docs(transformers): add docs section about transformers

Fixes: #1276

.dockerignore

@@ -1,11 +1,6 @@
 .idea
-.github
-.vscode
 models
 examples/chatbot-ui/models
 examples/rwkv/models
 examples/**/models
-Dockerfile*
-
-# SonarQube
-.scannerwork
+Dockerfile

.editorconfig

@@ -1,31 +0,0 @@
-
-root = true
-
-[*]
-indent_style = space
-indent_size = 2
-end_of_line = lf
-charset = utf-8
-trim_trailing_whitespace = true
-insert_final_newline = true
-
-[*.go]
-indent_style = tab
-
-[Makefile]
-indent_style = tab
-
-[*.proto]
-indent_size = 2
-
-[*.py]
-indent_size = 4
-
-[*.js]
-indent_size = 2
-
-[*.yaml]
-indent_size = 2
-
-[*.md]
-trim_trailing_whitespace = false

.env

@@ -1,33 +1,33 @@
 ## Set number of threads.
 ## Note: prefer the number of physical cores. Overbooking the CPU degrades performance notably.
-# LOCALAI_THREADS=14
+# THREADS=14
 
 ## Specify a different bind address (defaults to ":8080")
-# LOCALAI_ADDRESS=127.0.0.1:8080
+# ADDRESS=127.0.0.1:8080
 
 ## Default models context size
-# LOCALAI_CONTEXT_SIZE=512
+# CONTEXT_SIZE=512
 #
 ## Define galleries.
 ## models will to install will be visible in `/models/available`
-# LOCALAI_GALLERIES=[{"name":"model-gallery", "url":"github:go-skynet/model-gallery/index.yaml"}]
+# GALLERIES=[{"name":"model-gallery", "url":"github:go-skynet/model-gallery/index.yaml"}]
 
 ## CORS settings
-# LOCALAI_CORS=true
-# LOCALAI_CORS_ALLOW_ORIGINS=*
+# CORS=true
+# CORS_ALLOW_ORIGINS=*
 
 ## Default path for models
 #
-# LOCALAI_MODELS_PATH=/models
+# MODELS_PATH=/models
 
 ## Enable debug mode
-# LOCALAI_LOG_LEVEL=debug
+# DEBUG=true
 
 ## Disables COMPEL (Diffusers)
 # COMPEL=0
 
 ## Enable/Disable single backend (useful if only one GPU is available)
-# LOCALAI_SINGLE_ACTIVE_BACKEND=true
+# SINGLE_ACTIVE_BACKEND=true
 
 ## Specify a build type. Available: cublas, openblas, clblas.
 ## cuBLAS: This is a GPU-accelerated version of the complete standard BLAS (Basic Linear Algebra Subprograms) library. It's provided by Nvidia and is part of their CUDA toolkit.
@@ -46,13 +46,13 @@
 # GO_TAGS=stablediffusion
 
 ## Path where to store generated images
-# LOCALAI_IMAGE_PATH=/tmp/generated/images
+# IMAGE_PATH=/tmp
 
 ## Specify a default upload limit in MB (whisper)
-# LOCALAI_UPLOAD_LIMIT=15
+# UPLOAD_LIMIT
 
 ## List of external GRPC backends (note on the container image this variable is already set to use extra backends available in extra/)
-# LOCALAI_EXTERNAL_GRPC_BACKENDS=my-backend:127.0.0.1:9000,my-backend2:/usr/bin/backend.py
+# EXTERNAL_GRPC_BACKENDS=my-backend:127.0.0.1:9000,my-backend2:/usr/bin/backend.py
 
 ### Advanced settings ###
 ### Those are not really used by LocalAI, but from components in the stack ###
@@ -72,18 +72,18 @@
 # LLAMACPP_PARALLEL=1
 
 ### Enable to run parallel requests
-# LOCALAI_PARALLEL_REQUESTS=true
+# PARALLEL_REQUESTS=true
 
 ### Watchdog settings
 ###
 # Enables watchdog to kill backends that are inactive for too much time
-# LOCALAI_WATCHDOG_IDLE=true
-#
-# Time in duration format (e.g. 1h30m) after which a backend is considered idle
-# LOCALAI_WATCHDOG_IDLE_TIMEOUT=5m
+# WATCHDOG_IDLE=true
 #
 # Enables watchdog to kill backends that are busy for too much time
-# LOCALAI_WATCHDOG_BUSY=true
+# WATCHDOG_BUSY=true
+#
+# Time in duration format (e.g. 1h30m) after which a backend is considered idle
+# WATCHDOG_IDLE_TIMEOUT=5m
 #
 # Time in duration format (e.g. 1h30m) after which a backend is considered busy
-# LOCALAI_WATCHDOG_BUSY_TIMEOUT=5m
+# WATCHDOG_BUSY_TIMEOUT=5m

.github/dependabot.yml

@@ -1,25 +0,0 @@
-# https://docs.github.com/en/code-security/dependabot/dependabot-version-updates/configuration-options-for-the-dependabot.yml-file
-version: 2
-updates:
-  - package-ecosystem: "gomod"
-    directory: "/"
-    schedule:
-      interval: "weekly"
-  - package-ecosystem: "github-actions"
-    # Workflow files stored in the default location of `.github/workflows`. (You don't need to specify `/.github/workflows` for `directory`. You can use `directory: "/"`.)
-    directory: "/"
-    schedule:
-      # Check for updates to GitHub Actions every weekday
-      interval: "weekly"
-  - package-ecosystem: "pip"
-    # Workflow files stored in the default location of `.github/workflows`. (You don't need to specify `/.github/workflows` for `directory`. You can use `directory: "/"`.)
-    directory: "/"
-    schedule:
-      # Check for updates to GitHub Actions every weekday
-      interval: "weekly"
-  - package-ecosystem: "docker"
-    # Workflow files stored in the default location of `.github/workflows`. (You don't need to specify `/.github/workflows` for `directory`. You can use `directory: "/"`.)
-    directory: "/"
-    schedule:
-      # Check for updates to GitHub Actions every weekday
-      interval: "weekly"

.github/labeler.yml

@@ -1,19 +0,0 @@
-enhancements:
- - head-branch: ['^feature', 'feature']
-
-kind/documentation:
-- any:
-  - changed-files:
-    - any-glob-to-any-file: 'docs/*'
-  - changed-files:
-    - any-glob-to-any-file: '*.md'
-
-examples:
-- any:
-  - changed-files:
-    - any-glob-to-any-file: 'examples/*'
-
-ci:
-- any:
-  - changed-files:
-    - any-glob-to-any-file: '.github/*'

.github/release.yml

@@ -12,23 +12,13 @@ changelog:
     - title: "Bug fixes :bug:"
       labels:
         - bug
-        - regression
     - title: Exciting New Features 🎉
       labels:
         - Semver-Minor
         - enhancement
-        - ux
-        - roadmap
-    - title: 🧠 Models
-      labels:
-        - area/ai-model
-    - title: 📖 Documentation and examples
-      labels:
-        - kind/documentation
-        - examples
     - title: 👒 Dependencies
       labels:
         - dependencies
     - title: Other Changes
       labels:
-        - "*"
+        - "*"

.github/workflows/bump_deps.yaml

@@ -49,7 +49,7 @@ jobs:
         run: |
           bash .github/bump_deps.sh ${{ matrix.repository }} ${{ matrix.branch }} ${{ matrix.variable }}
       - name: Create Pull Request
-        uses: peter-evans/create-pull-request@v6
+        uses: peter-evans/create-pull-request@v5
         with:
           token: ${{ secrets.UPDATE_BOT_TOKEN }}
           push-to-fork: ci-forks/LocalAI

.github/workflows/bump_docs.yaml

@@ -17,7 +17,7 @@ jobs:
         run: |
           bash .github/bump_docs.sh ${{ matrix.repository }}
       - name: Create Pull Request
-        uses: peter-evans/create-pull-request@v6
+        uses: peter-evans/create-pull-request@v5
         with:
           token: ${{ secrets.UPDATE_BOT_TOKEN }}
           push-to-fork: ci-forks/LocalAI

.github/workflows/dependabot_auto.yml

@@ -1,43 +0,0 @@
-name: Dependabot auto-merge
-on:
-- pull_request_target
-
-permissions:
-  contents: write
-  pull-requests: write
-  packages: read
-
-jobs:
-  dependabot:
-    runs-on: ubuntu-latest
-    if: ${{ github.actor == 'dependabot[bot]' }}
-    steps:
-      - name: Dependabot metadata
-        id: metadata
-        uses: dependabot/fetch-metadata@v2.0.0
-        with:
-          github-token: "${{ secrets.GITHUB_TOKEN }}"
-          skip-commit-verification: true
-
-      - name: Checkout repository
-        uses: actions/checkout@v4
-
-      - name: Approve a PR if not already approved
-        run: |
-          gh pr checkout "$PR_URL"
-            if [ "$(gh pr status --json reviewDecision -q .currentBranch.reviewDecision)" != "APPROVED" ];
-          then
-            gh pr review --approve "$PR_URL"
-          else
-            echo "PR already approved.";
-          fi
-        env:
-          PR_URL: ${{github.event.pull_request.html_url}}
-          GITHUB_TOKEN: ${{secrets.GITHUB_TOKEN}}
-
-      - name: Enable auto-merge for Dependabot PRs
-        if: ${{ contains(github.event.pull_request.title, 'bump')}}
-        run: gh pr merge --auto --squash "$PR_URL"
-        env:
-          PR_URL: ${{github.event.pull_request.html_url}}
-          GITHUB_TOKEN: ${{secrets.GITHUB_TOKEN}}

.github/workflows/generate_grpc_cache.yaml

@@ -1,90 +0,0 @@
-name: 'generate and publish GRPC docker caches'
-
-on:
-- workflow_dispatch
-
-concurrency:
-  group: grpc-cache-${{ github.head_ref || github.ref }}-${{ github.repository }}
-  cancel-in-progress: true
-
-jobs:
-  generate_caches:
-    strategy:
-      matrix:
-        include:
-          - grpc-base-image: ubuntu:22.04
-            runs-on: 'ubuntu-latest'
-            platforms: 'linux/amd64'
-    runs-on: ${{matrix.runs-on}}
-    steps:
-      - name: Release space from worker
-        if: matrix.runs-on == 'ubuntu-latest'
-        run: |
-          echo "Listing top largest packages"
-          pkgs=$(dpkg-query -Wf '${Installed-Size}\t${Package}\t${Status}\n' | awk '$NF == "installed"{print $1 "\t" $2}' | sort -nr)
-          head -n 30 <<< "${pkgs}"
-          echo
-          df -h
-          echo
-          sudo apt-get remove -y '^llvm-.*|^libllvm.*' || true
-          sudo apt-get remove --auto-remove android-sdk-platform-tools || true
-          sudo apt-get purge --auto-remove android-sdk-platform-tools || true
-          sudo rm -rf /usr/local/lib/android
-          sudo apt-get remove -y '^dotnet-.*|^aspnetcore-.*' || true
-          sudo rm -rf /usr/share/dotnet
-          sudo apt-get remove -y '^mono-.*' || true
-          sudo apt-get remove -y '^ghc-.*' || true
-          sudo apt-get remove -y '.*jdk.*|.*jre.*' || true
-          sudo apt-get remove -y 'php.*' || true
-          sudo apt-get remove -y hhvm powershell firefox monodoc-manual msbuild || true
-          sudo apt-get remove -y '^google-.*' || true
-          sudo apt-get remove -y azure-cli || true
-          sudo apt-get remove -y '^mongo.*-.*|^postgresql-.*|^mysql-.*|^mssql-.*' || true
-          sudo apt-get remove -y '^gfortran-.*' || true
-          sudo apt-get remove -y microsoft-edge-stable || true
-          sudo apt-get remove -y firefox || true
-          sudo apt-get remove -y powershell || true
-          sudo apt-get remove -y r-base-core || true
-          sudo apt-get autoremove -y
-          sudo apt-get clean
-          echo
-          echo "Listing top largest packages"
-          pkgs=$(dpkg-query -Wf '${Installed-Size}\t${Package}\t${Status}\n' | awk '$NF == "installed"{print $1 "\t" $2}' | sort -nr)
-          head -n 30 <<< "${pkgs}"
-          echo
-          sudo rm -rfv build || true
-          sudo rm -rf /usr/share/dotnet || true
-          sudo rm -rf /opt/ghc || true
-          sudo rm -rf "/usr/local/share/boost" || true
-          sudo rm -rf "$AGENT_TOOLSDIRECTORY" || true
-          df -h
-
-      - name: Set up QEMU
-        uses: docker/setup-qemu-action@master
-        with:
-          platforms: all
-
-      - name: Set up Docker Buildx
-        id: buildx
-        uses: docker/setup-buildx-action@master
-
-      - name: Checkout
-        uses: actions/checkout@v4
-
-      - name: Cache GRPC
-        uses: docker/build-push-action@v5
-        with:
-          builder: ${{ steps.buildx.outputs.name }}
-          # The build-args MUST be an EXACT match between the image cache and other workflow steps that want to use that cache.
-          # This means that even the MAKEFLAGS have to be an EXACT match.
-          # If the build-args are not an EXACT match, it will result in a cache miss, which will require GRPC to be built from scratch.
-          build-args: |
-            GRPC_BASE_IMAGE=${{ matrix.grpc-base-image }}
-            MAKEFLAGS=--jobs=4 --output-sync=target
-            GRPC_VERSION=v1.58.0
-          context: .
-          file: ./Dockerfile
-          cache-to: type=gha,ignore-error=true
-          target: grpc
-          platforms: ${{ matrix.platforms }}
-          push: false

.github/workflows/image-pr.yml

@@ -22,8 +22,6 @@ jobs:
       platforms: ${{ matrix.platforms }}
       runs-on: ${{ matrix.runs-on }}
       base-image: ${{ matrix.base-image }}
-      grpc-base-image: ${{ matrix.grpc-base-image }}
-      makeflags: ${{ matrix.makeflags }}
     secrets:
       dockerUsername: ${{ secrets.DOCKERHUB_USERNAME }}
       dockerPassword: ${{ secrets.DOCKERHUB_PASSWORD }}
@@ -43,7 +41,6 @@ jobs:
             image-type: 'extras'
             runs-on: 'arc-runner-set'
             base-image: "ubuntu:22.04"
-            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "12"
             cuda-minor-version: "1"
@@ -54,7 +51,6 @@ jobs:
             image-type: 'extras'
             runs-on: 'arc-runner-set'
             base-image: "ubuntu:22.04"
-            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'hipblas'
             platforms: 'linux/amd64'
             tag-latest: 'false'
@@ -62,19 +58,15 @@ jobs:
             ffmpeg: 'false'
             image-type: 'extras'
             base-image: "rocm/dev-ubuntu-22.04:6.0-complete"
-            grpc-base-image: "ubuntu:22.04"
             runs-on: 'arc-runner-set'
-            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'sycl_f16'
             platforms: 'linux/amd64'
             tag-latest: 'false'
             base-image: "intel/oneapi-basekit:2024.0.1-devel-ubuntu22.04"
-            grpc-base-image: "ubuntu:22.04"
             tag-suffix: 'sycl-f16-ffmpeg'
             ffmpeg: 'true'
             image-type: 'extras'
             runs-on: 'arc-runner-set'
-            makeflags: "--jobs=3 --output-sync=target"
   core-image-build:
     uses: ./.github/workflows/image_build.yml
     with:
@@ -88,8 +80,6 @@ jobs:
       platforms: ${{ matrix.platforms }}
       runs-on: ${{ matrix.runs-on }}
       base-image: ${{ matrix.base-image }}
-      grpc-base-image: ${{ matrix.grpc-base-image }}
-      makeflags: ${{ matrix.makeflags }}
     secrets:
       dockerUsername: ${{ secrets.DOCKERHUB_USERNAME }}
       dockerPassword: ${{ secrets.DOCKERHUB_PASSWORD }}
@@ -106,17 +96,14 @@ jobs:
             image-type: 'core'
             runs-on: 'ubuntu-latest'
             base-image: "ubuntu:22.04"
-            makeflags: "--jobs=4 --output-sync=target"
           - build-type: 'sycl_f16'
             platforms: 'linux/amd64'
             tag-latest: 'false'
             base-image: "intel/oneapi-basekit:2024.0.1-devel-ubuntu22.04"
-            grpc-base-image: "ubuntu:22.04"
             tag-suffix: 'sycl-f16-ffmpeg-core'
             ffmpeg: 'true'
             image-type: 'core'
             runs-on: 'arc-runner-set'
-            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "12"
             cuda-minor-version: "1"
@@ -126,5 +113,4 @@ jobs:
             ffmpeg: 'true'
             image-type: 'core'
             runs-on: 'ubuntu-latest'
-            base-image: "ubuntu:22.04"
-            makeflags: "--jobs=4 --output-sync=target"
+            base-image: "ubuntu:22.04"

.github/workflows/image.yml

@@ -26,11 +26,6 @@ jobs:
       platforms: ${{ matrix.platforms }}
       runs-on: ${{ matrix.runs-on }}
       base-image: ${{ matrix.base-image }}
-      grpc-base-image: ${{ matrix.grpc-base-image }}
-      aio: ${{ matrix.aio }}
-      makeflags: ${{ matrix.makeflags }}
-      latest-image: ${{ matrix.latest-image }}
-      latest-image-aio: ${{ matrix.latest-image-aio }}
     secrets:
       dockerUsername: ${{ secrets.DOCKERHUB_USERNAME }}
       dockerPassword: ${{ secrets.DOCKERHUB_PASSWORD }}
@@ -52,16 +47,14 @@ jobs:
             image-type: 'extras'
             runs-on: 'arc-runner-set'
             base-image: "ubuntu:22.04"
-            makeflags: "--jobs=3 --output-sync=target"
           - build-type: ''
             platforms: 'linux/amd64'
-            tag-latest: 'auto'
+            tag-latest: 'false'
             tag-suffix: '-ffmpeg'
             ffmpeg: 'true'
             image-type: 'extras'
             runs-on: 'arc-runner-set'
             base-image: "ubuntu:22.04"
-            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "11"
             cuda-minor-version: "7"
@@ -72,7 +65,6 @@ jobs:
             image-type: 'extras'
             runs-on: 'arc-runner-set'
             base-image: "ubuntu:22.04"
-            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "12"
             cuda-minor-version: "1"
@@ -83,35 +75,26 @@ jobs:
             image-type: 'extras'
             runs-on: 'arc-runner-set'
             base-image: "ubuntu:22.04"
-            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "11"
             cuda-minor-version: "7"
             platforms: 'linux/amd64'
-            tag-latest: 'auto'
+            tag-latest: 'false'
             tag-suffix: '-cublas-cuda11-ffmpeg'
             ffmpeg: 'true'
             image-type: 'extras'
             runs-on: 'arc-runner-set'
             base-image: "ubuntu:22.04"
-            aio: "-aio-gpu-nvidia-cuda-11"
-            latest-image: 'latest-gpu-nvidia-cuda-11'
-            latest-image-aio: 'latest-aio-gpu-nvidia-cuda-11'
-            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "12"
             cuda-minor-version: "1"
             platforms: 'linux/amd64'
-            tag-latest: 'auto'
+            tag-latest: 'false'
             tag-suffix: '-cublas-cuda12-ffmpeg'
             ffmpeg: 'true'
             image-type: 'extras'
             runs-on: 'arc-runner-set'
             base-image: "ubuntu:22.04"
-            aio: "-aio-gpu-nvidia-cuda-12"
-            latest-image: 'latest-gpu-nvidia-cuda-12'
-            latest-image-aio: 'latest-aio-gpu-nvidia-cuda-12'
-            makeflags: "--jobs=3 --output-sync=target"
           - build-type: ''
             #platforms: 'linux/amd64,linux/arm64'
             platforms: 'linux/amd64'
@@ -121,20 +104,14 @@ jobs:
             image-type: 'extras'
             base-image: "ubuntu:22.04"
             runs-on: 'arc-runner-set'
-            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'hipblas'
             platforms: 'linux/amd64'
-            tag-latest: 'auto'
+            tag-latest: 'false'
             tag-suffix: '-hipblas-ffmpeg'
             ffmpeg: 'true'
             image-type: 'extras'
-            aio: "-aio-gpu-hipblas"
             base-image: "rocm/dev-ubuntu-22.04:6.0-complete"
-            grpc-base-image: "ubuntu:22.04"
-            latest-image: 'latest-gpu-hipblas'
-            latest-image-aio: 'latest-aio-gpu-hipblas'
             runs-on: 'arc-runner-set'
-            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'hipblas'
             platforms: 'linux/amd64'
             tag-latest: 'false'
@@ -142,76 +119,56 @@ jobs:
             ffmpeg: 'false'
             image-type: 'extras'
             base-image: "rocm/dev-ubuntu-22.04:6.0-complete"
-            grpc-base-image: "ubuntu:22.04"
             runs-on: 'arc-runner-set'
-            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'sycl_f16'
             platforms: 'linux/amd64'
-            tag-latest: 'auto'
+            tag-latest: 'false'
             base-image: "intel/oneapi-basekit:2024.0.1-devel-ubuntu22.04"
-            grpc-base-image: "ubuntu:22.04"
             tag-suffix: '-sycl-f16-ffmpeg'
             ffmpeg: 'true'
             image-type: 'extras'
             runs-on: 'arc-runner-set'
-            aio: "-aio-gpu-intel-f16"
-            latest-image: 'latest-gpu-intel-f16'
-            latest-image-aio: 'latest-aio-gpu-intel-f16'
-            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'sycl_f32'
             platforms: 'linux/amd64'
-            tag-latest: 'auto'
+            tag-latest: 'false'
             base-image: "intel/oneapi-basekit:2024.0.1-devel-ubuntu22.04"
-            grpc-base-image: "ubuntu:22.04"
             tag-suffix: '-sycl-f32-ffmpeg'
             ffmpeg: 'true'
             image-type: 'extras'
             runs-on: 'arc-runner-set'
-            aio: "-aio-gpu-intel-f32"
-            latest-image: 'latest-gpu-intel-f32'
-            latest-image-aio: 'latest-aio-gpu-intel-f32'
-            makeflags: "--jobs=3 --output-sync=target"
           # Core images
           - build-type: 'sycl_f16'
             platforms: 'linux/amd64'
             tag-latest: 'false'
             base-image: "intel/oneapi-basekit:2024.0.1-devel-ubuntu22.04"
-            grpc-base-image: "ubuntu:22.04"
             tag-suffix: '-sycl-f16-core'
             ffmpeg: 'false'
             image-type: 'core'
             runs-on: 'arc-runner-set'
-            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'sycl_f32'
             platforms: 'linux/amd64'
             tag-latest: 'false'
             base-image: "intel/oneapi-basekit:2024.0.1-devel-ubuntu22.04"
-            grpc-base-image: "ubuntu:22.04"
             tag-suffix: '-sycl-f32-core'
             ffmpeg: 'false'
             image-type: 'core'
             runs-on: 'arc-runner-set'
-            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'sycl_f16'
             platforms: 'linux/amd64'
             tag-latest: 'false'
             base-image: "intel/oneapi-basekit:2024.0.1-devel-ubuntu22.04"
-            grpc-base-image: "ubuntu:22.04"
             tag-suffix: '-sycl-f16-ffmpeg-core'
             ffmpeg: 'true'
             image-type: 'core'
             runs-on: 'arc-runner-set'
-            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'sycl_f32'
             platforms: 'linux/amd64'
             tag-latest: 'false'
             base-image: "intel/oneapi-basekit:2024.0.1-devel-ubuntu22.04"
-            grpc-base-image: "ubuntu:22.04"
             tag-suffix: '-sycl-f32-ffmpeg-core'
             ffmpeg: 'true'
             image-type: 'core'
             runs-on: 'arc-runner-set'
-            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'hipblas'
             platforms: 'linux/amd64'
             tag-latest: 'false'
@@ -219,9 +176,7 @@ jobs:
             ffmpeg: 'true'
             image-type: 'core'
             base-image: "rocm/dev-ubuntu-22.04:6.0-complete"
-            grpc-base-image: "ubuntu:22.04"
             runs-on: 'arc-runner-set'
-            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'hipblas'
             platforms: 'linux/amd64'
             tag-latest: 'false'
@@ -229,9 +184,7 @@ jobs:
             ffmpeg: 'false'
             image-type: 'core'
             base-image: "rocm/dev-ubuntu-22.04:6.0-complete"
-            grpc-base-image: "ubuntu:22.04"
             runs-on: 'arc-runner-set'
-            makeflags: "--jobs=3 --output-sync=target"
   
   core-image-build:
     uses: ./.github/workflows/image_build.yml
@@ -245,12 +198,7 @@ jobs:
       cuda-minor-version: ${{ matrix.cuda-minor-version }}
       platforms: ${{ matrix.platforms }}
       runs-on: ${{ matrix.runs-on }}
-      aio: ${{ matrix.aio }}
       base-image: ${{ matrix.base-image }}
-      grpc-base-image: ${{ matrix.grpc-base-image }}
-      makeflags: ${{ matrix.makeflags }}
-      latest-image: ${{ matrix.latest-image }}
-      latest-image-aio: ${{ matrix.latest-image-aio }}
     secrets:
       dockerUsername: ${{ secrets.DOCKERHUB_USERNAME }}
       dockerPassword: ${{ secrets.DOCKERHUB_PASSWORD }}
@@ -261,16 +209,12 @@ jobs:
         include:
           - build-type: ''
             platforms: 'linux/amd64'
-            tag-latest: 'auto'
+            tag-latest: 'false'
             tag-suffix: '-ffmpeg-core'
             ffmpeg: 'true'
             image-type: 'core'
             base-image: "ubuntu:22.04"
             runs-on: 'ubuntu-latest'
-            aio: "-aio-cpu"
-            latest-image: 'latest-cpu'
-            latest-image-aio: 'latest-aio-cpu'
-            makeflags: "--jobs=4 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "11"
             cuda-minor-version: "7"
@@ -281,7 +225,6 @@ jobs:
             image-type: 'core'
             base-image: "ubuntu:22.04"
             runs-on: 'ubuntu-latest'
-            makeflags: "--jobs=4 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "12"
             cuda-minor-version: "1"
@@ -292,7 +235,6 @@ jobs:
             image-type: 'core'
             base-image: "ubuntu:22.04"
             runs-on: 'ubuntu-latest'
-            makeflags: "--jobs=4 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "11"
             cuda-minor-version: "7"
@@ -303,7 +245,6 @@ jobs:
             image-type: 'core'
             runs-on: 'ubuntu-latest'
             base-image: "ubuntu:22.04"
-            makeflags: "--jobs=4 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "12"
             cuda-minor-version: "1"
@@ -314,4 +255,3 @@ jobs:
             image-type: 'core'
             runs-on: 'ubuntu-latest'
             base-image: "ubuntu:22.04"
-            makeflags: "--jobs=4 --output-sync=target"

.github/workflows/image_build.yml

@@ -6,10 +6,6 @@ on:
     inputs:
       base-image:
         description: 'Base image'
-        required: true
-        type: string
-      grpc-base-image:
-        description: 'GRPC Base image, must be a compatible image with base-image'
         required: false
         default: ''
         type: string
@@ -33,14 +29,6 @@ on:
         description: 'Tag latest'
         default: ''
         type: string
-      latest-image:
-          description: 'Tag latest'
-          default: ''
-          type: string
-      latest-image-aio:
-          description: 'Tag latest'
-          default: ''
-          type: string
       tag-suffix:
         description: 'Tag suffix'
         default: ''
@@ -58,16 +46,6 @@ on:
         required: true
         default: ''
         type: string
-      makeflags:
-        description: 'Make Flags'
-        required: false
-        default: '--jobs=4 --output-sync=target'
-        type: string
-      aio:
-        description: 'AIO Image Name'
-        required: false
-        default: ''
-        type: string
     secrets:
       dockerUsername:
         required: true
@@ -91,7 +69,6 @@ jobs:
           && sudo apt-get install -y git
       - name: Checkout
         uses: actions/checkout@v4
-
       - name: Release space from worker
         if: inputs.runs-on == 'ubuntu-latest'
         run: |
@@ -133,7 +110,6 @@ jobs:
           sudo rm -rf "/usr/local/share/boost" || true
           sudo rm -rf "$AGENT_TOOLSDIRECTORY" || true
           df -h
-
       - name: Docker meta
         id: meta
         uses: docker/metadata-action@v5
@@ -149,34 +125,6 @@ jobs:
             latest=${{ inputs.tag-latest }}
             suffix=${{ inputs.tag-suffix }}
 
-      - name: Docker meta AIO (quay.io)
-        if: inputs.aio != ''
-        id: meta_aio
-        uses: docker/metadata-action@v5
-        with:
-          images: |
-            quay.io/go-skynet/local-ai
-          tags: |
-            type=ref,event=branch
-            type=semver,pattern={{raw}}
-          flavor: |
-            latest=${{ inputs.tag-latest }}
-            suffix=${{ inputs.aio }}
-
-      - name: Docker meta AIO (dockerhub)
-        if: inputs.aio != ''
-        id: meta_aio_dockerhub
-        uses: docker/metadata-action@v5
-        with:
-          images: |
-            localai/localai
-          tags: |
-            type=ref,event=branch
-            type=semver,pattern={{raw}}
-          flavor: |
-            latest=${{ inputs.tag-latest }}
-            suffix=${{ inputs.aio }}
-
       - name: Set up QEMU
         uses: docker/setup-qemu-action@master
         with:
@@ -201,26 +149,6 @@ jobs:
           username: ${{ secrets.quayUsername }}
           password: ${{ secrets.quayPassword }}
 
-      - name: Cache GRPC
-        uses: docker/build-push-action@v5
-        with:
-          builder: ${{ steps.buildx.outputs.name }}
-          # The build-args MUST be an EXACT match between the image cache and other workflow steps that want to use that cache.
-          # This means that even the MAKEFLAGS have to be an EXACT match.
-          # If the build-args are not an EXACT match, it will result in a cache miss, which will require GRPC to be built from scratch.
-          build-args: |
-            GRPC_BASE_IMAGE=${{ inputs.grpc-base-image || inputs.base-image }}
-            MAKEFLAGS=--jobs=4 --output-sync=target
-            GRPC_VERSION=v1.58.0
-          context: .
-          file: ./Dockerfile
-          cache-from: type=gha
-          target: grpc
-          platforms: ${{ inputs.platforms }}
-          push: false
-          tags: ${{ steps.meta.outputs.tags }}
-          labels: ${{ steps.meta.outputs.labels }}
-
       - name: Build and push
         uses: docker/build-push-action@v5
         with:
@@ -232,79 +160,12 @@ jobs:
             FFMPEG=${{ inputs.ffmpeg }}
             IMAGE_TYPE=${{ inputs.image-type }}
             BASE_IMAGE=${{ inputs.base-image }}
-            MAKEFLAGS=${{ inputs.makeflags }}
           context: .
           file: ./Dockerfile
-          cache-from: type=gha
           platforms: ${{ inputs.platforms }}
           push: ${{ github.event_name != 'pull_request' }}
           tags: ${{ steps.meta.outputs.tags }}
           labels: ${{ steps.meta.outputs.labels }}
-
-      - name: Inspect image
-        if: github.event_name != 'pull_request'
-        run: |
-          docker pull localai/localai:${{ steps.meta.outputs.version }}
-          docker image inspect localai/localai:${{ steps.meta.outputs.version }}
-          docker pull quay.io/go-skynet/local-ai:${{ steps.meta.outputs.version }}
-          docker image inspect quay.io/go-skynet/local-ai:${{ steps.meta.outputs.version }}
-
-      - name: Build and push AIO image
-        if: inputs.aio != ''
-        uses: docker/build-push-action@v5
-        with:
-          builder: ${{ steps.buildx.outputs.name }}
-          build-args: |
-            BASE_IMAGE=quay.io/go-skynet/local-ai:${{ steps.meta.outputs.version }}
-            MAKEFLAGS=${{ inputs.makeflags }}
-          context: .
-          file: ./Dockerfile.aio
-          platforms: ${{ inputs.platforms }}
-          push: ${{ github.event_name != 'pull_request' }}
-          tags: ${{ steps.meta_aio.outputs.tags }}
-          labels: ${{ steps.meta_aio.outputs.labels }}
-
-      - name: Build and push AIO image (dockerhub)
-        if: inputs.aio != ''
-        uses: docker/build-push-action@v5
-        with:
-          builder: ${{ steps.buildx.outputs.name }}
-          build-args: |
-            BASE_IMAGE=localai/localai:${{ steps.meta.outputs.version }}
-            MAKEFLAGS=${{ inputs.makeflags }}
-          context: .
-          file: ./Dockerfile.aio
-          platforms: ${{ inputs.platforms }}
-          push: ${{ github.event_name != 'pull_request' }}
-          tags: ${{ steps.meta_aio_dockerhub.outputs.tags }}
-          labels: ${{ steps.meta_aio_dockerhub.outputs.labels }}
-
-      - name: Latest tag
-        # run this on branches, when it is a tag and there is a latest-image defined
-        if: github.event_name != 'pull_request' && inputs.latest-image != ''  && github.ref_type == 'tag'
-        run: |
-          docker pull localai/localai:${{ steps.meta.outputs.version }}
-          docker tag localai/localai:${{ steps.meta.outputs.version }} localai/localai:${{ inputs.latest-image }}
-          docker push localai/localai:${{ inputs.latest-image }}
-          docker pull quay.io/go-skynet/local-ai:${{ steps.meta.outputs.version }}
-          docker tag quay.io/go-skynet/local-ai:${{ steps.meta.outputs.version }} quay.io/go-skynet/local-ai:${{ inputs.latest-image }}
-          docker push quay.io/go-skynet/local-ai:${{ inputs.latest-image }}
-      - name: Latest AIO tag
-        # run this on branches, when it is a tag and there is a latest-image defined
-        if: github.event_name != 'pull_request' && inputs.latest-image-aio != ''  && github.ref_type == 'tag'
-        run: |
-          docker pull localai/localai:${{ steps.meta_aio_dockerhub.outputs.version }}
-          docker tag localai/localai:${{ steps.meta_aio_dockerhub.outputs.version }} localai/localai:${{ inputs.latest-image-aio }}
-          docker push localai/localai:${{ inputs.latest-image-aio }}
-          docker pull quay.io/go-skynet/local-ai:${{ steps.meta_aio.outputs.version }}
-          docker tag quay.io/go-skynet/local-ai:${{ steps.meta_aio.outputs.version }} quay.io/go-skynet/local-ai:${{ inputs.latest-image-aio }}
-          docker push quay.io/go-skynet/local-ai:${{ inputs.latest-image-aio }}
-  
       - name: job summary
         run: |
           echo "Built image: ${{ steps.meta.outputs.labels }}" >> $GITHUB_STEP_SUMMARY
-
-      - name: job summary(AIO)
-        if: inputs.aio != ''
-        run: |
-          echo "Built image: ${{ steps.meta_aio.outputs.labels }}" >> $GITHUB_STEP_SUMMARY

