    for (size_t id = id_dense_start + 1; id < nd; id++) {
+        const int64_t projected_margin = dmds_full[id].free - mem[id];
+        LOG_INF(
+            "%s:   - %s: %2" PRIu32 " layers (%2" PRIu32 " overflowing), %6" PRId64 " MiB used, %6" PRId64 " MiB free\n",
+            __func__, dev_names[id].c_str(), ngl_per_device[id].n_layer, ngl_per_device[id].n_part, mem[id]/MiB, projected_margin/MiB);
+    }
+
+    set_ngl_tensor_split_tbo(ngl_per_device, overflow_bufts, *mparams);
+}
+
+enum common_params_fit_status common_fit_params(
+        const char * path_model,
+        llama_model_params * mparams,
+        llama_context_params * cparams,
+        float * tensor_split,
+        llama_model_tensor_buft_override * tensor_buft_overrides,
+        size_t * margins,
+        uint32_t n_ctx_min,
+        ggml_log_level log_level) {
+    const int64_t t0_us = llama_time_us();
+    common_params_fit_status status = COMMON_PARAMS_FIT_STATUS_SUCCESS;
+    try {
+        common_params_fit_impl(path_model, mparams, cparams, tensor_split, tensor_buft_overrides, margins, n_ctx_min, log_level);
+        LOG_INF("%s: successfully fit params to free device memory\n", __func__);
+    } catch (const common_params_fit_exception & e) {
+        LOG_WRN("%s: failed to fit params to free device memory: %s\n", __func__, e.what());
+        status = COMMON_PARAMS_FIT_STATUS_FAILURE;
+    } catch (const std::runtime_error & e) {
+        LOG_ERR("%s: encountered an error while trying to fit params to free device memory: %s\n", __func__, e.what());
+        status = COMMON_PARAMS_FIT_STATUS_ERROR;
+    }
+    const int64_t t1_us = llama_time_us();
+    LOG_INF("%s: fitting params to free memory took %.2f seconds\n", __func__, (t1_us - t0_us) * 1e-6);
+    return status;
+}
+
+void common_memory_breakdown_print(const struct llama_context * ctx) {
+    //const auto & devices = ctx->get_model().devices;
+    const auto * model = llama_get_model(ctx);
+
+    std::vector<ggml_backend_dev_t> devices;
+    for (int i = 0; i < llama_model_n_devices(model); i++) {
+        devices.push_back(llama_model_get_device(model, i));
+    }
+
+    llama_memory_breakdown memory_breakdown = llama_get_memory_breakdown(ctx);
+
+    std::vector<std::array<std::string, 9>> table_data;
+    table_data.reserve(devices.size());
+    const std::string template_header = "%s: | %s | %s   %s    %s   %s   %s   %s    %s |\n";
+    const std::string template_gpu    = "%s: | %s | %s = %s + (%s = %s + %s + %s) + %s |\n";
+    const std::string template_other  = "%s: | %s | %s   %s    %s = %s + %s + %s    %s |\n";
+
+    table_data.push_back({template_header, "memory breakdown [MiB]", "total", "free", "self", "model", "context", "compute", "unaccounted"});
+
+    constexpr size_t MiB = 1024 * 1024;
+    const std::vector<std::string> desc_prefixes_strip = {"NVIDIA ", "GeForce ", "Tesla ", "AMD ", "Radeon ", "Instinct "};
+
+    // track seen buffer types to avoid double counting:
+    std::set<ggml_backend_buffer_type_t> seen_buffer_types;
+
+    // accumulative memory breakdown for each device and for host:
+    std::vector<llama_memory_breakdown_data> mb_dev(devices.size());
+    llama_memory_breakdown_data              mb_host;
+
+    for (const auto & buft_mb : memory_breakdown) {
+        ggml_backend_buffer_type_t          buft = buft_mb.first;
+        const llama_memory_breakdown_data & mb   = buft_mb.second;
+        if (ggml_backend_buft_is_host(buft)) {
+            mb_host.model   += mb.model;
+            mb_host.context += mb.context;
+            mb_host.compute += mb.compute;
+            seen_buffer_types.insert(buft);
+            continue;
+        }
+        ggml_backend_dev_t dev = ggml_backend_buft_get_device(buft);
+        if (dev) {
+            int i_dev = -1;
+            for (size_t i = 0; i < devices.size(); i++) {
+                if (devices[i] == dev) {
+                    i_dev = i;
+                    break;
+                }
+            }
+            if (i_dev != -1) {
+                mb_dev[i_dev].model   += mb.model;
+                mb_dev[i_dev].context += mb.context;
+                mb_dev[i_dev].compute += mb.compute;
+                seen_buffer_types.insert(buft);
+                continue;
+            }
+        }
+    }
+
+    // print memory breakdown for each device:
+    for (size_t i = 0; i < devices.size(); i++) {
+        ggml_backend_dev_t dev = devices[i];
+        llama_memory_breakdown_data mb = mb_dev[i];
+
+        const std::string name = ggml_backend_dev_name(dev);
+        std::string desc = ggml_backend_dev_description(dev);
+        for (const std::string & prefix : desc_prefixes_strip) {
+            if (desc.length() >= prefix.length() && desc.substr(0, prefix.length()) == prefix) {
+                desc = desc.substr(prefix.length());
+            }
+        }
+
+        size_t free, total;
+        ggml_backend_dev_memory(dev, &free, &total);
+
+        const size_t self = mb.model + mb.context + mb.compute;
+        const int64_t unaccounted = static_cast<int64_t>(total) - static_cast<int64_t>(free) - static_cast<int64_t>(self);
+
+        table_data.push_back({
+            template_gpu,
+            "  - " + name + " (" + desc + ")",
+            std::to_string(total / MiB),
+            std::to_string(free / MiB),
+            std::to_string(self / MiB),
+            std::to_string(mb.model / MiB),
+            std::to_string(mb.context / MiB),
+            std::to_string(mb.compute / MiB),
+            std::to_string(unaccounted / static_cast<int64_t>(MiB))});
+    }
+
+    // print memory breakdown for host:
+    {
+        const size_t self = mb_host.model + mb_host.context + mb_host.compute;
+        table_data.push_back({
+            template_other,
+            "  - Host",
+            "", // total
+            "", // free
+            std::to_string(self / MiB),
+            std::to_string(mb_host.model / MiB),
+            std::to_string(mb_host.context / MiB),
+            std::to_string(mb_host.compute / MiB),
+            ""}); // unaccounted
+    }
+
+    // print memory breakdown for all remaining buffer types:
+    for (const auto & buft_mb : memory_breakdown) {
+        ggml_backend_buffer_type_t          buft = buft_mb.first;
+        const llama_memory_breakdown_data & mb   = buft_mb.second;
+        if (seen_buffer_types.count(buft) == 1) {
+            continue;
+        }
+        const std::string name = ggml_backend_buft_name(buft);
+        const size_t self = mb.model + mb.context + mb.compute;
+        table_data.push_back({
+            template_other,
+            "  - " + name,
+            "", // total
+            "", // free
+            std::to_string(self / MiB),
+            std::to_string(mb.model / MiB),
+            std::to_string(mb.context / MiB),
+            std::to_string(mb.compute / MiB),
+            ""}); // unaccounted
+        seen_buffer_types.insert(buft);
+    }
+
+    for (size_t j = 1; j < table_data[0].size(); j++) {
+        size_t max_len = 0;
+        for (const auto & td : table_data) {
+            max_len = std::max(max_len, td[j].length());
+        }
+        for (auto & td : table_data) {
+            td[j].insert(j == 1 ? td[j].length() : 0, max_len - td[j].length(), ' ');
+        }
+    }
+    for (const auto & td : table_data) {
+        LOG_INF(td[0].c_str(),
+            __func__, td[1].c_str(), td[2].c_str(), td[3].c_str(), td[4].c_str(), td[5].c_str(),
+            td[6].c_str(), td[7].c_str(), td[8].c_str());
+    }
+}
+
+void common_fit_print(
+        const char * path_model,
+        llama_model_params * mparams,
+        llama_context_params * cparams) {
+    std::vector<ggml_backend_dev_t> devs;
+    uint32_t hp_ngl = 0; // hparams.n_gpu_layers
+    uint32_t hp_nct = 0; // hparams.n_ctx_train
+    uint32_t hp_nex = 0; // hparams.n_expert
+
+    auto dmd = common_get_device_memory_data(path_model, mparams, cparams, devs, hp_ngl, hp_nct, hp_nex, GGML_LOG_LEVEL_ERROR);
+    GGML_ASSERT(dmd.size() == devs.size() + 1);
+
+    for (size_t id = 0; id < devs.size(); id++) {
+        printf("%s ",  ggml_backend_dev_name(devs[id]));
+        printf("%zu ", dmd[id].mb.model/1024/1024);
+        printf("%zu ", dmd[id].mb.context/1024/1024);
+        printf("%zu ", dmd[id].mb.compute/1024/1024);
+        printf("\n");
+    }
+
+    printf("Host ");
+    printf("%zu ", dmd.back().mb.model/1024/1024);
+    printf("%zu ", dmd.back().mb.context/1024/1024);
+    printf("%zu ", dmd.back().mb.compute/1024/1024);
+    printf("\n");
+}
--- a/common/fit.h
+++ b/common/fit.h
@@ -0,0 +1,32 @@
+#pragma once
+
+#include "ggml.h"
+
+enum common_params_fit_status {
+    COMMON_PARAMS_FIT_STATUS_SUCCESS = 0, // found allocations that are projected to fit
+    COMMON_PARAMS_FIT_STATUS_FAILURE = 1, // could not find allocations that are projected to fit
+    COMMON_PARAMS_FIT_STATUS_ERROR   = 2, // a hard error occurred, e.g. because no model could be found at the specified path
+};
+
+// fits mparams and cparams to free device memory (assumes system memory is unlimited)
+//   - returns true if the parameters could be successfully modified to fit device memory
+//   - this function is NOT thread safe because it modifies the global llama logger state
+//   - only parameters that have the same value as in llama_default_model_params are modified
+//     with the exception of the context size which is modified if and only if equal to 0
+enum common_params_fit_status common_fit_params(
+                               const char   * path_model,
+                struct llama_model_params   * mparams,
+                struct llama_context_params * cparams,
+                                      float * tensor_split,          // writable buffer for tensor split, needs at least llama_max_devices elements
+    struct llama_model_tensor_buft_override * tensor_buft_overrides, // writable buffer for overrides, needs at least llama_max_tensor_buft_overrides elements
+                                     size_t * margins,               // margins of memory to leave per device in bytes
+                                   uint32_t   n_ctx_min,             // minimum context size to set when trying to reduce memory use
+                        enum ggml_log_level   log_level);            // minimum log level to print during fitting, lower levels go to debug log
+
+// print estimated memory to stdout
+void common_fit_print(
+                               const char   * path_model,
+                struct llama_model_params   * mparams,
+                struct llama_context_params * cparams);
+
+void common_memory_breakdown_print(const struct llama_context * ctx);
--- a/common/hf-cache.cpp
+++ b/common/hf-cache.cpp
@@ -1,5 +1,6 @@
 #include "hf-cache.h"

+#include "build-info.h"
 #include "common.h"
 #include "log.h"
 #include "http.h"
@@ -200,7 +201,7 @@ static nl::json api_get(const std::string & url,
    auto [cli, parts] = common_http_client(url);

    httplib::Headers headers = {
-        {"User-Agent", "llama-cpp/" + build_info},
+        {"User-Agent", "llama-cpp/" + std::string(llama_build_info())},
        {"Accept", "application/json"}
    };

@@ -229,7 +230,7 @@ static nl::json api_get(const std::string & url,
 static std::string get_repo_commit(const std::string & repo_id,
                                   const std::string & token) {
    try {
-        auto endpoint = get_model_endpoint();
+        auto endpoint = common_get_model_endpoint();
        auto json = api_get(endpoint + "api/models/" + repo_id + "/refs", token);

        if (!json.is_object() ||
@@ -307,7 +308,7 @@ hf_files get_repo_files(const std::string & repo_id,
    hf_files files;

    try {
-        auto endpoint = get_model_endpoint();
+        auto endpoint = common_get_model_endpoint();
        auto json = api_get(endpoint + "api/models/" + repo_id + "/tree/" + commit + "?recursive=true", token);

        if (!json.is_array()) {
--- a/common/jinja/caps.cpp
+++ b/common/jinja/caps.cpp
@@ -1,4 +1,3 @@
-#include "log.h"
 #include "value.h"
 #include "runtime.h"
 #include "caps.h"
--- a/common/jinja/runtime.h
+++ b/common/jinja/runtime.h
@@ -106,10 +106,16 @@ struct statement {
    size_t pos; // position in source, for debugging
    virtual ~statement() = default;
    virtual std::string type() const { return "Statement"; }
+
    // execute_impl must be overridden by derived classes
-    virtual value execute_impl(context &) { throw std::runtime_error("cannot exec " + type()); }
+    virtual value execute_impl(context &) { throw_exec_error(); }
    // execute is the public method to execute a statement with error handling
    value execute(context &);
+
+private:
+    [[noreturn]] void throw_exec_error() const {
+        throw std::runtime_error("cannot exec " + type());
+    }
 };

 // Type Checking Utilities
@@ -143,7 +149,7 @@ struct program : public statement {
    program() = default;
    explicit program(statements && body) : body(std::move(body)) {}
    std::string type() const override { return "Program"; }
-    value execute_impl(context &) override {
+    [[noreturn]] value execute_impl(context &) override {
        throw std::runtime_error("Cannot execute program directly, use jinja::runtime instead");
    }
 };
@@ -195,7 +201,7 @@ struct break_statement : public statement {
        }
    };

-    value execute_impl(context &) override {
+    [[noreturn]] value execute_impl(context &) override {
        throw break_statement::signal();
    }
 };
@@ -209,7 +215,7 @@ struct continue_statement : public statement {
        }
    };

-    value execute_impl(context &) override {
+    [[noreturn]] value execute_impl(context &) override {
        throw continue_statement::signal();
    }
 };
@@ -509,7 +515,7 @@ struct slice_expression : public expression {
        chk_type<expression>(this->step_expr);
    }
    std::string type() const override { return "SliceExpression"; }
-    value execute_impl(context &) override {
+    [[noreturn]] value execute_impl(context &) override {
        throw std::runtime_error("must be handled by MemberExpression");
    }
 };
--- a/common/jinja/value.cpp
+++ b/common/jinja/value.cpp
@@ -590,6 +590,10 @@ static bool string_endswith(const std::string & str, const std::string & suffix)
    return str.compare(str.length() - suffix.length(), suffix.length(), suffix) == 0;
 }

+[[noreturn]] static value string_join_not_implemented(const func_args &) {
+    throw not_implemented_exception("String join builtin not implemented");
+}
+
 const func_builtins & value_string_t::get_builtins() const {
    static const func_builtins builtins = {
        {"default", default_value},
@@ -851,9 +855,7 @@ const func_builtins & value_string_t::get_builtins() const {
            res->val_str.mark_input_based_on(val_input->as_string());
            return res;
        }},
-        {"join", [](const func_args &) -> value {
-            throw not_implemented_exception("String join builtin not implemented");
-        }},
+        {"join", string_join_not_implemented},
    };
    return builtins;
 }
@@ -884,6 +886,9 @@ const func_builtins & value_bool_t::get_builtins() const {
    return builtins;
 }

+[[noreturn]] static value array_unique_not_implemented(const func_args &) {
+    throw not_implemented_exception("Array unique builtin not implemented");
+}

 const func_builtins & value_array_t::get_builtins() const {
    static const func_builtins builtins = {
@@ -1084,13 +1089,14 @@ const func_builtins & value_array_t::get_builtins() const {
            std::reverse(arr.begin(), arr.end());
            return is_val<value_tuple>(val) ? mk_val<value_tuple>(std::move(arr)) : mk_val<value_array>(std::move(arr));
        }},
-        {"unique", [](const func_args &) -> value {
-            throw not_implemented_exception("Array unique builtin not implemented");
-        }},
+        {"unique", array_unique_not_implemented},
    };
    return builtins;
 }

+[[noreturn]] static value object_join_not_implemented(const func_args &) {
+    throw not_implemented_exception("object join not implemented");
+}

 const func_builtins & value_object_t::get_builtins() const {
    if (!has_builtins) {
@@ -1183,9 +1189,7 @@ const func_builtins & value_object_t::get_builtins() const {
            });
            return result;
        }},
-        {"join", [](const func_args &) -> value {
-            throw not_implemented_exception("object join not implemented");
-        }},
+        {"join", object_join_not_implemented},
    };
    return builtins;
 }
--- a/common/jinja/value.h
+++ b/common/jinja/value.h
@@ -129,27 +129,25 @@ struct value_t {
    // Note: only for debugging and error reporting purposes
    virtual std::string type() const { return ""; }

-    virtual int64_t as_int() const { throw std::runtime_error(type() + " is not an int value"); }
-    virtual double as_float() const { throw std::runtime_error(type() + " is not a float value"); }
-    virtual string as_string() const { throw std::runtime_error(type() + " is not a string value"); }
-    virtual bool as_bool() const { throw std::runtime_error(type() + " is not a bool value"); }
-    virtual const std::vector<value> & as_array() const { throw std::runtime_error(type() + " is not an array value"); }
-    virtual const std::vector<std::pair<value, value>> & as_ordered_object() const { throw std::runtime_error(type() + " is not an object value"); }
-    virtual value invoke(const func_args &) const { throw std::runtime_error(type() + " is not a function value"); }
+    virtual int64_t as_int() const { throw_type_error("is not an int value"); }
+    virtual double as_float() const { throw_type_error("is not a float value"); }
+    virtual string as_string() const { throw_type_error("is not a string value"); }
+    virtual bool as_bool() const { throw_type_error("is not a bool value"); }
+    virtual const std::vector<value> & as_array() const { throw_type_error("is not an array value"); }
+    virtual const std::vector<std::pair<value, value>> & as_ordered_object() const { throw_type_error("is not an object value"); }
+    virtual value invoke(const func_args &) const { throw_type_error("is not a function value"); }
    virtual bool is_none() const { return false; }
    virtual bool is_undefined() const { return false; }
-    virtual const func_builtins & get_builtins() const {
-        throw std::runtime_error("No builtins available for type " + type());
-    }
+    virtual const func_builtins & get_builtins() const { throw_type_error("has no builtins"); }

-    virtual bool has_key(const value &) { throw std::runtime_error(type() + " is not an object value"); }
-    virtual void insert(const value & /* key */, const value & /* val */) { throw std::runtime_error(type() + " is not an object value"); }
-    virtual value & at(const value & /* key */, value & /* default_val */) { throw std::runtime_error(type() + " is not an object value"); }
-    virtual value & at(const value & /* key */) { throw std::runtime_error(type() + " is not an object value"); }
-    virtual value & at(const std::string & /* key */, value & /* default_val */) { throw std::runtime_error(type() + " is not an object value"); }
-    virtual value & at(const std::string & /* key */) { throw std::runtime_error(type() + " is not an object value"); }
-    virtual value & at(int64_t /* idx */, value & /* default_val */) { throw std::runtime_error(type() + " is not an array value"); }
-    virtual value & at(int64_t /* idx */) { throw std::runtime_error(type() + " is not an array value"); }
+    virtual bool has_key(const value &) { throw_type_error("is not an object value"); }
+    virtual void insert(const value & /* key */, const value & /* val */) { throw_type_error("is not an object value"); }
+    virtual value & at(const value & /* key */, value & /* default_val */) { throw_type_error("is not an object value"); }
+    virtual value & at(const value & /* key */) { throw_type_error("is not an object value"); }
+    virtual value & at(const std::string & /* key */, value & /* default_val */) { throw_type_error("is not an object value"); }
+    virtual value & at(const std::string & /* key */) { throw_type_error("is not an object value"); }
+    virtual value & at(int64_t /* idx */, value & /* default_val */) { throw_type_error("is not an array value"); }
+    virtual value & at(int64_t /* idx */) { throw_type_error("is not an array value"); }

    virtual bool is_numeric() const { return false; }
    virtual bool is_hashable() const { return false; }
@@ -163,6 +161,11 @@ struct value_t {
    // Note: only for debugging purposes
    virtual std::string as_repr() const { return as_string().str(); }

+private:
+    [[noreturn]] void throw_type_error(const char* expected) const {
+        throw std::runtime_error(type() + " " + expected);
+    }
+
 protected:
    virtual bool equivalent(const value_t &) const = 0;
    virtual bool nonequal(const value_t & other) const { return !equivalent(other); }
--- a/common/log.cpp
+++ b/common/log.cpp
@@ -23,6 +23,10 @@

 int common_log_verbosity_thold = LOG_DEFAULT_LLAMA;

+int common_log_get_verbosity_thold(void) {
+    return common_log_verbosity_thold;
+}
+
 void common_log_set_verbosity_thold(int verbosity) {
    common_log_verbosity_thold = verbosity;
 }
--- a/common/log.h
+++ b/common/log.h
@@ -38,7 +38,7 @@ enum log_colors {

 // needed by the LOG_TMPL macro to avoid computing log arguments if the verbosity lower
 // set via common_log_set_verbosity()
-extern int common_log_verbosity_thold;
+int  common_log_get_verbosity_thold(void);

 void common_log_set_verbosity_thold(int verbosity); // not thread-safe

@@ -98,7 +98,7 @@ void common_log_flush         (struct common_log * log);                    // f

 #define LOG_TMPL(level, verbosity, ...) \
    do { \
-        if ((verbosity) <= common_log_verbosity_thold) { \
+        if ((verbosity) <= common_log_get_verbosity_thold()) { \
            common_log_add(common_log_main(), (level), __VA_ARGS__); \
        } \
    } while (0)
--- a/common/ngram-map.cpp
+++ b/common/ngram-map.cpp
@@ -208,7 +208,7 @@ void common_ngram_map_begin(
                count_keys, count_keys_del, count_values_del, count_map_entries_upd);
    }

-    map.idx_last_check = (map.size_last_begin > 0) ? map.size_last_begin - 1 : 0;
+    map.idx_last_check = size_begin;
    map.size_last_begin = size_begin;
 }

@@ -231,7 +231,7 @@ void common_ngram_map_draft(common_ngram_map & map,
        GGML_ABORT("%s: cur_len exceeds UINT32_MAX: %zu", __func__, cur_len);
    }

-    if (map.idx_last_check  > cur_len) {
+    if (map.idx_last_check > cur_len) {
        // Should not happen because of common_ngram_map_begin().
        GGML_ABORT("%s: map.idx_last_check > cur_len: %zu > %zu", __func__, map.idx_last_check, cur_len);
    }
@@ -386,7 +386,7 @@ void common_ngram_map_draft(common_ngram_map & map,
        LOG_DBG("%s: key_idx = %zu, key_offset = %zu, key_num = %d, draft.size = %zu\n", __func__,
                curr_key.key_idx, key_offset, curr_key.key_num, draft.size());

-        map.last_draft_created   = false;
+        map.last_draft_created   = true;
        map.last_draft_key_idx   = key_offset;
        map.last_draft_value_idx = 0; // value 0 is used for simple mode
        return;
@@ -524,7 +524,7 @@ void common_ngram_map_accept(common_ngram_map & map, uint16_t n_accepted) {
    struct common_ngram_map_value & curr_value = curr_key.values[val_idx]; // value used for draft generation.

    // update the value statistics
-    LOG_INF("common_ngram_map_send_accepted: n_accepted = %d, prev value_num = %d\n",
+    LOG_DBG("common_ngram_map_send_accepted: n_accepted = %d, prev value_num = %d\n",
            n_accepted, curr_value.n_accepted);
    curr_value.n_accepted = n_accepted;
 }
--- a/common/peg-parser.cpp
+++ b/common/peg-parser.cpp
@@ -890,6 +890,10 @@ struct parser_executor {
        }
        return result;
    }
+
+    common_peg_parse_result operator()(const common_peg_gbnf_parser & p) {
+        return arena.parse(p.child, ctx, start_pos);
+    }
 };

 common_peg_parse_result common_peg_arena::parse(common_peg_parse_context & ctx, size_t start) const {
@@ -957,7 +961,8 @@ void common_peg_arena::resolve_refs() {
                                 std::is_same_v<T, common_peg_and_parser> ||
                                 std::is_same_v<T, common_peg_not_parser> ||
                                 std::is_same_v<T, common_peg_tag_parser> ||
-                                 std::is_same_v<T, common_peg_atomic_parser>) {
+                                 std::is_same_v<T, common_peg_atomic_parser> ||
+                                 std::is_same_v<T, common_peg_gbnf_parser>) {
                p.child = resolve_ref(p.child);
            } else if constexpr (std::is_same_v<T, common_peg_rule_parser>) {
                p.child = resolve_ref(p.child);
@@ -1036,6 +1041,8 @@ std::string common_peg_arena::dump_impl(common_peg_parser_id
            return "Not(" + dump_impl(p.child, visited) + ")";
        } else if constexpr (std::is_same_v<T, common_peg_atomic_parser>) {
            return "Atomic(" + dump_impl(p.child, visited) + ")";
+        } else if constexpr (std::is_same_v<T, common_peg_gbnf_parser>) {
+            return "Gbnf(" + p.grammar + ", " + dump_impl(p.child, visited) + ")";
        } else if constexpr (std::is_same_v<T, common_peg_any_parser>) {
            return "Any";
        } else if constexpr (std::is_same_v<T, common_peg_space_parser>) {
@@ -1565,6 +1572,7 @@ static std::unordered_set<std::string> collect_reachable_rules(
                                 std::is_same_v<T, common_peg_not_parser> ||
                                 std::is_same_v<T, common_peg_tag_parser> ||
                                 std::is_same_v<T, common_peg_atomic_parser> ||
+                                 std::is_same_v<T, common_peg_gbnf_parser> ||
                                 std::is_same_v<T, common_peg_schema_parser>) {
                visit(p.child);
            } else if constexpr (std::is_same_v<T, common_peg_rule_parser>) {
@@ -1651,10 +1659,13 @@ void common_peg_arena::build_grammar(const common_grammar_builder & builder, boo
            } else if constexpr (std::is_same_v<T, common_peg_sequence_parser>) {
                std::string s;
                for (const auto & child : p.children) {
+                    auto child_gbnf = to_gbnf(child);
+                    if (child_gbnf.empty()) {
+                        continue;
+                    }
                    if (!s.empty()) {
                        s += " ";
                    }
-                    auto child_gbnf = to_gbnf(child);
                    const auto & child_parser = effective_parser(child);
                    if (std::holds_alternative<common_peg_choice_parser>(child_parser) ||
                        std::holds_alternative<common_peg_sequence_parser>(child_parser)) {
@@ -1754,6 +1765,8 @@ void common_peg_arena::build_grammar(const common_grammar_builder & builder, boo
                return to_gbnf(p.child);
            } else if constexpr (std::is_same_v<T, common_peg_atomic_parser>) {
                return to_gbnf(p.child);
+            } else if constexpr (std::is_same_v<T, common_peg_gbnf_parser>) {
+                return p.grammar;
            } else {
                static_assert(is_always_false_v<T>);
            }
@@ -1888,6 +1901,8 @@ static nlohmann::json serialize_parser_variant(const common_peg_parser_variant &
                {"child", p.child},
                {"tag", p.tag}
            };
+        } else if constexpr (std::is_same_v<T, common_peg_gbnf_parser>) {
+            return json{{"type", "gbnf"}, {"child", p.child}, {"grammar", p.grammar}};
        }
    }, variant);
 }
@@ -2050,6 +2065,16 @@ static common_peg_parser_variant deserialize_parser_variant(const nlohmann::json
        };
    }

+    if (type == "gbnf") {
+        if (!j.contains("child") || !j.contains("grammar")) {
+            throw std::runtime_error("gbnf parser missing required fields");
+        }
+        return common_peg_gbnf_parser{
+            j["child"].get<common_peg_parser_id>(),
+            j["grammar"].get<std::string>(),
+        };
+    }
+
    throw std::runtime_error("Unknown parser type: " + type);
 }

--- a/common/peg-parser.h
+++ b/common/peg-parser.h
@@ -270,6 +270,11 @@ struct common_peg_tag_parser {
    std::string tag;
 };

+struct common_peg_gbnf_parser {
+    common_peg_parser_id child;
+    std::string grammar;
+};
+
 // Variant holding all parser types
 using common_peg_parser_variant = std::variant<
    common_peg_epsilon_parser,
@@ -290,7 +295,8 @@ using common_peg_parser_variant = std::variant<
    common_peg_rule_parser,
    common_peg_ref_parser,
    common_peg_atomic_parser,
-    common_peg_tag_parser
+    common_peg_tag_parser,
+    common_peg_gbnf_parser
 >;

 class common_peg_arena {
@@ -504,6 +510,10 @@ class common_peg_parser_builder {
    // Unlike rules, you can tag multiple nodes with the same tag.
    common_peg_parser tag(const std::string & tag, const common_peg_parser & p) { return add(common_peg_tag_parser{p.id(), tag}); }

+    // Wraps a child parser but emits a custom GBNF grammar string instead of
+    // the child's grammar. Parsing delegates entirely to the child.
+    common_peg_parser gbnf(const common_peg_parser & p, const std::string & grammar) { return add(common_peg_gbnf_parser{p, grammar}); }
+
    void set_root(const common_peg_parser & p);

    common_peg_arena build();
--- a/common/preset.cpp
+++ b/common/preset.cpp
@@ -43,7 +43,7 @@ static std::set<std::string> get_remote_preset_whitelist(const std::map<std::str
    for (const auto & it : key_to_opt) {
        const std::string & key = it.first;
        const common_arg & opt = it.second;
-        if (allowed_options.find(key) != allowed_options.end() || opt.is_sparam) {
+        if (allowed_options.find(key) != allowed_options.end() || opt.is_sampling) {
            allowed_keys.insert(key);
            // also add variant keys (args without leading dashes and env vars)
            for (const auto & arg : opt.get_args()) {
--- a/common/sampling.cpp
+++ b/common/sampling.cpp
@@ -1,10 +1,12 @@
 #include "sampling.h"

 #include "common.h"
-#include "ggml.h"
+#include "fit.h"
 #include "log.h"
 #include "reasoning-budget.h"

+#include "ggml.h"
+
 #include <algorithm>
 #include <cctype>
 #include <climits>
@@ -287,8 +289,8 @@ struct common_sampler * common_sampler_init(const struct llama_model * model, st
        }
    }

-    // reasoning budget sampler
-    if (!params.reasoning_budget_start.empty() && !params.reasoning_budget_end.empty()) {
+    // reasoning budget sampler (skip when budget is unlimited unless a lazy grammar is active, which needs rbudget for thinking-block suppression)
+    if (!params.reasoning_budget_start.empty() && !params.reasoning_budget_end.empty() && (params.grammar_lazy || params.reasoning_budget_tokens >= 0)) {
        rbudget = common_reasoning_budget_init(
            vocab,
            params.reasoning_budget_start,
@@ -511,7 +513,7 @@ void common_perf_print(const struct llama_context * ctx, const struct common_sam
        LOG_INF("%s: unaccounted time = %10.2f ms / %5.1f %%      (total - sampling - prompt eval - eval) / (total)\n", __func__, t_unacc_ms, t_unacc_pc);
        LOG_INF("%s:    graphs reused = %10d\n", __func__, data.n_reused);

-        llama_memory_breakdown_print(ctx);
+        common_memory_breakdown_print(ctx);
    }
 }

--- a/common/speculative.cpp
+++ b/common/speculative.cpp
@@ -13,6 +13,7 @@
 #include <cstring>
 #include <iomanip>
 #include <map>
+#include <cinttypes>

 #define SPEC_VOCAB_MAX_SIZE_DIFFERENCE  128
 #define SPEC_VOCAB_CHECK_START_TOKEN_ID 5
@@ -60,18 +61,26 @@ static bool common_speculative_are_compatible(
    LOG_DBG("%s: vocab_type dft: %d\n", __func__, vocab_type_dft);

    if (vocab_type_tgt != vocab_type_dft) {
-        LOG_DBG("%s: draft model vocab type must match target model to use speculation but ", __func__);
-        LOG_DBG("vocab_type_dft = %d while vocab_type_tgt = %d\n", vocab_type_dft, vocab_type_tgt);
+        LOG_WRN("%s: draft model vocab type must match target model to use speculation but "
+                "vocab_type_dft = %d while vocab_type_tgt = %d\n", __func__, vocab_type_dft, vocab_type_tgt);
        return false;
    }

-    if (
-        llama_vocab_get_add_bos(vocab_tgt) != llama_vocab_get_add_bos(vocab_dft) ||
-        llama_vocab_get_add_eos(vocab_tgt) != llama_vocab_get_add_eos(vocab_dft) ||
-        llama_vocab_bos(vocab_tgt) != llama_vocab_bos(vocab_dft) ||
-        llama_vocab_eos(vocab_tgt) != llama_vocab_eos(vocab_dft)
-    ) {
-        LOG_DBG("%s: draft model special tokens must match target model to use speculation\n", __func__);
+    if (llama_vocab_get_add_bos(vocab_tgt) != llama_vocab_get_add_bos(vocab_dft) ||
+        (llama_vocab_get_add_bos(vocab_tgt) && llama_vocab_bos(vocab_tgt) != llama_vocab_bos(vocab_dft))) {
+        LOG_WRN("%s: draft model bos tokens must match target model to use speculation. add: %d - %d, id: %d - %d)\n",
+                __func__,
+                llama_vocab_get_add_bos(vocab_tgt), llama_vocab_get_add_bos(vocab_dft),
+                llama_vocab_bos(vocab_tgt), llama_vocab_bos(vocab_dft));
+        return false;
+    }
+
+    if (llama_vocab_get_add_eos(vocab_tgt) != llama_vocab_get_add_eos(vocab_dft) ||
+        (llama_vocab_get_add_eos(vocab_tgt) && llama_vocab_eos(vocab_tgt) != llama_vocab_eos(vocab_dft))) {
+        LOG_WRN("%s: draft model eos tokens must match target model to use speculation. add: %d - %d, id: %d - %d)\n",
+                __func__,
+                llama_vocab_get_add_eos(vocab_tgt), llama_vocab_get_add_eos(vocab_dft),
+                llama_vocab_eos(vocab_tgt), llama_vocab_eos(vocab_dft));
        return false;
    }