.github/workflows/labeler.yml

@@ -1,12 +0,0 @@
-name: "Pull Request Labeler"
-on:
-- pull_request_target
-
-jobs:
-  labeler:
-    permissions:
-      contents: read
-      pull-requests: write
-    runs-on: ubuntu-latest
-    steps:
-    - uses: actions/labeler@v5

.github/workflows/localaibot_automerge.yml

@@ -1,35 +0,0 @@
-name: LocalAI-bot auto-merge
-on:
-- pull_request_target
-
-permissions:
-  contents: write
-  pull-requests: write
-  packages: read
-
-jobs:
-  dependabot:
-    runs-on: ubuntu-latest
-    if: ${{ github.actor == 'localai-bot' }}
-    steps:
-      - name: Checkout repository
-        uses: actions/checkout@v4
-
-      - name: Approve a PR if not already approved
-        run: |
-          gh pr checkout "$PR_URL"
-            if [ "$(gh pr status --json reviewDecision -q .currentBranch.reviewDecision)" != "APPROVED" ];
-          then
-            gh pr review --approve "$PR_URL"
-          else
-            echo "PR already approved.";
-          fi
-        env:
-          PR_URL: ${{github.event.pull_request.html_url}}
-          GITHUB_TOKEN: ${{secrets.GITHUB_TOKEN}}
-
-      - name: Enable auto-merge for LocalAIBot PRs
-        run: gh pr merge --auto --squash "$PR_URL"
-        env:
-          PR_URL: ${{github.event.pull_request.html_url}}
-          GITHUB_TOKEN: ${{secrets.GITHUB_TOKEN}}

.github/workflows/release.yaml

@@ -1,11 +1,6 @@
 name: Build and Release
 
-on: 
-- push
-- pull_request
-
-env:
-  GRPC_VERSION: v1.58.0
+on: push
 
 permissions:
   contents: write
@@ -35,14 +30,13 @@ jobs:
         uses: actions/checkout@v4
         with:
           submodules: true
-      - uses: actions/setup-go@v5
+      - uses: actions/setup-go@v4
         with:
-          go-version: '1.21.x'
-          cache: false
+          go-version: '>=1.21.0'
       - name: Dependencies
         run: |
           sudo apt-get update
-          sudo apt-get install build-essential ffmpeg protobuf-compiler
+          sudo apt-get install build-essential ffmpeg
       - name: Install CUDA Dependencies
         if: ${{ matrix.build == 'cuda12' || matrix.build == 'cuda11' }}
         run: |
@@ -57,29 +51,26 @@ jobs:
           sudo apt-get install -y cuda-nvcc-${CUDA_VERSION} libcublas-dev-${CUDA_VERSION}
       - name: Cache grpc
         id: cache-grpc
-        uses: actions/cache@v4
+        uses: actions/cache@v3
         with:
           path: grpc
-          key: ${{ runner.os }}-grpc-${{ env.GRPC_VERSION }}
+          key: ${{ runner.os }}-grpc
       - name: Build grpc
         if: steps.cache-grpc.outputs.cache-hit != 'true'
         run: |
-          git clone --recurse-submodules -b ${{ env.GRPC_VERSION }} --depth 1 --shallow-submodules https://github.com/grpc/grpc && \
+          git clone --recurse-submodules -b v1.58.0 --depth 1 --shallow-submodules https://github.com/grpc/grpc && \
           cd grpc && mkdir -p cmake/build && cd cmake/build && cmake -DgRPC_INSTALL=ON \
             -DgRPC_BUILD_TESTS=OFF \
-            ../.. && sudo make --jobs 5 --output-sync=target
+            ../.. && sudo make -j12
       - name: Install gRPC
         run: |
-          cd grpc && cd cmake/build && sudo make --jobs 5 --output-sync=target install
+          cd grpc && cd cmake/build && sudo make -j12 install
       - name: Build
         id: build
         env:
           CMAKE_ARGS: "${{ matrix.defines }}"
           BUILD_ID: "${{ matrix.build }}"
         run: |
-          go install google.golang.org/grpc/cmd/protoc-gen-go-grpc@latest
-          go install google.golang.org/protobuf/cmd/protoc-gen-go@latest
-          export PATH=$PATH:$GOPATH/bin
           if [ "${{ matrix.build }}" == "cuda12" ] || [ "${{ matrix.build }}" == "cuda11" ]; then
             export BUILD_TYPE=cublas
             export PATH=/usr/local/cuda/bin:$PATH
@@ -87,12 +78,12 @@ jobs:
           else
             STATIC=true make dist
           fi
-      - uses: actions/upload-artifact@v4
+      - uses: actions/upload-artifact@v3
         with:
-          name: LocalAI-linux-${{ matrix.build }}
+          name: ${{ matrix.build }}
           path: release/
       - name: Release
-        uses: softprops/action-gh-release@v2
+        uses: softprops/action-gh-release@v1
         if: startsWith(github.ref, 'refs/tags/')
         with:
           files: |
@@ -105,24 +96,27 @@ jobs:
         uses: actions/checkout@v4
         with:
           submodules: true
-      - uses: actions/setup-go@v5
+      - uses: actions/setup-go@v4
         with:
-          go-version: '1.21.x'
-          cache: false
+          go-version: '>=1.21.0'
       - name: Dependencies
         run: |
-          sudo apt-get install -y --no-install-recommends libopencv-dev protobuf-compiler
-          go install google.golang.org/grpc/cmd/protoc-gen-go-grpc@latest
-          go install google.golang.org/protobuf/cmd/protoc-gen-go@latest
+          sudo apt-get install -y --no-install-recommends libopencv-dev
+          sudo ln -s /usr/include/opencv4/opencv2 /usr/include/opencv2
       - name: Build stablediffusion
         run: |
-          export PATH=$PATH:$GOPATH/bin
           make backend-assets/grpc/stablediffusion
           mkdir -p release && cp backend-assets/grpc/stablediffusion release
-      - uses: actions/upload-artifact@v4
+      - uses: actions/upload-artifact@v3
         with:
           name: stablediffusion
           path: release/
+      - name: Release
+        uses: softprops/action-gh-release@v1
+        if: startsWith(github.ref, 'refs/tags/')
+        with:
+          files: |
+            release/*
 
   build-macOS:
     strategy:
@@ -140,15 +134,12 @@ jobs:
         uses: actions/checkout@v4
         with:
           submodules: true
-      - uses: actions/setup-go@v5
+      - uses: actions/setup-go@v4
         with:
-          go-version: '1.21.x'
-          cache: false
+          go-version: '>=1.21.0'
       - name: Dependencies
         run: |
           brew install protobuf grpc
-          go install google.golang.org/grpc/cmd/protoc-gen-go-grpc@latest
-          go install google.golang.org/protobuf/cmd/protoc-gen-go@latest
       - name: Build
         id: build
         env:
@@ -157,61 +148,13 @@ jobs:
         run: |
           export C_INCLUDE_PATH=/usr/local/include
           export CPLUS_INCLUDE_PATH=/usr/local/include
-          export PATH=$PATH:$GOPATH/bin
           make dist
-      - uses: actions/upload-artifact@v4
+      - uses: actions/upload-artifact@v3
         with:
-          name: LocalAI-MacOS-${{ matrix.build }}
+          name: ${{ matrix.build }}
           path: release/
       - name: Release
-        uses: softprops/action-gh-release@v2
-        if: startsWith(github.ref, 'refs/tags/')
-        with:
-          files: |
-            release/*
-
-
-  build-macOS-arm64:
-    strategy:
-      matrix:
-        include:
-          - build: 'avx2'
-            defines: ''
-          - build: 'avx'
-            defines: '-DLLAMA_AVX2=OFF'
-          - build: 'avx512'
-            defines: '-DLLAMA_AVX512=ON'
-    runs-on: macos-14
-    steps:
-      - name: Clone
-        uses: actions/checkout@v4
-        with:
-          submodules: true
-      - uses: actions/setup-go@v5
-        with:
-          go-version: '1.21.x'
-          cache: false
-      - name: Dependencies
-        run: |
-          brew install protobuf grpc
-          go install google.golang.org/grpc/cmd/protoc-gen-go-grpc@latest
-          go install google.golang.org/protobuf/cmd/protoc-gen-go@latest
-      - name: Build
-        id: build
-        env:
-          CMAKE_ARGS: "${{ matrix.defines }}"
-          BUILD_ID: "${{ matrix.build }}"
-        run: |
-          export C_INCLUDE_PATH=/usr/local/include
-          export CPLUS_INCLUDE_PATH=/usr/local/include
-          export PATH=$PATH:$GOPATH/bin
-          make dist
-      - uses: actions/upload-artifact@v4
-        with:
-          name: LocalAI-MacOS-arm64-${{ matrix.build }}
-          path: release/
-      - name: Release
-        uses: softprops/action-gh-release@v2
+        uses: softprops/action-gh-release@v1
         if: startsWith(github.ref, 'refs/tags/')
         with:
           files: |

.github/workflows/secscan.yaml

@@ -1,30 +0,0 @@
-name: "Security Scan"
-
-# Run workflow each time code is pushed to your repository and on a schedule.
-# The scheduled workflow runs every at 00:00 on Sunday UTC time.
-on:
-  push:
-  schedule:
-  - cron: '0 0 * * 0'
-
-jobs:
-  tests:
-    runs-on: ubuntu-latest
-    env:
-      GO111MODULE: on
-    steps:
-      - name: Checkout Source
-        uses: actions/checkout@v4
-        if: ${{ github.actor != 'dependabot[bot]' }}
-      - name: Run Gosec Security Scanner
-        if: ${{ github.actor != 'dependabot[bot]' }}
-        uses: securego/gosec@master
-        with:
-          # we let the report trigger content trigger a failure using the GitHub Security features.
-          args: '-no-fail -fmt sarif -out results.sarif ./...'
-      - name: Upload SARIF file
-        if: ${{ github.actor != 'dependabot[bot]' }}
-        uses: github/codeql-action/upload-sarif@v3
-        with:
-          # Path to SARIF file relative to the root of the repository
-          sarif_file: results.sarif

.github/workflows/test-extra.yml

@@ -32,17 +32,16 @@ jobs:
              sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" | tee -a /etc/apt/sources.list.d/conda.list' && \
              sudo apt-get update && \
              sudo apt-get install -y conda
-          sudo apt-get install -y ca-certificates cmake curl patch python3-pip
-          sudo apt-get install -y libopencv-dev
-          pip install --user grpcio-tools
+          sudo apt-get install -y ca-certificates cmake curl patch
+          sudo apt-get install -y libopencv-dev && sudo ln -s /usr/include/opencv4/opencv2 /usr/include/opencv2
           
           sudo rm -rfv /usr/bin/conda || true
 
       - name: Test transformers
         run: |
            export PATH=$PATH:/opt/conda/bin
-           make --jobs=5 --output-sync=target -C backend/python/transformers
-           make --jobs=5 --output-sync=target -C backend/python/transformers test
+           make -C backend/python/transformers
+           make -C backend/python/transformers test
 
   tests-sentencetransformers:
     runs-on: ubuntu-latest
@@ -62,48 +61,16 @@ jobs:
              sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" | tee -a /etc/apt/sources.list.d/conda.list' && \
              sudo apt-get update && \
              sudo apt-get install -y conda
-          sudo apt-get install -y ca-certificates cmake curl patch python3-pip
-          sudo apt-get install -y libopencv-dev
-          pip install --user grpcio-tools
+          sudo apt-get install -y ca-certificates cmake curl patch
+          sudo apt-get install -y libopencv-dev && sudo ln -s /usr/include/opencv4/opencv2 /usr/include/opencv2
           
           sudo rm -rfv /usr/bin/conda || true
 
       - name: Test sentencetransformers
         run: |
            export PATH=$PATH:/opt/conda/bin
-           make --jobs=5 --output-sync=target -C backend/python/sentencetransformers
-           make --jobs=5 --output-sync=target -C backend/python/sentencetransformers test
-
-
-  tests-rerankers:
-    runs-on: ubuntu-latest
-    steps:
-      - name: Clone
-        uses: actions/checkout@v4
-        with: 
-          submodules: true
-      - name: Dependencies
-        run: |
-          sudo apt-get update
-          sudo apt-get install build-essential ffmpeg
-          curl https://repo.anaconda.com/pkgs/misc/gpgkeys/anaconda.asc | gpg --dearmor > conda.gpg && \
-             sudo install -o root -g root -m 644 conda.gpg /usr/share/keyrings/conda-archive-keyring.gpg && \
-              gpg --keyring /usr/share/keyrings/conda-archive-keyring.gpg --no-default-keyring --fingerprint 34161F5BF5EB1D4BFBBB8F0A8AEB4F8B29D82806 && \
-             sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" > /etc/apt/sources.list.d/conda.list' && \
-             sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" | tee -a /etc/apt/sources.list.d/conda.list' && \
-             sudo apt-get update && \
-             sudo apt-get install -y conda
-          sudo apt-get install -y ca-certificates cmake curl patch python3-pip
-          sudo apt-get install -y libopencv-dev
-          pip install --user grpcio-tools
-          
-          sudo rm -rfv /usr/bin/conda || true
-
-      - name: Test rerankers
-        run: |
-           export PATH=$PATH:/opt/conda/bin
-           make --jobs=5 --output-sync=target -C backend/python/rerankers
-           make --jobs=5 --output-sync=target -C backend/python/rerankers test
+           make -C backend/python/sentencetransformers
+           make -C backend/python/sentencetransformers test
 
   tests-diffusers:
     runs-on: ubuntu-latest
@@ -123,47 +90,17 @@ jobs:
              sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" | tee -a /etc/apt/sources.list.d/conda.list' && \
              sudo apt-get update && \
              sudo apt-get install -y conda
-          sudo apt-get install -y ca-certificates cmake curl patch python3-pip
-          sudo apt-get install -y libopencv-dev
-          pip install --user grpcio-tools
+          sudo apt-get install -y ca-certificates cmake curl patch
+          sudo apt-get install -y libopencv-dev && sudo ln -s /usr/include/opencv4/opencv2 /usr/include/opencv2
           
           sudo rm -rfv /usr/bin/conda || true
 
       - name: Test diffusers
         run: |
            export PATH=$PATH:/opt/conda/bin
-           make --jobs=5 --output-sync=target -C backend/python/diffusers
-           make --jobs=5 --output-sync=target -C backend/python/diffusers test
+           make -C backend/python/diffusers
+           make -C backend/python/diffusers test
 
-  tests-parler-tts:
-    runs-on: ubuntu-latest
-    steps:
-      - name: Clone
-        uses: actions/checkout@v4
-        with: 
-          submodules: true
-      - name: Dependencies
-        run: |
-          sudo apt-get update
-          sudo apt-get install build-essential ffmpeg
-          curl https://repo.anaconda.com/pkgs/misc/gpgkeys/anaconda.asc | gpg --dearmor > conda.gpg && \
-             sudo install -o root -g root -m 644 conda.gpg /usr/share/keyrings/conda-archive-keyring.gpg && \
-              gpg --keyring /usr/share/keyrings/conda-archive-keyring.gpg --no-default-keyring --fingerprint 34161F5BF5EB1D4BFBBB8F0A8AEB4F8B29D82806 && \
-             sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" > /etc/apt/sources.list.d/conda.list' && \
-             sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" | tee -a /etc/apt/sources.list.d/conda.list' && \
-             sudo apt-get update && \
-             sudo apt-get install -y conda
-          sudo apt-get install -y ca-certificates cmake curl patch python3-pip
-          sudo apt-get install -y libopencv-dev
-          pip install --user grpcio-tools
-          
-          sudo rm -rfv /usr/bin/conda || true
-
-      - name: Test parler-tts
-        run: |
-           export PATH=$PATH:/opt/conda/bin
-           make --jobs=5 --output-sync=target -C backend/python/parler-tts
-           make --jobs=5 --output-sync=target -C backend/python/parler-tts test
 
   tests-transformers-musicgen:
     runs-on: ubuntu-latest
@@ -183,49 +120,47 @@ jobs:
              sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" | tee -a /etc/apt/sources.list.d/conda.list' && \
              sudo apt-get update && \
              sudo apt-get install -y conda
-          sudo apt-get install -y ca-certificates cmake curl patch python3-pip
-          sudo apt-get install -y libopencv-dev
-          pip install --user grpcio-tools
+          sudo apt-get install -y ca-certificates cmake curl patch
+          sudo apt-get install -y libopencv-dev && sudo ln -s /usr/include/opencv4/opencv2 /usr/include/opencv2
           
           sudo rm -rfv /usr/bin/conda || true
 
       - name: Test transformers-musicgen
         run: |
            export PATH=$PATH:/opt/conda/bin
-           make --jobs=5 --output-sync=target -C backend/python/transformers-musicgen
-           make --jobs=5 --output-sync=target -C backend/python/transformers-musicgen test
+           make -C backend/python/transformers-musicgen
+           make -C backend/python/transformers-musicgen test
 
 
 
-  # tests-petals:
-  #   runs-on: ubuntu-latest
-  #   steps:
-  #     - name: Clone
-  #       uses: actions/checkout@v4
-  #       with: 
-  #         submodules: true
-  #     - name: Dependencies
-  #       run: |
-  #         sudo apt-get update
-  #         sudo apt-get install build-essential ffmpeg
-  #         curl https://repo.anaconda.com/pkgs/misc/gpgkeys/anaconda.asc | gpg --dearmor > conda.gpg && \
-  #            sudo install -o root -g root -m 644 conda.gpg /usr/share/keyrings/conda-archive-keyring.gpg && \
-  #             gpg --keyring /usr/share/keyrings/conda-archive-keyring.gpg --no-default-keyring --fingerprint 34161F5BF5EB1D4BFBBB8F0A8AEB4F8B29D82806 && \
-  #            sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" > /etc/apt/sources.list.d/conda.list' && \
-  #            sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" | tee -a /etc/apt/sources.list.d/conda.list' && \
-  #            sudo apt-get update && \
-  #            sudo apt-get install -y conda
-  #         sudo apt-get install -y ca-certificates cmake curl patch python3-pip
-  #         sudo apt-get install -y libopencv-dev
-  #         pip install --user grpcio-tools
+  tests-petals:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Clone
+        uses: actions/checkout@v4
+        with: 
+          submodules: true
+      - name: Dependencies
+        run: |
+          sudo apt-get update
+          sudo apt-get install build-essential ffmpeg
+          curl https://repo.anaconda.com/pkgs/misc/gpgkeys/anaconda.asc | gpg --dearmor > conda.gpg && \
+             sudo install -o root -g root -m 644 conda.gpg /usr/share/keyrings/conda-archive-keyring.gpg && \
+              gpg --keyring /usr/share/keyrings/conda-archive-keyring.gpg --no-default-keyring --fingerprint 34161F5BF5EB1D4BFBBB8F0A8AEB4F8B29D82806 && \
+             sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" > /etc/apt/sources.list.d/conda.list' && \
+             sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" | tee -a /etc/apt/sources.list.d/conda.list' && \
+             sudo apt-get update && \
+             sudo apt-get install -y conda
+          sudo apt-get install -y ca-certificates cmake curl patch
+          sudo apt-get install -y libopencv-dev && sudo ln -s /usr/include/opencv4/opencv2 /usr/include/opencv2
           
-  #         sudo rm -rfv /usr/bin/conda || true
+          sudo rm -rfv /usr/bin/conda || true
 
-  #     - name: Test petals
-  #       run: |
-  #          export PATH=$PATH:/opt/conda/bin
-  #          make --jobs=5 --output-sync=target -C backend/python/petals
-  #          make --jobs=5 --output-sync=target -C backend/python/petals test
+      - name: Test petals
+        run: |
+           export PATH=$PATH:/opt/conda/bin
+           make -C backend/python/petals
+           make -C backend/python/petals test
 
            
 
@@ -287,17 +222,16 @@ jobs:
   #            sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" | tee -a /etc/apt/sources.list.d/conda.list' && \
   #            sudo apt-get update && \
   #            sudo apt-get install -y conda
-  #         sudo apt-get install -y ca-certificates cmake curl patch python3-pip
-  #         sudo apt-get install -y libopencv-dev
-  #         pip install --user grpcio-tools
+  #         sudo apt-get install -y ca-certificates cmake curl patch
+  #         sudo apt-get install -y libopencv-dev && sudo ln -s /usr/include/opencv4/opencv2 /usr/include/opencv2
           
   #         sudo rm -rfv /usr/bin/conda || true
 
   #     - name: Test bark
   #       run: |
   #          export PATH=$PATH:/opt/conda/bin
-  #          make --jobs=5 --output-sync=target -C backend/python/bark
-  #          make --jobs=5 --output-sync=target -C backend/python/bark test
+  #          make -C backend/python/bark
+  #          make -C backend/python/bark test
 
            
   # Below tests needs GPU. Commented out for now
@@ -320,15 +254,14 @@ jobs:
   #            sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" | tee -a /etc/apt/sources.list.d/conda.list' && \
   #            sudo apt-get update && \
   #            sudo apt-get install -y conda
-  #         sudo apt-get install -y ca-certificates cmake curl patch python3-pip
-  #         sudo apt-get install -y libopencv-dev
-  #         pip install --user grpcio-tools
+  #         sudo apt-get install -y ca-certificates cmake curl patch
+  #         sudo apt-get install -y libopencv-dev && sudo ln -s /usr/include/opencv4/opencv2 /usr/include/opencv2
   #         sudo rm -rfv /usr/bin/conda || true
   #     - name: Test vllm
   #       run: |
   #          export PATH=$PATH:/opt/conda/bin
-  #          make --jobs=5 --output-sync=target -C backend/python/vllm
-  #          make --jobs=5 --output-sync=target -C backend/python/vllm test
+  #          make -C backend/python/vllm
+  #          make -C backend/python/vllm test
   tests-vallex:
     runs-on: ubuntu-latest
     steps:
@@ -347,15 +280,14 @@ jobs:
              sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" | tee -a /etc/apt/sources.list.d/conda.list' && \
              sudo apt-get update && \
              sudo apt-get install -y conda
-          sudo apt-get install -y ca-certificates cmake curl patch python3-pip
-          sudo apt-get install -y libopencv-dev
-          pip install --user grpcio-tools
+          sudo apt-get install -y ca-certificates cmake curl patch
+          sudo apt-get install -y libopencv-dev && sudo ln -s /usr/include/opencv4/opencv2 /usr/include/opencv2    
           sudo rm -rfv /usr/bin/conda || true
       - name: Test vall-e-x
         run: |
            export PATH=$PATH:/opt/conda/bin
-           make --jobs=5 --output-sync=target -C backend/python/vall-e-x
-           make --jobs=5 --output-sync=target -C backend/python/vall-e-x test
+           make -C backend/python/vall-e-x
+           make -C backend/python/vall-e-x test
 
   tests-coqui:
     runs-on: ubuntu-latest
@@ -375,12 +307,11 @@ jobs:
              sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" | tee -a /etc/apt/sources.list.d/conda.list' && \
              sudo apt-get update && \
              sudo apt-get install -y conda
-          sudo apt-get install -y ca-certificates cmake curl patch espeak espeak-ng python3-pip
-          pip install --user grpcio-tools
+          sudo apt-get install -y ca-certificates cmake curl patch espeak espeak-ng          
           sudo rm -rfv /usr/bin/conda || true
 
       - name: Test coqui
         run: |
            export PATH=$PATH:/opt/conda/bin
-           make --jobs=5 --output-sync=target -C backend/python/coqui
-           make --jobs=5 --output-sync=target -C backend/python/coqui test
+           make -C backend/python/coqui
+           make -C backend/python/coqui test

.github/workflows/test.yml

@@ -9,9 +9,6 @@ on:
     tags:
       - '*'
 
-env:
-  GRPC_VERSION: v1.58.0
-
 concurrency:
   group: ci-tests-${{ github.head_ref || github.ref }}-${{ github.repository }}
   cancel-in-progress: true
@@ -60,37 +57,26 @@ jobs:
         with: 
           submodules: true
       - name: Setup Go ${{ matrix.go-version }}
-        uses: actions/setup-go@v5
+        uses: actions/setup-go@v4
         with:
           go-version: ${{ matrix.go-version }}
-          cache: false
       # You can test your matrix by printing the current Go version
       - name: Display Go version
         run: go version
       - name: Dependencies
         run: |
           sudo apt-get update
-          sudo apt-get install build-essential curl ffmpeg
+          sudo apt-get install build-essential ffmpeg
           curl https://repo.anaconda.com/pkgs/misc/gpgkeys/anaconda.asc | gpg --dearmor > conda.gpg && \
              sudo install -o root -g root -m 644 conda.gpg /usr/share/keyrings/conda-archive-keyring.gpg && \
-             gpg --keyring /usr/share/keyrings/conda-archive-keyring.gpg --no-default-keyring --fingerprint 34161F5BF5EB1D4BFBBB8F0A8AEB4F8B29D82806 && \
+              gpg --keyring /usr/share/keyrings/conda-archive-keyring.gpg --no-default-keyring --fingerprint 34161F5BF5EB1D4BFBBB8F0A8AEB4F8B29D82806 && \
              sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" > /etc/apt/sources.list.d/conda.list' && \
              sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" | tee -a /etc/apt/sources.list.d/conda.list' && \
              sudo apt-get update && \
              sudo apt-get install -y conda
-          sudo apt-get install -y ca-certificates cmake patch python3-pip unzip
-          sudo apt-get install -y libopencv-dev
-
-          curl -L -s https://github.com/protocolbuffers/protobuf/releases/download/v26.1/protoc-26.1-linux-x86_64.zip -o protoc.zip && \
-          unzip -j -d /usr/local/bin protoc.zip bin/protoc && \
-          rm protoc.zip
-
-          go install google.golang.org/protobuf/cmd/protoc-gen-go@latest
-          go install google.golang.org/grpc/cmd/protoc-gen-go-grpc@latest
-
-          # The python3-grpc-tools package in 22.04 is too old
-          pip install --user grpcio-tools
-
+          sudo apt-get install -y ca-certificates cmake curl patch
+          sudo apt-get install -y libopencv-dev && sudo ln -s /usr/include/opencv4/opencv2 /usr/include/opencv2
+          
           sudo rm -rfv /usr/bin/conda || true
           PATH=$PATH:/opt/conda/bin make -C backend/python/sentencetransformers
 
@@ -99,91 +85,29 @@ jobs:
           GO_TAGS="tts" make -C sources/go-piper piper.o && \
           sudo cp -rfv sources/go-piper/piper-phonemize/pi/lib/. /usr/lib/ && \
           # Pre-build stable diffusion before we install a newer version of abseil (not compatible with stablediffusion-ncn)
-          PATH="$PATH:/root/go/bin" GO_TAGS="stablediffusion tts" GRPC_BACKENDS=backend-assets/grpc/stablediffusion make build
+          GO_TAGS="stablediffusion tts" GRPC_BACKENDS=backend-assets/grpc/stablediffusion make build
       - name: Cache grpc
         id: cache-grpc
-        uses: actions/cache@v4
+        uses: actions/cache@v3
         with:
           path: grpc
-          key: ${{ runner.os }}-grpc-${{ env.GRPC_VERSION }}
+          key: ${{ runner.os }}-grpc
       - name: Build grpc
         if: steps.cache-grpc.outputs.cache-hit != 'true'
         run: |
-          git clone --recurse-submodules -b ${{ env.GRPC_VERSION }} --depth 1 --jobs 5 --shallow-submodules https://github.com/grpc/grpc && \
+          git clone --recurse-submodules -b v1.58.0 --depth 1 --shallow-submodules https://github.com/grpc/grpc && \
           cd grpc && mkdir -p cmake/build && cd cmake/build && cmake -DgRPC_INSTALL=ON \
             -DgRPC_BUILD_TESTS=OFF \
-            ../.. && sudo make --jobs 5
+            ../.. && sudo make -j12
       - name: Install gRPC
         run: |
-          cd grpc && cd cmake/build && sudo make --jobs 5 install
+          cd grpc && cd cmake/build && sudo make -j12 install
       - name: Test
         run: |
-          PATH="$PATH:/root/go/bin" GO_TAGS="stablediffusion tts" make --jobs 5 --output-sync=target test
-      - name: Setup tmate session if tests fail
-        if: ${{ failure() }}
-        uses: mxschmitt/action-tmate@v3.18
-        with:
-          detached: true
-          connect-timeout-seconds: 180
-          limit-access-to-actor: true
-
-  tests-aio-container:
-    runs-on: ubuntu-latest
-    steps:
-      - name: Release space from worker
-        run: |
-          echo "Listing top largest packages"
-          pkgs=$(dpkg-query -Wf '${Installed-Size}\t${Package}\t${Status}\n' | awk '$NF == "installed"{print $1 "\t" $2}' | sort -nr)
-          head -n 30 <<< "${pkgs}"
-          echo
-          df -h
-          echo
-          sudo apt-get remove -y '^llvm-.*|^libllvm.*' || true
-          sudo apt-get remove --auto-remove android-sdk-platform-tools || true
-          sudo apt-get purge --auto-remove android-sdk-platform-tools || true
-          sudo rm -rf /usr/local/lib/android
-          sudo apt-get remove -y '^dotnet-.*|^aspnetcore-.*' || true
-          sudo rm -rf /usr/share/dotnet
-          sudo apt-get remove -y '^mono-.*' || true
-          sudo apt-get remove -y '^ghc-.*' || true
-          sudo apt-get remove -y '.*jdk.*|.*jre.*' || true
-          sudo apt-get remove -y 'php.*' || true
-          sudo apt-get remove -y hhvm powershell firefox monodoc-manual msbuild || true
-          sudo apt-get remove -y '^google-.*' || true
-          sudo apt-get remove -y azure-cli || true
-          sudo apt-get remove -y '^mongo.*-.*|^postgresql-.*|^mysql-.*|^mssql-.*' || true
-          sudo apt-get remove -y '^gfortran-.*' || true
-          sudo apt-get autoremove -y
-          sudo apt-get clean
-          echo
-          echo "Listing top largest packages"
-          pkgs=$(dpkg-query -Wf '${Installed-Size}\t${Package}\t${Status}\n' | awk '$NF == "installed"{print $1 "\t" $2}' | sort -nr)
-          head -n 30 <<< "${pkgs}"
-          echo
-          sudo rm -rfv build || true
-          df -h
-      - name: Clone
-        uses: actions/checkout@v4
-        with: 
-          submodules: true
-      - name: Build images
-        run: |
-          docker build --build-arg FFMPEG=true --build-arg IMAGE_TYPE=core --build-arg MAKEFLAGS="--jobs=5 --output-sync=target" -t local-ai:tests -f Dockerfile .
-          BASE_IMAGE=local-ai:tests DOCKER_AIO_IMAGE=local-ai-aio:test make docker-aio
-      - name: Test
-        run: |
-          LOCALAI_MODELS_DIR=$PWD/models LOCALAI_IMAGE_TAG=test LOCALAI_IMAGE=local-ai-aio \
-            make run-e2e-aio
-      - name: Setup tmate session if tests fail
-        if: ${{ failure() }}
-        uses: mxschmitt/action-tmate@v3.18
-        with:
-          detached: true
-          connect-timeout-seconds: 180
-          limit-access-to-actor: true
+          GO_TAGS="stablediffusion tts" make test
 
   tests-apple:
-    runs-on: macOS-14
+    runs-on: macOS-latest
     strategy:
       matrix:
         go-version: ['1.21.x']
@@ -193,28 +117,17 @@ jobs:
         with: 
           submodules: true
       - name: Setup Go ${{ matrix.go-version }}
-        uses: actions/setup-go@v5
+        uses: actions/setup-go@v4
         with:
           go-version: ${{ matrix.go-version }}
-          cache: false
       # You can test your matrix by printing the current Go version
       - name: Display Go version
         run: go version
       - name: Dependencies
         run: |
-          brew install protobuf grpc make protoc-gen-go protoc-gen-go-grpc
-          pip install --user grpcio-tools
+          brew install protobuf grpc
       - name: Test
         run: |
           export C_INCLUDE_PATH=/usr/local/include
           export CPLUS_INCLUDE_PATH=/usr/local/include
-          # Used to run the newer GNUMake version from brew that supports --output-sync
-          export PATH="/opt/homebrew/opt/make/libexec/gnubin:$PATH"
-          BUILD_TYPE="GITHUB_CI_HAS_BROKEN_METAL" CMAKE_ARGS="-DLLAMA_F16C=OFF -DLLAMA_AVX512=OFF -DLLAMA_AVX2=OFF -DLLAMA_FMA=OFF" make --jobs 4 --output-sync=target test
-      - name: Setup tmate session if tests fail
-        if: ${{ failure() }}
-        uses: mxschmitt/action-tmate@v3.18
-        with:
-          detached: true
-          connect-timeout-seconds: 180
-          limit-access-to-actor: true
+          CMAKE_ARGS="-DLLAMA_F16C=OFF -DLLAMA_AVX512=OFF -DLLAMA_AVX2=OFF -DLLAMA_FMA=OFF" make test

.github/workflows/update_swagger.yaml

@@ -1,31 +0,0 @@
-name: Update swagger
-on:
-  schedule:
-    - cron: 0 20 * * *
-  workflow_dispatch:
-jobs:
-  swagger:
-    strategy:
-      fail-fast: false
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-go@v5
-        with:
-          go-version: 'stable'
-      - run: |
-          go install github.com/swaggo/swag/cmd/swag@latest
-      - name: Bump swagger 🔧
-        run: |
-          make swagger
-      - name: Create Pull Request
-        uses: peter-evans/create-pull-request@v6
-        with:
-          token: ${{ secrets.UPDATE_BOT_TOKEN }}
-          push-to-fork: ci-forks/LocalAI
-          commit-message: 'feat(swagger): update swagger'
-          title: 'feat(swagger): update swagger'
-          branch: "update/swagger"
-          body:  Update swagger
-          signoff: true
-

.github/workflows/yaml-check.yml

@@ -1,18 +0,0 @@
-name: 'Yamllint GitHub Actions'
-on:
-  - pull_request
-jobs:
-  yamllint:
-    name: 'Yamllint'
-    runs-on: ubuntu-latest
-    steps:
-      - name: 'Checkout'
-        uses: actions/checkout@master
-      - name: 'Yamllint'
-        uses: karancode/yamllint-github-action@master
-        with:
-          yamllint_file_or_dir: 'gallery'
-          yamllint_strict: false
-          yamllint_comment: true
-        env:
-          GITHUB_ACCESS_TOKEN: ${{ secrets.GITHUB_TOKEN }}

.gitignore

@@ -39,11 +39,3 @@ backend-assets/*
 !backend-assets/.keep
 prepare
 /ggml-metal.metal
-
-# Protobuf generated files
-*.pb.go
-*pb2.py
-*pb2_grpc.py
-
-# SonarQube
-.scannerwork

.vscode/extensions.json

@@ -1,5 +0,0 @@
-{
-    "recommendations": [
-        "golang.go"
-    ]
-}

CONTRIBUTING.md

@@ -1,4 +1,4 @@
-# Contributing to LocalAI
+# Contributing to localAI
 
 Thank you for your interest in contributing to LocalAI! We appreciate your time and effort in helping to improve our project. Before you get started, please take a moment to review these guidelines.
 
@@ -29,9 +29,8 @@ Thank you for your interest in contributing to LocalAI! We appreciate your time
 
 1. Clone the repository: `git clone https://github.com/go-skynet/LocalAI.git`
 2. Navigate to the project directory: `cd LocalAI`
-3. Install the required dependencies ( see https://localai.io/basics/build/#build-localai-locally )
-4. Build LocalAI: `make build`
-5. Run LocalAI: `./local-ai`
+3. Install the required dependencies: `make prepare`
+4. Run LocalAI: `make run`
 
 ## Contributing
 
@@ -60,29 +59,14 @@ If you find a bug, have a feature request, or encounter any issues, please check
 
 `make test` cannot handle all the model now. Please be sure to add a test case for the new features or the part was changed.
 
-### Running AIO tests
-
-All-In-One images has a set of tests that automatically verifies that most of the endpoints works correctly, a flow can be :
-
-```bash
-# Build the LocalAI docker image
-make DOCKER_IMAGE=local-ai docker
-
-# Build the corresponding AIO image
-BASE_IMAGE=local-ai DOCKER_AIO_IMAGE=local-ai-aio:test make docker-aio
-
-# Run the AIO e2e tests
-LOCALAI_IMAGE_TAG=test LOCALAI_IMAGE=local-ai-aio make run-e2e-aio
-```
-
 ## Documentation
 
-We are welcome the contribution of the documents, please open new PR or create a new issue. The documentation is available under `docs/` https://github.com/mudler/LocalAI/tree/master/docs
- 
+- We are welcome the contribution of the documents, please open new PR in the official document repo [localai-website](https://github.com/go-skynet/localai-website)
+
 ## Community and Communication
 
 - You can reach out via the Github issue tracker.
 - Open a new discussion at [Discussion](https://github.com/go-skynet/LocalAI/discussions)
 - Join the Discord channel [Discord](https://discord.gg/uJAeKSAGDy)
 
----
+---

Dockerfile

@@ -1,9 +1,8 @@
 ARG IMAGE_TYPE=extras
 ARG BASE_IMAGE=ubuntu:22.04
-ARG GRPC_BASE_IMAGE=${BASE_IMAGE}
 
 # extras or core
-FROM ${BASE_IMAGE} AS requirements-core
+FROM ${BASE_IMAGE} as requirements-core
 
 USER root
 
@@ -16,30 +15,17 @@ ARG TARGETVARIANT
 
 ENV BUILD_TYPE=${BUILD_TYPE}
 ENV DEBIAN_FRONTEND=noninteractive
-ENV EXTERNAL_GRPC_BACKENDS="coqui:/build/backend/python/coqui/run.sh,huggingface-embeddings:/build/backend/python/sentencetransformers/run.sh,petals:/build/backend/python/petals/run.sh,transformers:/build/backend/python/transformers/run.sh,sentencetransformers:/build/backend/python/sentencetransformers/run.sh,rerankers:/build/backend/python/rerankers/run.sh,autogptq:/build/backend/python/autogptq/run.sh,bark:/build/backend/python/bark/run.sh,diffusers:/build/backend/python/diffusers/run.sh,exllama:/build/backend/python/exllama/run.sh,vall-e-x:/build/backend/python/vall-e-x/run.sh,vllm:/build/backend/python/vllm/run.sh,mamba:/build/backend/python/mamba/run.sh,exllama2:/build/backend/python/exllama2/run.sh,transformers-musicgen:/build/backend/python/transformers-musicgen/run.sh,parler-tts:/build/backend/python/parler-tts/run.sh"
+ENV EXTERNAL_GRPC_BACKENDS="coqui:/build/backend/python/coqui/run.sh,huggingface-embeddings:/build/backend/python/sentencetransformers/run.sh,petals:/build/backend/python/petals/run.sh,transformers:/build/backend/python/transformers/run.sh,sentencetransformers:/build/backend/python/sentencetransformers/run.sh,autogptq:/build/backend/python/autogptq/run.sh,bark:/build/backend/python/bark/run.sh,diffusers:/build/backend/python/diffusers/run.sh,exllama:/build/backend/python/exllama/run.sh,vall-e-x:/build/backend/python/vall-e-x/run.sh,vllm:/build/backend/python/vllm/run.sh,mamba:/build/backend/python/mamba/run.sh,exllama2:/build/backend/python/exllama2/run.sh,transformers-musicgen:/build/backend/python/transformers-musicgen/run.sh"
 
 ARG GO_TAGS="stablediffusion tinydream tts"
 
 RUN apt-get update && \
-    apt-get install -y ca-certificates curl python3-pip unzip && apt-get clean
+    apt-get install -y ca-certificates curl patch pip cmake git && apt-get clean
 
 # Install Go
-RUN curl -L -s https://go.dev/dl/go${GO_VERSION}.linux-${TARGETARCH}.tar.gz | tar -C /usr/local -xz
+RUN curl -L -s https://go.dev/dl/go$GO_VERSION.linux-$TARGETARCH.tar.gz | tar -C /usr/local -xz
 ENV PATH $PATH:/usr/local/go/bin
 
-# Install grpc compilers
-ENV PATH $PATH:/root/go/bin
-RUN go install google.golang.org/protobuf/cmd/protoc-gen-go@latest && \
-    go install google.golang.org/grpc/cmd/protoc-gen-go-grpc@latest
-
-# Install protobuf (the version in 22.04 is too old)
-RUN curl -L -s https://github.com/protocolbuffers/protobuf/releases/download/v26.1/protoc-26.1-linux-x86_64.zip -o protoc.zip && \
-    unzip -j -d /usr/local/bin protoc.zip bin/protoc && \
-    rm protoc.zip
-
-# Install grpcio-tools (the version in 22.04 is too old)
-RUN pip install --user grpcio-tools
-
 COPY --chmod=644 custom-ca-certs/* /usr/local/share/ca-certificates/
 RUN update-ca-certificates
 
@@ -77,13 +63,10 @@ WORKDIR /build
 RUN test -n "$TARGETARCH" \
     || (echo 'warn: missing $TARGETARCH, either set this `ARG` manually, or run using `docker buildkit`')
 
-###################################
-###################################
+# Extras requirements
+FROM requirements-core as requirements-extras
 
-FROM requirements-core AS requirements-extras
-
-RUN apt install -y gpg && \
-    curl https://repo.anaconda.com/pkgs/misc/gpgkeys/anaconda.asc | gpg --dearmor > conda.gpg && \
+RUN curl https://repo.anaconda.com/pkgs/misc/gpgkeys/anaconda.asc | gpg --dearmor > conda.gpg && \
     install -o root -g root -m 644 conda.gpg /usr/share/keyrings/conda-archive-keyring.gpg && \
     gpg --keyring /usr/share/keyrings/conda-archive-keyring.gpg --no-default-keyring --fingerprint 34161F5BF5EB1D4BFBBB8F0A8AEB4F8B29D82806 && \
     echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" > /etc/apt/sources.list.d/conda.list && \
@@ -105,39 +88,13 @@ RUN if [ ! -e /usr/bin/python ]; then \
 ###################################
 ###################################
 
-FROM ${GRPC_BASE_IMAGE} AS grpc
-
-ARG MAKEFLAGS
-ARG GRPC_VERSION=v1.58.0
-
-ENV MAKEFLAGS=${MAKEFLAGS}
-
-WORKDIR /build
-
-RUN apt-get update && \
-    apt-get install -y build-essential cmake git  && \
-    apt-get clean && \
-    rm -rf /var/lib/apt/lists/*
-
-RUN git clone --recurse-submodules --jobs 4 -b ${GRPC_VERSION} --depth 1 --shallow-submodules https://github.com/grpc/grpc
-
-WORKDIR /build/grpc/cmake/build
-
-RUN cmake -DgRPC_INSTALL=ON -DgRPC_BUILD_TESTS=OFF ../.. && \
-    make
-
-###################################
-###################################
-
-FROM requirements-${IMAGE_TYPE} AS builder
+FROM requirements-${IMAGE_TYPE} as builder
 
 ARG GO_TAGS="stablediffusion tts"
 ARG GRPC_BACKENDS
-ARG MAKEFLAGS
-
+ARG BUILD_GRPC=true
 ENV GRPC_BACKENDS=${GRPC_BACKENDS}
 ENV GO_TAGS=${GO_TAGS}
-ENV MAKEFLAGS=${MAKEFLAGS}
 ENV NVIDIA_DRIVER_CAPABILITIES=compute,utility
 ENV NVIDIA_REQUIRE_CUDA="cuda>=${CUDA_MAJOR_VERSION}.0"
 ENV NVIDIA_VISIBLE_DEVICES=all
@@ -146,13 +103,6 @@ WORKDIR /build
 
 COPY . .
 COPY .git .
-RUN echo "GO_TAGS: $GO_TAGS"
-
-RUN apt-get update && \
-    apt-get install -y build-essential cmake git  && \
-    apt-get clean && \
-    rm -rf /var/lib/apt/lists/*
-
 RUN make prepare
 
 # If we are building with clblas support, we need the libraries for the builds
@@ -165,13 +115,14 @@ RUN if [ "${BUILD_TYPE}" = "clblas" ]; then \
 # stablediffusion does not tolerate a newer version of abseil, build it first
 RUN GRPC_BACKENDS=backend-assets/grpc/stablediffusion make build
 
-COPY --from=grpc /build/grpc ./grpc/
-
-WORKDIR /build/grpc/cmake/build
-RUN make install
+RUN if [ "${BUILD_GRPC}" = "true" ]; then \
+    git clone --recurse-submodules -b v1.58.0 --depth 1 --shallow-submodules https://github.com/grpc/grpc && \
+    cd grpc && mkdir -p cmake/build && cd cmake/build && cmake -DgRPC_INSTALL=ON \
+      -DgRPC_BUILD_TESTS=OFF \
+       ../.. && make -j12 install \
+    ; fi
 
 # Rebuild with defaults backends
-WORKDIR /build
 RUN make build
 
 RUN if [ ! -d "/build/sources/go-piper/piper-phonemize/pi/lib/" ]; then \
@@ -188,12 +139,10 @@ ARG FFMPEG
 ARG BUILD_TYPE
 ARG TARGETARCH
 ARG IMAGE_TYPE=extras
-ARG MAKEFLAGS
 
 ENV BUILD_TYPE=${BUILD_TYPE}
 ENV REBUILD=false
 ENV HEALTHCHECK_ENDPOINT=http://localhost:8080/readyz
-ENV MAKEFLAGS=${MAKEFLAGS}
 
 ARG CUDA_MAJOR_VERSION=11
 ENV NVIDIA_DRIVER_CAPABILITIES=compute,utility
@@ -213,11 +162,6 @@ RUN if [ "${BUILD_TYPE}" = "clblas" ]; then \
     apt-get clean \
     ; fi
 
-RUN apt-get update && \
-    apt-get install -y cmake git  && \
-    apt-get clean && \
-    rm -rf /var/lib/apt/lists/*
-
 WORKDIR /build
 
 # we start fresh & re-copy all assets because `make build` does not clean up nicely after itself
@@ -227,9 +171,9 @@ WORKDIR /build
 COPY . .
 
 COPY --from=builder /build/sources ./sources/
-COPY --from=grpc /build/grpc ./grpc/
+COPY --from=builder /build/grpc ./grpc/
 
-RUN make prepare-sources && cd /build/grpc/cmake/build && make install && rm -rf /build/grpc
+RUN make prepare-sources && cd /build/grpc/cmake/build && make install && rm -rf grpc
 
 # Copy the binary
 COPY --from=builder /build/local-ai ./
@@ -242,49 +186,43 @@ COPY --from=builder /build/backend-assets/grpc/stablediffusion ./backend-assets/
 
 ## Duplicated from Makefile to avoid having a big layer that's hard to push
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-    make -C backend/python/autogptq \
+	 make -C backend/python/autogptq \
     ; fi
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-    make -C backend/python/bark \
+	 make -C backend/python/bark \
     ; fi
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-    make -C backend/python/diffusers \
+	 make -C backend/python/diffusers \
     ; fi
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-    make -C backend/python/vllm \
+	 make -C backend/python/vllm \
     ; fi
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-    make -C backend/python/mamba \
+	 make -C backend/python/mamba \
     ; fi
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-    make -C backend/python/sentencetransformers \
+	 make -C backend/python/sentencetransformers \
     ; fi
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-    make -C backend/python/rerankers \
+	 make -C backend/python/transformers \
     ; fi
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-    make -C backend/python/transformers \
+	 make -C backend/python/vall-e-x \
     ; fi
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-    make -C backend/python/vall-e-x \
+	 make -C backend/python/exllama \
     ; fi
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-    make -C backend/python/exllama \
+     make -C backend/python/exllama2 \
     ; fi
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-    make -C backend/python/exllama2 \
+	 make -C backend/python/petals \
     ; fi
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-    make -C backend/python/petals \
+	 make -C backend/python/transformers-musicgen \
     ; fi
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-    make -C backend/python/transformers-musicgen \
-    ; fi
-RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-    make -C backend/python/parler-tts \
-    ; fi
-RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-    make -C backend/python/coqui \
+	 make -C backend/python/coqui \
     ; fi
 
 # Make sure the models directory exists
@@ -292,8 +230,7 @@ RUN mkdir -p /build/models
 
 # Define the health check command
 HEALTHCHECK --interval=1m --timeout=10m --retries=10 \
-  CMD curl -f ${HEALTHCHECK_ENDPOINT} || exit 1
-  
-VOLUME /build/models
+  CMD curl -f $HEALTHCHECK_ENDPOINT || exit 1
+
 EXPOSE 8080
 ENTRYPOINT [ "/build/entrypoint.sh" ]

Dockerfile.aio

@@ -1,8 +0,0 @@
-ARG BASE_IMAGE=ubuntu:22.04
-
-FROM ${BASE_IMAGE} 
-
-RUN apt-get update && apt-get install -y pciutils && apt-get clean
-
-COPY aio/ /aio
-ENTRYPOINT [ "/aio/entrypoint.sh" ]

Makefile

@@ -4,8 +4,11 @@ GOVET=$(GOCMD) vet
 BINARY_NAME=local-ai
 
 # llama.cpp versions
-GOLLAMA_STABLE_VERSION?=2b57a8ae43e4699d3dc5d1496a1ccd42922993be
-CPPLLAMA_VERSION?=784e11dea1f5ce9638851b2b0dddb107e2a609c8
+GOLLAMA_VERSION?=aeba71ee842819da681ea537e78846dc75949ac0
+
+GOLLAMA_STABLE_VERSION?=50cee7712066d9e38306eccadcfbb44ea87df4b7
+
+CPPLLAMA_VERSION?=19885d205e768579ab090d1e99281cae58c21b54
 
 # gpt4all version
 GPT4ALL_REPO?=https://github.com/nomic-ai/gpt4all
@@ -16,26 +19,25 @@ RWKV_REPO?=https://github.com/donomii/go-rwkv.cpp
 RWKV_VERSION?=661e7ae26d442f5cfebd2a0881b44e8c55949ec6
 
 # whisper.cpp version
-WHISPER_CPP_VERSION?=858452d58dba3acdc3431c9bced2bb8cfd9bf418
+WHISPER_CPP_VERSION?=37a709f6558c6d9783199e2b8cbb136e1c41d346
 
 # bert.cpp version
 BERT_VERSION?=6abe312cded14042f6b7c3cd8edf082713334a4d
 
 # go-piper version
-PIPER_VERSION?=9d0100873a7dbb0824dfea40e8cec70a1b110759
+PIPER_VERSION?=d6b6275ba037dabdba4a8b65dfdf6b2a73a67f07
 
 # stablediffusion version
 STABLEDIFFUSION_VERSION?=362df9da29f882dbf09ade61972d16a1f53c3485
 
 # tinydream version
-TINYDREAM_VERSION?=22a12a4bc0ac5455856f28f3b771331a551a4293
+TINYDREAM_VERSION?=772a9c0d9aaf768290e63cca3c904fe69faf677a
 
 export BUILD_TYPE?=
 export STABLE_BUILD_TYPE?=$(BUILD_TYPE)
 export CMAKE_ARGS?=
 
 CGO_LDFLAGS?=
-CGO_LDFLAGS_WHISPER?=
 CUDA_LIBPATH?=/usr/local/cuda/lib64/
 GO_TAGS?=
 BUILD_ID?=git
@@ -70,7 +72,7 @@ UNAME_S := $(shell uname -s)
 endif
 
 ifeq ($(OS),Darwin)
-	
+	CGO_LDFLAGS += -lcblas -framework Accelerate
 	ifeq ($(OSX_SIGNING_IDENTITY),)
 		OSX_SIGNING_IDENTITY := $(shell security find-identity -v -p codesigning | grep '"' | head -n 1 | sed -E 's/.*"(.*)"/\1/')
 	endif
@@ -81,12 +83,6 @@ ifeq ($(OS),Darwin)
 	# disable metal if on Darwin and any other value is explicitly passed.
 	else ifneq ($(BUILD_TYPE),metal)
 		CMAKE_ARGS+=-DLLAMA_METAL=OFF
-		export LLAMA_NO_ACCELERATE=1
-	endif
-
-	ifeq ($(BUILD_TYPE),metal)
-#			-lcblas 	removed: it seems to always be listed as a duplicate flag.
-		CGO_LDFLAGS += -framework Accelerate
 	endif
 endif
 
@@ -95,12 +91,10 @@ ifeq ($(BUILD_TYPE),openblas)
 	export WHISPER_OPENBLAS=1
 endif
 
-
 ifeq ($(BUILD_TYPE),cublas)
 	CGO_LDFLAGS+=-lcublas -lcudart -L$(CUDA_LIBPATH)
 	export LLAMA_CUBLAS=1
 	export WHISPER_CUBLAS=1
-	CGO_LDFLAGS_WHISPER+=-L$(CUDA_LIBPATH)/stubs/ -lcuda
 endif
 
 ifeq ($(BUILD_TYPE),hipblas)
@@ -154,12 +148,12 @@ endif
 
 ALL_GRPC_BACKENDS=backend-assets/grpc/langchain-huggingface
 ALL_GRPC_BACKENDS+=backend-assets/grpc/bert-embeddings
+ALL_GRPC_BACKENDS+=backend-assets/grpc/llama
 ALL_GRPC_BACKENDS+=backend-assets/grpc/llama-cpp
 ALL_GRPC_BACKENDS+=backend-assets/grpc/llama-ggml
 ALL_GRPC_BACKENDS+=backend-assets/grpc/gpt4all
 ALL_GRPC_BACKENDS+=backend-assets/grpc/rwkv
 ALL_GRPC_BACKENDS+=backend-assets/grpc/whisper
-ALL_GRPC_BACKENDS+=backend-assets/grpc/local-store
 ALL_GRPC_BACKENDS+=$(OPTIONAL_GRPC)
 
 GRPC_BACKENDS?=$(ALL_GRPC_BACKENDS) $(OPTIONAL_GRPC)
@@ -174,115 +168,126 @@ ifeq ($(BUILD_API_ONLY),true)
 	GRPC_BACKENDS=
 endif
 
-.PHONY: all test build vendor get-sources prepare-sources prepare
+.PHONY: all test build vendor
 
 all: help
 
-## BERT embeddings
-sources/go-bert.cpp:
-	git clone --recurse-submodules https://github.com/go-skynet/go-bert.cpp sources/go-bert.cpp
-	cd sources/go-bert.cpp && git checkout -b build $(BERT_VERSION) && git submodule update --init --recursive --depth 1
-
-sources/go-bert.cpp/libgobert.a: sources/go-bert.cpp
-	$(MAKE) -C sources/go-bert.cpp libgobert.a
-
-## go-llama.cpp
-sources/go-llama.cpp:
-	git clone --recurse-submodules https://github.com/go-skynet/go-llama.cpp sources/go-llama.cpp
-	cd sources/go-llama.cpp && git checkout -b build $(GOLLAMA_STABLE_VERSION) && git submodule update --init --recursive --depth 1
-
-sources/go-llama.cpp/libbinding.a: sources/go-llama.cpp
-	$(MAKE) -C sources/go-llama.cpp BUILD_TYPE=$(STABLE_BUILD_TYPE) libbinding.a
-
-## go-piper
-sources/go-piper:
-	git clone --recurse-submodules https://github.com/mudler/go-piper sources/go-piper
-	cd sources/go-piper && git checkout -b build $(PIPER_VERSION) && git submodule update --init --recursive --depth 1
-
-sources/go-piper/libpiper_binding.a: sources/go-piper
-	$(MAKE) -C sources/go-piper libpiper_binding.a example/main piper.o
-
 ## GPT4ALL
 sources/gpt4all:
 	git clone --recurse-submodules $(GPT4ALL_REPO) sources/gpt4all
 	cd sources/gpt4all && git checkout -b build $(GPT4ALL_VERSION) && git submodule update --init --recursive --depth 1
 
-sources/gpt4all/gpt4all-bindings/golang/libgpt4all.a: sources/gpt4all
-	$(MAKE) -C sources/gpt4all/gpt4all-bindings/golang/ libgpt4all.a
+## go-piper
+sources/go-piper:
+	git clone --recurse-submodules https://github.com/mudler/go-piper sources/go-piper
+	cd sources/go-piper && git checkout -b build $(PIPER_VERSION) && git submodule update --init --recursive --depth 1
 
-## RWKV
-sources/go-rwkv.cpp:
-	git clone --recurse-submodules $(RWKV_REPO) sources/go-rwkv.cpp
-	cd sources/go-rwkv.cpp && git checkout -b build $(RWKV_VERSION) && git submodule update --init --recursive --depth 1
-
-sources/go-rwkv.cpp/librwkv.a: sources/go-rwkv.cpp
-	cd sources/go-rwkv.cpp && cd rwkv.cpp &&	cmake . -DRWKV_BUILD_SHARED_LIBRARY=OFF &&	cmake --build . && 	cp librwkv.a ..
+## BERT embeddings
+sources/go-bert:
+	git clone --recurse-submodules https://github.com/go-skynet/go-bert.cpp sources/go-bert
+	cd sources/go-bert && git checkout -b build $(BERT_VERSION) && git submodule update --init --recursive --depth 1
 
 ## stable diffusion
 sources/go-stable-diffusion:
 	git clone --recurse-submodules https://github.com/mudler/go-stable-diffusion sources/go-stable-diffusion
 	cd sources/go-stable-diffusion && git checkout -b build $(STABLEDIFFUSION_VERSION) && git submodule update --init --recursive --depth 1
 
-sources/go-stable-diffusion/libstablediffusion.a: sources/go-stable-diffusion
-	CPATH="$(CPATH):/usr/include/opencv4" $(MAKE) -C sources/go-stable-diffusion libstablediffusion.a
+sources/go-stable-diffusion/libstablediffusion.a:
+	$(MAKE) -C sources/go-stable-diffusion libstablediffusion.a
 
 ## tiny-dream
 sources/go-tiny-dream:
 	git clone --recurse-submodules https://github.com/M0Rf30/go-tiny-dream sources/go-tiny-dream
 	cd sources/go-tiny-dream && git checkout -b build $(TINYDREAM_VERSION) && git submodule update --init --recursive --depth 1
 
-sources/go-tiny-dream/libtinydream.a: sources/go-tiny-dream
+sources/go-tiny-dream/libtinydream.a:
 	$(MAKE) -C sources/go-tiny-dream libtinydream.a
 
-## whisper
+## RWKV
+sources/go-rwkv:
+	git clone --recurse-submodules $(RWKV_REPO) sources/go-rwkv
+	cd sources/go-rwkv && git checkout -b build $(RWKV_VERSION) && git submodule update --init --recursive --depth 1
+
+sources/go-rwkv/librwkv.a: sources/go-rwkv
+	cd sources/go-rwkv && cd rwkv.cpp &&	cmake . -DRWKV_BUILD_SHARED_LIBRARY=OFF &&	cmake --build . && 	cp librwkv.a ..
+
+sources/go-bert/libgobert.a: sources/go-bert
+	$(MAKE) -C sources/go-bert libgobert.a
+
+backend-assets/gpt4all: sources/gpt4all/gpt4all-bindings/golang/libgpt4all.a
+	mkdir -p backend-assets/gpt4all
+	@cp sources/gpt4all/gpt4all-bindings/golang/buildllm/*.so backend-assets/gpt4all/ || true
+	@cp sources/gpt4all/gpt4all-bindings/golang/buildllm/*.dylib backend-assets/gpt4all/ || true
+	@cp sources/gpt4all/gpt4all-bindings/golang/buildllm/*.dll backend-assets/gpt4all/ || true
+
+backend-assets/espeak-ng-data: sources/go-piper
+	mkdir -p backend-assets/espeak-ng-data
+	$(MAKE) -C sources/go-piper piper.o
+	@cp -rf sources/go-piper/piper-phonemize/pi/share/espeak-ng-data/. backend-assets/espeak-ng-data
+
+sources/gpt4all/gpt4all-bindings/golang/libgpt4all.a: sources/gpt4all
+	$(MAKE) -C sources/gpt4all/gpt4all-bindings/golang/ libgpt4all.a
+
 sources/whisper.cpp:
-	git clone https://github.com/ggerganov/whisper.cpp sources/whisper.cpp
+	git clone https://github.com/ggerganov/whisper.cpp.git sources/whisper.cpp
 	cd sources/whisper.cpp && git checkout -b build $(WHISPER_CPP_VERSION) && git submodule update --init --recursive --depth 1
 
 sources/whisper.cpp/libwhisper.a: sources/whisper.cpp
 	cd sources/whisper.cpp && make libwhisper.a
 
-get-sources: sources/go-llama.cpp sources/gpt4all sources/go-piper sources/go-rwkv.cpp sources/whisper.cpp sources/go-bert.cpp sources/go-stable-diffusion sources/go-tiny-dream
+sources/go-llama:
+	git clone --recurse-submodules https://github.com/go-skynet/go-llama.cpp sources/go-llama
+	cd sources/go-llama && git checkout -b build $(GOLLAMA_VERSION) && git submodule update --init --recursive --depth 1
+
+sources/go-llama-ggml:
+	git clone --recurse-submodules https://github.com/go-skynet/go-llama.cpp sources/go-llama-ggml
+	cd sources/go-llama-ggml && git checkout -b build $(GOLLAMA_STABLE_VERSION) && git submodule update --init --recursive --depth 1
+
+sources/go-llama/libbinding.a: sources/go-llama
+	$(MAKE) -C sources/go-llama BUILD_TYPE=$(BUILD_TYPE) libbinding.a
+
+sources/go-llama-ggml/libbinding.a: sources/go-llama-ggml
+	$(MAKE) -C sources/go-llama-ggml BUILD_TYPE=$(STABLE_BUILD_TYPE) libbinding.a
+
+sources/go-piper/libpiper_binding.a: sources/go-piper
+	$(MAKE) -C sources/go-piper libpiper_binding.a example/main
+
+backend/cpp/llama/llama.cpp:
+	LLAMA_VERSION=$(CPPLLAMA_VERSION) $(MAKE) -C backend/cpp/llama llama.cpp
+
+get-sources: backend/cpp/llama/llama.cpp sources/go-llama sources/go-llama-ggml sources/gpt4all sources/go-piper sources/go-rwkv sources/whisper.cpp sources/go-bert sources/go-stable-diffusion sources/go-tiny-dream
+	touch $@
 
 replace:
-	$(GOCMD) mod edit -replace github.com/donomii/go-rwkv.cpp=$(CURDIR)/sources/go-rwkv.cpp
+	$(GOCMD) mod edit -replace github.com/nomic-ai/gpt4all/gpt4all-bindings/golang=$(CURDIR)/sources/gpt4all/gpt4all-bindings/golang
+	$(GOCMD) mod edit -replace github.com/donomii/go-rwkv.cpp=$(CURDIR)/sources/go-rwkv
 	$(GOCMD) mod edit -replace github.com/ggerganov/whisper.cpp=$(CURDIR)/sources/whisper.cpp
 	$(GOCMD) mod edit -replace github.com/ggerganov/whisper.cpp/bindings/go=$(CURDIR)/sources/whisper.cpp/bindings/go
-	$(GOCMD) mod edit -replace github.com/go-skynet/go-bert.cpp=$(CURDIR)/sources/go-bert.cpp
+	$(GOCMD) mod edit -replace github.com/go-skynet/go-bert.cpp=$(CURDIR)/sources/go-bert
+	$(GOCMD) mod edit -replace github.com/mudler/go-stable-diffusion=$(CURDIR)/sources/go-stable-diffusion
 	$(GOCMD) mod edit -replace github.com/M0Rf30/go-tiny-dream=$(CURDIR)/sources/go-tiny-dream
 	$(GOCMD) mod edit -replace github.com/mudler/go-piper=$(CURDIR)/sources/go-piper
-	$(GOCMD) mod edit -replace github.com/mudler/go-stable-diffusion=$(CURDIR)/sources/go-stable-diffusion
-	$(GOCMD) mod edit -replace github.com/nomic-ai/gpt4all/gpt4all-bindings/golang=$(CURDIR)/sources/gpt4all/gpt4all-bindings/golang
-	$(GOCMD) mod edit -replace github.com/go-skynet/go-llama.cpp=$(CURDIR)/sources/go-llama.cpp
-
-dropreplace:
-	$(GOCMD) mod edit -dropreplace github.com/donomii/go-rwkv.cpp
-	$(GOCMD) mod edit -dropreplace github.com/ggerganov/whisper.cpp
-	$(GOCMD) mod edit -dropreplace github.com/ggerganov/whisper.cpp/bindings/go
-	$(GOCMD) mod edit -dropreplace github.com/go-skynet/go-bert.cpp
-	$(GOCMD) mod edit -dropreplace github.com/M0Rf30/go-tiny-dream
-	$(GOCMD) mod edit -dropreplace github.com/mudler/go-piper
-	$(GOCMD) mod edit -dropreplace github.com/mudler/go-stable-diffusion
-	$(GOCMD) mod edit -dropreplace github.com/nomic-ai/gpt4all/gpt4all-bindings/golang
-	$(GOCMD) mod edit -dropreplace github.com/go-skynet/go-llama.cpp
 
 prepare-sources: get-sources replace
 	$(GOCMD) mod download
+	touch $@
 
 ## GENERIC
 rebuild: ## Rebuilds the project
 	$(GOCMD) clean -cache
-	$(MAKE) -C sources/go-llama.cpp clean
+	$(MAKE) -C sources/go-llama clean
+	$(MAKE) -C sources/go-llama-ggml clean
 	$(MAKE) -C sources/gpt4all/gpt4all-bindings/golang/ clean
-	$(MAKE) -C sources/go-rwkv.cpp clean
+	$(MAKE) -C sources/go-rwkv clean
 	$(MAKE) -C sources/whisper.cpp clean
 	$(MAKE) -C sources/go-stable-diffusion clean
-	$(MAKE) -C sources/go-bert.cpp clean
+	$(MAKE) -C sources/go-bert clean
 	$(MAKE) -C sources/go-piper clean
 	$(MAKE) -C sources/go-tiny-dream clean
 	$(MAKE) build
 
 prepare: prepare-sources $(OPTIONAL_TARGETS)
+	touch $@
 
 clean: ## Remove build related file
 	$(GOCMD) clean -cache
@@ -290,32 +295,19 @@ clean: ## Remove build related file
 	rm -rf ./sources
 	rm -rf $(BINARY_NAME)
 	rm -rf release/
-	rm -rf backend-assets/*
+	rm -rf backend-assets
 	$(MAKE) -C backend/cpp/grpc clean
 	$(MAKE) -C backend/cpp/llama clean
-	$(MAKE) dropreplace
-	$(MAKE) protogen-clean
-	rmdir pkg/grpc/proto || true
-
-clean-tests:
-	rm -rf test-models
-	rm -rf test-dir
-	rm -rf core/http/backend-assets
 