@@ -142,12 +151,33 @@ struct common_speculative_state {
            llama_tokens & result) = 0;

    virtual void accept(uint16_t n_accepted) = 0;
+
+    virtual int32_t n_max(const common_params_speculative & params) const = 0;
+    virtual int32_t n_min(const common_params_speculative & params) const = 0;
+};
+
+struct common_speculative_checkpoint {
+    llama_pos pos_min  = 0;
+    llama_pos pos_max  = 0;
+
+    int64_t   n_tokens = 0;
+
+    std::vector<uint8_t> data;
+
+    size_t size() const {
+        return data.size();
+    }
+
+    size_t ckpt_size   = 0;
 };

 struct common_speculative_state_draft : public common_speculative_state {
    llama_context * ctx_tgt; // only used for retokenizing from ctx_dft
    llama_context * ctx_dft;

+    bool use_ckpt = false;
+    struct common_speculative_checkpoint ckpt;
+
    common_sampler * smpl;

    llama_batch  batch;
@@ -160,10 +190,12 @@ struct common_speculative_state_draft : public common_speculative_state {
            enum common_speculative_type type,
            llama_context * ctx_tgt,
            llama_context * ctx_dft,
-            const std::vector<std::pair<std::string, std::string>> & replacements)
+            const std::vector<std::pair<std::string, std::string>> & replacements,
+            bool use_ckpt)
        : common_speculative_state(type)
        , ctx_tgt(ctx_tgt)
        , ctx_dft(ctx_dft)
+        , use_ckpt(use_ckpt)
    {
        batch = llama_batch_init(llama_n_batch(ctx_dft), 0, 1);
        smpl = nullptr;
@@ -218,7 +250,48 @@ struct common_speculative_state_draft : public common_speculative_state {
    }

    void begin(const llama_tokens & prompt) override {
-        GGML_UNUSED(prompt);
+        if (use_ckpt && ckpt.size() > 0) {
+            // delete checkpoint
+            LOG_DBG("%s: delete checkpoint, prompt.size=%zu, pos_min=%d, pos_max=%d, n_tokens=%" PRId64 ", size=%.3f MiB\n",
+                    __func__, prompt.size(), ckpt.pos_min, ckpt.pos_max, ckpt.n_tokens, (float) ckpt.data.size() / 1024 / 1024);
+            ckpt.pos_min   = 0;
+            ckpt.pos_max   = 0;
+            ckpt.n_tokens  = 0;
+            ckpt.ckpt_size = 0;
+            ckpt.data.clear();
+        }
+    }
+
+    size_t draft_create_checkpoint(int n_tokens_prompt, int n_tokens_batch) {
+        int slot_id = 0;
+        const size_t checkpoint_size = llama_state_seq_get_size_ext(ctx_dft, slot_id, LLAMA_STATE_SEQ_FLAGS_PARTIAL_ONLY);
+
+        ckpt.pos_min  = llama_memory_seq_pos_min(llama_get_memory(ctx_dft), slot_id);
+        ckpt.pos_max  = llama_memory_seq_pos_max(llama_get_memory(ctx_dft), slot_id);
+        ckpt.n_tokens = n_tokens_prompt - n_tokens_batch;
+        ckpt.data.resize(checkpoint_size);
+
+        const size_t n = llama_state_seq_get_data_ext(ctx_dft, ckpt.data.data(), checkpoint_size, slot_id, LLAMA_STATE_SEQ_FLAGS_PARTIAL_ONLY);
+        if (n != checkpoint_size) {
+            GGML_ABORT("checkpoint size mismatch: expected %zu, got %zu\n", checkpoint_size, n);
+        }
+
+        LOG_DBG("%s: pos_min = %d, pos_max = %d, size = %.3f MiB\n", __func__,
+                ckpt.pos_min, ckpt.pos_max, (float) ckpt.data.size() / 1024 / 1024);
+        return n;
+    }
+
+    size_t draft_restore_checkpoint(size_t ckpt_size_part_expected) {
+        int slot_id = 0;
+        LOG_DBG("%s: pos_min = %d, pos_max = %d\n", __func__, ckpt.pos_min, ckpt.pos_max);
+        const size_t n = llama_state_seq_set_data_ext(ctx_dft, ckpt.data.data(), ckpt.size(), slot_id, LLAMA_STATE_SEQ_FLAGS_PARTIAL_ONLY);
+        if (n != ckpt_size_part_expected) {
+            GGML_ABORT("%s: failed to restore context checkpoint (pos_min=%d, pos_max=%d, size=%zu, get_data_ext->%zu, set_data_ext->%zu",
+                        __func__, ckpt.pos_min, ckpt.pos_max, ckpt.size(), ckpt_size_part_expected, n);
+        }
+        llama_memory_seq_rm(llama_get_memory(ctx_dft), slot_id, ckpt.pos_max + 1, -1);
+
+        return n;
    }

    void draft(
@@ -226,6 +299,8 @@ struct common_speculative_state_draft : public common_speculative_state {
            const llama_tokens & prompt_tgt,
            llama_token id_last,
            llama_tokens & result) override {
+        const auto & sparams = params.draft;
+
        auto * spec = this;

        auto & batch      = spec->batch;
@@ -236,10 +311,10 @@ struct common_speculative_state_draft : public common_speculative_state {

        auto * mem_dft = llama_get_memory(ctx_dft);

-        int reuse_i = 0;
-        int reuse_n = 0;
+        int reuse_i = 0; // index of part to be reused in prompt_dft
+        int reuse_n = 0; // length of part to be reused in prompt_dft

-        const int n_ctx = llama_n_ctx(ctx_dft) - params.n_max;
+        const int n_ctx = llama_n_ctx(ctx_dft) - sparams.n_max;

        llama_tokens prompt_cnv;
        if (!spec->vocab_cmpt) {
@@ -287,22 +362,30 @@ struct common_speculative_state_draft : public common_speculative_state {
            }
        }

-        LOG_DBG("%s: reuse_i = %d, reuse_n = %d, prompt = %d\n", __func__, reuse_i, reuse_n, (int) prompt_dft.size());
+        LOG_DBG("%s: reuse_i = %d, reuse_n = %d, #prompt_dft = %zu, #prompt_cur = %zu\n",
+                __func__, reuse_i, reuse_n, prompt_dft.size(), prompt_cur.size());
+        if (use_ckpt && ckpt.ckpt_size == 0 && reuse_n > 0) {
+            LOG_DBG("%s: no checkpoint available, no reuse, (reuse_i=%d, reuse_n=%d) -> (0, 0)\n",
+                    __func__, reuse_i, reuse_n);
+            reuse_i = 0;
+            reuse_n = 0;
+        }

        result.clear();
-        result.reserve(params.n_max);
+        result.reserve(sparams.n_max);

-        if (reuse_n == 0) {
+        bool needs_ckpt = use_ckpt && prompt_dft.size() > 0;
+        if (reuse_n == 0 || (use_ckpt && reuse_i > 0)) {
            llama_memory_clear(mem_dft, false);
            prompt_dft.clear();
        } else {
            // this happens when a previous draft has been discarded (for example, due to being too small), but the
            // target model agreed with it. in this case, we simply pass back the previous results to save compute
-            if (reuse_i + reuse_n < (int) prompt_dft.size() && prompt_dft[reuse_i + reuse_n] == id_last) {
+            if (reuse_i + reuse_n < (int64_t) prompt_dft.size() && prompt_dft[reuse_i + reuse_n] == id_last) {
                for (int i = reuse_i + reuse_n + 1; i < (int) prompt_dft.size(); ++i) {
                    result.push_back(prompt_dft[i]);

-                    if (params.n_max <= (int) result.size()) {
+                    if (sparams.n_max <= (int) result.size()) {
                        break;
                    }
                }
@@ -310,19 +393,50 @@ struct common_speculative_state_draft : public common_speculative_state {
                return;
            }

+            bool do_restore = false;
+            if (prompt_dft.size() > prompt_cur.size() && reuse_i + reuse_n < (int64_t) prompt_dft.size()) {
+                // This can happen after a partial acceptance (speculative decoding with checkpoints)
+                LOG_DBG("%s: #prompt_dft=%zu, #prompt_cur=%zu, shorten draft\n",
+                        __func__, prompt_dft.size(), prompt_cur.size());
+                prompt_dft.resize(prompt_cur.size());
+                do_restore = true;
+            }
+
            if (reuse_i > 0) {
-                llama_memory_seq_rm (mem_dft, 0, 0, reuse_i);
+                bool is_removed = llama_memory_seq_rm (mem_dft, 0, 0, reuse_i);
+                if (!is_removed) {
+                    LOG_ERR("%s: llama_memory_seq_rm failed, reuse_i=%d\n", __func__, reuse_i);
+                }
                llama_memory_seq_add(mem_dft, 0, reuse_i, -1, -reuse_i);

                prompt_dft.erase(prompt_dft.begin(), prompt_dft.begin() + reuse_i);
            }

-            if (reuse_n < (int) prompt_dft.size()) {
-                llama_memory_seq_rm (mem_dft, 0, reuse_n, -1);
-                prompt_dft.erase(prompt_dft.begin() + reuse_n, prompt_dft.end());
+            if (reuse_n < (int) prompt_dft.size() || do_restore) {
+                if (use_ckpt) {
+                    if (ckpt.n_tokens > (int64_t) prompt_dft.size()) {
+                        LOG_INF("%s: checkpoint is too large, prompt_tgt.size=%zu, ckpt.n_tokens=%" PRId64 ", reuse_n=%d, prompt_dft.size=%zu\n",
+                                __func__, prompt_tgt.size(), ckpt.n_tokens, reuse_n, prompt_dft.size());
+                    }
+                    draft_restore_checkpoint(ckpt.ckpt_size);
+                    reuse_n = ckpt.n_tokens;
+                    prompt_dft.resize(reuse_n);
+                    needs_ckpt = false;
+                } else {
+                    bool is_removed = llama_memory_seq_rm (mem_dft, 0, reuse_n, -1);
+                    if (!is_removed) {
+                        LOG_ERR("%s: llama_memory_seq_rm failed, reuse_n=%d, prompt_dft.size=%zu\n",
+                                __func__, reuse_n, prompt_dft.size());
+                    }
+                    prompt_dft.erase(prompt_dft.begin() + reuse_n, prompt_dft.end());
+                }
            }
        }

+        if (needs_ckpt) {
+            ckpt.ckpt_size = draft_create_checkpoint(prompt_dft.size(), batch.n_tokens);
+        }
+
        // prepare a batch to evaluate any new tokens in the prompt
        common_batch_clear(batch);

@@ -337,7 +451,11 @@ struct common_speculative_state_draft : public common_speculative_state {
        if (batch.n_tokens > 0) {
            //LOG_DBG("%s: draft prompt batch: %s\n", __func__, string_from(ctx, batch).c_str());

-            llama_decode(ctx_dft, batch);
+            int ret = llama_decode(ctx_dft, batch);
+            if (ret != 0 && ret != 1) {
+                LOG_WRN("%s: llama_decode returned %d, prompt_cur.size=%zu\n",
+                        __func__, ret, prompt_cur.size());
+            }
        }

        const llama_pos n_past = prompt_dft.size();
@@ -351,12 +469,16 @@ struct common_speculative_state_draft : public common_speculative_state {

        LOG_DBG("%s: draft prompt: %s\n", __func__, string_from(ctx_dft, prompt_dft).c_str());

-        llama_decode(ctx_dft, batch);
+        int ret = llama_decode(ctx_dft, batch);
+        if (ret != 0 && ret != 1) {
+            LOG_WRN("%s: llama_decode returned %d, prompt_cur.size=%zu, prompt_dft.size=%zu\n",
+                    __func__, ret, prompt_cur.size(), prompt_dft.size());
+        }

        common_sampler_reset(smpl);

        // sample n_draft tokens from the draft model
-        for (int i = 0; i < params.n_max; ++i) {
+        for (int i = 0; i < sparams.n_max; ++i) {
            common_batch_clear(batch);

            common_sampler_sample(smpl, ctx_dft, 0, true);
@@ -375,19 +497,23 @@ struct common_speculative_state_draft : public common_speculative_state {

            result.push_back(id);

-            if (params.n_max <= (int) result.size()) {
+            if (sparams.n_max <= (int) result.size()) {
                break;
            }

            // only collect very high-confidence draft tokens
-            if (cur_p->data[0].p < params.p_min) {
+            if (cur_p->data[0].p < sparams.p_min) {
                break;
            }

            common_batch_add(batch, id, n_past + i + 1, { 0 }, true);

            // evaluate the drafted tokens on the draft model
-            llama_decode(ctx_dft, batch);
+            ret = llama_decode(ctx_dft, batch);
+            if (ret != 0) {
+                LOG_WRN("%s: llama_decode[%d] returned %d, prompt_cur.size=%zu, prompt_dft.size=%zu\n",
+                        __func__, i, ret, prompt_cur.size(), prompt_dft.size());
+            }

            prompt_dft.push_back(id);
        }
@@ -397,10 +523,14 @@ struct common_speculative_state_draft : public common_speculative_state {
            detokenized = replace_to_tgt(detokenized);
            LOG_DBG("draft->main detokenized string: '%s'\n", detokenized.c_str());
            result = common_tokenize(ctx_tgt, detokenized, false, true);
-            if (result.size() > (size_t)params.n_max) {
-                result.resize(params.n_max);
+            if (result.size() > (size_t) sparams.n_max) {
+                result.resize(sparams.n_max);
            }
        }
+
+        if (result.size() < (size_t) sparams.n_min) {
+            result.clear();
+        }
    }

    void accept(uint16_t n_accepted) override {
@@ -408,6 +538,14 @@ struct common_speculative_state_draft : public common_speculative_state {
        GGML_UNUSED(n_accepted);
    }

+    int32_t n_max(const common_params_speculative & params) const override {
+        return params.draft.n_max;
+    }
+
+    int32_t n_min(const common_params_speculative & params) const override {
+        return params.draft.n_min;
+    }
+
    std::string replace_to_dft(const std::string & input) const {
        std::string result = input;

@@ -460,6 +598,14 @@ struct common_speculative_state_eagle3 : public common_speculative_state {
        // noop
        GGML_UNUSED(n_accepted);
    }
+
+    int32_t n_max(const common_params_speculative & params) const override {
+        return params.draft.n_max;
+    }
+
+    int32_t n_min(const common_params_speculative & params) const override {
+        return params.draft.n_min;
+    }
 };

 // state of self-speculation (simple implementation, not ngram-map)
@@ -489,19 +635,27 @@ struct common_speculative_state_ngram_simple : public common_speculative_state {
        // noop
        GGML_UNUSED(n_accepted);
    }
+
+    int32_t n_max(const common_params_speculative & /*params*/) const override {
+        return config.size_mgram;
+    }
+
+    int32_t n_min(const common_params_speculative & /*params*/) const override {
+        return config.size_mgram;
+    }
 };

 struct common_speculative_state_ngram_map_k : public common_speculative_state {
    // draft ngram map for speculative decoding without draft model
-    common_ngram_map map;
+    common_ngram_map config;

    common_speculative_state_ngram_map_k(
            enum common_speculative_type type,
-            common_ngram_map map)
-        : common_speculative_state(type), map(std::move(map)) {}
+            common_ngram_map config)
+        : common_speculative_state(type), config(std::move(config)) {}

    void begin(const llama_tokens & prompt) override {
-        common_ngram_map_begin(map, prompt);
+        common_ngram_map_begin(config, prompt);
    }

    void draft(
@@ -509,12 +663,20 @@ struct common_speculative_state_ngram_map_k : public common_speculative_state {
            const llama_tokens & prompt_tgt,
            llama_token id_last,
            llama_tokens & result) override {
-        common_ngram_map_draft(map, prompt_tgt, id_last, result);
+        common_ngram_map_draft(config, prompt_tgt, id_last, result);
        GGML_UNUSED(params);
    }

    void accept(uint16_t n_accepted) override {
-        common_ngram_map_accept(map, n_accepted);
+        common_ngram_map_accept(config, n_accepted);
+    }
+
+    int32_t n_max(const common_params_speculative & /*params*/) const override {
+        return config.size_value;
+    }
+
+    int32_t n_min(const common_params_speculative & /*params*/) const override {
+        return config.size_value;
    }
 };

@@ -571,7 +733,7 @@ struct common_speculative_state_ngram_mod : public common_speculative_state {
            const llama_tokens & prompt_tgt,
            llama_token id_last,
            llama_tokens & result) override {
-        GGML_UNUSED(params);
+        const auto & sparams = params.ngram_mod;

        n_draft_last = 0;

@@ -591,16 +753,16 @@ struct common_speculative_state_ngram_mod : public common_speculative_state {
            i_last = cur_len - n;
        }

-        result.resize(n + params.n_max);
+        result.resize(n + sparams.n_max);
        for (size_t i = 0; i < n - 1; ++i) {
            result[i] = prompt_tgt[cur_len - n + 1 + i];
        }
        result[n - 1] = id_last;