 ## Build:
-build: prepare backend-assets grpcs ## Build the project
+
+build: backend-assets grpcs prepare ## Build the project
 	$(info ${GREEN}I local-ai build info:${RESET})
 	$(info ${GREEN}I BUILD_TYPE: ${YELLOW}$(BUILD_TYPE)${RESET})
 	$(info ${GREEN}I GO_TAGS: ${YELLOW}$(GO_TAGS)${RESET})
 	$(info ${GREEN}I LD_FLAGS: ${YELLOW}$(LD_FLAGS)${RESET})
 	CGO_LDFLAGS="$(CGO_LDFLAGS)" $(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o $(BINARY_NAME) ./
 
-build-minimal:
-	BUILD_GRPC_FOR_BACKEND_LLAMA=true GRPC_BACKENDS=backend-assets/grpc/llama-cpp GO_TAGS=none $(MAKE) build
-
-build-api:
-	BUILD_GRPC_FOR_BACKEND_LLAMA=true BUILD_API_ONLY=true GO_TAGS=none $(MAKE) build
-
 dist: build
 	mkdir -p release
 	cp $(BINARY_NAME) release/$(BINARY_NAME)-$(BUILD_ID)-$(OS)-$(ARCH)
@@ -327,10 +319,10 @@ osx-signed: build
 run: prepare ## run local-ai
 	CGO_LDFLAGS="$(CGO_LDFLAGS)" $(GOCMD) run ./
 
-test-models/testmodel.ggml:
+test-models/testmodel:
 	mkdir test-models
 	mkdir test-dir
-	wget -q https://huggingface.co/TheBloke/orca_mini_3B-GGML/resolve/main/orca-mini-3b.ggmlv3.q4_0.bin -O test-models/testmodel.ggml
+	wget -q https://huggingface.co/TheBloke/orca_mini_3B-GGML/resolve/main/orca-mini-3b.ggmlv3.q4_0.bin -O test-models/testmodel
 	wget -q https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-base.en.bin -O test-models/whisper-en
 	wget -q https://huggingface.co/mudler/all-MiniLM-L6-v2/resolve/main/ggml-model-q4_0.bin -O test-models/bert
 	wget -q https://cdn.openai.com/whisper/draft-20220913a/micro-machines.wav -O test-dir/audio.wav
@@ -342,9 +334,9 @@ prepare-test: grpcs
 	cp -rf backend-assets core/http
 	cp tests/models_fixtures/* test-models
 
-test: prepare test-models/testmodel.ggml grpcs
+test: prepare test-models/testmodel grpcs
 	@echo 'Running tests'
-	export GO_TAGS="tts stablediffusion debug"
+	export GO_TAGS="tts stablediffusion"
 	$(MAKE) prepare-test
 	HUGGINGFACE_GRPC=$(abspath ./)/backend/python/sentencetransformers/run.sh TEST_DIR=$(abspath ./)/test-dir/ FIXTURES=$(abspath ./)/tests/fixtures CONFIG_FILE=$(abspath ./)/test-models/config.yaml MODELS_PATH=$(abspath ./)/test-models \
 	$(GOCMD) run github.com/onsi/ginkgo/v2/ginkgo --label-filter="!gpt4all && !llama && !llama-gguf"  --flake-attempts $(TEST_FLAKES) --fail-fast -v -r $(TEST_PATHS)
@@ -358,16 +350,12 @@ prepare-e2e:
 	mkdir -p $(TEST_DIR)
 	cp -rfv $(abspath ./tests/e2e-fixtures)/gpu.yaml $(TEST_DIR)/gpu.yaml
 	test -e $(TEST_DIR)/ggllm-test-model.bin || wget -q https://huggingface.co/TheBloke/CodeLlama-7B-Instruct-GGUF/resolve/main/codellama-7b-instruct.Q2_K.gguf -O $(TEST_DIR)/ggllm-test-model.bin
-	docker build --build-arg GRPC_BACKENDS="$(GRPC_BACKENDS)" --build-arg IMAGE_TYPE=core --build-arg BUILD_TYPE=$(BUILD_TYPE) --build-arg CUDA_MAJOR_VERSION=11 --build-arg CUDA_MINOR_VERSION=7 --build-arg FFMPEG=true -t localai-tests .
+	docker build --build-arg BUILD_GRPC=true --build-arg GRPC_BACKENDS="$(GRPC_BACKENDS)" --build-arg IMAGE_TYPE=core --build-arg BUILD_TYPE=$(BUILD_TYPE) --build-arg CUDA_MAJOR_VERSION=11 --build-arg CUDA_MINOR_VERSION=7 --build-arg FFMPEG=true -t localai-tests .
 
 run-e2e-image:
 	ls -liah $(abspath ./tests/e2e-fixtures)
 	docker run -p 5390:8080 -e MODELS_PATH=/models -e THREADS=1 -e DEBUG=true -d --rm -v $(TEST_DIR):/models --gpus all --name e2e-tests-$(RANDOM) localai-tests
 
-run-e2e-aio:
-	@echo 'Running e2e AIO tests'
-	$(GOCMD) run github.com/onsi/ginkgo/v2/ginkgo --flake-attempts 5 -v -r ./tests/e2e-aio
-
 test-e2e:
 	@echo 'Running e2e tests'
 	BUILD_TYPE=$(BUILD_TYPE) \
@@ -398,11 +386,6 @@ test-stablediffusion: prepare-test
 	TEST_DIR=$(abspath ./)/test-dir/ FIXTURES=$(abspath ./)/tests/fixtures CONFIG_FILE=$(abspath ./)/test-models/config.yaml MODELS_PATH=$(abspath ./)/test-models \
 	$(GOCMD) run github.com/onsi/ginkgo/v2/ginkgo --label-filter="stablediffusion" --flake-attempts 1 -v -r $(TEST_PATHS)
 
-test-stores: backend-assets/grpc/local-store
-	mkdir -p tests/integration/backend-assets/grpc
-	cp -f backend-assets/grpc/local-store tests/integration/backend-assets/grpc/
-	$(GOCMD) run github.com/onsi/ginkgo/v2/ginkgo --label-filter="stores" --flake-attempts 1 -v -r tests/integration
-
 test-container:
 	docker build --target requirements -t local-ai-test-container .
 	docker run -ti --rm --entrypoint /bin/bash -ti -v $(abspath ./):/build local-ai-test-container
@@ -419,152 +402,30 @@ help: ## Show this help.
 		else if (/^## .*$$/) {printf "  ${CYAN}%s${RESET}\n", substr($$1,4)} \
 		}' $(MAKEFILE_LIST)
 
-.PHONY: protogen
 protogen: protogen-go protogen-python
 
-.PHONY: protogen-clean
-protogen-clean: protogen-go-clean protogen-python-clean
-
-.PHONY: protogen-go
 protogen-go:
-	mkdir -p pkg/grpc/proto
 	protoc -Ibackend/ --go_out=pkg/grpc/proto/ --go_opt=paths=source_relative --go-grpc_out=pkg/grpc/proto/ --go-grpc_opt=paths=source_relative \
     backend/backend.proto
 
-.PHONY: protogen-go-clean
-protogen-go-clean:
-	$(RM) pkg/grpc/proto/backend.pb.go pkg/grpc/proto/backend_grpc.pb.go
-	$(RM) bin/*
-
-.PHONY: protogen-python
-protogen-python: autogptq-protogen bark-protogen coqui-protogen diffusers-protogen exllama-protogen exllama2-protogen mamba-protogen petals-protogen rerankers-protogen sentencetransformers-protogen transformers-protogen parler-tts-protogen transformers-musicgen-protogen vall-e-x-protogen vllm-protogen
-
-.PHONY: protogen-python-clean
-protogen-python-clean: autogptq-protogen-clean bark-protogen-clean coqui-protogen-clean diffusers-protogen-clean exllama-protogen-clean exllama2-protogen-clean mamba-protogen-clean petals-protogen-clean sentencetransformers-protogen-clean rerankers-protogen-clean transformers-protogen-clean transformers-musicgen-protogen-clean parler-tts-protogen-clean vall-e-x-protogen-clean vllm-protogen-clean
-
-.PHONY: autogptq-protogen
-autogptq-protogen:
-	$(MAKE) -C backend/python/autogptq protogen
-
-.PHONY: autogptq-protogen-clean
-autogptq-protogen-clean:
-	$(MAKE) -C backend/python/autogptq protogen-clean
-
-.PHONY: bark-protogen
-bark-protogen:
-	$(MAKE) -C backend/python/bark protogen
-
-.PHONY: bark-protogen-clean
-bark-protogen-clean:
-	$(MAKE) -C backend/python/bark protogen-clean
-
-.PHONY: coqui-protogen
-coqui-protogen:
-	$(MAKE) -C backend/python/coqui protogen
-
-.PHONY: coqui-protogen-clean
-coqui-protogen-clean:
-	$(MAKE) -C backend/python/coqui protogen-clean
-
-.PHONY: diffusers-protogen
-diffusers-protogen:
-	$(MAKE) -C backend/python/diffusers protogen
-
-.PHONY: diffusers-protogen-clean
-diffusers-protogen-clean:
-	$(MAKE) -C backend/python/diffusers protogen-clean
-
-.PHONY: exllama-protogen
-exllama-protogen:
-	$(MAKE) -C backend/python/exllama protogen
-
-.PHONY: exllama-protogen-clean
-exllama-protogen-clean:
-	$(MAKE) -C backend/python/exllama protogen-clean
-
-.PHONY: exllama2-protogen
-exllama2-protogen:
-	$(MAKE) -C backend/python/exllama2 protogen
-
-.PHONY: exllama2-protogen-clean
-exllama2-protogen-clean:
-	$(MAKE) -C backend/python/exllama2 protogen-clean
-
-.PHONY: mamba-protogen
-mamba-protogen:
-	$(MAKE) -C backend/python/mamba protogen
-
-.PHONY: mamba-protogen-clean
-mamba-protogen-clean:
-	$(MAKE) -C backend/python/mamba protogen-clean
-
-.PHONY: petals-protogen
-petals-protogen:
-	$(MAKE) -C backend/python/petals protogen
-
-.PHONY: petals-protogen-clean
-petals-protogen-clean:
-	$(MAKE) -C backend/python/petals protogen-clean
-
-.PHONY: rerankers-protogen
-rerankers-protogen:
-	$(MAKE) -C backend/python/rerankers protogen
-
-.PHONY: rerankers-protogen-clean
-rerankers-protogen-clean:
-	$(MAKE) -C backend/python/rerankers protogen-clean
-
-.PHONY: sentencetransformers-protogen
-sentencetransformers-protogen:
-	$(MAKE) -C backend/python/sentencetransformers protogen
-
-.PHONY: sentencetransformers-protogen-clean
-sentencetransformers-protogen-clean:
-	$(MAKE) -C backend/python/sentencetransformers protogen-clean
-
-.PHONY: transformers-protogen
-transformers-protogen:
-	$(MAKE) -C backend/python/transformers protogen
-
-.PHONY: transformers-protogen-clean
-transformers-protogen-clean:
-	$(MAKE) -C backend/python/transformers protogen-clean
-
-.PHONY: parler-tts-protogen
-parler-tts-protogen:
-	$(MAKE) -C backend/python/parler-tts protogen
-
-.PHONY: parler-tts-protogen-clean
-parler-tts-protogen-clean:
-	$(MAKE) -C backend/python/parler-tts protogen-clean
-
-.PHONY: transformers-musicgen-protogen
-transformers-musicgen-protogen:
-	$(MAKE) -C backend/python/transformers-musicgen protogen
-
-.PHONY: transformers-musicgen-protogen-clean
-transformers-musicgen-protogen-clean:
-	$(MAKE) -C backend/python/transformers-musicgen protogen-clean
-
-.PHONY: vall-e-x-protogen
-vall-e-x-protogen:
-	$(MAKE) -C backend/python/vall-e-x protogen
-
-.PHONY: vall-e-x-protogen-clean
-vall-e-x-protogen-clean:
-	$(MAKE) -C backend/python/vall-e-x protogen-clean
-
-.PHONY: vllm-protogen
-vllm-protogen:
-	$(MAKE) -C backend/python/vllm protogen
-
-.PHONY: vllm-protogen-clean
-vllm-protogen-clean:
-	$(MAKE) -C backend/python/vllm protogen-clean
+protogen-python:
+	python3 -m grpc_tools.protoc -Ibackend/ --python_out=backend/python/sentencetransformers/ --grpc_python_out=backend/python/sentencetransformers/ backend/backend.proto
+	python3 -m grpc_tools.protoc -Ibackend/ --python_out=backend/python/transformers/ --grpc_python_out=backend/python/transformers/ backend/backend.proto
+	python3 -m grpc_tools.protoc -Ibackend/ --python_out=backend/python/transformers-musicgen/ --grpc_python_out=backend/python/transformers-musicgen/ backend/backend.proto
+	python3 -m grpc_tools.protoc -Ibackend/ --python_out=backend/python/autogptq/ --grpc_python_out=backend/python/autogptq/ backend/backend.proto
+	python3 -m grpc_tools.protoc -Ibackend/ --python_out=backend/python/exllama/ --grpc_python_out=backend/python/exllama/ backend/backend.proto
+	python3 -m grpc_tools.protoc -Ibackend/ --python_out=backend/python/bark/ --grpc_python_out=backend/python/bark/ backend/backend.proto
+	python3 -m grpc_tools.protoc -Ibackend/ --python_out=backend/python/diffusers/ --grpc_python_out=backend/python/diffusers/ backend/backend.proto
+	python3 -m grpc_tools.protoc -Ibackend/ --python_out=backend/python/coqui/ --grpc_python_out=backend/python/coqui/ backend/backend.proto
+	python3 -m grpc_tools.protoc -Ibackend/ --python_out=backend/python/vall-e-x/ --grpc_python_out=backend/python/vall-e-x/ backend/backend.proto
+	python3 -m grpc_tools.protoc -Ibackend/ --python_out=backend/python/vllm/ --grpc_python_out=backend/python/vllm/ backend/backend.proto
+	python3 -m grpc_tools.protoc -Ibackend/ --python_out=backend/python/petals/ --grpc_python_out=backend/python/petals/ backend/backend.proto
+	python3 -m grpc_tools.protoc -Ibackend/ --python_out=backend/python/mamba/ --grpc_python_out=backend/python/mamba/ backend/backend.proto
+	python3 -m grpc_tools.protoc -Ibackend/ --python_out=backend/python/exllama2/ --grpc_python_out=backend/python/exllama2/ backend/backend.proto
 
 ## GRPC
 # Note: it is duplicated in the Dockerfile
-prepare-extra-conda-environments: protogen-python
+prepare-extra-conda-environments:
 	$(MAKE) -C backend/python/autogptq
 	$(MAKE) -C backend/python/bark
 	$(MAKE) -C backend/python/coqui
@@ -572,16 +433,14 @@ prepare-extra-conda-environments: protogen-python
 	$(MAKE) -C backend/python/vllm
 	$(MAKE) -C backend/python/mamba
 	$(MAKE) -C backend/python/sentencetransformers
-	$(MAKE) -C backend/python/rerankers
 	$(MAKE) -C backend/python/transformers
 	$(MAKE) -C backend/python/transformers-musicgen
-	$(MAKE) -C backend/python/parler-tts
 	$(MAKE) -C backend/python/vall-e-x
 	$(MAKE) -C backend/python/exllama
 	$(MAKE) -C backend/python/petals
 	$(MAKE) -C backend/python/exllama2
 
-prepare-test-extra: protogen-python
+prepare-test-extra:
 	$(MAKE) -C backend/python/transformers
 	$(MAKE) -C backend/python/diffusers
 
@@ -595,93 +454,92 @@ ifeq ($(BUILD_API_ONLY),true)
 	touch backend-assets/keep
 endif
 
-backend-assets/espeak-ng-data: sources/go-piper sources/go-piper/libpiper_binding.a
-	mkdir -p backend-assets/espeak-ng-data
-	@cp -rf sources/go-piper/piper-phonemize/pi/share/espeak-ng-data/. backend-assets/espeak-ng-data
-
-backend-assets/gpt4all: sources/gpt4all sources/gpt4all/gpt4all-bindings/golang/libgpt4all.a
-	mkdir -p backend-assets/gpt4all
-	@cp sources/gpt4all/gpt4all-bindings/golang/buildllm/*.so backend-assets/gpt4all/ || true
-	@cp sources/gpt4all/gpt4all-bindings/golang/buildllm/*.dylib backend-assets/gpt4all/ || true
-	@cp sources/gpt4all/gpt4all-bindings/golang/buildllm/*.dll backend-assets/gpt4all/ || true
-
-backend-assets/grpc: protogen-go replace
+backend-assets/grpc:
 	mkdir -p backend-assets/grpc
 
-backend-assets/grpc/bert-embeddings: sources/go-bert.cpp sources/go-bert.cpp/libgobert.a backend-assets/grpc
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(CURDIR)/sources/go-bert.cpp LIBRARY_PATH=$(CURDIR)/sources/go-bert.cpp \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/bert-embeddings ./backend/go/llm/bert/
-
-backend-assets/grpc/gpt4all: sources/gpt4all sources/gpt4all/gpt4all-bindings/golang/libgpt4all.a backend-assets/gpt4all backend-assets/grpc
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(CURDIR)/sources/gpt4all/gpt4all-bindings/golang/ LIBRARY_PATH=$(CURDIR)/sources/gpt4all/gpt4all-bindings/golang/ \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/gpt4all ./backend/go/llm/gpt4all/
-
-backend-assets/grpc/langchain-huggingface: backend-assets/grpc
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/langchain-huggingface ./backend/go/llm/langchain/
-
-backend/cpp/llama/llama.cpp:
-	LLAMA_VERSION=$(CPPLLAMA_VERSION) $(MAKE) -C backend/cpp/llama llama.cpp
+backend-assets/grpc/llama: backend-assets/grpc sources/go-llama/libbinding.a
+	$(GOCMD) mod edit -replace github.com/go-skynet/go-llama.cpp=$(CURDIR)/sources/go-llama
+	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(CURDIR)/sources/go-llama LIBRARY_PATH=$(CURDIR)/sources/go-llama \
+	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/llama ./backend/go/llm/llama/
+# TODO: every binary should have its own folder instead, so can have different  implementations
+ifeq ($(BUILD_TYPE),metal)
+	cp backend/cpp/llama/llama.cpp/ggml-metal.metal backend-assets/grpc/
+endif
 
+## BACKEND CPP LLAMA START
+# Sets the variables in case it has to build the gRPC locally.
 INSTALLED_PACKAGES=$(CURDIR)/backend/cpp/grpc/installed_packages
 INSTALLED_LIB_CMAKE=$(INSTALLED_PACKAGES)/lib/cmake
 ADDED_CMAKE_ARGS=-Dabsl_DIR=${INSTALLED_LIB_CMAKE}/absl \
-				 -DProtobuf_DIR=${INSTALLED_LIB_CMAKE}/protobuf \
-				 -Dutf8_range_DIR=${INSTALLED_LIB_CMAKE}/utf8_range \
-				 -DgRPC_DIR=${INSTALLED_LIB_CMAKE}/grpc \
-				 -DCMAKE_CXX_STANDARD_INCLUDE_DIRECTORIES=${INSTALLED_PACKAGES}/include
+                 -DProtobuf_DIR=${INSTALLED_LIB_CMAKE}/protobuf \
+                 -Dutf8_range_DIR=${INSTALLED_LIB_CMAKE}/utf8_range \
+                 -DgRPC_DIR=${INSTALLED_LIB_CMAKE}/grpc \
+                 -DCMAKE_CXX_STANDARD_INCLUDE_DIRECTORIES=${INSTALLED_PACKAGES}/include
+
 backend/cpp/llama/grpc-server:
-# Conditionally build grpc for the llama backend to use if needed
 ifdef BUILD_GRPC_FOR_BACKEND_LLAMA
 	$(MAKE) -C backend/cpp/grpc build
-	_PROTOBUF_PROTOC=${INSTALLED_PACKAGES}/bin/proto \
-	_GRPC_CPP_PLUGIN_EXECUTABLE=${INSTALLED_PACKAGES}/bin/grpc_cpp_plugin \
-	PATH="${INSTALLED_PACKAGES}/bin:${PATH}" \
-	CMAKE_ARGS="${CMAKE_ARGS} ${ADDED_CMAKE_ARGS}" \
-	LLAMA_VERSION=$(CPPLLAMA_VERSION) \
-	$(MAKE) -C backend/cpp/llama grpc-server
+	export _PROTOBUF_PROTOC=${INSTALLED_PACKAGES}/bin/proto && \
+	export _GRPC_CPP_PLUGIN_EXECUTABLE=${INSTALLED_PACKAGES}/bin/grpc_cpp_plugin && \
+	export PATH="${INSTALLED_PACKAGES}/bin:${PATH}" && \
+	CMAKE_ARGS="${CMAKE_ARGS} ${ADDED_CMAKE_ARGS}" LLAMA_VERSION=$(CPPLLAMA_VERSION) $(MAKE) -C backend/cpp/llama grpc-server
 else
 	echo "BUILD_GRPC_FOR_BACKEND_LLAMA is not defined."
 	LLAMA_VERSION=$(CPPLLAMA_VERSION) $(MAKE) -C backend/cpp/llama grpc-server
 endif
+## BACKEND CPP LLAMA END
 
+##
 backend-assets/grpc/llama-cpp: backend-assets/grpc backend/cpp/llama/grpc-server
 	cp -rfv backend/cpp/llama/grpc-server backend-assets/grpc/llama-cpp
 # TODO: every binary should have its own folder instead, so can have different metal implementations
 ifeq ($(BUILD_TYPE),metal)
-	cp backend/cpp/llama/llama.cpp/build/bin/default.metallib backend-assets/grpc/
+	cp backend/cpp/llama/llama.cpp/build/bin/ggml-metal.metal backend-assets/grpc/
 endif
 
-backend-assets/grpc/llama-ggml: sources/go-llama.cpp sources/go-llama.cpp/libbinding.a backend-assets/grpc
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(CURDIR)/sources/go-llama.cpp LIBRARY_PATH=$(CURDIR)/sources/go-llama.cpp \
+backend-assets/grpc/llama-ggml: backend-assets/grpc sources/go-llama-ggml/libbinding.a
+	$(GOCMD) mod edit -replace github.com/go-skynet/go-llama.cpp=$(CURDIR)/sources/go-llama-ggml
+	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(CURDIR)/sources/go-llama-ggml LIBRARY_PATH=$(CURDIR)/sources/go-llama-ggml \
 	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/llama-ggml ./backend/go/llm/llama-ggml/
 
-backend-assets/grpc/piper: sources/go-piper sources/go-piper/libpiper_binding.a backend-assets/grpc backend-assets/espeak-ng-data
-	CGO_CXXFLAGS="$(PIPER_CGO_CXXFLAGS)" CGO_LDFLAGS="$(PIPER_CGO_LDFLAGS)" LIBRARY_PATH=$(CURDIR)/sources/go-piper \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/piper ./backend/go/tts/
+backend-assets/grpc/gpt4all: backend-assets/grpc backend-assets/gpt4all sources/gpt4all/gpt4all-bindings/golang/libgpt4all.a
+	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(CURDIR)/sources/gpt4all/gpt4all-bindings/golang/ LIBRARY_PATH=$(CURDIR)/sources/gpt4all/gpt4all-bindings/golang/ \
+	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/gpt4all ./backend/go/llm/gpt4all/
 
-backend-assets/grpc/rwkv: sources/go-rwkv.cpp sources/go-rwkv.cpp/librwkv.a backend-assets/grpc
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(CURDIR)/sources/go-rwkv.cpp LIBRARY_PATH=$(CURDIR)/sources/go-rwkv.cpp \
+backend-assets/grpc/rwkv: backend-assets/grpc sources/go-rwkv/librwkv.a
+	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(CURDIR)/sources/go-rwkv LIBRARY_PATH=$(CURDIR)/sources/go-rwkv \
 	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/rwkv ./backend/go/llm/rwkv
 
-backend-assets/grpc/stablediffusion: sources/go-stable-diffusion sources/go-stable-diffusion/libstablediffusion.a backend-assets/grpc
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" CPATH="$(CPATH):$(CURDIR)/sources/go-stable-diffusion/:/usr/include/opencv4" LIBRARY_PATH=$(CURDIR)/sources/go-stable-diffusion/ \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/stablediffusion ./backend/go/image/stablediffusion
+backend-assets/grpc/bert-embeddings: backend-assets/grpc sources/go-bert/libgobert.a
+	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(CURDIR)/sources/go-bert LIBRARY_PATH=$(CURDIR)/sources/go-bert \
+	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/bert-embeddings ./backend/go/llm/bert/
 
-backend-assets/grpc/tinydream: sources/go-tiny-dream sources/go-tiny-dream/libtinydream.a backend-assets/grpc
+backend-assets/grpc/langchain-huggingface: backend-assets/grpc
+	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/langchain-huggingface ./backend/go/llm/langchain/
+
+backend-assets/grpc/stablediffusion: backend-assets/grpc
+	if [ ! -f backend-assets/grpc/stablediffusion ]; then \
+		$(MAKE) sources/go-stable-diffusion; \
+		$(MAKE) sources/go-stable-diffusion/libstablediffusion.a; \
+		CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(CURDIR)/sources/go-stable-diffusion/ LIBRARY_PATH=$(CURDIR)/sources/go-stable-diffusion/ \
+		$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/stablediffusion ./backend/go/image/stablediffusion; \
+	fi
+
+backend-assets/grpc/tinydream: backend-assets/grpc sources/go-tiny-dream/libtinydream.a
 	CGO_LDFLAGS="$(CGO_LDFLAGS)" LIBRARY_PATH=$(CURDIR)/go-tiny-dream \
 	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/tinydream ./backend/go/image/tinydream
 
-backend-assets/grpc/whisper: sources/whisper.cpp sources/whisper.cpp/libwhisper.a backend-assets/grpc
-	CGO_LDFLAGS="$(CGO_LDFLAGS) $(CGO_LDFLAGS_WHISPER)" C_INCLUDE_PATH=$(CURDIR)/sources/whisper.cpp LIBRARY_PATH=$(CURDIR)/sources/whisper.cpp \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/whisper ./backend/go/transcribe/
+backend-assets/grpc/piper: backend-assets/grpc backend-assets/espeak-ng-data sources/go-piper/libpiper_binding.a
+	CGO_CXXFLAGS="$(PIPER_CGO_CXXFLAGS)" CGO_LDFLAGS="$(PIPER_CGO_LDFLAGS)" LIBRARY_PATH=$(CURDIR)/sources/go-piper \
+	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/piper ./backend/go/tts/
 
-backend-assets/grpc/local-store: backend-assets/grpc
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/local-store ./backend/go/stores/
+backend-assets/grpc/whisper: backend-assets/grpc sources/whisper.cpp/libwhisper.a
+	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(CURDIR)/sources/whisper.cpp LIBRARY_PATH=$(CURDIR)/sources/whisper.cpp \
+	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/whisper ./backend/go/transcribe/
 
 grpcs: prepare $(GRPC_BACKENDS)
 
 DOCKER_IMAGE?=local-ai
-DOCKER_AIO_IMAGE?=local-ai-aio
 IMAGE_TYPE?=core
 BASE_IMAGE?=ubuntu:22.04
 
@@ -689,28 +547,15 @@ docker:
 	docker build \
 		--build-arg BASE_IMAGE=$(BASE_IMAGE) \
 		--build-arg IMAGE_TYPE=$(IMAGE_TYPE) \
-		--build-arg GO_TAGS="$(GO_TAGS)" \
-		--build-arg MAKEFLAGS="$(DOCKER_MAKEFLAGS)" \
+		--build-arg GO_TAGS=$(GO_TAGS) \
 		--build-arg BUILD_TYPE=$(BUILD_TYPE) \
 		-t $(DOCKER_IMAGE) .
-	
-docker-aio:
-	@echo "Building AIO image with base $(BASE_IMAGE) as $(DOCKER_AIO_IMAGE)"
-	docker build \
-		--build-arg BASE_IMAGE=$(BASE_IMAGE) \
-		--build-arg MAKEFLAGS="$(DOCKER_MAKEFLAGS)" \
-		-t $(DOCKER_AIO_IMAGE) -f Dockerfile.aio .
-
-docker-aio-all:
-	$(MAKE) docker-aio DOCKER_AIO_SIZE=cpu
-	$(MAKE) docker-aio DOCKER_AIO_SIZE=cpu
 
 docker-image-intel:
 	docker build \
 		--build-arg BASE_IMAGE=intel/oneapi-basekit:2024.0.1-devel-ubuntu22.04 \
 		--build-arg IMAGE_TYPE=$(IMAGE_TYPE) \
 		--build-arg GO_TAGS="none" \
-		--build-arg MAKEFLAGS="$(DOCKER_MAKEFLAGS)" \
 		--build-arg BUILD_TYPE=sycl_f32 -t $(DOCKER_IMAGE) .
 
 docker-image-intel-xpu:
@@ -718,9 +563,4 @@ docker-image-intel-xpu:
 		--build-arg BASE_IMAGE=intel/oneapi-basekit:2024.0.1-devel-ubuntu22.04 \
 		--build-arg IMAGE_TYPE=$(IMAGE_TYPE) \
 		--build-arg GO_TAGS="none" \
-		--build-arg MAKEFLAGS="$(DOCKER_MAKEFLAGS)" \
-		--build-arg BUILD_TYPE=sycl_f32 -t $(DOCKER_IMAGE) .
-
-.PHONY: swagger
-swagger:
-	swag init -g core/http/app.go --output swagger
+		--build-arg BUILD_TYPE=sycl_f32 -t $(DOCKER_IMAGE) .

README.md

@@ -20,14 +20,14 @@
 </a>
 </p>
 
-<p align="center">
-<a href="https://hub.docker.com/r/localai/localai" target="blank">
-<img src="https://img.shields.io/badge/dockerhub-images-important.svg?logo=Docker" alt="LocalAI Docker hub"/>
-</a>
-<a href="https://quay.io/repository/go-skynet/local-ai?tab=tags&tag=latest" target="blank">
-<img src="https://img.shields.io/badge/quay.io-images-important.svg?" alt="LocalAI Quay.io"/>
-</a>
-</p>
+[<img src="https://img.shields.io/badge/dockerhub-images-important.svg?logo=Docker">](https://hub.docker.com/r/localai/localai)
+[<img src="https://img.shields.io/badge/quay.io-images-important.svg?">](https://quay.io/repository/go-skynet/local-ai?tab=tags&tag=latest)
+
+> :bulb: Get help - [❓FAQ](https://localai.io/faq/) [💭Discussions](https://github.com/go-skynet/LocalAI/discussions) [:speech_balloon: Discord](https://discord.gg/uJAeKSAGDy) [:book: Documentation website](https://localai.io/)
+>
+> [💻 Quickstart](https://localai.io/basics/getting_started/) [📣 News](https://localai.io/basics/news/) [ 🛫 Examples ](https://github.com/go-skynet/LocalAI/tree/master/examples/) [ 🖼️ Models ](https://localai.io/models/) [ 🚀 Roadmap ](https://github.com/mudler/LocalAI/issues?q=is%3Aissue+is%3Aopen+label%3Aroadmap)
+
+[![tests](https://github.com/go-skynet/LocalAI/actions/workflows/test.yml/badge.svg)](https://github.com/go-skynet/LocalAI/actions/workflows/test.yml)[![Build and Release](https://github.com/go-skynet/LocalAI/actions/workflows/release.yaml/badge.svg)](https://github.com/go-skynet/LocalAI/actions/workflows/release.yaml)[![build container images](https://github.com/go-skynet/LocalAI/actions/workflows/image.yml/badge.svg)](https://github.com/go-skynet/LocalAI/actions/workflows/image.yml)[![Bump dependencies](https://github.com/go-skynet/LocalAI/actions/workflows/bump_deps.yaml/badge.svg)](https://github.com/go-skynet/LocalAI/actions/workflows/bump_deps.yaml)[![Artifact Hub](https://img.shields.io/endpoint?url=https://artifacthub.io/badge/repository/localai)](https://artifacthub.io/packages/search?repo=localai)
 
 <p align="center">
 <a href="https://twitter.com/LocalAI_API" target="blank">
@@ -36,27 +36,24 @@
 <a href="https://discord.gg/uJAeKSAGDy" target="blank">
 <img src="https://dcbadge.vercel.app/api/server/uJAeKSAGDy?style=flat-square&theme=default-inverted" alt="Join LocalAI Discord Community"/>
 </a>
-</p>
 
-> :bulb: Get help - [❓FAQ](https://localai.io/faq/) [💭Discussions](https://github.com/go-skynet/LocalAI/discussions) [:speech_balloon: Discord](https://discord.gg/uJAeKSAGDy) [:book: Documentation website](https://localai.io/)
->
-> [💻 Quickstart](https://localai.io/basics/getting_started/) [📣 News](https://localai.io/basics/news/) [ 🛫 Examples ](https://github.com/go-skynet/LocalAI/tree/master/examples/) [ 🖼️ Models ](https://localai.io/models/) [ 🚀 Roadmap ](https://github.com/mudler/LocalAI/issues?q=is%3Aissue+is%3Aopen+label%3Aroadmap)
-
-[![tests](https://github.com/go-skynet/LocalAI/actions/workflows/test.yml/badge.svg)](https://github.com/go-skynet/LocalAI/actions/workflows/test.yml)[![Build and Release](https://github.com/go-skynet/LocalAI/actions/workflows/release.yaml/badge.svg)](https://github.com/go-skynet/LocalAI/actions/workflows/release.yaml)[![build container images](https://github.com/go-skynet/LocalAI/actions/workflows/image.yml/badge.svg)](https://github.com/go-skynet/LocalAI/actions/workflows/image.yml)[![Bump dependencies](https://github.com/go-skynet/LocalAI/actions/workflows/bump_deps.yaml/badge.svg)](https://github.com/go-skynet/LocalAI/actions/workflows/bump_deps.yaml)[![Artifact Hub](https://img.shields.io/endpoint?url=https://artifacthub.io/badge/repository/localai)](https://artifacthub.io/packages/search?repo=localai)
-
-**LocalAI** is the free, Open Source OpenAI alternative. LocalAI act as a drop-in replacement REST API that’s compatible with OpenAI (Elevenlabs, Anthropic... ) API specifications for local AI inferencing. It allows you to run LLMs, generate images, audio (and not only) locally or on-prem with consumer grade hardware, supporting multiple model families. Does not require GPU. It is created and maintained by [Ettore Di Giacinto](https://github.com/mudler).
+**LocalAI** is the free, Open Source OpenAI alternative. LocalAI act as a drop-in replacement REST API that’s compatible with OpenAI API specifications for local inferencing. It allows you to run LLMs, generate images, audio (and not only) locally or on-prem with consumer grade hardware, supporting multiple model families. Does not require GPU.
 
 ## 🔥🔥 Hot topics / Roadmap
 
 [Roadmap](https://github.com/mudler/LocalAI/issues?q=is%3Aissue+is%3Aopen+label%3Aroadmap)
 
-- Reranker API: https://github.com/mudler/LocalAI/pull/2121
-- Gallery WebUI: https://github.com/mudler/LocalAI/pull/2104
-- llama3: https://github.com/mudler/LocalAI/discussions/2076
-- Parler-TTS: https://github.com/mudler/LocalAI/pull/2027
-- Openvino support: https://github.com/mudler/LocalAI/pull/1892
-- Vector store: https://github.com/mudler/LocalAI/pull/1795
-- All-in-one container image: https://github.com/mudler/LocalAI/issues/1855
+- Parallel function calling: https://github.com/mudler/LocalAI/pull/1726
+- Upload file API: https://github.com/mudler/LocalAI/pull/1703
+- Tools API support: https://github.com/mudler/LocalAI/pull/1715
+- LLaVa 1.6: https://github.com/mudler/LocalAI/pull/1714
+- ROCm container images: https://github.com/mudler/LocalAI/pull/1595
+- Intel GPU support (sycl, transformers, diffusers): https://github.com/mudler/LocalAI/issues/1653
+- Deprecation of old backends: https://github.com/mudler/LocalAI/issues/1651
+- Mamba support: https://github.com/mudler/LocalAI/pull/1589
+- Start and share models with config file: https://github.com/mudler/LocalAI/pull/1522
+- 🐸 Coqui: https://github.com/mudler/LocalAI/pull/1489
+- Img2vid https://github.com/mudler/LocalAI/pull/1442
 
 Hot topics (looking for contributors):
 - Backends v2: https://github.com/mudler/LocalAI/issues/1126
@@ -69,14 +66,10 @@ If you want to help and contribute, issues up for grabs: https://github.com/mudl
 
 ## 💻 [Getting started](https://localai.io/basics/getting_started/index.html)
 
-For a detailed step-by-step introduction, refer to the [Getting Started](https://localai.io/basics/getting_started/index.html) guide. 
+For a detailed step-by-step introduction, refer to the [Getting Started](https://localai.io/basics/getting_started/index.html) guide. For those in a hurry, here's a straightforward one-liner to launch a LocalAI instance with [phi-2](https://huggingface.co/microsoft/phi-2) using `docker`:
 