-        for (int i = 0; i < params.n_max; ++i) {
+        for (int i = 0; i < sparams.n_max; ++i) {
            const llama_token token = mod.get(result.data() + i);
            if (token == common_ngram_mod::EMPTY) {
-                if (i < params.n_min) {
+                if (i < sparams.n_min) {
                    result.clear();
                    return;
                }
@@ -636,12 +798,21 @@ struct common_speculative_state_ngram_mod : public common_speculative_state {

                    mod.reset();
                    n_low = 0;
+                    i_last = 0;
                }
            } else {
                n_low = 0;
            }
        }
    }
+
+    int32_t n_max(const common_params_speculative & params) const override {
+        return params.ngram_mod.n_max;
+    }
+
+    int32_t n_min(const common_params_speculative & params) const override {
+        return params.ngram_mod.n_min;
+    }
 };

 struct common_speculative_state_ngram_cache : public common_speculative_state {
@@ -735,18 +906,29 @@ struct common_speculative_state_ngram_cache : public common_speculative_state {
        // TODO: noop
        GGML_UNUSED(n_accepted);
    }
+
+    int32_t n_max(const common_params_speculative & /*params*/) const override {
+        return n_draft;
+    }
+
+    int32_t n_min(const common_params_speculative & /*params*/) const override {
+        return 0;
+    }
 };

 struct common_speculative {
    std::vector<std::unique_ptr<common_speculative_state>> impls; // list of implementations to use and their states
+
    common_speculative_state * curr_impl = nullptr; // current implementation in use (for stats)
 };

-static common_ngram_map get_common_ngram_map(const common_speculative_config & config) {
-    uint16_t size_key   = config.params.ngram_size_n;
-    uint16_t size_value = config.params.ngram_size_m;
-    bool     key_only   = (config.type == COMMON_SPECULATIVE_TYPE_NGRAM_MAP_K);
-    uint16_t min_hits   = config.params.ngram_min_hits;
+static common_ngram_map get_common_ngram_map(
+        common_speculative_type type,
+        const common_params_speculative_ngram_map & config) {
+    uint16_t size_key   = config.size_n;
+    uint16_t size_value = config.size_m;
+    bool     key_only   = type == COMMON_SPECULATIVE_TYPE_NGRAM_MAP_K;
+    uint16_t min_hits   = config.min_hits;

    return common_ngram_map(size_key, size_value, key_only, min_hits);
 }
@@ -798,50 +980,14 @@ enum common_speculative_type common_speculative_type_from_name(const std::string
    return it->second;
 }

-bool common_speculative_is_compat(llama_context * ctx_tgt) {
-    auto * mem = llama_get_memory(ctx_tgt);
-    if (mem == nullptr) {
-        return false;
-    }
-
-    bool res = true;
-
-    llama_memory_clear(mem, true);
-
-    // eval 2 tokens to check if the context is compatible
-    std::vector<llama_token> tmp;
-    tmp.push_back(0);
-    tmp.push_back(0);
-
-    int ret = llama_decode(ctx_tgt, llama_batch_get_one(tmp.data(), tmp.size()));
-    if (ret != 0) {
-        LOG_ERR("%s: llama_decode() failed: %d\n", __func__, ret);
-        res = false;
-        goto done;
-    }
-
-    // try to remove the last tokens
-    if (!llama_memory_seq_rm(mem, 0, 1, -1)) {
-        LOG_WRN("%s: the target context does not support partial sequence removal\n", __func__);
-        res = false;
-        goto done;
-    }
-
-done:
-    llama_memory_clear(mem, true);
-    llama_synchronize(ctx_tgt);
-
-    return res;
-}
-
 // initialization of the speculative decoding system
 //
 common_speculative * common_speculative_init(
        common_params_speculative & params,
        llama_context             * ctx_tgt) {
    llama_context * ctx_dft = nullptr;
-    if (params.model_dft) {
-        ctx_dft = llama_init_from_model(params.model_dft, params.cparams_dft);
+    if (params.draft.model) {
+        ctx_dft = llama_init_from_model(params.draft.model, params.draft.cparams);
        if (ctx_dft == nullptr) {
            LOG_ERR("%s", "failed to create draft context\n");
            return nullptr;
@@ -851,7 +997,7 @@ common_speculative * common_speculative_init(
    // Compute the implementations to use based on the config and their order of preference
    std::vector<common_speculative_config> configs = {}; // list of speculative configs to try
    {
-        bool has_draft = !params.mparams_dft.path.empty();
+        bool has_draft = !params.draft.mparams.path.empty();
        bool has_draft_eagle3 = false; // TODO PR-18039: if params.speculative.eagle3

        bool has_ngram_cache   = (params.type == COMMON_SPECULATIVE_TYPE_NGRAM_CACHE);
@@ -874,16 +1020,17 @@ common_speculative * common_speculative_init(
            configs.push_back(common_speculative_config(COMMON_SPECULATIVE_TYPE_NGRAM_MAP_K4V, params));
        }
        if (has_ngram_mod) {
-            // shared instance for all speculative decoding contexts
-            if (!params.ngram_mod) {
-                params.ngram_mod = std::make_shared<common_ngram_mod>(params.ngram_size_n, 4*1024*1024);
+            auto & sparams = params.ngram_mod;

-                LOG_INF("%s: initialized ngram_mod with n=%d, size=%zu (%.3f MB)\n", __func__,
-                        params.ngram_size_n, params.ngram_mod->size(),
-                        (float)(params.ngram_mod->size_bytes())/1024/1024);
+            if (!sparams.obj) {
+                sparams.obj = std::make_shared<common_ngram_mod>(sparams.n_match, 4*1024*1024);

-                if (params.ngram_size_n < 16) {
-                    LOG_WRN("%s: ngram_mod n=%d is too small - poor quality is possible, see: https://github.com/ggml-org/llama.cpp/pull/19164\n", __func__, params.ngram_size_n);
+                LOG_INF("%s: initialized ngram_mod with n_match=%d, size=%zu (%.3f MB)\n", __func__,
+                        sparams.n_match, sparams.obj->size(), (float)(sparams.obj->size_bytes())/1024/1024);
+
+                if (sparams.n_match < 16) {
+                    LOG_WRN("%s: ngram_mod n_match=%d is too small - poor quality is possible, "
+                            "see: https://github.com/ggml-org/llama.cpp/pull/19164\n", __func__, sparams.n_match);
                }
            }

@@ -908,10 +1055,13 @@ common_speculative * common_speculative_init(
            case COMMON_SPECULATIVE_TYPE_NONE:
                break;
            case COMMON_SPECULATIVE_TYPE_DRAFT: {
+                const bool use_ckpt = common_context_can_seq_rm(ctx_dft) == COMMON_CONTEXT_SEQ_RM_TYPE_FULL;
+
                impls.push_back(std::make_unique<common_speculative_state_draft>(config.type,
                    /* .ctx_tgt      = */ ctx_tgt,
                    /* .ctx_dft      = */ ctx_dft,
-                    /* .replacements = */ params.replacements
+                    /* .replacements = */ params.draft.replacements,
+                    /* .use_ckpt     = */ use_ckpt
                ));
                break;
            }
@@ -920,18 +1070,18 @@ common_speculative * common_speculative_init(
                break;
            }
            case COMMON_SPECULATIVE_TYPE_NGRAM_SIMPLE: {
-                common_ngram_map ngram_map = get_common_ngram_map(config);
+                common_ngram_map ngram_map = get_common_ngram_map(config.type, config.params.ngram_simple);

                uint16_t ngram_size_key   = ngram_map.size_key;
                uint16_t mgram_size_value = ngram_map.size_value;

                auto config_simple = common_ngram_simple_config {
-                    /* .size_ngram      = */ ngram_size_key,
-                    /* .size_mgram      = */ mgram_size_value
+                    /* .size_ngram = */ ngram_size_key,
+                    /* .size_mgram = */ mgram_size_value
                };
                auto state = std::make_unique<common_speculative_state_ngram_simple>(
-                    /* .type            = */ config.type,
-                    /* .state           = */ config_simple
+                    /* .type  = */ config.type,
+                    /* .state = */ config_simple
                );
                impls.push_back(std::move(state));
                break;
@@ -940,18 +1090,17 @@ common_speculative * common_speculative_init(
            case COMMON_SPECULATIVE_TYPE_NGRAM_MAP_K4V: {
                impls.push_back(std::make_unique<common_speculative_state_ngram_map_k>(
                    (config.type),
-                    get_common_ngram_map(config)
+                    get_common_ngram_map(config.type, config.params.ngram_map_k)
                ));
                break;
            }
            case COMMON_SPECULATIVE_TYPE_NGRAM_MOD: {
-                GGML_ASSERT(config.params.ngram_mod);
-                impls.push_back(std::make_unique<common_speculative_state_ngram_mod>(config.type, *config.params.ngram_mod));
+                GGML_ASSERT(config.params.ngram_mod.obj);
+                impls.push_back(std::make_unique<common_speculative_state_ngram_mod>(config.type, *config.params.ngram_mod.obj));
                break;
            }
            case COMMON_SPECULATIVE_TYPE_NGRAM_CACHE: {
-                auto state = create_state_ngram_cache(
-                        params.lookup_cache_static, params.lookup_cache_dynamic, config);
+                auto state = create_state_ngram_cache(params.ngram_cache.lookup_cache_static, params.ngram_cache.lookup_cache_dynamic, config);
                impls.push_back(std::make_unique<common_speculative_state_ngram_cache>(state));
                break;
            }
@@ -966,7 +1115,8 @@ common_speculative * common_speculative_init(
    }

    auto * result = new common_speculative {
-        /* .impls = */ std::move(impls)
+        /* .impls     = */ std::move(impls),
+        /* .curr_impl = */ nullptr,
    };

    return result;
@@ -1008,6 +1158,15 @@ llama_tokens common_speculative_draft(
            impl->n_call_draft++;
        }

+        {
+            const int n_min = impl->n_min(params);
+
+            if (!result.empty() && (int) result.size() < n_min) {
+                LOG_DBG("%s: ignoring small draft: %d < %d\n", __func__, (int) result.size(), n_min);
+                result.clear();
+            }
+        }
+
        if (!result.empty()) {
            LOG_DBG("%s: called impl %s, hist size = %zu, call_count = %zu, gen = %zu\n", __func__,
                    common_speculative_type_to_str(impl.get()->type).c_str(), prompt_tgt.size(),
@@ -1017,7 +1176,7 @@ llama_tokens common_speculative_draft(
            impl->n_gen_drafts++;
            impl->n_gen_tokens += result.size();

-            break; // We have a draft, so break out of the loop and return it.
+            break; // we have a draft, so break out of the loop and return it.
        }
    }

@@ -1045,6 +1204,32 @@ void common_speculative_accept(common_speculative * spec, uint16_t n_accepted) {
    }
 }

+int32_t common_speculative_n_max(const common_speculative * spec, const common_params_speculative & params) {
+    if (spec == nullptr) {
+        return 0;
+    }
+
+    int32_t n_max = 0;
+    for (const auto & impl : spec->impls) {
+        n_max = std::max(n_max, impl->n_max(params));
+    }
+
+    return n_max;
+}
+
+int32_t common_speculative_n_min(const common_speculative * spec, const common_params_speculative & params) {
+    if (spec == nullptr) {
+        return 0;
+    }
+
+    int32_t n_min = 0;
+    for (const auto & impl : spec->impls) {
+        n_min = std::max(n_min, impl->n_min(params));
+    }
+
+    return n_min;
+}
+
 void common_speculative_print_stats(const common_speculative * spec) {
    if (spec == nullptr) {
        return;
--- a/common/speculative.h
+++ b/common/speculative.h
@@ -14,10 +14,6 @@ enum common_speculative_type common_speculative_type_from_name(const std::string
 // convert type to string
 std::string common_speculative_type_to_str(enum common_speculative_type type);

-// check if the llama_context is compatible for speculative decoding
-// note: clears the memory of the context
-bool common_speculative_is_compat(llama_context * ctx_tgt);
-
 common_speculative * common_speculative_init(
        common_params_speculative & params,
        llama_context             * ctx_tgt);
@@ -37,5 +33,14 @@ llama_tokens common_speculative_draft(
 // informs the speculative decoder that n_accepted tokens were accepted by the target model
 void common_speculative_accept(common_speculative * spec, uint16_t n_accepted);

+int32_t common_speculative_n_max(const common_speculative * spec, const common_params_speculative & params);
+int32_t common_speculative_n_min(const common_speculative * spec, const common_params_speculative & params);
+
 // print statistics about the speculative decoding
 void common_speculative_print_stats(const common_speculative * spec);
+
+struct common_speculative_deleter {
+    void operator()(common_speculative * s) { common_speculative_free(s); }
+};
+
+typedef std::unique_ptr<common_speculative, common_speculative_deleter> common_speculative_ptr;
--- a/convert_hf_to_gguf.py
+++ b/convert_hf_to_gguf.py
--- a/convert_hf_to_gguf_update.py
+++ b/convert_hf_to_gguf_update.py
@@ -296,7 +296,7 @@ for model in [*pre_computed_hashes, *all_models]:
        except Exception as e:
            raise OSError(f"Error loading tokenizer for model {name}.") from e

-        chktok = tokenizer.encode(CHK_TXT)
+        chktok = tokenizer.encode(CHK_TXT)  # ty: ignore[unresolved-attribute]
        chkhsh = sha256(str(chktok).encode()).hexdigest()

        logger.info(f"model: {name}")
@@ -468,7 +468,7 @@ for model in models:

    with open(f"models/ggml-vocab-{name}.gguf.out", "w") as f:
        for text in tests:
-            res = tokenizer.encode(text, add_special_tokens=False)
+            res = tokenizer.encode(text, add_special_tokens=False)  # ty: ignore[unresolved-attribute]
            for r in res:
                f.write(f" {r}")
            f.write("\n")
--- a/convert_lora_to_gguf.py
+++ b/convert_lora_to_gguf.py
@@ -402,7 +402,7 @@ if __name__ == '__main__':
                    # the invocation string includes the "<|start_of_turn|>"
                    # token, but the adapters themselves were trained to
                    # activate _after_ that first token, so we drop it here.
-                    alora_invocation_tokens = tokenizer(invocation_string)["input_ids"][1:]
+                    alora_invocation_tokens = tokenizer(invocation_string)["input_ids"][1:]  # ty: ignore[call-non-callable]
                if alora_invocation_tokens:
                    logger.debug("GGUF KV: %s = %s", gguf.Keys.Adapter.ALORA_INVOCATION_TOKENS, alora_invocation_tokens)
                    self.gguf_writer.add_key_value(
--- a/docs/backend/OPENVINO.md
+++ b/docs/backend/OPENVINO.md
@@ -3,7 +3,7 @@
 > [!NOTE]
 > Performance and memory optimizations, accuracy validation, broader quantization coverage, broader operator and model support are work in progress.

-[OpenVINO](https://docs.openvino.ai/) is an open-source toolkit for optimizing and deploying high-performance AI inference, specifically designed for Intel hardware, including CPUs, GPUs, and NPUs, in the cloud, on-premises, and on the edge. [OpenVINO backend for llama.cpp](../../src/ggml-openvino) enables hardware-accelerated inference on **Intel® CPUs, GPUs, and NPUs** while remaining compatible with the existing **GGUF model ecosystem**. The backend translates GGML compute graphs into OpenVINO graphs and leverages graph compilation, kernel fusion, and device-specific optimizations to improve inference performance on supported Intel hardware.
+[OpenVINO](https://docs.openvino.ai/) is an open-source toolkit for optimizing and deploying high-performance AI inference, specifically designed for Intel hardware, including CPUs, GPUs, and NPUs, in the cloud, on-premises, and on the edge. [OpenVINO backend for llama.cpp](../../ggml/src/ggml-openvino) enables hardware-accelerated inference on **Intel® CPUs, GPUs, and NPUs** while remaining compatible with the existing **GGUF model ecosystem**. The backend translates GGML compute graphs into OpenVINO graphs and leverages graph compilation, kernel fusion, and device-specific optimizations to improve inference performance on supported Intel hardware.

 The OpenVINO backend is implemented in `ggml/src/ggml-openvino` and provides a translation layer for core GGML operations. The OpenVINO backend replaces the standard GGML graph execution path with Intel's OpenVINO inference engine. This approach allows the same GGUF model file to run on Intel CPUs, Intel GPUs (integrated and discrete), and Intel NPUs without changes to the model or the rest of the llama.cpp stack. When a `ggml_cgraph` is dispatched to OpenVINO backend, it:

@@ -244,7 +244,6 @@ build\ReleaseOV\bin\llama-cli.exe -m "C:\models\Llama-3.2-1B-Instruct-Q4_0.gguf"
 - `-fa 1` is required when running llama-bench with the OpenVINO backend.
  - `GGML_OPENVINO_STATEFUL_EXECUTION=1 GGML_OPENVINO_DEVICE=GPU ./llama-bench -fa 1`
 - `llama-server` with OpenVINO backend supports only one chat session/thread, when `GGML_OPENVINO_STATEFUL_EXECUTION=1` is enabled.
- For Intel GPU, NPU detection in containers, GPU, NPU user-space drivers/libraries must be present inside the image. We will include in a future PR. Until then, you can use this reference Dockerfile: [openvino.Dockerfile](https://github.com/ravi9/llama.cpp/blob/ov-docker-update/.devops/openvino.Dockerfile)

 > [!NOTE]
 > The OpenVINO backend is actively under development. Fixes are underway, and this document will continue to be updated as issues are resolved.
@@ -274,8 +273,6 @@ docker build --build-arg http_proxy=$http_proxy --build-arg https_proxy=$https_p
 Run llama.cpp with OpenVINO backend Docker container.
 Save sample models in `~/models` as [shown above](#3-download-sample-model). It will be mounted to the container in the examples below.