-For those in a hurry, here's a straightforward one-liner to launch a LocalAI AIO(All-in-one) Image using `docker`:
-
-```bash
-docker run -ti --name local-ai -p 8080:8080 localai/localai:latest-aio-cpu
-# or, if you have an Nvidia GPU:
-# docker run -ti --name local-ai -p 8080:8080 --gpus all localai/localai:latest-aio-gpu-nvidia-cuda-12
+```
+docker run -ti -p 8080:8080 localai/localai:v2.9.0-ffmpeg-core phi-2
 ```
 
 ## 🚀 [Features](https://localai.io/features/)

aio/cpu/README.md

@@ -1,5 +0,0 @@
-## AIO CPU size
-
-Use this image with CPU-only.
-
-Please keep using only C++ backends so the base image is as small as possible (without CUDA, cuDNN, python, etc).

aio/cpu/embeddings.yaml

@@ -1,12 +0,0 @@
-name: text-embedding-ada-002
-backend: bert-embeddings
-parameters:
-  model: huggingface://mudler/all-MiniLM-L6-v2/ggml-model-q4_0.bin
-
-usage: |
-    You can test this model with curl like this:
-
-    curl http://localhost:8080/embeddings -X POST -H "Content-Type: application/json" -d '{
-      "input": "Your text string goes here",
-      "model": "text-embedding-ada-002"
-    }'

aio/cpu/image-gen.yaml

@@ -1,62 +0,0 @@
-name: stablediffusion
-backend: stablediffusion
-parameters:
-  model: stablediffusion_assets
-
-license: "BSD-3"
-urls:
-- https://github.com/EdVince/Stable-Diffusion-NCNN
-- https://github.com/EdVince/Stable-Diffusion-NCNN/blob/main/LICENSE
-
-description: |
-     Stable Diffusion in NCNN with c++, supported txt2img and img2img
-
-download_files:
-- filename: "stablediffusion_assets/AutoencoderKL-256-256-fp16-opt.param"
-  sha256: "18ca4b66685e21406bcf64c484b3b680b4949900415536d599cc876579c85c82"
-  uri: "https://raw.githubusercontent.com/EdVince/Stable-Diffusion-NCNN/main/x86/linux/assets/AutoencoderKL-256-256-fp16-opt.param"
-- filename: "stablediffusion_assets/AutoencoderKL-512-512-fp16-opt.param"
-  sha256: "cf45f63aacf3dbbab0f59ed92a6f2c14d9a1801314631cd3abe91e3c85639a20"
-  uri: "https://raw.githubusercontent.com/EdVince/Stable-Diffusion-NCNN/main/x86/linux/assets/AutoencoderKL-512-512-fp16-opt.param"
-- filename: "stablediffusion_assets/AutoencoderKL-base-fp16.param"
-  sha256: "0254a056dce61b0c27dc9ec1b78b53bcf55315c540f55f051eb841aa992701ba"
-  uri: "https://raw.githubusercontent.com/EdVince/Stable-Diffusion-NCNN/main/x86/linux/assets/AutoencoderKL-base-fp16.param"
-- filename: "stablediffusion_assets/AutoencoderKL-encoder-512-512-fp16.bin"
-  sha256: "ddcb79a9951b9f91e05e087739ed69da2c1c4ae30ba4168cce350b49d617c9fa"
-  uri: "https://github.com/EdVince/Stable-Diffusion-NCNN/releases/download/naifu/AutoencoderKL-encoder-512-512-fp16.bin"
-- filename: "stablediffusion_assets/AutoencoderKL-fp16.bin"
-  sha256: "f02e71f80e70252734724bbfaed5c4ddd3a8ed7e61bb2175ff5f53099f0e35dd"
-  uri: "https://github.com/EdVince/Stable-Diffusion-NCNN/releases/download/naifu/AutoencoderKL-fp16.bin"
-- filename: "stablediffusion_assets/FrozenCLIPEmbedder-fp16.bin"
-  sha256: "1c9a12f4e1dd1b295a388045f7f28a2352a4d70c3dc96a542189a3dd7051fdd6"
-  uri: "https://github.com/EdVince/Stable-Diffusion-NCNN/releases/download/naifu/FrozenCLIPEmbedder-fp16.bin"
-- filename: "stablediffusion_assets/FrozenCLIPEmbedder-fp16.param"
-  sha256: "471afbe678dd1fd3fe764ef9c6eccaccb0a7d7e601f27b462aa926b20eb368c9"
-  uri: "https://raw.githubusercontent.com/EdVince/Stable-Diffusion-NCNN/main/x86/linux/assets/FrozenCLIPEmbedder-fp16.param"
-- filename: "stablediffusion_assets/log_sigmas.bin"
-  sha256: "a2089f8aa4c61f9c200feaec541ab3f5c94233b28deb6d5e8bcd974fa79b68ac"
-  uri: "https://github.com/EdVince/Stable-Diffusion-NCNN/raw/main/x86/linux/assets/log_sigmas.bin"
-- filename: "stablediffusion_assets/UNetModel-256-256-MHA-fp16-opt.param"
-  sha256: "a58c380229f09491776df837b7aa7adffc0a87821dc4708b34535da2e36e3da1"
-  uri: "https://raw.githubusercontent.com/EdVince/Stable-Diffusion-NCNN/main/x86/linux/assets/UNetModel-256-256-MHA-fp16-opt.param"
-- filename: "stablediffusion_assets/UNetModel-512-512-MHA-fp16-opt.param"
-  sha256: "f12034067062827bd7f43d1d21888d1f03905401acf6c6eea22be23c259636fa"
-  uri: "https://raw.githubusercontent.com/EdVince/Stable-Diffusion-NCNN/main/x86/linux/assets/UNetModel-512-512-MHA-fp16-opt.param"
-- filename: "stablediffusion_assets/UNetModel-base-MHA-fp16.param"
-  sha256: "696f6975de49f4325b53ce32aff81861a6d6c07cd9ce3f0aae2cc405350af38d"
-  uri: "https://raw.githubusercontent.com/EdVince/Stable-Diffusion-NCNN/main/x86/linux/assets/UNetModel-base-MHA-fp16.param"
-- filename: "stablediffusion_assets/UNetModel-MHA-fp16.bin"
-  sha256: "d618918d011bfc1f644c0f2a33bf84931bd53b28a98492b0a8ed6f3a818852c3"
-  uri: "https://github.com/EdVince/Stable-Diffusion-NCNN/releases/download/naifu/UNetModel-MHA-fp16.bin"
-- filename: "stablediffusion_assets/vocab.txt"
-  sha256: "e30e57b6f1e47616982ef898d8922be24e535b4fa3d0110477b3a6f02ebbae7d"
-  uri: "https://raw.githubusercontent.com/EdVince/Stable-Diffusion-NCNN/main/x86/linux/assets/vocab.txt"
-
-usage: |
-        curl http://localhost:8080/v1/images/generations \
-          -H "Content-Type: application/json" \
-          -d '{
-            "prompt": "<positive prompt>|<negative prompt>",
-            "step": 25,
-            "size": "512x512"
-          }'

aio/cpu/rerank.yaml

@@ -1,27 +0,0 @@
-name: jina-reranker-v1-base-en
-backend: rerankers
-parameters:
-  model: cross-encoder
-
-usage: |
-    You can test this model with curl like this:
-
-    curl http://localhost:8080/v1/rerank \
-      -H "Content-Type: application/json" \
-      -d '{
-      "model": "jina-reranker-v1-base-en",
-      "query": "Organic skincare products for sensitive skin",
-      "documents": [
-        "Eco-friendly kitchenware for modern homes",
-        "Biodegradable cleaning supplies for eco-conscious consumers",
-        "Organic cotton baby clothes for sensitive skin",
-        "Natural organic skincare range for sensitive skin",
-        "Tech gadgets for smart homes: 2024 edition",
-        "Sustainable gardening tools and compost solutions",
-        "Sensitive skin-friendly facial cleansers and toners",
-        "Organic food wraps and storage solutions",
-        "All-natural pet food for dogs with allergies",
-        "Yoga mats made from recycled materials"
-      ],
-      "top_n": 3
-    }'

aio/cpu/speech-to-text.yaml

@@ -1,18 +0,0 @@
-name: whisper-1
-backend: whisper
-parameters:
-  model: ggml-whisper-base.bin
-
-usage: |
-    ## example audio file
-    wget --quiet --show-progress -O gb1.ogg https://upload.wikimedia.org/wikipedia/commons/1/1f/George_W_Bush_Columbia_FINAL.ogg
-
-    ## Send the example audio file to the transcriptions endpoint
-    curl http://localhost:8080/v1/audio/transcriptions \
-         -H "Content-Type: multipart/form-data" \
-         -F file="@$PWD/gb1.ogg" -F model="whisper-1"
-
-download_files:
-- filename: "ggml-whisper-base.bin"
-  sha256: "60ed5bc3dd14eea856493d334349b405782ddcaf0028d4b5df4088345fba2efe"
-  uri: "https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-base.bin"

aio/cpu/text-to-speech.yaml

@@ -1,15 +0,0 @@
-name: tts-1
-download_files:
-  - filename: voice-en-us-amy-low.tar.gz
-    uri: https://github.com/rhasspy/piper/releases/download/v0.0.2/voice-en-us-amy-low.tar.gz
-
-parameters:
-  model: en-us-amy-low.onnx
-
-usage: |
-    To test if this model works as expected, you can use the following curl command:
-
-    curl http://localhost:8080/tts -H "Content-Type: application/json" -d '{
-      "model":"voice-en-us-amy-low",
-      "input": "Hi, this is a test."
-    }'

aio/cpu/text-to-text.yaml

@@ -1,59 +0,0 @@
-name: gpt-4
-mmap: true
-parameters:
-  model: huggingface://NousResearch/Hermes-2-Pro-Mistral-7B-GGUF/Hermes-2-Pro-Mistral-7B.Q2_K.gguf
-
-template:
-  chat_message: |
-    <|im_start|>{{if eq .RoleName "assistant"}}assistant{{else if eq .RoleName "system"}}system{{else if eq .RoleName "tool"}}tool{{else if eq .RoleName "user"}}user{{end}}
-    {{- if .FunctionCall }}
-    <tool_call>
-    {{- else if eq .RoleName "tool" }}
-    <tool_response>
-    {{- end }}
-    {{- if .Content}}
-    {{.Content }}
-    {{- end }}
-    {{- if .FunctionCall}}
-    {{toJson .FunctionCall}}
-    {{- end }}
-    {{- if .FunctionCall }}
-    </tool_call>
-    {{- else if eq .RoleName "tool" }}
-    </tool_response>
-    {{- end }}<|im_end|>
-  # https://huggingface.co/NousResearch/Hermes-2-Pro-Mistral-7B-GGUF#prompt-format-for-function-calling
-  function: |
-    <|im_start|>system
-    You are a function calling AI model. You are provided with function signatures within <tools></tools> XML tags. You may call one or more functions to assist with the user query. Don't make assumptions about what values to plug into functions. Here are the available tools:
-    <tools>
-    {{range .Functions}}
-    {'type': 'function', 'function': {'name': '{{.Name}}', 'description': '{{.Description}}', 'parameters': {{toJson .Parameters}} }}
-    {{end}}
-    </tools>
-    Use the following pydantic model json schema for each tool call you will make:
-    {'title': 'FunctionCall', 'type': 'object', 'properties': {'arguments': {'title': 'Arguments', 'type': 'object'}, 'name': {'title': 'Name', 'type': 'string'}}, 'required': ['arguments', 'name']}
-    For each function call return a json object with function name and arguments within <tool_call></tool_call> XML tags as follows:
-    <tool_call>
-    {'arguments': <args-dict>, 'name': <function-name>}
-    </tool_call><|im_end|>
-    {{.Input -}}
-    <|im_start|>assistant
-    <tool_call>
-  chat: |
-    {{.Input -}}
-    <|im_start|>assistant
-  completion: |
-    {{.Input}}
-context_size: 4096
-f16: true
-stopwords:
-- <|im_end|>
-- <dummy32000>
-- "\n</tool_call>"
-- "\n\n\n"
-usage: |
-      curl http://localhost:8080/v1/chat/completions -H "Content-Type: application/json" -d '{
-          "model": "gpt-4",
-          "messages": [{"role": "user", "content": "How are you doing?", "temperature": 0.1}]
-      }'

aio/cpu/vision.yaml

@@ -1,31 +0,0 @@
-backend: llama-cpp
-context_size: 4096
-f16: true
-mmap: true
-name: gpt-4-vision-preview
-
-roles:
-  user: "USER:"
-  assistant: "ASSISTANT:"
-  system: "SYSTEM:"
-
-mmproj: bakllava-mmproj.gguf
-parameters:
-  model: bakllava.gguf
-
-template:
-  chat: |
-    A chat between a curious human and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the human's questions.
-    {{.Input}}
-    ASSISTANT:
-
-download_files:
-- filename: bakllava.gguf
-  uri: huggingface://mys/ggml_bakllava-1/ggml-model-q4_k.gguf
-- filename: bakllava-mmproj.gguf
-  uri: huggingface://mys/ggml_bakllava-1/mmproj-model-f16.gguf
-
-usage: |
-    curl http://localhost:8080/v1/chat/completions -H "Content-Type: application/json" -d '{
-        "model": "gpt-4-vision-preview",
-        "messages": [{"role": "user", "content": [{"type":"text", "text": "What is in the image?"}, {"type": "image_url", "image_url": {"url": "https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg" }}], "temperature": 0.9}]}'

aio/entrypoint.sh

@@ -1,138 +0,0 @@
-#!/bin/bash
-
-echo "===> LocalAI All-in-One (AIO) container starting..."
-
-GPU_ACCELERATION=false
-GPU_VENDOR=""
-
-function check_intel() {
-    if lspci | grep -E 'VGA|3D' | grep -iq intel; then
-        echo "Intel GPU detected"
-        if [ -d /opt/intel ]; then
-            GPU_ACCELERATION=true
-            GPU_VENDOR=intel
-        else
-            echo "Intel GPU detected, but Intel GPU drivers are not installed. GPU acceleration will not be available."
-        fi
-    fi
-}
-
-function check_nvidia_wsl() {
-    if lspci | grep -E 'VGA|3D' | grep -iq "Microsoft Corporation Device 008e"; then
-        # We make the assumption this WSL2 cars is NVIDIA, then check for nvidia-smi
-        # Make sure the container was run with `--gpus all` as the only required parameter
-        echo "NVIDIA GPU detected via WSL2"
-        # nvidia-smi should be installed in the container
-        if nvidia-smi; then
-            GPU_ACCELERATION=true
-            GPU_VENDOR=nvidia
-        else
-            echo "NVIDIA GPU detected via WSL2, but nvidia-smi is not installed. GPU acceleration will not be available."
-        fi
-    fi
-}
-
-function check_amd() {
-    if lspci | grep -E 'VGA|3D' | grep -iq amd; then
-        echo "AMD GPU detected"
-        # Check if ROCm is installed
-        if [ -d /opt/rocm ]; then
-            GPU_ACCELERATION=true
-            GPU_VENDOR=amd
-        else
-            echo "AMD GPU detected, but ROCm is not installed. GPU acceleration will not be available."
-        fi
-    fi
-}
-
-function check_nvidia() {
-    if lspci | grep -E 'VGA|3D' | grep -iq nvidia; then
-        echo "NVIDIA GPU detected"
-        # nvidia-smi should be installed in the container
-        if nvidia-smi; then
-            GPU_ACCELERATION=true
-            GPU_VENDOR=nvidia
-        else
-            echo "NVIDIA GPU detected, but nvidia-smi is not installed. GPU acceleration will not be available."
-        fi
-    fi
-}
-
-function check_metal() {
-    if system_profiler SPDisplaysDataType | grep -iq 'Metal'; then
-        echo "Apple Metal supported GPU detected"
-        GPU_ACCELERATION=true
-        GPU_VENDOR=apple
-    fi
-}
-
-function detect_gpu() {
-    case "$(uname -s)" in
-        Linux)
-            check_nvidia
-            check_amd
-            check_intel
-            check_nvidia_wsl
-            ;;
-        Darwin)
-            check_metal
-            ;;
-    esac
-}
-
-function detect_gpu_size() {
-    # Attempting to find GPU memory size for NVIDIA GPUs
-    if [ "$GPU_ACCELERATION" = true ] && [ "$GPU_VENDOR" = "nvidia" ]; then
-        echo "NVIDIA GPU detected. Attempting to find memory size..."
-        # Using head -n 1 to get the total memory of the 1st NVIDIA GPU detected.
-        # If handling multiple GPUs is required in the future, this is the place to do it
-        nvidia_sm=$(nvidia-smi --query-gpu=memory.total --format=csv,noheader,nounits | head -n 1)
-        if [ ! -z "$nvidia_sm" ]; then
-            echo "Total GPU Memory: $nvidia_sm MiB"
-            # if bigger than 8GB, use 16GB
-            #if [ "$nvidia_sm" -gt 8192 ]; then
-            #    GPU_SIZE=gpu-16g
-            #else
-            GPU_SIZE=gpu-8g
-            #fi
-        else
-            echo "Unable to determine NVIDIA GPU memory size. Falling back to CPU."
-            GPU_SIZE=gpu-8g
-        fi
-    elif [ "$GPU_ACCELERATION" = true ] && [ "$GPU_VENDOR" = "intel" ]; then
-        GPU_SIZE=intel
-    # Default to a generic GPU size until we implement GPU size detection for non NVIDIA GPUs
-    elif [ "$GPU_ACCELERATION" = true ]; then
-        echo "Non-NVIDIA GPU detected. Specific GPU memory size detection is not implemented."
-        GPU_SIZE=gpu-8g
-
-    # default to cpu if GPU_SIZE is not set
-    else
-        echo "GPU acceleration is not enabled or supported. Defaulting to CPU."
-        GPU_SIZE=cpu
-    fi
-}
-
-function check_vars() {
-    if [ -z "$MODELS" ]; then
-        echo "MODELS environment variable is not set. Please set it to a comma-separated list of model YAML files to load."
-        exit 1
-    fi
-
-    if [ -z "$PROFILE" ]; then
-        echo "PROFILE environment variable is not set. Please set it to one of the following: cpu, gpu-8g, gpu-16g, apple"
-        exit 1
-    fi
-}
-
-detect_gpu
-detect_gpu_size
-
-PROFILE="${PROFILE:-$GPU_SIZE}" # default to cpu
-export MODELS="${MODELS:-/aio/${PROFILE}/embeddings.yaml,/aio/${PROFILE}/rerank.yaml,/aio/${PROFILE}/text-to-speech.yaml,/aio/${PROFILE}/image-gen.yaml,/aio/${PROFILE}/text-to-text.yaml,/aio/${PROFILE}/speech-to-text.yaml,/aio/${PROFILE}/vision.yaml}"
-
-check_vars
-
-echo "===> Starting LocalAI[$PROFILE] with the following models: $MODELS"
-
-exec /build/entrypoint.sh "$@"

aio/gpu-8g/embeddings.yaml

@@ -1,12 +0,0 @@
-name: text-embedding-ada-002
-backend: sentencetransformers
-parameters:
-  model: all-MiniLM-L6-v2
-
-usage: |
-    You can test this model with curl like this:
-
-    curl http://localhost:8080/embeddings -X POST -H "Content-Type: application/json" -d '{
-      "input": "Your text string goes here",
-      "model": "text-embedding-ada-002"
-    }'

aio/gpu-8g/image-gen.yaml

@@ -1,25 +0,0 @@
-name: stablediffusion
-parameters:
-  model: DreamShaper_8_pruned.safetensors
-backend: diffusers
-step: 25
-f16: true
-
-diffusers:
-  pipeline_type: StableDiffusionPipeline
-  cuda: true
-  enable_parameters: "negative_prompt,num_inference_steps"
-  scheduler_type: "k_dpmpp_2m"
-
-download_files:
-- filename: DreamShaper_8_pruned.safetensors
-  uri: huggingface://Lykon/DreamShaper/DreamShaper_8_pruned.safetensors
-
-usage: |
-        curl http://localhost:8080/v1/images/generations \
-          -H "Content-Type: application/json" \
-          -d '{
-            "prompt": "<positive prompt>|<negative prompt>",
-            "step": 25,
-            "size": "512x512"
-          }'

aio/gpu-8g/rerank.yaml

@@ -1,27 +0,0 @@
-name: jina-reranker-v1-base-en
-backend: rerankers
-parameters:
-  model: cross-encoder
-
-usage: |
-    You can test this model with curl like this:
-
-    curl http://localhost:8080/v1/rerank \
-      -H "Content-Type: application/json" \
-      -d '{
-      "model": "jina-reranker-v1-base-en",
-      "query": "Organic skincare products for sensitive skin",
-      "documents": [
-        "Eco-friendly kitchenware for modern homes",
-        "Biodegradable cleaning supplies for eco-conscious consumers",
-        "Organic cotton baby clothes for sensitive skin",
-        "Natural organic skincare range for sensitive skin",
-        "Tech gadgets for smart homes: 2024 edition",
-        "Sustainable gardening tools and compost solutions",
-        "Sensitive skin-friendly facial cleansers and toners",
-        "Organic food wraps and storage solutions",
-        "All-natural pet food for dogs with allergies",
-        "Yoga mats made from recycled materials"
-      ],
-      "top_n": 3
-    }'

aio/gpu-8g/speech-to-text.yaml

@@ -1,18 +0,0 @@
-name: whisper-1
-backend: whisper
-parameters:
-  model: ggml-whisper-base.bin
-
-usage: |
-    ## example audio file
-    wget --quiet --show-progress -O gb1.ogg https://upload.wikimedia.org/wikipedia/commons/1/1f/George_W_Bush_Columbia_FINAL.ogg
-
-    ## Send the example audio file to the transcriptions endpoint
-    curl http://localhost:8080/v1/audio/transcriptions \
-         -H "Content-Type: multipart/form-data" \
-         -F file="@$PWD/gb1.ogg" -F model="whisper-1"
-
-download_files:
-- filename: "ggml-whisper-base.bin"
-  sha256: "60ed5bc3dd14eea856493d334349b405782ddcaf0028d4b5df4088345fba2efe"
-  uri: "https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-base.bin"

aio/gpu-8g/text-to-speech.yaml

@@ -1,15 +0,0 @@
-name: tts-1
-download_files:
-  - filename: voice-en-us-amy-low.tar.gz
-    uri: https://github.com/rhasspy/piper/releases/download/v0.0.2/voice-en-us-amy-low.tar.gz
-
-parameters:
-  model: en-us-amy-low.onnx
-
-usage: |
-    To test if this model works as expected, you can use the following curl command:
-
-    curl http://localhost:8080/tts -H "Content-Type: application/json" -d '{
-      "model":"tts-1",
-      "input": "Hi, this is a test."
-    }'

aio/gpu-8g/text-to-text.yaml

@@ -1,59 +0,0 @@
-name: gpt-4
-mmap: true
-parameters:
-  model: huggingface://NousResearch/Hermes-2-Pro-Mistral-7B-GGUF/Hermes-2-Pro-Mistral-7B.Q6_K.gguf
-
-template:
-  chat_message: |
-    <|im_start|>{{if eq .RoleName "assistant"}}assistant{{else if eq .RoleName "system"}}system{{else if eq .RoleName "tool"}}tool{{else if eq .RoleName "user"}}user{{end}}
-    {{- if .FunctionCall }}
-    <tool_call>
-    {{- else if eq .RoleName "tool" }}
-    <tool_response>
-    {{- end }}
-    {{- if .Content}}
-    {{.Content }}
-    {{- end }}
-    {{- if .FunctionCall}}
-    {{toJson .FunctionCall}}
-    {{- end }}
-    {{- if .FunctionCall }}
-    </tool_call>
-    {{- else if eq .RoleName "tool" }}
-    </tool_response>
-    {{- end }}<|im_end|>
-  # https://huggingface.co/NousResearch/Hermes-2-Pro-Mistral-7B-GGUF#prompt-format-for-function-calling
-  function: |
-    <|im_start|>system
-    You are a function calling AI model. You are provided with function signatures within <tools></tools> XML tags. You may call one or more functions to assist with the user query. Don't make assumptions about what values to plug into functions. Here are the available tools:
-    <tools>
-    {{range .Functions}}
-    {'type': 'function', 'function': {'name': '{{.Name}}', 'description': '{{.Description}}', 'parameters': {{toJson .Parameters}} }}
-    {{end}}
-    </tools>
-    Use the following pydantic model json schema for each tool call you will make:
-    {'title': 'FunctionCall', 'type': 'object', 'properties': {'arguments': {'title': 'Arguments', 'type': 'object'}, 'name': {'title': 'Name', 'type': 'string'}}, 'required': ['arguments', 'name']}
-    For each function call return a json object with function name and arguments within <tool_call></tool_call> XML tags as follows:
-    <tool_call>
-    {'arguments': <args-dict>, 'name': <function-name>}
-    </tool_call><|im_end|>
-    {{.Input -}}
-    <|im_start|>assistant
-    <tool_call>
-  chat: |
-    {{.Input -}}
-    <|im_start|>assistant
-  completion: |
-    {{.Input}}
-context_size: 4096
-f16: true
-stopwords:
-- <|im_end|>
-- <dummy32000>
-- "\n</tool_call>"
-- "\n\n\n"
-usage: |
-      curl http://localhost:8080/v1/chat/completions -H "Content-Type: application/json" -d '{
-          "model": "gpt-4",
-          "messages": [{"role": "user", "content": "How are you doing?", "temperature": 0.1}]
-      }'

aio/gpu-8g/vision.yaml

@@ -1,35 +0,0 @@
-backend: llama-cpp
-context_size: 4096
-f16: true
-mmap: true
-name: gpt-4-vision-preview
-
-roles:
-  user: "USER:"
-  assistant: "ASSISTANT:"
-  system: "SYSTEM:"
-
-mmproj: llava-v1.6-7b-mmproj-f16.gguf
-parameters:
-  model: llava-v1.6-mistral-7b.Q5_K_M.gguf
-  temperature: 0.2
-  top_k: 40
-  top_p: 0.95
-  seed: -1
-
-template:
-  chat: |
-    A chat between a curious human and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the human's questions.
-    {{.Input}}
-    ASSISTANT:
-
-download_files:
-- filename: llava-v1.6-mistral-7b.Q5_K_M.gguf
-  uri: huggingface://cjpais/llava-1.6-mistral-7b-gguf/llava-v1.6-mistral-7b.Q5_K_M.gguf
-- filename: llava-v1.6-7b-mmproj-f16.gguf
-  uri: huggingface://cjpais/llava-1.6-mistral-7b-gguf/mmproj-model-f16.gguf
-
-usage: |
-    curl http://localhost:8080/v1/chat/completions -H "Content-Type: application/json" -d '{
-        "model": "gpt-4-vision-preview",
-        "messages": [{"role": "user", "content": [{"type":"text", "text": "What is in the image?"}, {"type": "image_url", "image_url": {"url": "https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg" }}], "temperature": 0.9}]}'

aio/intel/embeddings.yaml

@@ -1,12 +0,0 @@
-name: text-embedding-ada-002
-backend: sentencetransformers
-parameters:
-  model: all-MiniLM-L6-v2
-
-usage: |
-    You can test this model with curl like this:
-
-    curl http://localhost:8080/embeddings -X POST -H "Content-Type: application/json" -d '{
-      "input": "Your text string goes here",
-      "model": "text-embedding-ada-002"
-    }'

aio/intel/image-gen.yaml

@@ -1,20 +0,0 @@
-name: stablediffusion
-parameters:
-  model: runwayml/stable-diffusion-v1-5
-backend: diffusers
-step: 25
-f16: true
-diffusers:
-  pipeline_type: StableDiffusionPipeline
-  cuda: true
-  enable_parameters: "negative_prompt,num_inference_steps"
-  scheduler_type: "k_dpmpp_2m"
-
-usage: |
-        curl http://localhost:8080/v1/images/generations \
-          -H "Content-Type: application/json" \
-          -d '{
-            "prompt": "<positive prompt>|<negative prompt>",
-            "step": 25,
-            "size": "512x512"
-          }'

aio/intel/rerank.yaml

@@ -1,27 +0,0 @@
-name: jina-reranker-v1-base-en
-backend: rerankers
-parameters:
-  model: cross-encoder
-
-usage: |
-    You can test this model with curl like this:
-
-    curl http://localhost:8080/v1/rerank \
-      -H "Content-Type: application/json" \
-      -d '{
-      "model": "jina-reranker-v1-base-en",
-      "query": "Organic skincare products for sensitive skin",
-      "documents": [
-        "Eco-friendly kitchenware for modern homes",
-        "Biodegradable cleaning supplies for eco-conscious consumers",
-        "Organic cotton baby clothes for sensitive skin",
-        "Natural organic skincare range for sensitive skin",
-        "Tech gadgets for smart homes: 2024 edition",
-        "Sustainable gardening tools and compost solutions",
-        "Sensitive skin-friendly facial cleansers and toners",
-        "Organic food wraps and storage solutions",
-        "All-natural pet food for dogs with allergies",
-        "Yoga mats made from recycled materials"
-      ],
-      "top_n": 3
-    }'

aio/intel/speech-to-text.yaml

@@ -1,18 +0,0 @@
-name: whisper-1
-backend: whisper
-parameters:
-  model: ggml-whisper-base.bin
-
-usage: |
-    ## example audio file
-    wget --quiet --show-progress -O gb1.ogg https://upload.wikimedia.org/wikipedia/commons/1/1f/George_W_Bush_Columbia_FINAL.ogg
-
-    ## Send the example audio file to the transcriptions endpoint
-    curl http://localhost:8080/v1/audio/transcriptions \
-         -H "Content-Type: multipart/form-data" \
-         -F file="@$PWD/gb1.ogg" -F model="whisper-1"
-
-download_files:
-- filename: "ggml-whisper-base.bin"
-  sha256: "60ed5bc3dd14eea856493d334349b405782ddcaf0028d4b5df4088345fba2efe"
-  uri: "https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-base.bin"

aio/intel/text-to-speech.yaml

@@ -1,15 +0,0 @@
-name: tts-1
-download_files:
-  - filename: voice-en-us-amy-low.tar.gz
-    uri: https://github.com/rhasspy/piper/releases/download/v0.0.2/voice-en-us-amy-low.tar.gz
-
-parameters:
-  model: en-us-amy-low.onnx
-
-usage: |
-    To test if this model works as expected, you can use the following curl command:
-
-    curl http://localhost:8080/tts -H "Content-Type: application/json" -d '{
-      "model":"tts-1",
-      "input": "Hi, this is a test."
-    }'

aio/intel/text-to-text.yaml

@@ -1,59 +0,0 @@
-name: gpt-4
-mmap: false
-f16: false
-parameters:
-  model: huggingface://NousResearch/Hermes-2-Pro-Mistral-7B-GGUF/Hermes-2-Pro-Mistral-7B.Q6_K.gguf
-
-template:
-  chat_message: |
-    <|im_start|>{{if eq .RoleName "assistant"}}assistant{{else if eq .RoleName "system"}}system{{else if eq .RoleName "tool"}}tool{{else if eq .RoleName "user"}}user{{end}}
-    {{- if .FunctionCall }}
-    <tool_call>
-    {{- else if eq .RoleName "tool" }}
-    <tool_response>
-    {{- end }}
-    {{- if .Content}}
-    {{.Content }}
-    {{- end }}
-    {{- if .FunctionCall}}
-    {{toJson .FunctionCall}}
-    {{- end }}
-    {{- if .FunctionCall }}
-    </tool_call>
-    {{- else if eq .RoleName "tool" }}
-    </tool_response>
-    {{- end }}<|im_end|>
-  # https://huggingface.co/NousResearch/Hermes-2-Pro-Mistral-7B-GGUF#prompt-format-for-function-calling
-  function: |
-    <|im_start|>system
-    You are a function calling AI model. You are provided with function signatures within <tools></tools> XML tags. You may call one or more functions to assist with the user query. Don't make assumptions about what values to plug into functions. Here are the available tools:
-    <tools>
-    {{range .Functions}}
-    {'type': 'function', 'function': {'name': '{{.Name}}', 'description': '{{.Description}}', 'parameters': {{toJson .Parameters}} }}
-    {{end}}
-    </tools>
-    Use the following pydantic model json schema for each tool call you will make:
-    {'title': 'FunctionCall', 'type': 'object', 'properties': {'arguments': {'title': 'Arguments', 'type': 'object'}, 'name': {'title': 'Name', 'type': 'string'}}, 'required': ['arguments', 'name']}
-    For each function call return a json object with function name and arguments within <tool_call></tool_call> XML tags as follows:
-    <tool_call>
-    {'arguments': <args-dict>, 'name': <function-name>}
-    </tool_call><|im_end|>
-    {{.Input -}}
-    <|im_start|>assistant
-    <tool_call>
-  chat: |
-    {{.Input -}}
-    <|im_start|>assistant
-  completion: |
-    {{.Input}}
-context_size: 4096
-stopwords:
-- <|im_end|>
-- "\n</tool_call>"
-- <dummy32000>
-- "\n\n\n"
-usage: |
-      curl http://localhost:8080/v1/chat/completions -H "Content-Type: application/json" -d '{
-          "model": "gpt-4",
-          "messages": [{"role": "user", "content": "How are you doing?", "temperature": 0.1}]
-      }'

aio/intel/vision.yaml

@@ -1,35 +0,0 @@
-backend: llama-cpp
-context_size: 4096
-mmap: false
-f16: false
-name: gpt-4-vision-preview
-
-roles:
-  user: "USER:"
-  assistant: "ASSISTANT:"
-  system: "SYSTEM:"
-
-mmproj: llava-v1.6-7b-mmproj-f16.gguf
-parameters:
-  model: llava-v1.6-mistral-7b.Q5_K_M.gguf
-  temperature: 0.2
-  top_k: 40
-  top_p: 0.95
-  seed: -1
-
-template:
-  chat: |
-    A chat between a curious human and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the human's questions.
-    {{.Input}}
-    ASSISTANT:
-
-download_files:
-- filename: llava-v1.6-mistral-7b.Q5_K_M.gguf
-  uri: huggingface://cjpais/llava-1.6-mistral-7b-gguf/llava-v1.6-mistral-7b.Q5_K_M.gguf
-- filename: llava-v1.6-7b-mmproj-f16.gguf
-  uri: huggingface://cjpais/llava-1.6-mistral-7b-gguf/mmproj-model-f16.gguf
-
-usage: |
-    curl http://localhost:8080/v1/chat/completions -H "Content-Type: application/json" -d '{
-        "model": "gpt-4-vision-preview",
-        "messages": [{"role": "user", "content": [{"type":"text", "text": "What is in the image?"}, {"type": "image_url", "image_url": {"url": "https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg" }}], "temperature": 0.9}]}'

backend/backend.proto

@@ -18,72 +18,6 @@ service Backend {
   rpc TTS(TTSRequest) returns (Result) {}
   rpc TokenizeString(PredictOptions) returns (TokenizationResponse) {}
   rpc Status(HealthMessage) returns (StatusResponse) {}
-
-  rpc StoresSet(StoresSetOptions) returns (Result) {}
-  rpc StoresDelete(StoresDeleteOptions) returns (Result) {}
-  rpc StoresGet(StoresGetOptions) returns (StoresGetResult) {}
-  rpc StoresFind(StoresFindOptions) returns (StoresFindResult) {}
-
-  rpc Rerank(RerankRequest) returns (RerankResult) {}
-}
-
-message RerankRequest {
-  string query = 1;
-  repeated string documents = 2;
-  int32 top_n = 3;
-}
-
-message RerankResult {
-  Usage usage = 1;
-  repeated DocumentResult results = 2;
-}
-
-message Usage {
-  int32 total_tokens = 1;
-  int32 prompt_tokens = 2;
-}
-
-message DocumentResult {
-  int32 index = 1;
-  string text = 2;
-  float relevance_score = 3;
-}
-
-message StoresKey {
-  repeated float Floats = 1;
-}
-
-message StoresValue {
-  bytes Bytes = 1;
-}
-
-message StoresSetOptions {
-  repeated StoresKey Keys = 1;
-  repeated StoresValue Values = 2;
-}
-
-message StoresDeleteOptions {
-  repeated StoresKey Keys = 1;
-}
-
-message StoresGetOptions {
-  repeated StoresKey Keys = 1;
-}
-
-message StoresGetResult {
-  repeated StoresKey Keys = 1;
-  repeated StoresValue Values = 2;
-}
-
-message StoresFindOptions {
-  StoresKey Key = 1;
-  int32 TopK = 2;
-}
-
-message StoresFindResult {
-  repeated StoresKey Keys = 1;
-  repeated StoresValue Values = 2;
-  repeated float Similarities = 3;
 }
 
 message HealthMessage {}
@@ -131,15 +65,11 @@ message PredictOptions {
   string NegativePrompt = 40;
   int32 NDraft = 41;
   repeated string Images = 42;
-  bool UseTokenizerTemplate = 43;
-  repeated Message Messages = 44;
 }
 
 // The response message containing the result
 message Reply {
   bytes message = 1;
-  int32 tokens = 2;
-  int32 prompt_tokens = 3;
 }
 
 message ModelOptions {
@@ -191,7 +121,7 @@ message ModelOptions {
 
   bool NoMulMatQ = 37;
   string DraftModel = 39;
-
+  
   string AudioPath = 38;
 
   // vllm
@@ -201,7 +131,6 @@ message ModelOptions {
   bool   EnforceEager = 52;
   int32  SwapSpace = 53;
   int32  MaxModelLen = 54;
-  int32  TensorParallelSize = 55;
 
   string MMProj = 41;
 
@@ -284,9 +213,4 @@ message StatusResponse {
   }
   State state = 1;
   MemoryUsageData memory = 2;
-}
-
-message Message {
-  string role = 1;
-  string content = 2;
 }

backend/backend_grpc.pb.go

@@ -0,0 +1,457 @@
+// Code generated by protoc-gen-go-grpc. DO NOT EDIT.
+// versions:
+// - protoc-gen-go-grpc v1.2.0
+// - protoc             v4.23.4
+// source: backend/backend.proto
+
+package proto
+
+import (
+	context "context"
+	grpc "google.golang.org/grpc"
+	codes "google.golang.org/grpc/codes"
+	status "google.golang.org/grpc/status"
+)
+
+// This is a compile-time assertion to ensure that this generated file
+// is compatible with the grpc package it is being compiled against.
+// Requires gRPC-Go v1.32.0 or later.
+const _ = grpc.SupportPackageIsVersion7
+
+// BackendClient is the client API for Backend service.
+//
+// For semantics around ctx use and closing/ending streaming RPCs, please refer to https://pkg.go.dev/google.golang.org/grpc/?tab=doc#ClientConn.NewStream.
+type BackendClient interface {
+	Health(ctx context.Context, in *HealthMessage, opts ...grpc.CallOption) (*Reply, error)
+	Predict(ctx context.Context, in *PredictOptions, opts ...grpc.CallOption) (*Reply, error)
+	LoadModel(ctx context.Context, in *ModelOptions, opts ...grpc.CallOption) (*Result, error)
+	PredictStream(ctx context.Context, in *PredictOptions, opts ...grpc.CallOption) (Backend_PredictStreamClient, error)
+	Embedding(ctx context.Context, in *PredictOptions, opts ...grpc.CallOption) (*EmbeddingResult, error)
+	GenerateImage(ctx context.Context, in *GenerateImageRequest, opts ...grpc.CallOption) (*Result, error)
+	AudioTranscription(ctx context.Context, in *TranscriptRequest, opts ...grpc.CallOption) (*TranscriptResult, error)
+	TTS(ctx context.Context, in *TTSRequest, opts ...grpc.CallOption) (*Result, error)
+	TokenizeString(ctx context.Context, in *PredictOptions, opts ...grpc.CallOption) (*TokenizationResponse, error)
+	Status(ctx context.Context, in *HealthMessage, opts ...grpc.CallOption) (*StatusResponse, error)
+}
+
+type backendClient struct {
+	cc grpc.ClientConnInterface
+}
+
+func NewBackendClient(cc grpc.ClientConnInterface) BackendClient {
+	return &backendClient{cc}
+}
+
+func (c *backendClient) Health(ctx context.Context, in *HealthMessage, opts ...grpc.CallOption) (*Reply, error) {
+	out := new(Reply)
+	err := c.cc.Invoke(ctx, "/backend.Backend/Health", in, out, opts...)
+	if err != nil {
+		return nil, err
+	}
+	return out, nil
+}
+
+func (c *backendClient) Predict(ctx context.Context, in *PredictOptions, opts ...grpc.CallOption) (*Reply, error) {
+	out := new(Reply)
+	err := c.cc.Invoke(ctx, "/backend.Backend/Predict", in, out, opts...)
+	if err != nil {
+		return nil, err
+	}
+	return out, nil
+}
+
+func (c *backendClient) LoadModel(ctx context.Context, in *ModelOptions, opts ...grpc.CallOption) (*Result, error) {
+	out := new(Result)
+	err := c.cc.Invoke(ctx, "/backend.Backend/LoadModel", in, out, opts...)
+	if err != nil {
+		return nil, err
+	}
+	return out, nil
+}
+
+func (c *backendClient) PredictStream(ctx context.Context, in *PredictOptions, opts ...grpc.CallOption) (Backend_PredictStreamClient, error) {
+	stream, err := c.cc.NewStream(ctx, &Backend_ServiceDesc.Streams[0], "/backend.Backend/PredictStream", opts...)
+	if err != nil {
+		return nil, err
+	}
+	x := &backendPredictStreamClient{stream}
+	if err := x.ClientStream.SendMsg(in); err != nil {
+		return nil, err
+	}
+	if err := x.ClientStream.CloseSend(); err != nil {
+		return nil, err
+	}
+	return x, nil
+}
+
+type Backend_PredictStreamClient interface {
+	Recv() (*Reply, error)
+	grpc.ClientStream
+}
+
+type backendPredictStreamClient struct {
+	grpc.ClientStream
+}
+
+func (x *backendPredictStreamClient) Recv() (*Reply, error) {
+	m := new(Reply)
+	if err := x.ClientStream.RecvMsg(m); err != nil {
+		return nil, err
+	}
+	return m, nil
+}
+
+func (c *backendClient) Embedding(ctx context.Context, in *PredictOptions, opts ...grpc.CallOption) (*EmbeddingResult, error) {
+	out := new(EmbeddingResult)
+	err := c.cc.Invoke(ctx, "/backend.Backend/Embedding", in, out, opts...)
+	if err != nil {
+		return nil, err
+	}
+	return out, nil
+}
+
+func (c *backendClient) GenerateImage(ctx context.Context, in *GenerateImageRequest, opts ...grpc.CallOption) (*Result, error) {
+	out := new(Result)
+	err := c.cc.Invoke(ctx, "/backend.Backend/GenerateImage", in, out, opts...)
+	if err != nil {
+		return nil, err
+	}
+	return out, nil
+}
+
+func (c *backendClient) AudioTranscription(ctx context.Context, in *TranscriptRequest, opts ...grpc.CallOption) (*TranscriptResult, error) {
+	out := new(TranscriptResult)
+	err := c.cc.Invoke(ctx, "/backend.Backend/AudioTranscription", in, out, opts...)
+	if err != nil {
+		return nil, err
+	}
+	return out, nil
+}
+
+func (c *backendClient) TTS(ctx context.Context, in *TTSRequest, opts ...grpc.CallOption) (*Result, error) {
+	out := new(Result)
+	err := c.cc.Invoke(ctx, "/backend.Backend/TTS", in, out, opts...)
+	if err != nil {
+		return nil, err
+	}
+	return out, nil
+}
+
+func (c *backendClient) TokenizeString(ctx context.Context, in *PredictOptions, opts ...grpc.CallOption) (*TokenizationResponse, error) {
+	out := new(TokenizationResponse)
+	err := c.cc.Invoke(ctx, "/backend.Backend/TokenizeString", in, out, opts...)
+	if err != nil {
+		return nil, err
+	}
+	return out, nil
+}
+
+func (c *backendClient) Status(ctx context.Context, in *HealthMessage, opts ...grpc.CallOption) (*StatusResponse, error) {
+	out := new(StatusResponse)
+	err := c.cc.Invoke(ctx, "/backend.Backend/Status", in, out, opts...)
+	if err != nil {
+		return nil, err
+	}
+	return out, nil
+}
+
+// BackendServer is the server API for Backend service.
+// All implementations must embed UnimplementedBackendServer
+// for forward compatibility
+type BackendServer interface {
+	Health(context.Context, *HealthMessage) (*Reply, error)
+	Predict(context.Context, *PredictOptions) (*Reply, error)
+	LoadModel(context.Context, *ModelOptions) (*Result, error)
+	PredictStream(*PredictOptions, Backend_PredictStreamServer) error
+	Embedding(context.Context, *PredictOptions) (*EmbeddingResult, error)
+	GenerateImage(context.Context, *GenerateImageRequest) (*Result, error)
+	AudioTranscription(context.Context, *TranscriptRequest) (*TranscriptResult, error)
+	TTS(context.Context, *TTSRequest) (*Result, error)
+	TokenizeString(context.Context, *PredictOptions) (*TokenizationResponse, error)
+	Status(context.Context, *HealthMessage) (*StatusResponse, error)
+	mustEmbedUnimplementedBackendServer()
+}
+
+// UnimplementedBackendServer must be embedded to have forward compatible implementations.
+type UnimplementedBackendServer struct {
+}
+
+func (UnimplementedBackendServer) Health(context.Context, *HealthMessage) (*Reply, error) {
+	return nil, status.Errorf(codes.Unimplemented, "method Health not implemented")
+}
+func (UnimplementedBackendServer) Predict(context.Context, *PredictOptions) (*Reply, error) {
+	return nil, status.Errorf(codes.Unimplemented, "method Predict not implemented")
+}
+func (UnimplementedBackendServer) LoadModel(context.Context, *ModelOptions) (*Result, error) {
+	return nil, status.Errorf(codes.Unimplemented, "method LoadModel not implemented")
+}
+func (UnimplementedBackendServer) PredictStream(*PredictOptions, Backend_PredictStreamServer) error {
+	return status.Errorf(codes.Unimplemented, "method PredictStream not implemented")
+}
+func (UnimplementedBackendServer) Embedding(context.Context, *PredictOptions) (*EmbeddingResult, error) {
+	return nil, status.Errorf(codes.Unimplemented, "method Embedding not implemented")
+}
+func (UnimplementedBackendServer) GenerateImage(context.Context, *GenerateImageRequest) (*Result, error) {
+	return nil, status.Errorf(codes.Unimplemented, "method GenerateImage not implemented")
+}
+func (UnimplementedBackendServer) AudioTranscription(context.Context, *TranscriptRequest) (*TranscriptResult, error) {
+	return nil, status.Errorf(codes.Unimplemented, "method AudioTranscription not implemented")
+}
+func (UnimplementedBackendServer) TTS(context.Context, *TTSRequest) (*Result, error) {
+	return nil, status.Errorf(codes.Unimplemented, "method TTS not implemented")
+}
+func (UnimplementedBackendServer) TokenizeString(context.Context, *PredictOptions) (*TokenizationResponse, error) {
+	return nil, status.Errorf(codes.Unimplemented, "method TokenizeString not implemented")
+}
+func (UnimplementedBackendServer) Status(context.Context, *HealthMessage) (*StatusResponse, error) {
+	return nil, status.Errorf(codes.Unimplemented, "method Status not implemented")
+}
+func (UnimplementedBackendServer) mustEmbedUnimplementedBackendServer() {}
+
+// UnsafeBackendServer may be embedded to opt out of forward compatibility for this service.
+// Use of this interface is not recommended, as added methods to BackendServer will
+// result in compilation errors.
+type UnsafeBackendServer interface {
+	mustEmbedUnimplementedBackendServer()
+}
+
+func RegisterBackendServer(s grpc.ServiceRegistrar, srv BackendServer) {
+	s.RegisterService(&Backend_ServiceDesc, srv)
+}
+
+func _Backend_Health_Handler(srv interface{}, ctx context.Context, dec func(interface{}) error, interceptor grpc.UnaryServerInterceptor) (interface{}, error) {
+	in := new(HealthMessage)
+	if err := dec(in); err != nil {
+		return nil, err
+	}
+	if interceptor == nil {
+		return srv.(BackendServer).Health(ctx, in)
+	}
+	info := &grpc.UnaryServerInfo{
+		Server:     srv,
+		FullMethod: "/backend.Backend/Health",
+	}
+	handler := func(ctx context.Context, req interface{}) (interface{}, error) {
+		return srv.(BackendServer).Health(ctx, req.(*HealthMessage))
+	}
+	return interceptor(ctx, in, info, handler)
+}
+
+func _Backend_Predict_Handler(srv interface{}, ctx context.Context, dec func(interface{}) error, interceptor grpc.UnaryServerInterceptor) (interface{}, error) {
+	in := new(PredictOptions)
+	if err := dec(in); err != nil {
+		return nil, err
+	}
+	if interceptor == nil {
+		return srv.(BackendServer).Predict(ctx, in)
+	}
+	info := &grpc.UnaryServerInfo{
+		Server:     srv,
+		FullMethod: "/backend.Backend/Predict",
+	}
+	handler := func(ctx context.Context, req interface{}) (interface{}, error) {
+		return srv.(BackendServer).Predict(ctx, req.(*PredictOptions))
+	}
+	return interceptor(ctx, in, info, handler)
+}
+
+func _Backend_LoadModel_Handler(srv interface{}, ctx context.Context, dec func(interface{}) error, interceptor grpc.UnaryServerInterceptor) (interface{}, error) {
+	in := new(ModelOptions)
+	if err := dec(in); err != nil {
+		return nil, err
+	}
+	if interceptor == nil {
+		return srv.(BackendServer).LoadModel(ctx, in)
+	}
+	info := &grpc.UnaryServerInfo{
+		Server:     srv,
+		FullMethod: "/backend.Backend/LoadModel",
+	}
+	handler := func(ctx context.Context, req interface{}) (interface{}, error) {
+		return srv.(BackendServer).LoadModel(ctx, req.(*ModelOptions))
+	}
+	return interceptor(ctx, in, info, handler)
+}
+
+func _Backend_PredictStream_Handler(srv interface{}, stream grpc.ServerStream) error {
+	m := new(PredictOptions)
+	if err := stream.RecvMsg(m); err != nil {
+		return err
+	}
+	return srv.(BackendServer).PredictStream(m, &backendPredictStreamServer{stream})
+}
+
+type Backend_PredictStreamServer interface {
+	Send(*Reply) error
+	grpc.ServerStream
+}
+
+type backendPredictStreamServer struct {
+	grpc.ServerStream
+}
+
+func (x *backendPredictStreamServer) Send(m *Reply) error {
+	return x.ServerStream.SendMsg(m)
+}
+
+func _Backend_Embedding_Handler(srv interface{}, ctx context.Context, dec func(interface{}) error, interceptor grpc.UnaryServerInterceptor) (interface{}, error) {
+	in := new(PredictOptions)
+	if err := dec(in); err != nil {
+		return nil, err
+	}
+	if interceptor == nil {
+		return srv.(BackendServer).Embedding(ctx, in)
+	}
+	info := &grpc.UnaryServerInfo{
+		Server:     srv,
+		FullMethod: "/backend.Backend/Embedding",
+	}
+	handler := func(ctx context.Context, req interface{}) (interface{}, error) {
+		return srv.(BackendServer).Embedding(ctx, req.(*PredictOptions))
+	}
+	return interceptor(ctx, in, info, handler)
+}
+
+func _Backend_GenerateImage_Handler(srv interface{}, ctx context.Context, dec func(interface{}) error, interceptor grpc.UnaryServerInterceptor) (interface{}, error) {
+	in := new(GenerateImageRequest)
+	if err := dec(in); err != nil {
+		return nil, err
+	}
+	if interceptor == nil {
+		return srv.(BackendServer).GenerateImage(ctx, in)
+	}
+	info := &grpc.UnaryServerInfo{
+		Server:     srv,
+		FullMethod: "/backend.Backend/GenerateImage",
+	}
+	handler := func(ctx context.Context, req interface{}) (interface{}, error) {
+		return srv.(BackendServer).GenerateImage(ctx, req.(*GenerateImageRequest))
+	}
+	return interceptor(ctx, in, info, handler)
+}
+
+func _Backend_AudioTranscription_Handler(srv interface{}, ctx context.Context, dec func(interface{}) error, interceptor grpc.UnaryServerInterceptor) (interface{}, error) {
+	in := new(TranscriptRequest)
+	if err := dec(in); err != nil {
+		return nil, err
+	}
+	if interceptor == nil {
+		return srv.(BackendServer).AudioTranscription(ctx, in)
+	}
+	info := &grpc.UnaryServerInfo{
+		Server:     srv,
+		FullMethod: "/backend.Backend/AudioTranscription",
+	}
+	handler := func(ctx context.Context, req interface{}) (interface{}, error) {
+		return srv.(BackendServer).AudioTranscription(ctx, req.(*TranscriptRequest))
+	}
+	return interceptor(ctx, in, info, handler)
+}
+
+func _Backend_TTS_Handler(srv interface{}, ctx context.Context, dec func(interface{}) error, interceptor grpc.UnaryServerInterceptor) (interface{}, error) {
+	in := new(TTSRequest)
+	if err := dec(in); err != nil {
+		return nil, err
+	}
+	if interceptor == nil {
+		return srv.(BackendServer).TTS(ctx, in)
+	}
+	info := &grpc.UnaryServerInfo{
+		Server:     srv,
+		FullMethod: "/backend.Backend/TTS",
+	}
+	handler := func(ctx context.Context, req interface{}) (interface{}, error) {
+		return srv.(BackendServer).TTS(ctx, req.(*TTSRequest))
+	}
+	return interceptor(ctx, in, info, handler)
+}
+
+func _Backend_TokenizeString_Handler(srv interface{}, ctx context.Context, dec func(interface{}) error, interceptor grpc.UnaryServerInterceptor) (interface{}, error) {
+	in := new(PredictOptions)
+	if err := dec(in); err != nil {
+		return nil, err
+	}
+	if interceptor == nil {
+		return srv.(BackendServer).TokenizeString(ctx, in)
+	}
+	info := &grpc.UnaryServerInfo{
+		Server:     srv,
+		FullMethod: "/backend.Backend/TokenizeString",
+	}
+	handler := func(ctx context.Context, req interface{}) (interface{}, error) {
+		return srv.(BackendServer).TokenizeString(ctx, req.(*PredictOptions))
+	}
+	return interceptor(ctx, in, info, handler)
+}
+
+func _Backend_Status_Handler(srv interface{}, ctx context.Context, dec func(interface{}) error, interceptor grpc.UnaryServerInterceptor) (interface{}, error) {
+	in := new(HealthMessage)
+	if err := dec(in); err != nil {
+		return nil, err
+	}
+	if interceptor == nil {
+		return srv.(BackendServer).Status(ctx, in)
+	}
+	info := &grpc.UnaryServerInfo{
+		Server:     srv,
+		FullMethod: "/backend.Backend/Status",
+	}
+	handler := func(ctx context.Context, req interface{}) (interface{}, error) {
+		return srv.(BackendServer).Status(ctx, req.(*HealthMessage))
+	}
+	return interceptor(ctx, in, info, handler)
+}
+
+// Backend_ServiceDesc is the grpc.ServiceDesc for Backend service.
+// It's only intended for direct use with grpc.RegisterService,
+// and not to be introspected or modified (even as a copy)
+var Backend_ServiceDesc = grpc.ServiceDesc{
+	ServiceName: "backend.Backend",
+	HandlerType: (*BackendServer)(nil),
+	Methods: []grpc.MethodDesc{
+		{
+			MethodName: "Health",
+			Handler:    _Backend_Health_Handler,
+		},
+		{
+			MethodName: "Predict",
+			Handler:    _Backend_Predict_Handler,
+		},
+		{
+			MethodName: "LoadModel",
+			Handler:    _Backend_LoadModel_Handler,
+		},
+		{
+			MethodName: "Embedding",
+			Handler:    _Backend_Embedding_Handler,
+		},
+		{
+			MethodName: "GenerateImage",
+			Handler:    _Backend_GenerateImage_Handler,
+		},
+		{
+			MethodName: "AudioTranscription",
+			Handler:    _Backend_AudioTranscription_Handler,
+		},
+		{
+			MethodName: "TTS",
+			Handler:    _Backend_TTS_Handler,
+		},
+		{
+			MethodName: "TokenizeString",
+			Handler:    _Backend_TokenizeString_Handler,
+		},
+		{
+			MethodName: "Status",
+			Handler:    _Backend_Status_Handler,
+		},
+	},
+	Streams: []grpc.StreamDesc{
+		{
+			StreamName:    "PredictStream",
+			Handler:       _Backend_PredictStream_Handler,
+			ServerStreams: true,
+		},
+	},
+	Metadata: "backend/backend.proto",
+}

backend/cpp/grpc/Makefile

@@ -5,6 +5,7 @@ SYSTEM ?= $(HOST_SYSTEM)
 TAG_LIB_GRPC?=v1.59.0
 GIT_REPO_LIB_GRPC?=https://github.com/grpc/grpc.git
 GIT_CLONE_DEPTH?=1
+NUM_BUILD_THREADS?=$(shell nproc --ignore=1)
 
 INSTALLED_PACKAGES=installed_packages
 GRPC_REPO=grpc_repo
@@ -47,11 +48,11 @@ $(INSTALLED_PACKAGES): grpc_build
 
 $(GRPC_REPO):
 	git clone --depth $(GIT_CLONE_DEPTH) -b $(TAG_LIB_GRPC) $(GIT_REPO_LIB_GRPC) $(GRPC_REPO)/grpc
-	cd $(GRPC_REPO)/grpc && git submodule update --jobs 2 --init --recursive --depth $(GIT_CLONE_DEPTH)
+	cd $(GRPC_REPO)/grpc && git submodule update --init --recursive --depth $(GIT_CLONE_DEPTH)
 
 $(GRPC_BUILD): $(GRPC_REPO)
 	mkdir -p $(GRPC_BUILD)
-	cd $(GRPC_BUILD) && cmake $(CMAKE_ARGS) ../$(GRPC_REPO)/grpc && cmake --build . && cmake --build . --target install
+	cd $(GRPC_BUILD) && cmake $(CMAKE_ARGS) ../$(GRPC_REPO)/grpc && cmake --build . -- -j ${NUM_BUILD_THREADS} && cmake --build . --target install -- -j ${NUM_BUILD_THREADS}
 
 build: $(INSTALLED_PACKAGES)
 

backend/cpp/llama/Makefile

@@ -18,12 +18,6 @@ else ifeq ($(BUILD_TYPE),clblas)
 # If it's hipblas we do have also to set CC=/opt/rocm/llvm/bin/clang CXX=/opt/rocm/llvm/bin/clang++ 
 else ifeq ($(BUILD_TYPE),hipblas)
 	CMAKE_ARGS+=-DLLAMA_HIPBLAS=ON
-# If it's OSX, DO NOT embed the metal library - -DLLAMA_METAL_EMBED_LIBRARY=ON requires further investigation
-# But if it's OSX without metal, disable it here
-else ifeq ($(OS),darwin)
-	ifneq ($(BUILD_TYPE),metal)
-		CMAKE_ARGS+=-DLLAMA_METAL=OFF
-	endif
 endif
 
 ifeq ($(BUILD_TYPE),sycl_f16)
@@ -41,7 +35,7 @@ llama.cpp:
 	fi
 	cd llama.cpp && git checkout -b build $(LLAMA_VERSION) && git submodule update --init --recursive --depth 1
 
-llama.cpp/examples/grpc-server: llama.cpp
+llama.cpp/examples/grpc-server:
 	mkdir -p llama.cpp/examples/grpc-server
 	cp -r $(abspath ./)/CMakeLists.txt llama.cpp/examples/grpc-server/
 	cp -r $(abspath ./)/grpc-server.cpp llama.cpp/examples/grpc-server/

backend/cpp/llama/grpc-server.cpp

@@ -1084,7 +1084,7 @@ struct llama_server_context
             slot.has_next_token = false;
         }
 
-        if (result.tok == llama_token_eos(model))
+        if (!slot.cache_tokens.empty() && result.tok == llama_token_eos(model))
         {
             slot.stopped_eos = true;
             slot.has_next_token = false;
@@ -2332,10 +2332,6 @@ public:
                 std::string completion_text = result.result_json.value("content", "");
 
                 reply.set_message(completion_text);
-                int32_t tokens_predicted = result.result_json.value("tokens_predicted", 0);
-                reply.set_tokens(tokens_predicted);
-                int32_t tokens_evaluated = result.result_json.value("tokens_evaluated", 0);
-                reply.set_prompt_tokens(tokens_evaluated);
 
                 // Send the reply
                 writer->Write(reply);
@@ -2361,10 +2357,6 @@ public:
         task_result result = llama.queue_results.recv(task_id);
         if (!result.error && result.stop) {
             completion_text = result.result_json.value("content", "");
-            int32_t tokens_predicted = result.result_json.value("tokens_predicted", 0);
-            int32_t tokens_evaluated = result.result_json.value("tokens_evaluated", 0);
-            reply->set_prompt_tokens(tokens_evaluated);
-            reply->set_tokens(tokens_predicted);
             reply->set_message(completion_text);
         }
         else

backend/go/stores/debug.go

@@ -1,14 +0,0 @@
-//go:build debug
-// +build debug
-
-package main
-
-import (
-	"github.com/rs/zerolog/log"
-)
-
-func assert(cond bool, msg string) {
-	if !cond {
-		log.Fatal().Stack().Msg(msg)
-	}
-}

backend/go/stores/main.go

@@ -1,26 +0,0 @@
-package main
-
-// Note: this is started internally by LocalAI and a server is allocated for each store
-
-import (
-	"flag"
-	"os"
-
-	grpc "github.com/go-skynet/LocalAI/pkg/grpc"
-	"github.com/rs/zerolog"
-	"github.com/rs/zerolog/log"
-)
-
-var (
-	addr = flag.String("addr", "localhost:50051", "the address to connect to")
-)
-
-func main() {
-	log.Logger = log.Output(zerolog.ConsoleWriter{Out: os.Stderr})
-
-	flag.Parse()
-
-	if err := grpc.StartServer(*addr, NewStore()); err != nil {
-		panic(err)
-	}
-}

backend/go/stores/production.go

@@ -1,7 +0,0 @@
-//go:build !debug
-// +build !debug
-
-package main
-
-func assert(cond bool, msg string) {
-}

backend/go/stores/store.go

@@ -1,507 +0,0 @@
-package main
-
-// This is a wrapper to statisfy the GRPC service interface
-// It is meant to be used by the main executable that is the server for the specific backend type (falcon, gpt3, etc)
-import (
-	"container/heap"
-	"fmt"
-	"math"
-	"slices"
-
-	"github.com/go-skynet/LocalAI/pkg/grpc/base"
-	pb "github.com/go-skynet/LocalAI/pkg/grpc/proto"
-
-	"github.com/rs/zerolog/log"
-)
-
-type Store struct {
-	base.SingleThread
-
-	// The sorted keys
-	keys [][]float32
-	// The sorted values
-	values [][]byte
-
-	// If for every K it holds that ||k||^2 = 1, then we can use the normalized distance functions
-	// TODO: Should we normalize incoming keys if they are not instead?
-	keysAreNormalized bool
-	// The first key decides the length of the keys
-	keyLen int
-}
-
-// TODO: Only used for sorting using Go's builtin implementation. The interfaces are columnar because
-// that's theoretically best for memory layout and cache locality, but this isn't optimized yet.
-type Pair struct {
-	Key   []float32
-	Value []byte
-}
-
-func NewStore() *Store {
-	return &Store{
-		keys:              make([][]float32, 0),
-		values:            make([][]byte, 0),
-		keysAreNormalized: true,
-		keyLen:            -1,
-	}
-}
-
-func compareSlices(k1, k2 []float32) int {
-	assert(len(k1) == len(k2), fmt.Sprintf("compareSlices: len(k1) = %d, len(k2) = %d", len(k1), len(k2)))
-
-	return slices.Compare(k1, k2)
-}
-
-func hasKey(unsortedSlice [][]float32, target []float32) bool {
-	return slices.ContainsFunc(unsortedSlice, func(k []float32) bool {
-		return compareSlices(k, target) == 0
-	})
-}
-
-func findInSortedSlice(sortedSlice [][]float32, target []float32) (int, bool) {
-	return slices.BinarySearchFunc(sortedSlice, target, func(k, t []float32) int {
-		return compareSlices(k, t)
-	})
-}
-
-func isSortedPairs(kvs []Pair) bool {
-	for i := 1; i < len(kvs); i++ {
-		if compareSlices(kvs[i-1].Key, kvs[i].Key) > 0 {
-			return false
-		}
-	}
-
-	return true
-}
-
-func isSortedKeys(keys [][]float32) bool {
-	for i := 1; i < len(keys); i++ {
-		if compareSlices(keys[i-1], keys[i]) > 0 {
-			return false
-		}
-	}
-
-	return true
-}
-
-func sortIntoKeySlicese(keys []*pb.StoresKey) [][]float32 {
-	ks := make([][]float32, len(keys))
-
-	for i, k := range keys {
-		ks[i] = k.Floats
-	}
-
-	slices.SortFunc(ks, compareSlices)
-
-	assert(len(ks) == len(keys), fmt.Sprintf("len(ks) = %d, len(keys) = %d", len(ks), len(keys)))
-	assert(isSortedKeys(ks), "keys are not sorted")
-
-	return ks
-}
-
-func (s *Store) Load(opts *pb.ModelOptions) error {
-	return nil
-}
-
-// Sort the incoming kvs and merge them with the existing sorted kvs
-func (s *Store) StoresSet(opts *pb.StoresSetOptions) error {
-	if len(opts.Keys) == 0 {
-		return fmt.Errorf("no keys to add")
-	}
-
-	if len(opts.Keys) != len(opts.Values) {
-		return fmt.Errorf("len(keys) = %d, len(values) = %d", len(opts.Keys), len(opts.Values))
-	}
-
-	if s.keyLen == -1 {
-		s.keyLen = len(opts.Keys[0].Floats)
-	} else {
-		if len(opts.Keys[0].Floats) != s.keyLen {
-			return fmt.Errorf("Try to add key with length %d when existing length is %d", len(opts.Keys[0].Floats), s.keyLen)
-		}
-	}
-
-	kvs := make([]Pair, len(opts.Keys))
-
-	for i, k := range opts.Keys {
-		if s.keysAreNormalized && !isNormalized(k.Floats) {
-			s.keysAreNormalized = false
-			var sample []float32
-			if len(s.keys) > 5 {
-				sample = k.Floats[:5]
-			} else {
-				sample = k.Floats
-			}
-			log.Debug().Msgf("Key is not normalized: %v", sample)
-		}
-
-		kvs[i] = Pair{
-			Key:   k.Floats,
-			Value: opts.Values[i].Bytes,
-		}
-	}
-
-	slices.SortFunc(kvs, func(a, b Pair) int {
-		return compareSlices(a.Key, b.Key)
-	})
-
-	assert(len(kvs) == len(opts.Keys), fmt.Sprintf("len(kvs) = %d, len(opts.Keys) = %d", len(kvs), len(opts.Keys)))
-	assert(isSortedPairs(kvs), "keys are not sorted")
-
-	l := len(kvs) + len(s.keys)
-	merge_ks := make([][]float32, 0, l)
-	merge_vs := make([][]byte, 0, l)
-
-	i, j := 0, 0
-	for {
-		if i+j >= l {
-			break
-		}
-
-		if i >= len(kvs) {
-			merge_ks = append(merge_ks, s.keys[j])
-			merge_vs = append(merge_vs, s.values[j])
-			j++
-			continue
-		}
-
-		if j >= len(s.keys) {
-			merge_ks = append(merge_ks, kvs[i].Key)
-			merge_vs = append(merge_vs, kvs[i].Value)
-			i++
-			continue
-		}
-
-		c := compareSlices(kvs[i].Key, s.keys[j])
-		if c < 0 {
-			merge_ks = append(merge_ks, kvs[i].Key)
-			merge_vs = append(merge_vs, kvs[i].Value)
-			i++
-		} else if c > 0 {
-			merge_ks = append(merge_ks, s.keys[j])
-			merge_vs = append(merge_vs, s.values[j])
-			j++
-		} else {
-			merge_ks = append(merge_ks, kvs[i].Key)
-			merge_vs = append(merge_vs, kvs[i].Value)
-			i++
-			j++
-		}
-	}
-
-	assert(len(merge_ks) == l, fmt.Sprintf("len(merge_ks) = %d, l = %d", len(merge_ks), l))
-	assert(isSortedKeys(merge_ks), "merge keys are not sorted")
-
-	s.keys = merge_ks
-	s.values = merge_vs
-
-	return nil
-}
-
-func (s *Store) StoresDelete(opts *pb.StoresDeleteOptions) error {
-	if len(opts.Keys) == 0 {
-		return fmt.Errorf("no keys to delete")
-	}
-
-	if len(opts.Keys) == 0 {
-		return fmt.Errorf("no keys to add")
-	}
-
-	if s.keyLen == -1 {
-		s.keyLen = len(opts.Keys[0].Floats)
-	} else {
-		if len(opts.Keys[0].Floats) != s.keyLen {
-			return fmt.Errorf("Trying to delete key with length %d when existing length is %d", len(opts.Keys[0].Floats), s.keyLen)
-		}
-	}
-
-	ks := sortIntoKeySlicese(opts.Keys)
-
-	l := len(s.keys) - len(ks)
-	merge_ks := make([][]float32, 0, l)
-	merge_vs := make([][]byte, 0, l)
-
-	tail_ks := s.keys
-	tail_vs := s.values
-	for _, k := range ks {
-		j, found := findInSortedSlice(tail_ks, k)
-
-		if found {
-			merge_ks = append(merge_ks, tail_ks[:j]...)
-			merge_vs = append(merge_vs, tail_vs[:j]...)
-			tail_ks = tail_ks[j+1:]
-			tail_vs = tail_vs[j+1:]
-		} else {
-			assert(!hasKey(s.keys, k), fmt.Sprintf("Key exists, but was not found: t=%d, %v", len(tail_ks), k))
-		}
-
-		log.Debug().Msgf("Delete: found = %v, t = %d, j = %d, len(merge_ks) = %d, len(merge_vs) = %d", found, len(tail_ks), j, len(merge_ks), len(merge_vs))
-	}
-
-	merge_ks = append(merge_ks, tail_ks...)
-	merge_vs = append(merge_vs, tail_vs...)
-
-	assert(len(merge_ks) <= len(s.keys), fmt.Sprintf("len(merge_ks) = %d, len(s.keys) = %d", len(merge_ks), len(s.keys)))
-
-	s.keys = merge_ks
-	s.values = merge_vs
-
-	assert(len(s.keys) >= l, fmt.Sprintf("len(s.keys) = %d, l = %d", len(s.keys), l))
-	assert(isSortedKeys(s.keys), "keys are not sorted")
-	assert(func() bool {
-		for _, k := range ks {
-			if _, found := findInSortedSlice(s.keys, k); found {
-				return false
-			}
-		}
-		return true
-	}(), "Keys to delete still present")
-
-	if len(s.keys) != l {
-		log.Debug().Msgf("Delete: Some keys not found: len(s.keys) = %d, l = %d", len(s.keys), l)
-	}
-
-	return nil
-}
-
-func (s *Store) StoresGet(opts *pb.StoresGetOptions) (pb.StoresGetResult, error) {
-	pbKeys := make([]*pb.StoresKey, 0, len(opts.Keys))
-	pbValues := make([]*pb.StoresValue, 0, len(opts.Keys))
-	ks := sortIntoKeySlicese(opts.Keys)
-
-	if len(s.keys) == 0 {
-		log.Debug().Msgf("Get: No keys in store")
-	}
-
-	if s.keyLen == -1 {
-		s.keyLen = len(opts.Keys[0].Floats)
-	} else {
-		if len(opts.Keys[0].Floats) != s.keyLen {
-			return pb.StoresGetResult{}, fmt.Errorf("Try to get a key with length %d when existing length is %d", len(opts.Keys[0].Floats), s.keyLen)
-		}
-	}
-
-	tail_k := s.keys
-	tail_v := s.values
-	for i, k := range ks {
-		j, found := findInSortedSlice(tail_k, k)
-
-		if found {
-			pbKeys = append(pbKeys, &pb.StoresKey{
-				Floats: k,
-			})
-			pbValues = append(pbValues, &pb.StoresValue{
-				Bytes: tail_v[j],
-			})
-
-			tail_k = tail_k[j+1:]
-			tail_v = tail_v[j+1:]
-		} else {
-			assert(!hasKey(s.keys, k), fmt.Sprintf("Key exists, but was not found: i=%d, %v", i, k))
-		}
-	}
-
-	if len(pbKeys) != len(opts.Keys) {
-		log.Debug().Msgf("Get: Some keys not found: len(pbKeys) = %d, len(opts.Keys) = %d, len(s.Keys) = %d", len(pbKeys), len(opts.Keys), len(s.keys))
-	}
-
-	return pb.StoresGetResult{
-		Keys:   pbKeys,
-		Values: pbValues,
-	}, nil
-}
-
-func isNormalized(k []float32) bool {
-	var sum float32
-	for _, v := range k {
-		sum += v
-	}
-
-	return sum == 1.0
-}
-
-// TODO: This we could replace with handwritten SIMD code
-func normalizedCosineSimilarity(k1, k2 []float32) float32 {
-	assert(len(k1) == len(k2), fmt.Sprintf("normalizedCosineSimilarity: len(k1) = %d, len(k2) = %d", len(k1), len(k2)))
-
-	var dot float32
-	for i := 0; i < len(k1); i++ {
-		dot += k1[i] * k2[i]
-	}
-
-	assert(dot >= -1 && dot <= 1, fmt.Sprintf("dot = %f", dot))
-
-	// 2.0 * (1.0 - dot) would be the Euclidean distance
-	return dot
-}
-
-type PriorityItem struct {
-	Similarity float32
-	Key        []float32
-	Value      []byte
-}
-
-type PriorityQueue []*PriorityItem
-
-func (pq PriorityQueue) Len() int { return len(pq) }
-
-func (pq PriorityQueue) Less(i, j int) bool {
-	// Inverted because the most similar should be at the top
-	return pq[i].Similarity < pq[j].Similarity
-}
-
-func (pq PriorityQueue) Swap(i, j int) {
-	pq[i], pq[j] = pq[j], pq[i]
-}
-
-func (pq *PriorityQueue) Push(x any) {
-	item := x.(*PriorityItem)
-	*pq = append(*pq, item)
-}
-
-func (pq *PriorityQueue) Pop() any {
-	old := *pq
-	n := len(old)
-	item := old[n-1]
-	*pq = old[0 : n-1]
-	return item
-}
-
-func (s *Store) StoresFindNormalized(opts *pb.StoresFindOptions) (pb.StoresFindResult, error) {
-	tk := opts.Key.Floats
-	top_ks := make(PriorityQueue, 0, int(opts.TopK))
-	heap.Init(&top_ks)
-
-	for i, k := range s.keys {
-		sim := normalizedCosineSimilarity(tk, k)
-		heap.Push(&top_ks, &PriorityItem{
-			Similarity: sim,
-			Key:        k,
-			Value:      s.values[i],
-		})
-
-		if top_ks.Len() > int(opts.TopK) {
-			heap.Pop(&top_ks)
-		}
-	}
-
-	similarities := make([]float32, top_ks.Len())
-	pbKeys := make([]*pb.StoresKey, top_ks.Len())
-	pbValues := make([]*pb.StoresValue, top_ks.Len())
-
-	for i := top_ks.Len() - 1; i >= 0; i-- {
-		item := heap.Pop(&top_ks).(*PriorityItem)
-
-		similarities[i] = item.Similarity
-		pbKeys[i] = &pb.StoresKey{
-			Floats: item.Key,
-		}
-		pbValues[i] = &pb.StoresValue{
-			Bytes: item.Value,
-		}
-	}
-
-	return pb.StoresFindResult{
-		Keys:         pbKeys,
-		Values:       pbValues,
-		Similarities: similarities,
-	}, nil
-}
-
-func cosineSimilarity(k1, k2 []float32, mag1 float64) float32 {
-	assert(len(k1) == len(k2), fmt.Sprintf("cosineSimilarity: len(k1) = %d, len(k2) = %d", len(k1), len(k2)))
-
-	var dot, mag2 float64
-	for i := 0; i < len(k1); i++ {
-		dot += float64(k1[i] * k2[i])
-		mag2 += float64(k2[i] * k2[i])
-	}
-
-	sim := float32(dot / (mag1 * math.Sqrt(mag2)))
-
-	assert(sim >= -1 && sim <= 1, fmt.Sprintf("sim = %f", sim))
-
-	return sim
-}
-
-func (s *Store) StoresFindFallback(opts *pb.StoresFindOptions) (pb.StoresFindResult, error) {
-	tk := opts.Key.Floats
-	top_ks := make(PriorityQueue, 0, int(opts.TopK))
-	heap.Init(&top_ks)
-
-	var mag1 float64
-	for _, v := range tk {
-		mag1 += float64(v * v)
-	}
-	mag1 = math.Sqrt(mag1)
-
-	for i, k := range s.keys {
-		dist := cosineSimilarity(tk, k, mag1)
-		heap.Push(&top_ks, &PriorityItem{
-			Similarity: dist,
-			Key:        k,
-			Value:      s.values[i],
-		})
-
-		if top_ks.Len() > int(opts.TopK) {
-			heap.Pop(&top_ks)
-		}
-	}
-
-	similarities := make([]float32, top_ks.Len())
-	pbKeys := make([]*pb.StoresKey, top_ks.Len())
-	pbValues := make([]*pb.StoresValue, top_ks.Len())
-
-	for i := top_ks.Len() - 1; i >= 0; i-- {
-		item := heap.Pop(&top_ks).(*PriorityItem)
-
-		similarities[i] = item.Similarity
-		pbKeys[i] = &pb.StoresKey{
-			Floats: item.Key,
-		}
-		pbValues[i] = &pb.StoresValue{
-			Bytes: item.Value,
-		}
-	}
-
-	return pb.StoresFindResult{
-		Keys:         pbKeys,
-		Values:       pbValues,
-		Similarities: similarities,
-	}, nil
-}
-
-func (s *Store) StoresFind(opts *pb.StoresFindOptions) (pb.StoresFindResult, error) {
-	tk := opts.Key.Floats
-
-	if len(tk) != s.keyLen {
-		return pb.StoresFindResult{}, fmt.Errorf("Try to find key with length %d when existing length is %d", len(tk), s.keyLen)
-	}
-
-	if opts.TopK < 1 {
-		return pb.StoresFindResult{}, fmt.Errorf("opts.TopK = %d, must be >= 1", opts.TopK)
-	}
-
-	if s.keyLen == -1 {
-		s.keyLen = len(opts.Key.Floats)
-	} else {
-		if len(opts.Key.Floats) != s.keyLen {
-			return pb.StoresFindResult{}, fmt.Errorf("Try to add key with length %d when existing length is %d", len(opts.Key.Floats), s.keyLen)
-		}
-	}
-
-	if s.keysAreNormalized && isNormalized(tk) {
-		return s.StoresFindNormalized(opts)
-	} else {
-		if s.keysAreNormalized {
-			var sample []float32
-			if len(s.keys) > 5 {
-				sample = tk[:5]
-			} else {
-				sample = tk
-			}
-			log.Debug().Msgf("Trying to compare non-normalized key with normalized keys: %v", sample)
-		}
-
-		return s.StoresFindFallback(opts)
-	}
-}

backend/python/autogptq/Makefile

@@ -1,13 +1,4 @@
 .PHONY: autogptq
-autogptq: protogen
+autogptq:
 	$(MAKE) -C ../common-env/transformers
 
-.PHONY: protogen
-protogen: backend_pb2_grpc.py backend_pb2.py
-
-.PHONY: protogen-clean
-protogen-clean:
-	$(RM) backend_pb2_grpc.py backend_pb2.py
-
-backend_pb2_grpc.py backend_pb2.py:
-	python3 -m grpc_tools.protoc -I../.. --python_out=. --grpc_python_out=. backend.proto

backend/python/autogptq/autogptq.py

@@ -5,14 +5,12 @@ import signal
 import sys
 import os
 import time
-import base64
 
 import grpc
 import backend_pb2
 import backend_pb2_grpc
-
 from auto_gptq import AutoGPTQForCausalLM
-from transformers import AutoTokenizer, AutoModelForCausalLM
+from transformers import AutoTokenizer
 from transformers import TextGenerationPipeline
 
 _ONE_DAY_IN_SECONDS = 60 * 60 * 24
@@ -30,18 +28,9 @@ class BackendServicer(backend_pb2_grpc.BackendServicer):
             if request.Device != "":
                 device = request.Device
 
-            # support loading local model files
-            model_path = os.path.join(os.environ.get('MODELS_PATH', './'), request.Model)
-            tokenizer = AutoTokenizer.from_pretrained(model_path, use_fast=True, trust_remote_code=request.TrustRemoteCode)
+            tokenizer = AutoTokenizer.from_pretrained(request.Model, use_fast=request.UseFastTokenizer)
 
-            # support model `Qwen/Qwen-VL-Chat-Int4`
-            if "qwen-vl" in request.Model.lower():
-                self.model_name = "Qwen-VL-Chat"
-                model = AutoModelForCausalLM.from_pretrained(model_path, 
-                    trust_remote_code=request.TrustRemoteCode,
-                    device_map="auto").eval()
-            else:
-                model = AutoGPTQForCausalLM.from_quantized(model_path,
+            model = AutoGPTQForCausalLM.from_quantized(request.Model,
                     model_basename=request.ModelBaseName,
                     use_safetensors=True,
                     trust_remote_code=request.TrustRemoteCode,
@@ -66,11 +55,6 @@ class BackendServicer(backend_pb2_grpc.BackendServicer):
         if request.TopP != 0.0:
             top_p = request.TopP
 
-        
-        prompt_images = self.recompile_vl_prompt(request)
-        compiled_prompt = prompt_images[0]
-        print(f"Prompt: {compiled_prompt}", file=sys.stderr)
-
         # Implement Predict RPC
         pipeline = TextGenerationPipeline(
             model=self.model, 
@@ -80,17 +64,10 @@ class BackendServicer(backend_pb2_grpc.BackendServicer):
             top_p=top_p,
             repetition_penalty=penalty,
             )
-        t = pipeline(compiled_prompt)[0]["generated_text"]
-        print(f"generated_text: {t}", file=sys.stderr)
-        
-        if compiled_prompt in t:
-            t = t.replace(compiled_prompt, "")
-        # house keeping. Remove the image files from /tmp folder
-        for img_path in prompt_images[1]:
-            try:
-                os.remove(img_path)
-            except Exception as e:
-                print(f"Error removing image file: {img_path}, {e}", file=sys.stderr)
+        t = pipeline(request.Prompt)[0]["generated_text"]
+        # Remove prompt from response if present
+        if request.Prompt in t:
+            t = t.replace(request.Prompt, "")
 
         return backend_pb2.Result(message=bytes(t, encoding='utf-8'))
 
@@ -101,24 +78,6 @@ class BackendServicer(backend_pb2_grpc.BackendServicer):
         # Not implemented yet
         return self.Predict(request, context)
 
-    def recompile_vl_prompt(self, request):
-        prompt = request.Prompt
-        image_paths = []
-
-        if "qwen-vl" in self.model_name.lower():
-            # request.Images is an array which contains base64 encoded images. Iterate the request.Images array, decode and save each image to /tmp folder with a random filename.
-            # Then, save the image file paths to an array "image_paths".
-            # read "request.Prompt", replace "[img-%d]" with the image file paths in the order they appear in "image_paths". Save the new prompt to "prompt".
-            for i, img in enumerate(request.Images):
-                timestamp = str(int(time.time() * 1000))  # Generate timestamp
-                img_path = f"/tmp/vl-{timestamp}.jpg"  # Use timestamp in filename
-                with open(img_path, "wb") as f:
-                    f.write(base64.b64decode(img))
-                image_paths.append(img_path)
-                prompt = prompt.replace(f"[img-{i}]", "<img>" + img_path + "</img>,")
-        else:
-            prompt = request.Prompt
-        return (prompt, image_paths)
 
 def serve(address):
     server = grpc.server(futures.ThreadPoolExecutor(max_workers=MAX_WORKERS))

backend/python/autogptq/autogptq.yml

@@ -1,7 +1,3 @@
-####
-# Attention! This file is abandoned. 
-# Please use the ../common-env/transformers/transformers.yml file to manage dependencies.
-###
 name: autogptq
 channels:
   - defaults
@@ -28,12 +24,12 @@ dependencies:
   - xz=5.4.2=h5eee18b_0
   - zlib=1.2.13=h5eee18b_0
   - pip:
-      - accelerate==0.27.0
+      - accelerate==0.23.0
       - aiohttp==3.8.5
       - aiosignal==1.3.1
       - async-timeout==4.0.3
       - attrs==23.1.0
-      - auto-gptq==0.7.1
+      - auto-gptq==0.4.2
       - certifi==2023.7.22
       - charset-normalizer==3.3.0
       - datasets==2.14.5
@@ -63,7 +59,6 @@ dependencies:
       - nvidia-nccl-cu12==2.18.1
       - nvidia-nvjitlink-cu12==12.2.140
       - nvidia-nvtx-cu12==12.1.105
-      - optimum==1.17.1
       - packaging==23.2
       - pandas==2.1.1
       - peft==0.5.0
@@ -80,11 +75,9 @@ dependencies:
       - six==1.16.0
       - sympy==1.12
       - tokenizers==0.14.0
+      - torch==2.1.0
       - tqdm==4.66.1
-      - torch==2.2.1
-      - torchvision==0.17.1
       - transformers==4.34.0
-      - transformers_stream_generator==0.0.5
       - triton==2.1.0
       - typing-extensions==4.8.0
       - tzdata==2023.3

backend/python/autogptq/backend_pb2_grpc.py

@@ -0,0 +1,363 @@
+# Generated by the gRPC Python protocol compiler plugin. DO NOT EDIT!
+"""Client and server classes corresponding to protobuf-defined services."""
+import grpc
+
+import backend_pb2 as backend__pb2
+
+
+class BackendStub(object):
+    """Missing associated documentation comment in .proto file."""
+
+    def __init__(self, channel):
+        """Constructor.
+
+        Args:
+            channel: A grpc.Channel.
+        """
+        self.Health = channel.unary_unary(
+                '/backend.Backend/Health',
+                request_serializer=backend__pb2.HealthMessage.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.Predict = channel.unary_unary(
+                '/backend.Backend/Predict',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.LoadModel = channel.unary_unary(
+                '/backend.Backend/LoadModel',
+                request_serializer=backend__pb2.ModelOptions.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.PredictStream = channel.unary_stream(
+                '/backend.Backend/PredictStream',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.Embedding = channel.unary_unary(
+                '/backend.Backend/Embedding',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.EmbeddingResult.FromString,
+                )
+        self.GenerateImage = channel.unary_unary(
+                '/backend.Backend/GenerateImage',
+                request_serializer=backend__pb2.GenerateImageRequest.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.AudioTranscription = channel.unary_unary(
+                '/backend.Backend/AudioTranscription',
+                request_serializer=backend__pb2.TranscriptRequest.SerializeToString,
+                response_deserializer=backend__pb2.TranscriptResult.FromString,
+                )
+        self.TTS = channel.unary_unary(
+                '/backend.Backend/TTS',
+                request_serializer=backend__pb2.TTSRequest.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.TokenizeString = channel.unary_unary(
+                '/backend.Backend/TokenizeString',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.TokenizationResponse.FromString,
+                )
+        self.Status = channel.unary_unary(
+                '/backend.Backend/Status',
+                request_serializer=backend__pb2.HealthMessage.SerializeToString,
+                response_deserializer=backend__pb2.StatusResponse.FromString,
+                )
+
+
+class BackendServicer(object):
+    """Missing associated documentation comment in .proto file."""
+
+    def Health(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Predict(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def LoadModel(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def PredictStream(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Embedding(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def GenerateImage(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def AudioTranscription(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def TTS(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def TokenizeString(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Status(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+
+def add_BackendServicer_to_server(servicer, server):
+    rpc_method_handlers = {
+            'Health': grpc.unary_unary_rpc_method_handler(
+                    servicer.Health,
+                    request_deserializer=backend__pb2.HealthMessage.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'Predict': grpc.unary_unary_rpc_method_handler(
+                    servicer.Predict,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'LoadModel': grpc.unary_unary_rpc_method_handler(
+                    servicer.LoadModel,
+                    request_deserializer=backend__pb2.ModelOptions.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'PredictStream': grpc.unary_stream_rpc_method_handler(
+                    servicer.PredictStream,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'Embedding': grpc.unary_unary_rpc_method_handler(
+                    servicer.Embedding,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.EmbeddingResult.SerializeToString,
+            ),
+            'GenerateImage': grpc.unary_unary_rpc_method_handler(
+                    servicer.GenerateImage,
+                    request_deserializer=backend__pb2.GenerateImageRequest.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'AudioTranscription': grpc.unary_unary_rpc_method_handler(
+                    servicer.AudioTranscription,
+                    request_deserializer=backend__pb2.TranscriptRequest.FromString,
+                    response_serializer=backend__pb2.TranscriptResult.SerializeToString,
+            ),
+            'TTS': grpc.unary_unary_rpc_method_handler(
+                    servicer.TTS,
+                    request_deserializer=backend__pb2.TTSRequest.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'TokenizeString': grpc.unary_unary_rpc_method_handler(
+                    servicer.TokenizeString,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.TokenizationResponse.SerializeToString,
+            ),
+            'Status': grpc.unary_unary_rpc_method_handler(
+                    servicer.Status,
+                    request_deserializer=backend__pb2.HealthMessage.FromString,
+                    response_serializer=backend__pb2.StatusResponse.SerializeToString,
+            ),
+    }
+    generic_handler = grpc.method_handlers_generic_handler(
+            'backend.Backend', rpc_method_handlers)
+    server.add_generic_rpc_handlers((generic_handler,))
+
+
+ # This class is part of an EXPERIMENTAL API.
+class Backend(object):
+    """Missing associated documentation comment in .proto file."""
+
+    @staticmethod
+    def Health(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Health',
+            backend__pb2.HealthMessage.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Predict(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Predict',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def LoadModel(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/LoadModel',
+            backend__pb2.ModelOptions.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def PredictStream(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_stream(request, target, '/backend.Backend/PredictStream',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Embedding(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Embedding',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.EmbeddingResult.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def GenerateImage(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/GenerateImage',
+            backend__pb2.GenerateImageRequest.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def AudioTranscription(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/AudioTranscription',
+            backend__pb2.TranscriptRequest.SerializeToString,
+            backend__pb2.TranscriptResult.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def TTS(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/TTS',
+            backend__pb2.TTSRequest.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def TokenizeString(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/TokenizeString',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.TokenizationResponse.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Status(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Status',
+            backend__pb2.HealthMessage.SerializeToString,
+            backend__pb2.StatusResponse.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)

backend/python/bark/Makefile

@@ -1,25 +1,15 @@
 .PHONY: ttsbark
-ttsbark: protogen
+ttsbark:
 	$(MAKE) -C ../common-env/transformers
 
 .PHONY: run
-run: protogen
+run:
 	@echo "Running bark..."
 	bash run.sh
 	@echo "bark run."
 
 .PHONY: test
-test: protogen
+test:
 	@echo "Testing bark..."
 	bash test.sh
 	@echo "bark tested."
-
-.PHONY: protogen
-protogen: backend_pb2_grpc.py backend_pb2.py
-
-.PHONY: protogen-clean
-protogen-clean:
-	$(RM) backend_pb2_grpc.py backend_pb2.py
-
-backend_pb2_grpc.py backend_pb2.py:
-	python3 -m grpc_tools.protoc -I../.. --python_out=. --grpc_python_out=. backend.proto

backend/python/bark/backend_pb2_grpc.py

@@ -0,0 +1,363 @@
+# Generated by the gRPC Python protocol compiler plugin. DO NOT EDIT!
+"""Client and server classes corresponding to protobuf-defined services."""
+import grpc
+
+import backend_pb2 as backend__pb2
+
+
+class BackendStub(object):
+    """Missing associated documentation comment in .proto file."""
+
+    def __init__(self, channel):
+        """Constructor.
+
+        Args:
+            channel: A grpc.Channel.
+        """
+        self.Health = channel.unary_unary(
+                '/backend.Backend/Health',
+                request_serializer=backend__pb2.HealthMessage.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.Predict = channel.unary_unary(
+                '/backend.Backend/Predict',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.LoadModel = channel.unary_unary(
+                '/backend.Backend/LoadModel',
+                request_serializer=backend__pb2.ModelOptions.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.PredictStream = channel.unary_stream(
+                '/backend.Backend/PredictStream',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.Embedding = channel.unary_unary(
+                '/backend.Backend/Embedding',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.EmbeddingResult.FromString,
+                )
+        self.GenerateImage = channel.unary_unary(
+                '/backend.Backend/GenerateImage',
+                request_serializer=backend__pb2.GenerateImageRequest.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.AudioTranscription = channel.unary_unary(
+                '/backend.Backend/AudioTranscription',
+                request_serializer=backend__pb2.TranscriptRequest.SerializeToString,
+                response_deserializer=backend__pb2.TranscriptResult.FromString,
+                )
+        self.TTS = channel.unary_unary(
+                '/backend.Backend/TTS',
+                request_serializer=backend__pb2.TTSRequest.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.TokenizeString = channel.unary_unary(
+                '/backend.Backend/TokenizeString',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.TokenizationResponse.FromString,
+                )
+        self.Status = channel.unary_unary(
+                '/backend.Backend/Status',
+                request_serializer=backend__pb2.HealthMessage.SerializeToString,
+                response_deserializer=backend__pb2.StatusResponse.FromString,
+                )
+
+
+class BackendServicer(object):
+    """Missing associated documentation comment in .proto file."""
+
+    def Health(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Predict(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def LoadModel(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def PredictStream(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Embedding(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def GenerateImage(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def AudioTranscription(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def TTS(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def TokenizeString(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Status(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+