-> [!NOTE]
-> Intel GPU, NPU detection in containers will be included in a future PR. Until then, you can use this reference Dockerfile: [openvino.Dockerfile](https://github.com/ravi9/llama.cpp/blob/ov-docker-update/.devops/openvino.Dockerfile).

 ```bash
 #  Run Docker container
--- a/docs/backend/SYCL.md
+++ b/docs/backend/SYCL.md
@@ -31,6 +31,8 @@ SYCL cross-platform capabilities enable support for other vendor GPUs as well.

 ## Recommended Release

+### Windows
+
 The following releases are verified and recommended:

 |Commit ID|Tag|Release|Verified  Platform| Update date|
@@ -39,9 +41,22 @@ The following releases are verified and recommended:
 |3bcd40b3c593d14261fb2abfabad3c0fb5b9e318|b4040 |[llama-b4040-bin-win-sycl-x64.zip](https://github.com/ggml-org/llama.cpp/releases/download/b4040/llama-b4040-bin-win-sycl-x64.zip) |Arc A770/Linux/oneAPI 2024.1<br>MTL Arc GPU/Windows 11/oneAPI 2024.1| 2024-11-19|
 |fb76ec31a9914b7761c1727303ab30380fd4f05c|b3038 |[llama-b3038-bin-win-sycl-x64.zip](https://github.com/ggml-org/llama.cpp/releases/download/b3038/llama-b3038-bin-win-sycl-x64.zip) |Arc A770/Linux/oneAPI 2024.1<br>MTL Arc GPU/Windows 11/oneAPI 2024.1||

+### Ubuntu 24.04
+
+The release packages for Ubuntu 24.04 x64 (FP32/FP16) only include the binary files of the llama.cpp SYCL backend. They require the target machine to have pre-installed Intel GPU drivers and oneAPI packages that are the same version as the build package. To get the version and installation info, refer to release.yml: ubuntu-24-sycl -> Download & Install oneAPI.
+
+It is recommended to use them with Intel Docker.
+
+The packages for FP32 and FP16 would have different accuracy and performance on LLMs. Please choose it acording to the test result.

 ## News

+- 2026.04
+
+  - Optimize mul_mat by reorder feature for data type: Q4_K, Q5_K, Q_K, Q8_0.
+  - Fused MoE.
+  - Upgrate CI and built package for oneAPI 2025.3.3, support Ubuntu 24.04 built package.
+
 - 2026.03
  - Support Flash-Attention: less memory usage, performance impact depends on LLM.

@@ -229,6 +244,7 @@ Upon a successful installation, SYCL is enabled for the available intel devices,

 |Verified release|
 |-|
+|2025.3.3 |
 |2025.2.1|
 |2025.1|
 |2024.1|
@@ -339,6 +355,12 @@ Choose one of following methods to run.
 ./examples/sycl/test.sh
 ```

+- Run llama-server:
+
+```sh
+./examples/sycl/start-svr.sh -m PATH/MODEL_FILE
+```
+
 2. Command line
 Launch inference

@@ -627,10 +649,18 @@ Choose one of following methods to run.

 1. Script

+- Run test:
+
 ```
 examples\sycl\win-test.bat
 ```

+- Run llama-server:
+
+```
+examples\sycl\win-start-svr.bat -m PATH\MODEL_FILE
+```
+
 2. Command line

 Launch inference
@@ -689,6 +719,7 @@ use 1 SYCL GPUs: [0] with Max compute units:512
 | GGML_SYCL_F16      | OFF *(default)* \|ON *(optional)*     | Enable FP16 build with SYCL code path. (1.) |
 | GGML_SYCL_GRAPH    | OFF *(default)* \|ON *(Optional)*     | Enable build with [SYCL Graph extension](https://github.com/intel/llvm/blob/sycl/sycl/doc/extensions/experimental/sycl_ext_oneapi_graph.asciidoc). |
 | GGML_SYCL_DNN      | ON *(default)* \|OFF *(Optional)*     | Enable build with oneDNN.                   |
+| GGML_SYCL_HOST_MEM_FALLBACK | ON *(default)* \|OFF *(Optional)* | Allow host memory fallback when device memory is full during quantized weight reorder. Enables inference to continue at reduced speed (reading over PCIe) instead of failing. Requires Linux kernel 6.8+. |
 | CMAKE_C_COMPILER   | `icx` *(Linux)*, `icx/cl` *(Windows)* | Set `icx` compiler for SYCL code path.      |
 | CMAKE_CXX_COMPILER | `icpx` *(Linux)*, `icx` *(Windows)*   | Set `icpx/icx` compiler for SYCL code path. |

--- a/docs/backend/snapdragon/CMakeUserPresets.json
+++ b/docs/backend/snapdragon/CMakeUserPresets.json
@@ -52,10 +52,39 @@
        }
    },

+    {
+        "name": "arm64-linux-snapdragon",
+        "hidden": true,
+        "architecture": { "value": "arm64",       "strategy": "external" },
+        "toolset":      { "value": "host=x86_64", "strategy": "external" },
+        "cacheVariables": {
+            "CMAKE_TOOLCHAIN_FILE": "cmake/arm64-linux-clang.cmake",
+            "CMAKE_C_FLAGS":   "-march=armv8 -fno-finite-math-only -flto -D_GNU_SOURCE",
+            "CMAKE_CXX_FLAGS": "-march=armv8 -fno-finite-math-only -flto -D_GNU_SOURCE",
+            "CMAKE_C_FLAGS_RELEASE":          "-O3 -DNDEBUG",
+            "CMAKE_CXX_FLAGS_RELEASE":        "-O3 -DNDEBUG",
+            "CMAKE_C_FLAGS_RELWITHDEBINFO":   "-O3 -DNDEBUG -g",
+            "CMAKE_CXX_FLAGS_RELWITHDEBINFO": "-O3 -DNDEBUG -g",
+            "CMAKE_PREFIX_PATH":  "$env{OPENCL_SDK_ROOT}",
+            "HEXAGON_SDK_ROOT":   "$env{HEXAGON_SDK_ROOT}",
+            "HEXAGON_TOOLS_ROOT": "$env{HEXAGON_TOOLS_ROOT}",
+            "PREBUILT_LIB_DIR": "linux_aarch64",
+            "GGML_OPENMP":      "OFF",
+            "GGML_LLAMAFILE":   "OFF",
+            "GGML_OPENCL":      "OFF",
+            "GGML_HEXAGON":     "ON",
+            "GGML_HEXAGON_FP32_QUANTIZE_GROUP_SIZE": "128",
+            "LLAMA_OPENSSL":    "OFF"
+        }
+    },
+
    { "name": "arm64-android-snapdragon-debug"  , "inherits": [ "base", "arm64-android-snapdragon", "debug" ] },
    { "name": "arm64-android-snapdragon-release", "inherits": [ "base", "arm64-android-snapdragon", "release" ] },

    { "name": "arm64-windows-snapdragon-debug"  , "inherits": [ "base", "arm64-windows-snapdragon", "debug" ] },
-    { "name": "arm64-windows-snapdragon-release", "inherits": [ "base", "arm64-windows-snapdragon", "release" ] }
+    { "name": "arm64-windows-snapdragon-release", "inherits": [ "base", "arm64-windows-snapdragon", "release" ] },
+
+    { "name": "arm64-linux-snapdragon-debug"  , "inherits": [ "base", "arm64-linux-snapdragon", "debug" ] },
+    { "name": "arm64-linux-snapdragon-release", "inherits": [ "base", "arm64-linux-snapdragon", "release" ] }
  ]
 }
--- a/docs/backend/snapdragon/README.md
+++ b/docs/backend/snapdragon/README.md
@@ -236,10 +236,6 @@ build: 6a8cf8914 (6733)
  Controls whether the Hexagon backend allocates host buffers. By default, all buffers except for REPACK are host buffers.
  This option is required for testing Ops that require REPACK buffers (MUL_MAT and MUL_MAT_ID).

- `GGML_HEXAGON_EXPERIMENTAL=1`
-  Controls whether the Hexagon backend enables experimental features.
-  This option is required for enabling/testing experimental Ops (FLASH_ATTN_EXT).
-
 - `GGML_HEXAGON_VERBOSE=1`
  Enables verbose logging of Ops from the backend. Example output:

@@ -253,17 +249,32 @@ build: 6a8cf8914 (6733)
  ```

 - `GGML_HEXAGON_PROFILE=1`
-  Generates a host-side profile for the ggml-hexagon Ops.
+  Enables Op profiling:

- `GGML_HEXAGON_OPMASK=0x0`
-  Allows enabling specific stages of the processing pipeline:
+  - `1` Basic profile with per-op `usecs` and `cycles` counters
+  - `2` Extended profile with per-op `usecs`, `cycles` and default PMU counter data
+  - `0x1,...,0x8` Extended profile with per-op `usecs`, `cycles` and custom PMU counter data
+
+  The logging output can be either saved into a file for post-processing or it can be piped directly into the post-processing tool to generate the report.
+  Examples:
+
+      `GGML_HEXAGON_PROFILE=1 llama-completion ... |& ./scripts/snapdragon/ggml-hexagon-profile.py -`
+
+- `GGML_HEXAGON_OPSTAGE=0x0`
+  Allows enabling specific stages of the Op processing pipeline:

  - `0x1` Enable Op Queue (i.e., queuing Ops into NPU)
-  - `0x2` Enable Dynamic Quantizer (if needed for the Op)
-  - `0x4` Enable Op Compute (MUL_MAT, etc.)
+  - `0x2` Enable Op Compute (MUL_MAT, etc.)

  Examples:

-      `GGML_HEXAGON_OPMASK=0x1 llama-completion ...` - Ops are enqueued but NPU-side processing is stubbed out
-      `GGML_HEXAGON_OPMASK=0x3 llama-completion ...` - NPU performs dynamic quantization and skips the rest
-      `GGML_HEXAGON_OPMASK=0x7 llama-completion ...` - Full queuing and processing of Ops (default)
+      `GGML_HEXAGON_OPSTAGE=0x1 llama-completion ...` - Ops are enqueued to the NPU but dma & compute are disabled
+      `GGML_HEXAGON_OPSTAGE=0x3 llama-completion ...` - Full queuing and processing of Ops (default)
+
+- `GGML_HEXAGON_OPFILTER=regex`
+  Allows filtering (disabling) Ops that match the regex pattern:
+
+  Examples:
+
+      `GGML_HEXAGON_OPFILTER="FLASH_ATTN_EXT" llama-completion ...` - Disable Flash Attention on Hexagon (falls back to CPU or GPU)
+      `GGML_HEXAGON_OPFILTER="ADD\|SUB" llama-completion ...` - Disable ADD and SUB on Hexagon (fall back to CPU or GPU)
--- a/docs/backend/snapdragon/linux.md
+++ b/docs/backend/snapdragon/linux.md
@@ -0,0 +1,58 @@
+# Snapdragon-based Linux devices
+
+## Docker Setup
+
+The easiest way to build llama.cpp for a Snapdragon-based Linux device is using the toolchain Docker image (see [github.com/snapdragon-toolchain](https://github.com/snapdragon-toolchain)).
+This image includes OpenCL SDK, Hexagon SDK, CMake, and the ARM64 Linux cross-compilation toolchain.
+
+Cross-compilation is supported on **Linux X86** hosts. The resulting binaries are deployed to and run on the target **Qualcomm Snapdragon ARM64 Linux** device.
+
+```
+~/src/llama.cpp$ docker run -it -u $(id -u):$(id -g) --volume $(pwd):/workspace --platform linux/amd64 ghcr.io/snapdragon-toolchain/arm64-linux:v0.1
+[d]/> cd /workspace
+```
+
+Note: The rest of the **Linux** build process assumes that you're running inside the toolchain container.
+
+
+## How to Build
+
+Let's build llama.cpp with CPU, OpenCL, and Hexagon backends via CMake presets:
+
+```
+[d]/workspace> cp docs/backend/snapdragon/CMakeUserPresets.json .
+
+[d]/workspace> cmake --preset arm64-linux-snapdragon-release -B build-snapdragon
+
+[d]/workspace> cmake --build build-snapdragon -j $(nproc)
+```
+
+To generate an installable "package" simply use cmake --install, then zip it:
+
+```
+[d]/workspace> cmake --install build-snapdragon --prefix pkg-snapdragon
+[d]/workspace> zip -r pkg-snapdragon.zip pkg-snapdragon
+```
+
+## How to Install
+
+For this step, you will deploy the built binaries and libraries to the target Linux device. Transfer `pkg-snapdragon.zip` to the target device, then unzip it and set up the environment variables:
+
+```
+$ unzip pkg-snapdragon.zip
+$ cd pkg-snapdragon
+$ export LD_LIBRARY_PATH=./lib
+$ export ADSP_LIBRARY_PATH=./lib
+```
+
+At this point, you should also download some models onto the device:
+
+```
+$ wget https://huggingface.co/bartowski/Llama-3.2-3B-Instruct-GGUF/resolve/main/Llama-3.2-3B-Instruct-Q4_0.gguf
+```
+
+## How to Run
+Next, since we have setup the environment variables, we can run the llama-cli with the Hexagon backends:
+```
+$ ./bin/llama-cli -m Llama-3.2-3B-Instruct-Q4_0.gguf --device HTP0 -ngl 99 -p "what is the most popular cookie in the world?"
+```
--- a/docs/build.md
+++ b/docs/build.md
@@ -281,6 +281,12 @@ Use `GGML_CUDA_FORCE_CUBLAS_COMPUTE_16F` environment variable to force use FP16

 The environment variable `GGML_CUDA_ENABLE_UNIFIED_MEMORY=1` can be used to enable unified memory in Linux. This allows swapping to system RAM instead of crashing when the GPU VRAM is exhausted. In Windows this setting is available in the NVIDIA control panel as `System Memory Fallback`.

+### Peer Access
+
+The environment variable `GGML_CUDA_P2P` can be set to enable peer-to-peer access between multiple GPUs, allowing them to transfer data directly rather than to go through system memory.
+Requires driver support (usually restricted to workstation/datacenter GPUs).
+May cause crashes or corrupted outputs for some motherboards and BIOS settings (e.g. IOMMU).
+
 ### Performance Tuning

 The following compilation options are also available to tweak performance:
@@ -456,7 +462,8 @@ pacman -S git \
    mingw-w64-ucrt-x86_64-gcc \
    mingw-w64-ucrt-x86_64-cmake \
    mingw-w64-ucrt-x86_64-vulkan-devel \
-    mingw-w64-ucrt-x86_64-shaderc
+    mingw-w64-ucrt-x86_64-shaderc \
+    mingw-w64-ucrt-x86_64-spirv-headers
 ```

 Switch into the `llama.cpp` directory and build using CMake.
@@ -490,9 +497,11 @@ First, follow the official LunarG instructions for the installation and setup of

 On Debian / Ubuntu, you can install the required dependencies using:
 ```sh
-sudo apt-get install libvulkan-dev glslc
+sudo apt-get install libvulkan-dev glslc spirv-headers
 ```

+SPIRV-Headers (`spirv/unified1/spirv.hpp`) are required for the Vulkan backend and are **not** always pulled in by the Vulkan loader dev package alone. Other distros use names such as `spirv-headers` (Ubuntu / Debian / Arch), or `spirv-headers-devel` (Fedora / openSUSE). On Windows, the LunarG Vulkan SDK’s `Include` directory already contains these headers.
+
 #### Common steps

 Second, after verifying that you have followed all of the SDK installation/setup steps, use this command to make sure before proceeding:
--- a/docs/development/HOWTO-add-model.md
+++ b/docs/development/HOWTO-add-model.md
@@ -5,6 +5,7 @@ Adding a model requires few steps:
 1. Convert the model to GGUF
 2. Define the model architecture in `llama.cpp`
 3. Build the GGML graph implementation
+4. Optional: Add multimodal encoder implementation

 After following these steps, you can open PR.

@@ -114,6 +115,38 @@ Some `ggml` backends do not support all operations. Backend implementations can

 Note: to debug the inference graph: you can use [llama-eval-callback](/examples/eval-callback/).

+### 4. Optional: Add multimodal encoder implementation
+
+If the new model supports multimodal inputs, you will need to add a new encoder definition in `libmtmd`. You can find more information about llama.cpp's multimodal support in [the docs](../multimodal.md) and in the `tools/mtmd` source directory.
+
+1. In the conversion script, make sure you add a subclass that extends `MmprojModel` or another class that inherits from the same base class.
+2. Add the encoder definition in `clip.cpp`.
+3. Implement the preprocessor in `mtmd.cpp`. In most cases, you can reuse an existing preprocessor.
+4. Implement the encoder GGML graph, either in a dedicated file if the model is truly different from existing ones, or by reusing an existing implementation (for example: siglip, pixtral, or qwen) and adding a model-specific projector.
+
+Note:
+- Many multimodal encoders are based on models that are already supported. Make sure to read the existing encoder definitions in `tools/mtmd/models` before adding a new one. In `libmtmd`, it is generally better to extend an existing model than to duplicate code.
+- To debug the multimodal preprocessor and encoder, you can use [llama-mtmd-debug](tools/mtmd/debug/mtmd-debug.cpp).
+- Adding a model-specific API or CLI is an anti-pattern in `libmtmd`. The goal of `libmtmd` is to provide an easy-to-use, model-agnostic library for multimodal pipeline.
+- In most cases, `llama-mtmd-cli` should not be modified. If a model requires a specific prompt, either let the user provide it or bake it into the Jinja chat template.
+
+## Tips and tricks
+
+### Working with ggml_rope_ext
+
+PyTorch implementations usually prefer explicitly calculating `freq_cis`/`sin`/`cos` components. However, in llama.cpp, most RoPE operations can be handled via `ggml_rope_ext`, which does not require a sin/cos matrix. This saves memory while allowing the GGML RoPE kernel to be fused with other ops.
+
+However, since `ggml_rope_ext` only provides a subset of the RoPE implementations that models use, converting models from PyTorch to llama.cpp may require some creative adaptations.
+
+For more information about `ggml_rope_ext`, please refer to the in-code documentation in `ggml.h`.
+
+Examples:
+- `libmtmd` implements 2D RoPE with `GGML_ROPE_TYPE_NORMAL` ordering by splitting the input tensor in half, applying `ggml_rope_ext` separately to each half, then joining them back together using `ggml_concat`.
+- The [Kimi-K2.5](https://github.com/ggml-org/llama.cpp/pull/19170) vision encoder uses vision RoPE with interleaved frequencies. The weights must be permuted during conversion in order to reuse the `build_rope_2d()` function.
+- [Gemma 4](https://github.com/ggml-org/llama.cpp/pull/21309) uses "proportional" RoPE. We employ a trick where `rope_freqs` is set to a very large value in the last dimensions to prevent those dimensions from being rotated. See the `Gemma4Model` class in `convert_hf_to_gguf.py`.
+- Some models require scaling the input position. For example, `[0, 1, 2, ...]` becomes `[0, 0.5, 1, ...]`. In this case, you can provide the scaling via `freq_scale = 0.5f`.
+- Some models use learned RoPE frequencies instead of relying on `powf(freq_base, -2.0 * i / n_dims)`. In this case, you can provide the learned frequencies via the `rope_freqs` tensor (corresponding to the `c` argument in `ggml_rope_ext`), then set `freq_base = 1.0f`. An important note is that `rope_freqs` in GGML is the **inverse** (`theta = pos[i] / rope_freqs`), so you may need to invert `rope_freqs` during conversion.
+
 ## GGUF specification

 https://github.com/ggml-org/ggml/blob/master/docs/gguf.md
--- a/docs/multimodal.md
+++ b/docs/multimodal.md
@@ -37,6 +37,7 @@ llama-server -hf ggml-org/gemma-3-4b-it-GGUF --no-mmproj-offload
 > - PaddleOCR-VL: https://github.com/ggml-org/llama.cpp/pull/18825
 > - GLM-OCR: https://github.com/ggml-org/llama.cpp/pull/19677
 > - Deepseek-OCR: https://github.com/ggml-org/llama.cpp/pull/17400
+> - Dots.OCR: https://github.com/ggml-org/llama.cpp/pull/17575
 > - HunyuanOCR: https://github.com/ggml-org/llama.cpp/pull/21395

 ## Pre-quantized models
@@ -93,6 +94,11 @@ NOTE: some models may require large context window, for example: `-c 8192`
 # Moondream2 20250414 version
 (tool_name) -hf ggml-org/moondream2-20250414-GGUF

+# Gemma 4
+(tool_name) -hf ggml-org/gemma-4-E2B-it-GGUF
+(tool_name) -hf ggml-org/gemma-4-E4B-it-GGUF
+(tool_name) -hf ggml-org/gemma-4-26B-A4B-it-GGUF
+(tool_name) -hf ggml-org/gemma-4-31B-it-GGUF
 ```

 **Audio models**:
@@ -108,6 +114,10 @@ NOTE: some models may require large context window, for example: `-c 8192`

 # Mistral's Voxtral
 (tool_name) -hf ggml-org/Voxtral-Mini-3B-2507-GGUF
+
+# Qwen3-ASR
+(tool_name) -hf ggml-org/Qwen3-ASR-0.6B-GGUF
+(tool_name) -hf ggml-org/Qwen3-ASR-1.7B-GGUF
 ```

 **Mixed modalities**:
@@ -117,6 +127,16 @@ NOTE: some models may require large context window, for example: `-c 8192`
 # Capabilities: audio input, vision input
 (tool_name) -hf ggml-org/Qwen2.5-Omni-3B-GGUF
 (tool_name) -hf ggml-org/Qwen2.5-Omni-7B-GGUF
+
+# Qwen3 Omni
+# Capabilities: audio input, vision input
+(tool_name) -hf ggml-org/Qwen3-Omni-30B-A3B-Instruct-GGUF
+(tool_name) -hf ggml-org/Qwen3-Omni-30B-A3B-Thinking-GGUF
+
+# Gemma 4
+# Capabilities: audio input, vision input
+(tool_name) -hf ggml-org/gemma-4-E2B-it-GGUF
+(tool_name) -hf ggml-org/gemma-4-E4B-it-GGUF
 ```

 ## Finding more models:
--- a/docs/ops.md
+++ b/docs/ops.md
@@ -22,13 +22,13 @@ Legend:
 |                           ARANGE | ❌ | ✅ | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ | ❌ | ❌ | ❌ |
 |                           ARGMAX | ❌ | ✅ | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ | ✅ | ❌ | ❌ |
 |                          ARGSORT | ❌ | ✅ | ✅ | ✅ | ✅ | 🟡 | 🟡 | ✅ | ✅ | ❌ | ❌ |
-|                             CEIL | ❌ | ❌ | ✅ | 🟡 | ❌ | ❌ | ✅ | 🟡 | ✅ | ❌ | ❌ |
+|                             CEIL | ❌ | ❌ | ✅ | 🟡 | ✅ | ❌ | ✅ | 🟡 | ✅ | ❌ | ❌ |
 |                            CLAMP | ❌ | ✅ | ✅ | ✅ | ✅ | 🟡 | 🟡 | 🟡 | ✅ | ❌ | ❌ |
 |                           CONCAT | ❌ | ✅ | ✅ | 🟡 | ✅ | 🟡 | ✅ | ✅ | ✅ | ❌ | ❌ |
-|                             CONT | ❌ | 🟡 | ✅ | ✅ | 🟡 | 🟡 | 🟡 | ✅ | 🟡 | ❌ | ❌ |
-|                          CONV_2D | ❌ | ❌ | ✅ | ✅ | ✅ | ✅ | ❌ | ✅ | ❌ | ❌ | ❌ |
+|                             CONT | ❌ | 🟡 | ✅ | ✅ | ✅ | 🟡 | 🟡 | ✅ | 🟡 | ❌ | ❌ |
+|                          CONV_2D | ❌ | ❌ | ✅ | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ | ❌ | ❌ |
 |                       CONV_2D_DW | ❌ | ❌ | ✅ | ✅ | ❌ | ❌ | ❌ | ✅ | ❌ | ❌ | ❌ |
-|                          CONV_3D | ❌ | ❌ | ✅ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ |
+|                          CONV_3D | ❌ | ❌ | ✅ | ❌ | ✅ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ |
 |                CONV_TRANSPOSE_1D | ❌ | ✅ | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ | ❌ | ❌ | ❌ |
 |                CONV_TRANSPOSE_2D | ❌ | ❌ | ✅ | ✅ | ✅ | ❌ | ❌ | ✅ | ❌ | ❌ | ❌ |
 |                              COS | ❌ | ✅ | ✅ | ✅ | ✅ | ❌ | 🟡 | 🟡 | ✅ | ❌ | ❌ |
@@ -46,7 +46,7 @@ Legend:
 |                            EXPM1 | ❌ | ❌ | ✅ | 🟡 | ✅ | ❌ | ❌ | ❌ | ✅ | ❌ | ❌ |
 |                             FILL | ❌ | ❌ | ✅ | ✅ | ✅ | ❌ | ❌ | ✅ | ✅ | ❌ | ❌ |
 |                   FLASH_ATTN_EXT | ❌ | 🟡 | ✅ | 🟡 | 🟡 | 🟡 | 🟡 | 🟡 | 🟡 | ❌ | ❌ |
-|                            FLOOR | ❌ | ❌ | ✅ | 🟡 | ❌ | ❌ | 🟡 | 🟡 | ✅ | ❌ | ❌ |
+|                            FLOOR | ❌ | ❌ | ✅ | 🟡 | ✅ | ❌ | 🟡 | 🟡 | ✅ | ❌ | ❌ |
 |                  GATED_DELTA_NET | ❌ | ❌ | ✅ | ❌ | 🟡 | ❌ | ✅ | ❌ | ✅ | ❌ | ❌ |
 |                GATED_LINEAR_ATTN | ❌ | ✅ | ✅ | ✅ | ❌ | ❌ | ✅ | ❌ | ❌ | ❌ | ❌ |
 |                            GEGLU | ❌ | ✅ | ✅ | ✅ | 🟡 | ✅ | ✅ | 🟡 | ✅ | ❌ | ❌ |
@@ -60,7 +60,7 @@ Legend:
 |                       GROUP_NORM | ❌ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ❌ | ❌ | ❌ |
 |                      HARDSIGMOID | ❌ | ✅ | ✅ | 🟡 | ✅ | ❌ | ✅ | 🟡 | ✅ | ❌ | ❌ |
 |                        HARDSWISH | ❌ | ✅ | ✅ | 🟡 | ✅ | ❌ | ✅ | 🟡 | ✅ | ❌ | ❌ |
-|                           IM2COL | ❌ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ❌ | ❌ | ❌ |
+|                           IM2COL | ❌ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ❌ | ❌ |
 |                        IM2COL_3D | ❌ | ❌ | ✅ | ✅ | ❌ | ❌ | ❌ | ✅ | ❌ | ❌ | ❌ |
 |                          L2_NORM | ❌ | ✅ | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ | ✅ | ❌ | ❌ |
 |                       LEAKY_RELU | ❌ | ✅ | ✅ | ✅ | 🟡 | ❌ | ✅ | 🟡 | ❌ | ❌ | ❌ |
@@ -84,10 +84,10 @@ Legend:
 |                      REPEAT_BACK | ❌ | ❌ | ✅ | ✅ | ❌ | ❌ | ✅ | ✅ | ❌ | ❌ | ❌ |
 |                         RMS_NORM | ❌ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ❌ | ❌ |
 |                    RMS_NORM_BACK | ❌ | ❌ | ✅ | ✅ | ❌ | ❌ | ✅ | ✅ | ❌ | ❌ | ❌ |
-|                             ROLL | ❌ | ❌ | ✅ | ✅ | ❌ | ❌ | ✅ | ✅ | ❌ | ❌ | ❌ |
+|                             ROLL | ❌ | ❌ | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ | ❌ | ❌ | ❌ |
 |                             ROPE | ❌ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ❌ | ❌ |
 |                        ROPE_BACK | ❌ | ❌ | ✅ | ✅ | ❌ | ❌ | ✅ | ✅ | ❌ | ❌ | ❌ |
-|                            ROUND | ❌ | ❌ | ✅ | 🟡 | ❌ | ❌ | 🟡 | 🟡 | ✅ | ❌ | ❌ |
+|                            ROUND | ❌ | ❌ | ✅ | 🟡 | ✅ | ❌ | 🟡 | 🟡 | ✅ | ❌ | ❌ |
 |                        RWKV_WKV6 | ❌ | ❌ | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ | ❌ | ❌ | ❌ |
 |                        RWKV_WKV7 | ❌ | ❌ | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ | ❌ | ❌ | ❌ |
 |                            SCALE | ❌ | 🟡 | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ❌ | ❌ |
@@ -105,7 +105,7 @@ Legend:
 |                              SQR | ❌ | ✅ | ✅ | ✅ | ✅ | ✅ | 🟡 | 🟡 | ✅ | ❌ | ❌ |
 |                             SQRT | ❌ | ✅ | ✅ | ✅ | ✅ | ✅ | 🟡 | 🟡 | ✅ | ❌ | ❌ |
 |                         SSM_CONV | ❌ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ❌ | ❌ |
-|                         SSM_SCAN | ❌ | ❌ | ✅ | ✅ | ✅ | ❌ | ❌ | 🟡 | ❌ | ❌ | ❌ |
+|                         SSM_SCAN | ❌ | ❌ | ✅ | ✅ | ✅ | ❌ | ❌ | 🟡 | ✅ | ❌ | ❌ |
 |                             STEP | ❌ | ✅ | ✅ | 🟡 | ✅ | ❌ | ✅ | 🟡 | ✅ | ❌ | ❌ |
 |                              SUB | ❌ | ✅ | ✅ | ✅ | 🟡 | ✅ | ✅ | ✅ | ✅ | ❌ | ❌ |
 |                              SUM | ❌ | 🟡 | ✅ | 🟡 | 🟡 | ❌ | 🟡 | 🟡 | 🟡 | ❌ | ❌ |
@@ -116,6 +116,6 @@ Legend:
 |               TIMESTEP_EMBEDDING | ❌ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ❌ | ❌ | ❌ |
 |                            TOP_K | ❌ | ❌ | ✅ | ❌ | ✅ | ❌ | 🟡 | 🟡 | ✅ | ❌ | ❌ |
 |                              TRI | ❌ | ❌ | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ | ✅ | ❌ | ❌ |
-|                            TRUNC | ❌ | ❌ | ✅ | 🟡 | ❌ | ❌ | 🟡 | 🟡 | ✅ | ❌ | ❌ |
+|                            TRUNC | ❌ | ❌ | ✅ | 🟡 | ✅ | ❌ | 🟡 | 🟡 | ✅ | ❌ | ❌ |
 |                          UPSCALE | ❌ | 🟡 | ✅ | ✅ | ✅ | 🟡 | ✅ | ✅ | ❌ | ❌ | ❌ |
-|                            XIELU | ❌ | ❌ | ✅ | ❌ | ❌ | ❌ | ❌ | ✅ | ✅ | ❌ | ❌ |
+|                            XIELU | ❌ | ❌ | ✅ | ❌ | ✅ | ❌ | ❌ | ✅ | ✅ | ❌ | ❌ |
--- a/docs/ops/Metal.csv
+++ b/docs/ops/Metal.csv
--- a/docs/ops/WebGPU.csv
+++ b/docs/ops/WebGPU.csv
--- a/examples/batched/CMakeLists.txt
+++ b/examples/batched/CMakeLists.txt
@@ -1,5 +1,5 @@
 set(TARGET llama-batched)
 add_executable(${TARGET} batched.cpp)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE llama-common llama ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PRIVATE cxx_std_17)
--- a/examples/convert-llama2c-to-ggml/CMakeLists.txt
+++ b/examples/convert-llama2c-to-ggml/CMakeLists.txt
@@ -1,5 +1,5 @@
 set(TARGET llama-convert-llama2c-to-ggml)
 add_executable(${TARGET} convert-llama2c-to-ggml.cpp)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE llama-common llama ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PRIVATE cxx_std_17)
--- a/examples/debug/CMakeLists.txt
+++ b/examples/debug/CMakeLists.txt
@@ -1,5 +1,5 @@
 set(TARGET llama-debug)
 add_executable(${TARGET} debug.cpp)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE llama-common llama ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PRIVATE cxx_std_17)
--- a/examples/debug/debug.cpp
+++ b/examples/debug/debug.cpp
@@ -202,10 +202,14 @@ static bool run(llama_context * ctx, const common_params & params) {
    print_tokenized_prompt(ctx, tokens, params.prompt);

    if (params.save_logits) {
-        output_data output {ctx, model, params};
-        std::filesystem::path model_path{params.model.path};
-        std::string model_name{model_path.stem().string()};
-        save_output_data(output, model_name, params.logits_output_dir);
+        try {
+            output_data output {ctx, model, params};
+            std::filesystem::path model_path{params.model.path};
+            std::string model_name{model_path.stem().string()};
+            save_output_data(output, model_name, params.logits_output_dir);
+        } catch (const std::exception & e) {
+            LOG_ERR("%s : error saving logits: %s\n", __func__, e.what());
+        }
    }

    return true;
@@ -223,7 +227,7 @@ int main(int argc, char ** argv) {
    llama_backend_init();
    llama_numa_init(params.numa);

-    std::optional<base_callback_data> cb_data;
+    std::optional<common_debug_cb_user_data> cb_data;
    if (!params.save_logits) {
        cb_data.emplace(params, params.tensor_filter);
    }
--- a/examples/diffusion/CMakeLists.txt
+++ b/examples/diffusion/CMakeLists.txt
@@ -1,5 +1,5 @@
 set(TARGET llama-diffusion-cli)
 add_executable(${TARGET} diffusion-cli.cpp)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE llama common ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE llama llama-common ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PRIVATE cxx_std_17)
--- a/examples/diffusion/diffusion-cli.cpp
+++ b/examples/diffusion/diffusion-cli.cpp
@@ -602,8 +602,8 @@ int main(int argc, char ** argv) {

    int n_input = input_tokens.size();

-    if (n_input >= params.n_ctx) {
-        LOG_ERR("error: input too long (%d tokens), max context is %d\n", n_input, params.n_ctx);
+    if (static_cast<uint32_t>(n_input) >= llama_n_ctx(ctx)) {
+        LOG_ERR("error: input too long (%d tokens), max context is %d\n", n_input, llama_n_ctx(ctx));
        llama_free(ctx);
        llama_model_free(model);
        return 1;
--- a/examples/embedding/CMakeLists.txt
+++ b/examples/embedding/CMakeLists.txt
@@ -1,5 +1,5 @@
 set(TARGET llama-embedding)
 add_executable(${TARGET} embedding.cpp)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE llama-common llama ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PRIVATE cxx_std_17)
--- a/examples/eval-callback/CMakeLists.txt
+++ b/examples/eval-callback/CMakeLists.txt
@@ -1,7 +1,7 @@
 set(TARGET llama-eval-callback)
 add_executable(${TARGET} eval-callback.cpp)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE llama-common llama ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PRIVATE cxx_std_17)

 if(LLAMA_BUILD_TESTS)
--- a/examples/eval-callback/eval-callback.cpp
+++ b/examples/eval-callback/eval-callback.cpp
@@ -3,7 +3,6 @@
 #include "debug.h"
 #include "log.h"
 #include "llama.h"
-#include "llama-cpp.h"

 #include <clocale>
 #include <string>
@@ -38,7 +37,7 @@ static bool run(llama_context * ctx, const common_params & params) {
 int main(int argc, char ** argv) {
    std::setlocale(LC_NUMERIC, "C");

-    base_callback_data cb_data;
+    common_debug_cb_user_data cb_data;

    common_params params;

@@ -53,7 +52,7 @@ int main(int argc, char ** argv) {

    // pass the callback to the backend scheduler
    // it will be executed for each node during the graph computation
-    params.cb_eval = common_debug_cb_eval<false>;
+    params.cb_eval = common_debug_cb_eval;
    params.cb_eval_user_data = &cb_data;
    params.warmup = false;

--- a/examples/gen-docs/CMakeLists.txt
+++ b/examples/gen-docs/CMakeLists.txt
@@ -1,5 +1,5 @@
 set(TARGET llama-gen-docs)
 add_executable(${TARGET} gen-docs.cpp)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE llama-common llama ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PRIVATE cxx_std_17)
--- a/examples/gen-docs/gen-docs.cpp
+++ b/examples/gen-docs/gen-docs.cpp
@@ -73,12 +73,12 @@ static void write_help(std::ostringstream & ss, const md_file & md) {
    auto ctx_arg = common_params_parser_init(params, md.ex);

    std::vector<common_arg *> common_options;
-    std::vector<common_arg *> sparam_options;
+    std::vector<common_arg *> sampling_options;
    std::vector<common_arg *> specific_options;
    for (auto & opt : ctx_arg.options) {
        // in case multiple LLAMA_EXAMPLE_* are set, we prioritize the LLAMA_EXAMPLE_* matching current example
-        if (opt.is_sparam) {
-            sparam_options.push_back(&opt);
+        if (opt.is_sampling) {
+            sampling_options.push_back(&opt);
        } else if (opt.in_example(ctx_arg.ex)) {
            specific_options.push_back(&opt);
        } else {
@@ -93,7 +93,7 @@ static void write_help(std::ostringstream & ss, const md_file & md) {
    ss << "### Common params\n\n";
    write_table(ss, common_options);
    ss << "\n\n### Sampling params\n\n";
-    write_table(ss, sparam_options);
+    write_table(ss, sampling_options);
    ss << "\n\n### " << md.specific_section_header << "\n\n";
    write_table(ss, specific_options);

--- a/examples/idle/CMakeLists.txt
+++ b/examples/idle/CMakeLists.txt
@@ -1,5 +1,5 @@
 set(TARGET llama-idle)
 add_executable(${TARGET} idle.cpp)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE llama common ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE llama llama-common ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PRIVATE cxx_std_11)
--- a/examples/llama.android/lib/src/main/cpp/CMakeLists.txt
+++ b/examples/llama.android/lib/src/main/cpp/CMakeLists.txt
@@ -51,6 +51,6 @@ target_include_directories(${CMAKE_PROJECT_NAME} PRIVATE

 target_link_libraries(${CMAKE_PROJECT_NAME}
        llama
-        common
+        llama-common
        android
        log)
--- a/examples/lookahead/CMakeLists.txt
+++ b/examples/lookahead/CMakeLists.txt
@@ -1,5 +1,5 @@
 set(TARGET llama-lookahead)
 add_executable(${TARGET} lookahead.cpp)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE llama-common llama ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PRIVATE cxx_std_17)
--- a/examples/lookup/CMakeLists.txt
+++ b/examples/lookup/CMakeLists.txt
@@ -1,23 +1,23 @@
 set(TARGET llama-lookup)
 add_executable(${TARGET} lookup.cpp)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE llama-common llama ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PRIVATE cxx_std_17)

 set(TARGET llama-lookup-create)
 add_executable(${TARGET} lookup-create.cpp)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE llama-common llama ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PRIVATE cxx_std_17)

 set(TARGET llama-lookup-merge)
 add_executable(${TARGET} lookup-merge.cpp)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE llama-common llama ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PRIVATE cxx_std_17)

 set(TARGET llama-lookup-stats)
 add_executable(${TARGET} lookup-stats.cpp)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE llama-common llama ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PRIVATE cxx_std_17)
--- a/examples/lookup/lookup-create.cpp
+++ b/examples/lookup/lookup-create.cpp
@@ -37,9 +37,9 @@ int main(int argc, char ** argv){

    common_ngram_cache ngram_cache;
    common_ngram_cache_update(ngram_cache, LLAMA_NGRAM_STATIC, LLAMA_NGRAM_STATIC, inp, inp.size(), true);
-    fprintf(stderr, "%s: hashing done, writing file to %s\n", __func__, params.speculative.lookup_cache_static.c_str());
+    fprintf(stderr, "%s: hashing done, writing file to %s\n", __func__, params.speculative.ngram_cache.lookup_cache_static.c_str());

-    common_ngram_cache_save(ngram_cache, params.speculative.lookup_cache_static);
+    common_ngram_cache_save(ngram_cache, params.speculative.ngram_cache.lookup_cache_static);

    return 0;
 }
--- a/examples/lookup/lookup-stats.cpp
+++ b/examples/lookup/lookup-stats.cpp
@@ -24,7 +24,7 @@ int main(int argc, char ** argv){
        return 1;
    }

-    const int n_draft = params.speculative.n_max;
+    const int n_draft = params.speculative.draft.n_max;

    // init llama.cpp
    llama_backend_init();
@@ -49,18 +49,18 @@ int main(int argc, char ** argv){
    {
        const int64_t t_start_draft_us = ggml_time_us();

-        if (!params.speculative.lookup_cache_static.empty()) {
+        if (!params.speculative.ngram_cache.lookup_cache_static.empty()) {
            try {
-                ngram_cache_static = common_ngram_cache_load(params.speculative.lookup_cache_static);
+                ngram_cache_static = common_ngram_cache_load(params.speculative.ngram_cache.lookup_cache_static);
            } catch (std::ifstream::failure const &) {
-                LOG_ERR("failed to open static lookup cache: %s", params.speculative.lookup_cache_static.c_str());
+                LOG_ERR("failed to open static lookup cache: %s", params.speculative.ngram_cache.lookup_cache_static.c_str());
                exit(1);
            }
        }

-        if (!params.speculative.lookup_cache_dynamic.empty()) {
+        if (!params.speculative.ngram_cache.lookup_cache_dynamic.empty()) {
            try {
-                ngram_cache_dynamic = common_ngram_cache_load(params.speculative.lookup_cache_dynamic);
+                ngram_cache_dynamic = common_ngram_cache_load(params.speculative.ngram_cache.lookup_cache_dynamic);
            } catch (std::ifstream::failure const &) {} // if the file does not exist it will simply be created at the end of the program
        }

--- a/examples/lookup/lookup.cpp
+++ b/examples/lookup/lookup.cpp
@@ -25,7 +25,7 @@ int main(int argc, char ** argv){
    }

    // max. number of additional tokens to draft if match is found
-    const int n_draft = params.speculative.n_max;
+    const int n_draft = params.speculative.draft.n_max;

    // init llama.cpp
    llama_backend_init();
@@ -54,18 +54,18 @@ int main(int argc, char ** argv){
        const int64_t t_start_draft_us = ggml_time_us();
        common_ngram_cache_update(ngram_cache_context, LLAMA_NGRAM_MIN, LLAMA_NGRAM_MAX, inp, inp.size(), false);

-        if (!params.speculative.lookup_cache_static.empty()) {
+        if (!params.speculative.ngram_cache.lookup_cache_static.empty()) {
            try {
-                ngram_cache_static = common_ngram_cache_load(params.speculative.lookup_cache_static);
+                ngram_cache_static = common_ngram_cache_load(params.speculative.ngram_cache.lookup_cache_static);
            } catch (std::ifstream::failure const &) {
-                LOG_ERR("failed to open static lookup cache: %s", params.speculative.lookup_cache_static.c_str());
+                LOG_ERR("failed to open static lookup cache: %s", params.speculative.ngram_cache.lookup_cache_static.c_str());
                exit(1);
            }
        }

-        if (!params.speculative.lookup_cache_dynamic.empty()) {
+        if (!params.speculative.ngram_cache.lookup_cache_dynamic.empty()) {
            try {
-                ngram_cache_dynamic = common_ngram_cache_load(params.speculative.lookup_cache_dynamic);
+                ngram_cache_dynamic = common_ngram_cache_load(params.speculative.ngram_cache.lookup_cache_dynamic);
            } catch (std::ifstream::failure const &) {} // if the file does not exist it will simply be created at the end of the program
        }

@@ -213,7 +213,7 @@ int main(int argc, char ** argv){

    // Update dynamic ngram cache with context ngram cache and save it to disk:
    common_ngram_cache_merge(ngram_cache_dynamic, ngram_cache_context);
-    common_ngram_cache_save(ngram_cache_dynamic, params.speculative.lookup_cache_dynamic);
+    common_ngram_cache_save(ngram_cache_dynamic, params.speculative.ngram_cache.lookup_cache_dynamic);

    LOG("\n\n");

--- a/examples/model-conversion/scripts/causal/convert-model.sh
+++ b/examples/model-conversion/scripts/causal/convert-model.sh
@@ -25,7 +25,11 @@ MODEL_NAME="${MODEL_NAME:-$(basename "$MODEL_PATH")}"
 OUTPUT_DIR="${OUTPUT_DIR:-../../models}"
 TYPE="${OUTTYPE:-f16}"
 METADATA_OVERRIDE="${METADATA_OVERRIDE:-}"
-CONVERTED_MODEL="${OUTPUT_DIR}/${MODEL_NAME}.gguf"
+if [[ -n "$MMPROJ" ]]; then
+    CONVERTED_MODEL="${OUTPUT_DIR}/mmproj-${MODEL_NAME}.gguf"
+else
+    CONVERTED_MODEL="${OUTPUT_DIR}/${MODEL_NAME}.gguf"
+fi

 echo "Model path: ${MODEL_PATH}"
 echo "Model name: ${MODEL_NAME}"
@@ -38,6 +42,7 @@ if [[ -n "$DEBUG" ]]; then
 else
    CMD_ARGS=("python")
 fi
+
 CMD_ARGS+=("../../convert_hf_to_gguf.py" "--verbose")
 CMD_ARGS+=("${MODEL_PATH}")
 CMD_ARGS+=("--outfile" "${CONVERTED_MODEL}")
@@ -50,7 +55,3 @@ CMD_ARGS+=("--outtype" "${TYPE}")
 echo ""
 echo "The environment variable CONVERTED_MODEL can be set to this path using:"
 echo "export CONVERTED_MODEL=$(realpath ${CONVERTED_MODEL})"
-if [[ -n "$MMPROJ" ]]; then
-    mmproj_file="${OUTPUT_DIR}/mmproj-$(basename "${CONVERTED_MODEL}")"
-    echo "The mmproj model was created in $(realpath "$mmproj_file")"
-fi
--- a/examples/model-conversion/scripts/causal/run-casual-gen-embeddings-org.py
+++ b/examples/model-conversion/scripts/causal/run-casual-gen-embeddings-org.py
@@ -53,10 +53,10 @@ model_name = os.path.basename(model_path)
 print(f"Model name: {model_name}")

 prompt = "Hello world today"
-input_ids = tokenizer(prompt, return_tensors="pt").input_ids
+input_ids = tokenizer(prompt, return_tensors="pt").input_ids  # ty: ignore[call-non-callable]
 print(f"Input tokens: {input_ids}")
 print(f"Input text: {repr(prompt)}")
-print(f"Tokenized: {tokenizer.convert_ids_to_tokens(input_ids[0])}")
+print(f"Tokenized: {tokenizer.convert_ids_to_tokens(input_ids[0])}")  # ty: ignore[unresolved-attribute]

 with torch.no_grad():
    outputs = model(input_ids, output_hidden_states=True)
@@ -92,7 +92,7 @@ with torch.no_grad():

    # Print embeddings per token in the requested format
    print("\nToken embeddings:")
-    tokens = tokenizer.convert_ids_to_tokens(input_ids[0])
+    tokens = tokenizer.convert_ids_to_tokens(input_ids[0])  # ty: ignore[unresolved-attribute]
    for i, embedding in enumerate(token_embeddings):
        # Format: show first few values, ..., then last few values
        if len(embedding) > 10:
--- a/examples/model-conversion/scripts/utils/semantic_check.py
+++ b/examples/model-conversion/scripts/utils/semantic_check.py
@@ -207,8 +207,8 @@ def main():
        else:
            model = AutoModel.from_pretrained(args.model_path, trust_remote_code=True)

-    encoded = tokenizer(prompt, return_tensors="pt")
-    tokens = tokenizer.convert_ids_to_tokens(encoded['input_ids'][0])
+    encoded = tokenizer(prompt, return_tensors="pt")  # ty: ignore[call-non-callable]
+    tokens = tokenizer.convert_ids_to_tokens(encoded['input_ids'][0])  # ty: ignore[unresolved-attribute]
    n_tokens = len(tokens)
    print(f"n_tokens: {n_tokens}");
    print(f"hidden_size: {model.config.hidden_size}")
--- a/examples/parallel/CMakeLists.txt
+++ b/examples/parallel/CMakeLists.txt
@@ -1,5 +1,5 @@
 set(TARGET llama-parallel)
 add_executable(${TARGET} parallel.cpp)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE llama-common llama ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PRIVATE cxx_std_17)
--- a/examples/passkey/CMakeLists.txt
+++ b/examples/passkey/CMakeLists.txt
@@ -1,5 +1,5 @@
 set(TARGET llama-passkey)
 add_executable(${TARGET} passkey.cpp)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE llama-common llama ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PRIVATE cxx_std_17)
--- a/examples/retrieval/CMakeLists.txt
+++ b/examples/retrieval/CMakeLists.txt
@@ -1,5 +1,5 @@
 set(TARGET llama-retrieval)
 add_executable(${TARGET} retrieval.cpp)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE llama-common llama ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PRIVATE cxx_std_17)
--- a/examples/save-load-state/CMakeLists.txt
+++ b/examples/save-load-state/CMakeLists.txt
@@ -1,5 +1,5 @@
 set(TARGET llama-save-load-state)
 add_executable(${TARGET} save-load-state.cpp)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE llama-common llama ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PRIVATE cxx_std_17)
--- a/examples/speculative-simple/CMakeLists.txt
+++ b/examples/speculative-simple/CMakeLists.txt
@@ -1,5 +1,5 @@
 set(TARGET llama-speculative-simple)
 add_executable(${TARGET} speculative-simple.cpp)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE llama-common llama ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PRIVATE cxx_std_17)
--- a/examples/speculative-simple/speculative-simple.cpp
+++ b/examples/speculative-simple/speculative-simple.cpp
@@ -8,8 +8,24 @@
 #include <clocale>
 #include <cstdio>
 #include <cstring>
+#include <cinttypes>
 #include <string>
 #include <vector>
+#include <utility>
+
+struct spec_checkpoint {
+    int64_t n_tokens = 0;
+
+    std::vector<uint8_t> data;
+
+    size_t size() const {
+        return data.size();
+    }
+
+    bool empty() const {
+        return data.empty();
+    }
+};

 int main(int argc, char ** argv) {
    std::setlocale(LC_NUMERIC, "C");
@@ -27,7 +43,7 @@ int main(int argc, char ** argv) {
        return 1;
    }

-    if (params.speculative.mparams_dft.path.empty()) {
+    if (params.speculative.draft.mparams.path.empty()) {
        LOG_ERR("%s: --model-draft is required\n", __func__);
        return 1;
    }
@@ -46,6 +62,14 @@ int main(int argc, char ** argv) {
    model_tgt = llama_init_tgt->model();
    ctx_tgt   = llama_init_tgt->context();

+    // check if the context supports partial sequence removal
+    const auto ctx_seq_rm = common_context_can_seq_rm(ctx_tgt);
+    const bool use_ckpt = (ctx_seq_rm == COMMON_CONTEXT_SEQ_RM_TYPE_FULL);
+
+    if (use_ckpt) {
+        LOG_INF("speculative decoding will use checkpoints (context does not support partial sequence removal)\n");
+    }
+
    const llama_vocab * vocab = llama_model_get_vocab(model_tgt);

    // load the draft model
@@ -53,7 +77,7 @@ int main(int argc, char ** argv) {

    // TODO: simplify this logic
    {
-        const auto & params_spec = params.speculative;
+        const auto & params_spec = params.speculative.draft;

        auto params_dft = params;

@@ -61,15 +85,15 @@ int main(int argc, char ** argv) {
        params_dft.n_ctx        = params_spec.n_ctx;
        params_dft.n_batch      = llama_n_ctx_seq(ctx_tgt);
        params_dft.devices      = params_spec.devices;
-        params_dft.model        = params_spec.mparams_dft;
+        params_dft.model        = params_spec.mparams;
        params_dft.n_gpu_layers = params_spec.n_gpu_layers;

        if (params_spec.cpuparams.n_threads > 0) {
-            params_dft.cpuparams.n_threads       = params.speculative.cpuparams.n_threads;
-            params_dft.cpuparams_batch.n_threads = params.speculative.cpuparams_batch.n_threads;
+            params_dft.cpuparams.n_threads       = params.speculative.draft.cpuparams.n_threads;
+            params_dft.cpuparams_batch.n_threads = params.speculative.draft.cpuparams_batch.n_threads;
        }

-        params_dft.tensor_buft_overrides = params.speculative.tensor_buft_overrides;
+        params_dft.tensor_buft_overrides = params.speculative.draft.tensor_buft_overrides;

        auto mparams_dft = common_model_params_to_llama(params_dft);

@@ -79,8 +103,8 @@ int main(int argc, char ** argv) {
            return 1;
        }

-        params.speculative.model_dft = model_dft.get();
-        params.speculative.cparams_dft = common_context_params_to_llama(params_dft);
+        params.speculative.draft.model = model_dft.get();
+        params.speculative.draft.cparams = common_context_params_to_llama(params_dft);
    }

    // Tokenize the prompt
@@ -119,7 +143,7 @@ int main(int argc, char ** argv) {
    const auto t_enc_start = ggml_time_us();

    // target model sampling context
-    struct common_sampler * smpl = common_sampler_init(model_tgt, params.sampling);
+    common_sampler_ptr smpl(common_sampler_init(model_tgt, params.sampling));

    // eval the prompt
    llama_decode(ctx_tgt, llama_batch_get_one(inp.data(), inp.size() - 1));
@@ -142,21 +166,49 @@ int main(int argc, char ** argv) {

    llama_batch batch_tgt = llama_batch_init(llama_n_batch(ctx_tgt), 0, 1);

+    size_t n_draft = 0;
+
+    llama_tokens draft;
+    spec_checkpoint spec_ckpt;
+
    const auto t_enc_end = ggml_time_us();

    const auto t_dec_start = ggml_time_us();

    while (true) {
-        // optionally, generate draft tokens that can be appended to the target batch
+        // generate or reuse draft tokens
        //
        // this is the most important part of the speculation. the more probable tokens that are provided here
        // the better the performance will be. in theory, this computation can be performed asynchronously and even
        // offloaded to a remote device. it doesn't even have to be based on an LLM. instead, it can provide tokens
        // from a cache or lookup tables.
        //
-        llama_tokens draft = common_speculative_draft(spec, params_spec, prompt_tgt, id_last);
+        if (draft.empty()) {
+            // generate a new draft
+            draft = common_speculative_draft(spec, params_spec, prompt_tgt, id_last);

-        //LOG_DBG("draft: %s\n", string_from(ctx_dft, draft).c_str());
+            // save the original draft size
+            n_draft = draft.size();
+
+            // save a checkpoint of the target context before evaluating the draft
+            // this allows us to restore the state if partial draft acceptance occurs
+            if (!draft.empty() && use_ckpt) {
+                const size_t ckpt_size = llama_state_seq_get_size_ext(ctx_tgt, 0, LLAMA_STATE_SEQ_FLAGS_PARTIAL_ONLY);
+                spec_ckpt.data.resize(ckpt_size);
+
+                const size_t n = llama_state_seq_get_data_ext(ctx_tgt, spec_ckpt.data.data(), ckpt_size, 0, LLAMA_STATE_SEQ_FLAGS_PARTIAL_ONLY);
+                GGML_ASSERT(n == ckpt_size);
+
+                spec_ckpt.n_tokens = (int64_t) prompt_tgt.size();
+                LOG_DBG("created speculative checkpoint (n_tokens = %" PRId64 ", size = %.3f MiB)\n",
+                        spec_ckpt.n_tokens, (float) spec_ckpt.data.size() / 1024 / 1024);
+            }
+        } else {
+            // we have a previous (partial) draft to reuse from checkpoint restoration
+            if (use_ckpt) {
+                GGML_ASSERT(!spec_ckpt.empty());
+            }
+        }

        // always have a token to evaluate from before - id_last
        common_batch_clear(batch_tgt);
@@ -164,11 +216,6 @@ int main(int argc, char ** argv) {

        // evaluate the target model on [id_last, draft0, draft1, ..., draftN-1]
        {
-            // do not waste time on small drafts
-            if (draft.size() < (size_t) params_spec.n_min) {
-                draft.clear();
-            }
-
            for (size_t i = 0; i < draft.size(); ++i) {
                common_batch_add(batch_tgt, draft[i], n_past + i, { 0 }, true);
            }
@@ -178,6 +225,12 @@ int main(int argc, char ** argv) {
            llama_decode(ctx_tgt, batch_tgt);
        }

+        // only save the sampler sampler state if we use checkpoints
+        common_sampler_ptr smpl_save;
+        if (use_ckpt) {
+            smpl_save.reset(common_sampler_clone(smpl.get()));
+        }
+
        // sample from the full target batch and return the accepted tokens based on the target sampler
        //
        // for each token to be accepted, the sampler would have to sample that same token
@@ -185,14 +238,38 @@ int main(int argc, char ** argv) {
        // available logits from the batch and sample the next token until we run out of logits or the sampler
        // disagrees with the draft
        //
-        const auto ids = common_sampler_sample_and_accept_n(smpl, ctx_tgt, draft);
+        auto ids = common_sampler_sample_and_accept_n(smpl.get(), ctx_tgt, draft);

        //LOG_DBG("ids: %s\n", string_from(ctx_tgt, ids).c_str());

        GGML_ASSERT(ids.size() > 0); // there will always be at least one accepted token

+        // check for partial draft acceptance:
+        // if the context doesn't support partial sequence removal, restore the checkpoint
+        // and make the accepted tokens the new partial draft for the next iteration
+        if (use_ckpt && ids.size() - 1 < draft.size()) {
+            LOG_DBG("partial acceptance: %zu < %zu, restoring checkpoint\n", ids.size() - 1, draft.size());
+
+            draft = std::move(ids);
+
+            const size_t n = llama_state_seq_set_data_ext(ctx_tgt, spec_ckpt.data.data(), spec_ckpt.size(), 0, LLAMA_STATE_SEQ_FLAGS_PARTIAL_ONLY);
+            GGML_ASSERT(n == spec_ckpt.size());
+
+            llama_memory_seq_rm(llama_get_memory(ctx_tgt), 0, spec_ckpt.n_tokens, -1);
+
+            prompt_tgt.resize(spec_ckpt.n_tokens);
+            smpl = std::move(smpl_save);
+
+            n_past = (int) prompt_tgt.size();
+
+            continue;
+        }
+
+        common_speculative_accept(spec, ids.size() - 1);
+
+        // full acceptance: consume the draft and commit accepted tokens
        n_past    += ids.size() - 1;
-        n_drafted += draft.size(); // note: we ignore the discarded small drafts
+        n_drafted += n_draft; // note: we ignore the discarded small drafts
        n_accept  += ids.size() - 1;
        n_predict += ids.size();

@@ -222,6 +299,9 @@ int main(int argc, char ** argv) {

        LOG_DBG("accepted %d/%d draft tokens, the last target token is: (%d)\n", (int) ids.size() - 1, (int) draft.size(), id_last);

+        // clear the draft since it has been consumed
+        draft.clear();
+
        {
            LOG_DBG("clear kv cache from any extra tokens, n_past = %d\n", n_past);

@@ -243,7 +323,7 @@ int main(int argc, char ** argv) {
    LOG_INF("decoded %4d tokens in %8.3f seconds, speed: %8.3f t/s\n", n_predict, (t_dec_end - t_dec_start) / 1e6f, n_predict  / ((t_dec_end - t_dec_start) / 1e6f));

    LOG_INF("\n");
-    LOG_INF("n_draft   = %d\n", params_spec.n_max);
+    LOG_INF("n_draft   = %d\n", params_spec.draft.n_max);
    LOG_INF("n_predict = %d\n", n_predict);
    LOG_INF("n_drafted = %d\n", n_drafted);
    LOG_INF("n_accept  = %d\n", n_accept);
@@ -254,11 +334,10 @@ int main(int argc, char ** argv) {

    LOG_INF("\n");
    LOG_INF("target:\n\n");
-    common_perf_print(ctx_tgt, smpl);
+    common_perf_print(ctx_tgt, smpl.get());

    llama_batch_free(batch_tgt);

-    common_sampler_free(smpl);
    common_speculative_free(spec);

    llama_backend_free();
--- a/examples/speculative/CMakeLists.txt
+++ b/examples/speculative/CMakeLists.txt
@@ -1,5 +1,5 @@
 set(TARGET llama-speculative)
 add_executable(${TARGET} speculative.cpp)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE llama-common llama ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PRIVATE cxx_std_17)
--- a/examples/speculative/speculative.cpp
+++ b/examples/speculative/speculative.cpp
@@ -49,7 +49,7 @@ int main(int argc, char ** argv) {
        return 1;
    }

-    if (params.speculative.mparams_dft.path.empty()) {
+    if (params.speculative.draft.mparams.path.empty()) {
        LOG_ERR("%s: --model-draft is required\n", __func__);
        return 1;
    }
@@ -58,7 +58,7 @@ int main(int argc, char ** argv) {
    const int n_seq_dft = params.n_parallel;

    // probability threshold for splitting a draft branch (only for n_seq_dft > 1)
-    const float p_draft_split = params.speculative.p_split;
+    const float p_draft_split = params.speculative.draft.p_split;

    std::default_random_engine rng(params.sampling.seed == LLAMA_DEFAULT_SEED ? std::random_device()() : params.sampling.seed);
    std::uniform_real_distribution<> u_dist;
@@ -80,15 +80,15 @@ int main(int argc, char ** argv) {
    ctx_tgt   = llama_init_tgt->context();

    // load the draft model
-    params.devices = params.speculative.devices;
-    params.model = params.speculative.mparams_dft;
-    params.n_gpu_layers = params.speculative.n_gpu_layers;
-    if (params.speculative.cpuparams.n_threads > 0) {
-        params.cpuparams.n_threads = params.speculative.cpuparams.n_threads;
+    params.devices = params.speculative.draft.devices;
+    params.model = params.speculative.draft.mparams;
+    params.n_gpu_layers = params.speculative.draft.n_gpu_layers;
+    if (params.speculative.draft.cpuparams.n_threads > 0) {
+        params.cpuparams.n_threads = params.speculative.draft.cpuparams.n_threads;
    }

-    params.cpuparams_batch.n_threads = params.speculative.cpuparams_batch.n_threads;
-    params.tensor_buft_overrides     = params.speculative.tensor_buft_overrides;
+    params.cpuparams_batch.n_threads = params.speculative.draft.cpuparams_batch.n_threads;
+    params.tensor_buft_overrides     = params.speculative.draft.tensor_buft_overrides;

    auto llama_init_dft = common_init_from_params(params);

@@ -183,7 +183,7 @@ int main(int argc, char ** argv) {
    //GGML_ASSERT(n_vocab == llama_vocab_n_tokens(model_dft));

    // how many tokens to draft each time
-    int n_draft = params.speculative.n_max;
+    int n_draft = params.speculative.draft.n_max;

    int n_predict = 0;
    int n_drafted = 0;
--- a/examples/sycl/CMakeLists.txt
+++ b/examples/sycl/CMakeLists.txt
@@ -5,5 +5,5 @@
 set(TARGET llama-ls-sycl-device)
 add_executable(${TARGET} ls-sycl-device.cpp)
 install(TARGETS ${TARGET} RUNTIME)
-target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT})
+target_link_libraries(${TARGET} PRIVATE llama-common llama ${CMAKE_THREAD_LIBS_INIT})
 target_compile_features(${TARGET} PRIVATE cxx_std_17)
--- a/examples/sycl/start-svr.sh
+++ b/examples/sycl/start-svr.sh
@@ -0,0 +1,124 @@
+#!/bin/bash
+
+#  MIT license
+#  Copyright (C) 2024 Intel Corporation
+#  SPDX-License-Identifier: MIT
+
+Help() {
+  cat << EOF
+Usage: $(basename "$0") [OPTIONS]
+
+This script processes files with specified options.
+
+Options:
+  -h, --help    Display this help message and exit.
+  -c, --context <value>    Set context length. Bigger need more memory.
+  -p, --promote <value>    Prompt to start generation with.
+  -m, --model   <value>    Full model file path.
+  -mg,--main-gpu <value>   Set main GPU ID (0 - n) for single GPU mode.
+  -sm,--split-mode <value> How to split the model across multiple GPUs, one of:
+                            - none: use one GPU only
+                            - layer (default): split layers and KV across GPUs
+                            - row: split rows across GPUs
+  -ngl,--n-gpu-layers <value>  Max. number of layers to store in VRAM (default: -1)
+  -lv,--log-verbosity <value>  Set the verbosity threshold. Messages with a higher verbosity will be
+                               ignored. Values:
+                                - 0: generic output
+                                - 1: error
+                                - 2: warning
+                                - 3: info
+                                - 4: debug
+
+
+EOF
+}
+
+BIN_FILE=./build/bin/llama-server
+SEED=0
+GPUS_SETTING=""
+
+MODEL_FILE=../models/Qwen3.5-4B-Q4_0.gguf
+NGL=99
+CONTEXT=4096
+GGML_SYCL_DEVICE=-1
+SPLIT_MODE=layer
+LOG_VERBOSE=3
+while [[ $# -gt 0 ]]; do
+    case "$1" in
+        -c|--context)
+            CONTEXT=$2
+            # Shift twice to consume both the option flag and its value
+            shift
+            shift
+            ;;
+        -m|--model)
+            MODEL_FILE="$2"
+            # Shift twice to consume both the option flag and its value
+            shift
+            shift
+            ;;
+        -mg|--main-gpu)
+            GGML_SYCL_DEVICE=$2
+            SPLIT_MODE=none
+            # Shift twice to consume both the option flag and its value
+            shift
+            shift
+            ;;
+        -sm|--split-mode)
+            SPLIT_MODE=$2
+            # Shift twice to consume both the option flag and its value
+            shift
+            shift
+            ;;
+        -ngl|--n-gpu-layers)
+            NGL=$2
+            # Shift twice to consume both the option flag and its value
+            shift
+            shift
+            ;;
+        -lv|--log-verbosity)
+            LOG_VERBOSE=$2
+            # Shift twice to consume both the option flag and its value
+            shift
+            shift
+            ;;
+        -h|--help)
+            Help
+            exit 0
+            ;;
+        *)
+            # Handle unknown options or stop processing options
+            echo "Invalid option: $1"
+            # Optional: exit script or shift to treat remaining as positional args
+            exit 1
+            ;;
+    esac
+done
+
+
+
+source /opt/intel/oneapi/setvars.sh
+
+#export GGML_SYCL_DEBUG=1
+
+#ZES_ENABLE_SYSMAN=1, Support to get free memory of GPU by sycl::aspect::ext_intel_free_memory. Recommended to use when --split-mode = layer.
+
+#support malloc device memory more than 4GB.
+export UR_L0_ENABLE_RELAXED_ALLOCATION_LIMITS=1
+echo "UR_L0_ENABLE_RELAXED_ALLOCATION_LIMITS=${UR_L0_ENABLE_RELAXED_ALLOCATION_LIMITS}"
+
+if [ $GGML_SYCL_DEVICE -ne -1 ]; then
+    echo "Use $GGML_SYCL_DEVICE as main GPU"
+    #use signle GPU only
+    GPUS_SETTING="-mg $GGML_SYCL_DEVICE -sm ${SPLIT_MODE}"
+    export ONEAPI_DEVICE_SELECTOR="level_zero:${$GGML_SYCL_DEVICE}"
+    echo "ONEAPI_DEVICE_SELECTOR=${ONEAPI_DEVICE_SELECTOR}"
+else
+    echo "Use all Intel GPUs, including iGPU & dGPU"
+    GPUS_SETTING="-sm ${SPLIT_MODE}"
+ fi
+
+echo "run cmd: ZES_ENABLE_SYSMAN=1 ${BIN_FILE} -m ${MODEL_FILE} -no-cnv -p "${INPUT_PROMPT}" -n 200 -e -ngl ${NGL} -s ${SEED} -c ${CONTEXT} ${GPUS_SETTING} -lv ${LOG_VERBOSE}  --mmap "
+ZES_ENABLE_SYSMAN=1 ${BIN_FILE} -m ${MODEL_FILE} -ngl ${NGL} -s ${SEED} -c ${CONTEXT} ${GPUS_SETTING} -lv ${LOG_VERBOSE} --mmap --host 0.0.0.0 --port 8000
+
+
--- a/Show More
+++ b/Show More