+def add_BackendServicer_to_server(servicer, server):
+    rpc_method_handlers = {
+            'Health': grpc.unary_unary_rpc_method_handler(
+                    servicer.Health,
+                    request_deserializer=backend__pb2.HealthMessage.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'Predict': grpc.unary_unary_rpc_method_handler(
+                    servicer.Predict,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'LoadModel': grpc.unary_unary_rpc_method_handler(
+                    servicer.LoadModel,
+                    request_deserializer=backend__pb2.ModelOptions.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'PredictStream': grpc.unary_stream_rpc_method_handler(
+                    servicer.PredictStream,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'Embedding': grpc.unary_unary_rpc_method_handler(
+                    servicer.Embedding,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.EmbeddingResult.SerializeToString,
+            ),
+            'GenerateImage': grpc.unary_unary_rpc_method_handler(
+                    servicer.GenerateImage,
+                    request_deserializer=backend__pb2.GenerateImageRequest.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'AudioTranscription': grpc.unary_unary_rpc_method_handler(
+                    servicer.AudioTranscription,
+                    request_deserializer=backend__pb2.TranscriptRequest.FromString,
+                    response_serializer=backend__pb2.TranscriptResult.SerializeToString,
+            ),
+            'TTS': grpc.unary_unary_rpc_method_handler(
+                    servicer.TTS,
+                    request_deserializer=backend__pb2.TTSRequest.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'TokenizeString': grpc.unary_unary_rpc_method_handler(
+                    servicer.TokenizeString,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.TokenizationResponse.SerializeToString,
+            ),
+            'Status': grpc.unary_unary_rpc_method_handler(
+                    servicer.Status,
+                    request_deserializer=backend__pb2.HealthMessage.FromString,
+                    response_serializer=backend__pb2.StatusResponse.SerializeToString,
+            ),
+    }
+    generic_handler = grpc.method_handlers_generic_handler(
+            'backend.Backend', rpc_method_handlers)
+    server.add_generic_rpc_handlers((generic_handler,))
+
+
+ # This class is part of an EXPERIMENTAL API.
+class Backend(object):
+    """Missing associated documentation comment in .proto file."""
+
+    @staticmethod
+    def Health(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Health',
+            backend__pb2.HealthMessage.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Predict(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Predict',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def LoadModel(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/LoadModel',
+            backend__pb2.ModelOptions.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def PredictStream(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_stream(request, target, '/backend.Backend/PredictStream',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Embedding(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Embedding',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.EmbeddingResult.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def GenerateImage(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/GenerateImage',
+            backend__pb2.GenerateImageRequest.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def AudioTranscription(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/AudioTranscription',
+            backend__pb2.TranscriptRequest.SerializeToString,
+            backend__pb2.TranscriptResult.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def TTS(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/TTS',
+            backend__pb2.TTSRequest.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def TokenizeString(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/TokenizeString',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.TokenizationResponse.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Status(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Status',
+            backend__pb2.HealthMessage.SerializeToString,
+            backend__pb2.StatusResponse.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)

backend/python/common-env/transformers/install.sh

@@ -2,7 +2,6 @@
 set -ex
 
 SKIP_CONDA=${SKIP_CONDA:-0}
-REQUIREMENTS_FILE=$1
 
 # Check if environment exist
 conda_env_exists(){
@@ -15,7 +14,7 @@ else
     export PATH=$PATH:/opt/conda/bin
     if conda_env_exists "transformers" ; then
         echo "Creating virtual environment..."
-        conda env create --name transformers --file $REQUIREMENTS_FILE
+        conda env create --name transformers --file $1
         echo "Virtual environment created."
     else 
         echo "Virtual environment already exists."
@@ -26,19 +25,14 @@ if [ -d "/opt/intel" ]; then
     # Intel GPU: If the directory exists, we assume we are using the intel image
     # (no conda env)
     # https://github.com/intel/intel-extension-for-pytorch/issues/538
-    pip install intel-extension-for-transformers datasets sentencepiece tiktoken neural_speed optimum[openvino]
-fi
-
-# If we didn't skip conda, activate the environment
-# to install FlashAttention
-if [ $SKIP_CONDA -eq 0 ]; then
-    source activate transformers
-fi
-if [[ $REQUIREMENTS_FILE =~ -nvidia.yml$ ]]; then
-    #TODO: FlashAttention is supported on nvidia and ROCm, but ROCm install can't be done this easily
-    pip install flash-attn --no-build-isolation
+    pip install intel-extension-for-transformers datasets sentencepiece tiktoken neural_speed
 fi
 
 if [ "$PIP_CACHE_PURGE" = true ] ; then
+    if [ $SKIP_CONDA -eq 0 ]; then
+        # Activate conda environment
+        source activate transformers
+    fi
+
     pip cache purge
 fi

backend/python/common-env/transformers/transformers-nvidia.yml

@@ -24,14 +24,12 @@ dependencies:
   - xz=5.4.2=h5eee18b_0
   - zlib=1.2.13=h5eee18b_0
   - pip:
-      - accelerate==0.27.0
+      - accelerate==0.23.0
       - aiohttp==3.8.5
       - aiosignal==1.3.1
       - async-timeout==4.0.3
-      - auto-gptq==0.7.1
       - attrs==23.1.0
       - bark==0.1.5
-      - bitsandbytes==0.43.0
       - boto3==1.28.61
       - botocore==1.31.61
       - certifi==2023.7.22
@@ -70,7 +68,6 @@ dependencies:
       - nvidia-nccl-cu12==2.18.1
       - nvidia-nvjitlink-cu12==12.2.140
       - nvidia-nvtx-cu12==12.1.105
-      - optimum==1.17.1
       - packaging==23.2
       - pandas
       - peft==0.5.0
@@ -90,7 +87,6 @@ dependencies:
       - sympy==1.12
       - tokenizers
       - torch==2.1.2
-      - torchvision==0.16.2
       - torchaudio==2.1.2
       - tqdm==4.66.1
       - triton==2.1.0
@@ -98,6 +94,7 @@ dependencies:
       - tzdata==2023.3
       - urllib3==1.26.17
       - xxhash==3.4.1
+      - auto-gptq==0.6.0
       - yarl==1.9.2
       - soundfile
       - langid
@@ -116,10 +113,7 @@ dependencies:
       - sudachipy
       - sudachidict_core
       - vocos
-      - vllm>=0.4.0
+      - vllm==0.3.2
       - transformers>=4.38.2  # Updated Version
-      - transformers_stream_generator==0.0.5
       - xformers==0.0.23.post1  
-      - rerankers[transformers]
-      - pydantic
 prefix: /opt/conda/envs/transformers

backend/python/common-env/transformers/transformers-rocm.yml

@@ -26,8 +26,7 @@ dependencies:
   - pip:
       - --pre
       - --extra-index-url https://download.pytorch.org/whl/nightly/
-      - accelerate==0.27.0
-      - auto-gptq==0.7.1
+      - accelerate==0.23.0
       - aiohttp==3.8.5
       - aiosignal==1.3.1
       - async-timeout==4.0.3
@@ -83,6 +82,7 @@ dependencies:
       - triton==2.1.0
       - typing-extensions==4.8.0
       - tzdata==2023.3
+      - auto-gptq==0.6.0
       - urllib3==1.26.17
       - xxhash==3.4.1
       - yarl==1.9.2
@@ -90,7 +90,6 @@ dependencies:
       - langid
       - wget
       - unidecode
-      - optimum==1.17.1
       - pyopenjtalk-prebuilt
       - pypinyin
       - inflect
@@ -104,10 +103,7 @@ dependencies:
       - sudachipy
       - sudachidict_core
       - vocos
-      - vllm>=0.4.0
+      - vllm==0.3.2
       - transformers>=4.38.2  # Updated Version
-      - transformers_stream_generator==0.0.5
       - xformers==0.0.23.post1
-      - rerankers[transformers]
-      - pydantic
 prefix: /opt/conda/envs/transformers

backend/python/common-env/transformers/transformers.yml

@@ -24,17 +24,15 @@ dependencies:
   - xz=5.4.2=h5eee18b_0
   - zlib=1.2.13=h5eee18b_0
   - pip:
-      - accelerate==0.27.0
+      - accelerate==0.23.0
       - aiohttp==3.8.5
       - aiosignal==1.3.1
-      - auto-gptq==0.7.1
       - async-timeout==4.0.3
       - attrs==23.1.0
       - bark==0.1.5
       - boto3==1.28.61
       - botocore==1.31.61
       - certifi==2023.7.22
-      - coloredlogs==15.0.1
       - TTS==0.22.0
       - charset-normalizer==3.3.0
       - datasets==2.14.5
@@ -49,7 +47,6 @@ dependencies:
       - funcy==2.0
       - grpcio==1.59.0
       - huggingface-hub
-      - humanfriendly==10.0
       - idna==3.4
       - jinja2==3.1.2
       - jmespath==1.0.1
@@ -59,10 +56,6 @@ dependencies:
       - multiprocess==0.70.15
       - networkx
       - numpy==1.26.0
-      - onnx==1.15.0
-      - openvino==2024.0.0
-      - openvino-telemetry==2023.2.1
-      - optimum[openvino]==1.17.1
       - packaging==23.2
       - pandas
       - peft==0.5.0
@@ -82,12 +75,12 @@ dependencies:
       - sympy==1.12
       - tokenizers
       - torch==2.1.2
-      - torchvision==0.16.2
       - torchaudio==2.1.2
       - tqdm==4.66.1
       - triton==2.1.0
       - typing-extensions==4.8.0
       - tzdata==2023.3
+      - auto-gptq==0.6.0
       - urllib3==1.26.17
       - xxhash==3.4.1
       - yarl==1.9.2
@@ -108,10 +101,7 @@ dependencies:
       - sudachipy
       - sudachidict_core
       - vocos
-      - vllm>=0.4.0
+      - vllm==0.3.2
       - transformers>=4.38.2  # Updated Version
-      - transformers_stream_generator==0.0.5
-      - xformers==0.0.23.post1
-      - rerankers[transformers]
-      - pydantic
+      - xformers==0.0.23.post1  
 prefix: /opt/conda/envs/transformers

backend/python/coqui/Makefile

@@ -1,25 +1,15 @@
 .PHONY: coqui
-coqui: protogen
+coqui:
 	$(MAKE) -C ../common-env/transformers
 
 .PHONY: run
-run: protogen
+run:
 	@echo "Running coqui..."
 	bash run.sh
 	@echo "coqui run."
 
 .PHONY: test
-test: protogen
+test:
 	@echo "Testing coqui..."
 	bash test.sh
 	@echo "coqui tested."
-
-.PHONY: protogen
-protogen: backend_pb2_grpc.py backend_pb2.py
-
-.PHONY: protogen-clean
-protogen-clean:
-	$(RM) backend_pb2_grpc.py backend_pb2.py
-
-backend_pb2_grpc.py backend_pb2.py:
-	python3 -m grpc_tools.protoc -I../.. --python_out=. --grpc_python_out=. backend.proto

backend/python/coqui/backend_pb2_grpc.py

@@ -0,0 +1,363 @@
+# Generated by the gRPC Python protocol compiler plugin. DO NOT EDIT!
+"""Client and server classes corresponding to protobuf-defined services."""
+import grpc
+
+import backend_pb2 as backend__pb2
+
+
+class BackendStub(object):
+    """Missing associated documentation comment in .proto file."""
+
+    def __init__(self, channel):
+        """Constructor.
+
+        Args:
+            channel: A grpc.Channel.
+        """
+        self.Health = channel.unary_unary(
+                '/backend.Backend/Health',
+                request_serializer=backend__pb2.HealthMessage.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.Predict = channel.unary_unary(
+                '/backend.Backend/Predict',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.LoadModel = channel.unary_unary(
+                '/backend.Backend/LoadModel',
+                request_serializer=backend__pb2.ModelOptions.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.PredictStream = channel.unary_stream(
+                '/backend.Backend/PredictStream',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.Embedding = channel.unary_unary(
+                '/backend.Backend/Embedding',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.EmbeddingResult.FromString,
+                )
+        self.GenerateImage = channel.unary_unary(
+                '/backend.Backend/GenerateImage',
+                request_serializer=backend__pb2.GenerateImageRequest.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.AudioTranscription = channel.unary_unary(
+                '/backend.Backend/AudioTranscription',
+                request_serializer=backend__pb2.TranscriptRequest.SerializeToString,
+                response_deserializer=backend__pb2.TranscriptResult.FromString,
+                )
+        self.TTS = channel.unary_unary(
+                '/backend.Backend/TTS',
+                request_serializer=backend__pb2.TTSRequest.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.TokenizeString = channel.unary_unary(
+                '/backend.Backend/TokenizeString',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.TokenizationResponse.FromString,
+                )
+        self.Status = channel.unary_unary(
+                '/backend.Backend/Status',
+                request_serializer=backend__pb2.HealthMessage.SerializeToString,
+                response_deserializer=backend__pb2.StatusResponse.FromString,
+                )
+
+
+class BackendServicer(object):
+    """Missing associated documentation comment in .proto file."""
+
+    def Health(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Predict(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def LoadModel(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def PredictStream(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Embedding(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def GenerateImage(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def AudioTranscription(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def TTS(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def TokenizeString(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Status(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+
+def add_BackendServicer_to_server(servicer, server):
+    rpc_method_handlers = {
+            'Health': grpc.unary_unary_rpc_method_handler(
+                    servicer.Health,
+                    request_deserializer=backend__pb2.HealthMessage.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'Predict': grpc.unary_unary_rpc_method_handler(
+                    servicer.Predict,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'LoadModel': grpc.unary_unary_rpc_method_handler(
+                    servicer.LoadModel,
+                    request_deserializer=backend__pb2.ModelOptions.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'PredictStream': grpc.unary_stream_rpc_method_handler(
+                    servicer.PredictStream,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'Embedding': grpc.unary_unary_rpc_method_handler(
+                    servicer.Embedding,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.EmbeddingResult.SerializeToString,
+            ),
+            'GenerateImage': grpc.unary_unary_rpc_method_handler(
+                    servicer.GenerateImage,
+                    request_deserializer=backend__pb2.GenerateImageRequest.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'AudioTranscription': grpc.unary_unary_rpc_method_handler(
+                    servicer.AudioTranscription,
+                    request_deserializer=backend__pb2.TranscriptRequest.FromString,
+                    response_serializer=backend__pb2.TranscriptResult.SerializeToString,
+            ),
+            'TTS': grpc.unary_unary_rpc_method_handler(
+                    servicer.TTS,
+                    request_deserializer=backend__pb2.TTSRequest.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'TokenizeString': grpc.unary_unary_rpc_method_handler(
+                    servicer.TokenizeString,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.TokenizationResponse.SerializeToString,
+            ),
+            'Status': grpc.unary_unary_rpc_method_handler(
+                    servicer.Status,
+                    request_deserializer=backend__pb2.HealthMessage.FromString,
+                    response_serializer=backend__pb2.StatusResponse.SerializeToString,
+            ),
+    }
+    generic_handler = grpc.method_handlers_generic_handler(
+            'backend.Backend', rpc_method_handlers)
+    server.add_generic_rpc_handlers((generic_handler,))
+
+
+ # This class is part of an EXPERIMENTAL API.
+class Backend(object):
+    """Missing associated documentation comment in .proto file."""
+
+    @staticmethod
+    def Health(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Health',
+            backend__pb2.HealthMessage.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Predict(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Predict',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def LoadModel(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/LoadModel',
+            backend__pb2.ModelOptions.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def PredictStream(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_stream(request, target, '/backend.Backend/PredictStream',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Embedding(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Embedding',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.EmbeddingResult.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def GenerateImage(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/GenerateImage',
+            backend__pb2.GenerateImageRequest.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def AudioTranscription(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/AudioTranscription',
+            backend__pb2.TranscriptRequest.SerializeToString,
+            backend__pb2.TranscriptResult.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def TTS(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/TTS',
+            backend__pb2.TTSRequest.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def TokenizeString(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/TokenizeString',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.TokenizationResponse.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Status(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Status',
+            backend__pb2.HealthMessage.SerializeToString,
+            backend__pb2.StatusResponse.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)

backend/python/diffusers/Makefile

@@ -12,25 +12,15 @@ export SKIP_CONDA=1
 endif
 
 .PHONY: diffusers
-diffusers: protogen
+diffusers:
 	@echo "Installing $(CONDA_ENV_PATH)..."
 	bash install.sh $(CONDA_ENV_PATH)
 
 .PHONY: run
-run: protogen
+run:
 	@echo "Running diffusers..."
 	bash run.sh
 	@echo "Diffusers run."
 
-test: protogen
+test:
 	bash test.sh
-
-.PHONY: protogen
-protogen: backend_pb2_grpc.py backend_pb2.py
-
-.PHONY: protogen-clean
-protogen-clean:
-	$(RM) backend_pb2_grpc.py backend_pb2.py
-
-backend_pb2_grpc.py backend_pb2.py:
-	python3 -m grpc_tools.protoc -I../.. --python_out=. --grpc_python_out=. backend.proto

backend/python/diffusers/backend_pb2_grpc.py

@@ -0,0 +1,363 @@
+# Generated by the gRPC Python protocol compiler plugin. DO NOT EDIT!
+"""Client and server classes corresponding to protobuf-defined services."""
+import grpc
+
+import backend_pb2 as backend__pb2
+
+
+class BackendStub(object):
+    """Missing associated documentation comment in .proto file."""
+
+    def __init__(self, channel):
+        """Constructor.
+
+        Args:
+            channel: A grpc.Channel.
+        """
+        self.Health = channel.unary_unary(
+                '/backend.Backend/Health',
+                request_serializer=backend__pb2.HealthMessage.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.Predict = channel.unary_unary(
+                '/backend.Backend/Predict',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.LoadModel = channel.unary_unary(
+                '/backend.Backend/LoadModel',
+                request_serializer=backend__pb2.ModelOptions.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.PredictStream = channel.unary_stream(
+                '/backend.Backend/PredictStream',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.Embedding = channel.unary_unary(
+                '/backend.Backend/Embedding',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.EmbeddingResult.FromString,
+                )
+        self.GenerateImage = channel.unary_unary(
+                '/backend.Backend/GenerateImage',
+                request_serializer=backend__pb2.GenerateImageRequest.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.AudioTranscription = channel.unary_unary(
+                '/backend.Backend/AudioTranscription',
+                request_serializer=backend__pb2.TranscriptRequest.SerializeToString,
+                response_deserializer=backend__pb2.TranscriptResult.FromString,
+                )
+        self.TTS = channel.unary_unary(
+                '/backend.Backend/TTS',
+                request_serializer=backend__pb2.TTSRequest.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.TokenizeString = channel.unary_unary(
+                '/backend.Backend/TokenizeString',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.TokenizationResponse.FromString,
+                )
+        self.Status = channel.unary_unary(
+                '/backend.Backend/Status',
+                request_serializer=backend__pb2.HealthMessage.SerializeToString,
+                response_deserializer=backend__pb2.StatusResponse.FromString,
+                )
+
+
+class BackendServicer(object):
+    """Missing associated documentation comment in .proto file."""
+
+    def Health(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Predict(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def LoadModel(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def PredictStream(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Embedding(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def GenerateImage(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def AudioTranscription(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def TTS(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def TokenizeString(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Status(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+
+def add_BackendServicer_to_server(servicer, server):
+    rpc_method_handlers = {
+            'Health': grpc.unary_unary_rpc_method_handler(
+                    servicer.Health,
+                    request_deserializer=backend__pb2.HealthMessage.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'Predict': grpc.unary_unary_rpc_method_handler(
+                    servicer.Predict,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'LoadModel': grpc.unary_unary_rpc_method_handler(
+                    servicer.LoadModel,
+                    request_deserializer=backend__pb2.ModelOptions.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'PredictStream': grpc.unary_stream_rpc_method_handler(
+                    servicer.PredictStream,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'Embedding': grpc.unary_unary_rpc_method_handler(
+                    servicer.Embedding,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.EmbeddingResult.SerializeToString,
+            ),
+            'GenerateImage': grpc.unary_unary_rpc_method_handler(
+                    servicer.GenerateImage,
+                    request_deserializer=backend__pb2.GenerateImageRequest.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'AudioTranscription': grpc.unary_unary_rpc_method_handler(
+                    servicer.AudioTranscription,
+                    request_deserializer=backend__pb2.TranscriptRequest.FromString,
+                    response_serializer=backend__pb2.TranscriptResult.SerializeToString,
+            ),
+            'TTS': grpc.unary_unary_rpc_method_handler(
+                    servicer.TTS,
+                    request_deserializer=backend__pb2.TTSRequest.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'TokenizeString': grpc.unary_unary_rpc_method_handler(
+                    servicer.TokenizeString,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.TokenizationResponse.SerializeToString,
+            ),
+            'Status': grpc.unary_unary_rpc_method_handler(
+                    servicer.Status,
+                    request_deserializer=backend__pb2.HealthMessage.FromString,
+                    response_serializer=backend__pb2.StatusResponse.SerializeToString,
+            ),
+    }
+    generic_handler = grpc.method_handlers_generic_handler(
+            'backend.Backend', rpc_method_handlers)
+    server.add_generic_rpc_handlers((generic_handler,))
+
+
+ # This class is part of an EXPERIMENTAL API.
+class Backend(object):
+    """Missing associated documentation comment in .proto file."""
+
+    @staticmethod
+    def Health(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Health',
+            backend__pb2.HealthMessage.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Predict(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Predict',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def LoadModel(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/LoadModel',
+            backend__pb2.ModelOptions.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def PredictStream(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_stream(request, target, '/backend.Backend/PredictStream',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Embedding(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Embedding',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.EmbeddingResult.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def GenerateImage(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/GenerateImage',
+            backend__pb2.GenerateImageRequest.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def AudioTranscription(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/AudioTranscription',
+            backend__pb2.TranscriptRequest.SerializeToString,
+            backend__pb2.TranscriptResult.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def TTS(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/TTS',
+            backend__pb2.TTSRequest.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def TokenizeString(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/TokenizeString',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.TokenizationResponse.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Status(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Status',
+            backend__pb2.HealthMessage.SerializeToString,
+            backend__pb2.StatusResponse.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)

backend/python/diffusers/diffusers-rocm.yml

@@ -61,5 +61,4 @@ dependencies:
       - urllib3==2.0.6
       - zipp==3.17.0
       - torch
-      - opencv-python
 prefix: /opt/conda/envs/diffusers

backend/python/diffusers/diffusers.yml

@@ -71,5 +71,4 @@ dependencies:
       - typing-extensions==4.8.0
       - urllib3==2.0.6
       - zipp==3.17.0
-      - opencv-python
 prefix: /opt/conda/envs/diffusers

backend/python/exllama/Makefile

@@ -1,21 +1,11 @@
 export CONDA_ENV_PATH = "exllama.yml"
 
 .PHONY: exllama
-exllama: protogen
+exllama:
 	bash install.sh ${CONDA_ENV_PATH}
 
 .PHONY: run
-run: protogen
+run:
 	@echo "Running exllama..."
 	bash run.sh
 	@echo "exllama run."
-
-.PHONY: protogen
-protogen: backend_pb2_grpc.py backend_pb2.py
-
-.PHONY: protogen-clean
-protogen-clean:
-	$(RM) backend_pb2_grpc.py backend_pb2.py
-
-backend_pb2_grpc.py backend_pb2.py:
-	python3 -m grpc_tools.protoc -I../.. --python_out=. --grpc_python_out=. backend.proto

backend/python/exllama/backend_pb2_grpc.py

@@ -0,0 +1,363 @@
+# Generated by the gRPC Python protocol compiler plugin. DO NOT EDIT!
+"""Client and server classes corresponding to protobuf-defined services."""
+import grpc
+
+import backend_pb2 as backend__pb2
+
+
+class BackendStub(object):
+    """Missing associated documentation comment in .proto file."""
+
+    def __init__(self, channel):
+        """Constructor.
+
+        Args:
+            channel: A grpc.Channel.
+        """
+        self.Health = channel.unary_unary(
+                '/backend.Backend/Health',
+                request_serializer=backend__pb2.HealthMessage.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.Predict = channel.unary_unary(
+                '/backend.Backend/Predict',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.LoadModel = channel.unary_unary(
+                '/backend.Backend/LoadModel',
+                request_serializer=backend__pb2.ModelOptions.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.PredictStream = channel.unary_stream(
+                '/backend.Backend/PredictStream',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.Embedding = channel.unary_unary(
+                '/backend.Backend/Embedding',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.EmbeddingResult.FromString,
+                )
+        self.GenerateImage = channel.unary_unary(
+                '/backend.Backend/GenerateImage',
+                request_serializer=backend__pb2.GenerateImageRequest.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.AudioTranscription = channel.unary_unary(
+                '/backend.Backend/AudioTranscription',
+                request_serializer=backend__pb2.TranscriptRequest.SerializeToString,
+                response_deserializer=backend__pb2.TranscriptResult.FromString,
+                )
+        self.TTS = channel.unary_unary(
+                '/backend.Backend/TTS',
+                request_serializer=backend__pb2.TTSRequest.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.TokenizeString = channel.unary_unary(
+                '/backend.Backend/TokenizeString',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.TokenizationResponse.FromString,
+                )
+        self.Status = channel.unary_unary(
+                '/backend.Backend/Status',
+                request_serializer=backend__pb2.HealthMessage.SerializeToString,
+                response_deserializer=backend__pb2.StatusResponse.FromString,
+                )
+
+
+class BackendServicer(object):
+    """Missing associated documentation comment in .proto file."""
+
+    def Health(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Predict(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def LoadModel(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def PredictStream(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Embedding(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def GenerateImage(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def AudioTranscription(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def TTS(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def TokenizeString(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Status(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+
+def add_BackendServicer_to_server(servicer, server):
+    rpc_method_handlers = {
+            'Health': grpc.unary_unary_rpc_method_handler(
+                    servicer.Health,
+                    request_deserializer=backend__pb2.HealthMessage.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'Predict': grpc.unary_unary_rpc_method_handler(
+                    servicer.Predict,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'LoadModel': grpc.unary_unary_rpc_method_handler(
+                    servicer.LoadModel,
+                    request_deserializer=backend__pb2.ModelOptions.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'PredictStream': grpc.unary_stream_rpc_method_handler(
+                    servicer.PredictStream,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'Embedding': grpc.unary_unary_rpc_method_handler(
+                    servicer.Embedding,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.EmbeddingResult.SerializeToString,
+            ),
+            'GenerateImage': grpc.unary_unary_rpc_method_handler(
+                    servicer.GenerateImage,
+                    request_deserializer=backend__pb2.GenerateImageRequest.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'AudioTranscription': grpc.unary_unary_rpc_method_handler(
+                    servicer.AudioTranscription,
+                    request_deserializer=backend__pb2.TranscriptRequest.FromString,
+                    response_serializer=backend__pb2.TranscriptResult.SerializeToString,
+            ),
+            'TTS': grpc.unary_unary_rpc_method_handler(
+                    servicer.TTS,
+                    request_deserializer=backend__pb2.TTSRequest.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'TokenizeString': grpc.unary_unary_rpc_method_handler(
+                    servicer.TokenizeString,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.TokenizationResponse.SerializeToString,
+            ),
+            'Status': grpc.unary_unary_rpc_method_handler(
+                    servicer.Status,
+                    request_deserializer=backend__pb2.HealthMessage.FromString,
+                    response_serializer=backend__pb2.StatusResponse.SerializeToString,
+            ),
+    }
+    generic_handler = grpc.method_handlers_generic_handler(
+            'backend.Backend', rpc_method_handlers)
+    server.add_generic_rpc_handlers((generic_handler,))
+
+
+ # This class is part of an EXPERIMENTAL API.
+class Backend(object):
+    """Missing associated documentation comment in .proto file."""
+
+    @staticmethod
+    def Health(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Health',
+            backend__pb2.HealthMessage.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Predict(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Predict',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def LoadModel(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/LoadModel',
+            backend__pb2.ModelOptions.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def PredictStream(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_stream(request, target, '/backend.Backend/PredictStream',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Embedding(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Embedding',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.EmbeddingResult.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def GenerateImage(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/GenerateImage',
+            backend__pb2.GenerateImageRequest.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def AudioTranscription(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/AudioTranscription',
+            backend__pb2.TranscriptRequest.SerializeToString,
+            backend__pb2.TranscriptResult.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def TTS(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/TTS',
+            backend__pb2.TTSRequest.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def TokenizeString(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/TokenizeString',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.TokenizationResponse.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Status(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Status',
+            backend__pb2.HealthMessage.SerializeToString,
+            backend__pb2.StatusResponse.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)

backend/python/exllama2/Makefile

@@ -1,20 +1,10 @@
 .PHONY: exllama2
-exllama2: protogen
+exllama2:
 	$(MAKE) -C ../common-env/transformers
 	bash install.sh
 
 .PHONY: run
-run: protogen
+run:
 	@echo "Running exllama2..."
 	bash run.sh
 	@echo "exllama2 run."
-
-.PHONY: protogen
-protogen: backend_pb2_grpc.py backend_pb2.py
-
-.PHONY: protogen-clean
-protogen-clean:
-	$(RM) backend_pb2_grpc.py backend_pb2.py
-
-backend_pb2_grpc.py backend_pb2.py:
-	python3 -m grpc_tools.protoc -I../.. --python_out=. --grpc_python_out=. backend.proto

backend/python/exllama2/backend_pb2_grpc.py

@@ -0,0 +1,363 @@
+# Generated by the gRPC Python protocol compiler plugin. DO NOT EDIT!
+"""Client and server classes corresponding to protobuf-defined services."""
+import grpc
+
+import backend_pb2 as backend__pb2
+
+
+class BackendStub(object):
+    """Missing associated documentation comment in .proto file."""
+
+    def __init__(self, channel):
+        """Constructor.
+
+        Args:
+            channel: A grpc.Channel.
+        """
+        self.Health = channel.unary_unary(
+                '/backend.Backend/Health',
+                request_serializer=backend__pb2.HealthMessage.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.Predict = channel.unary_unary(
+                '/backend.Backend/Predict',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.LoadModel = channel.unary_unary(
+                '/backend.Backend/LoadModel',
+                request_serializer=backend__pb2.ModelOptions.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.PredictStream = channel.unary_stream(
+                '/backend.Backend/PredictStream',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.Embedding = channel.unary_unary(
+                '/backend.Backend/Embedding',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.EmbeddingResult.FromString,
+                )
+        self.GenerateImage = channel.unary_unary(
+                '/backend.Backend/GenerateImage',
+                request_serializer=backend__pb2.GenerateImageRequest.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.AudioTranscription = channel.unary_unary(
+                '/backend.Backend/AudioTranscription',
+                request_serializer=backend__pb2.TranscriptRequest.SerializeToString,
+                response_deserializer=backend__pb2.TranscriptResult.FromString,
+                )
+        self.TTS = channel.unary_unary(
+                '/backend.Backend/TTS',
+                request_serializer=backend__pb2.TTSRequest.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.TokenizeString = channel.unary_unary(
+                '/backend.Backend/TokenizeString',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.TokenizationResponse.FromString,
+                )
+        self.Status = channel.unary_unary(
+                '/backend.Backend/Status',
+                request_serializer=backend__pb2.HealthMessage.SerializeToString,
+                response_deserializer=backend__pb2.StatusResponse.FromString,
+                )
+
+
+class BackendServicer(object):
+    """Missing associated documentation comment in .proto file."""
+
+    def Health(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Predict(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def LoadModel(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def PredictStream(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Embedding(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def GenerateImage(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def AudioTranscription(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def TTS(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def TokenizeString(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Status(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+
+def add_BackendServicer_to_server(servicer, server):
+    rpc_method_handlers = {
+            'Health': grpc.unary_unary_rpc_method_handler(
+                    servicer.Health,
+                    request_deserializer=backend__pb2.HealthMessage.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'Predict': grpc.unary_unary_rpc_method_handler(
+                    servicer.Predict,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'LoadModel': grpc.unary_unary_rpc_method_handler(
+                    servicer.LoadModel,
+                    request_deserializer=backend__pb2.ModelOptions.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'PredictStream': grpc.unary_stream_rpc_method_handler(
+                    servicer.PredictStream,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'Embedding': grpc.unary_unary_rpc_method_handler(
+                    servicer.Embedding,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.EmbeddingResult.SerializeToString,
+            ),
+            'GenerateImage': grpc.unary_unary_rpc_method_handler(
+                    servicer.GenerateImage,
+                    request_deserializer=backend__pb2.GenerateImageRequest.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'AudioTranscription': grpc.unary_unary_rpc_method_handler(
+                    servicer.AudioTranscription,
+                    request_deserializer=backend__pb2.TranscriptRequest.FromString,
+                    response_serializer=backend__pb2.TranscriptResult.SerializeToString,
+            ),
+            'TTS': grpc.unary_unary_rpc_method_handler(
+                    servicer.TTS,
+                    request_deserializer=backend__pb2.TTSRequest.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'TokenizeString': grpc.unary_unary_rpc_method_handler(
+                    servicer.TokenizeString,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.TokenizationResponse.SerializeToString,
+            ),
+            'Status': grpc.unary_unary_rpc_method_handler(
+                    servicer.Status,
+                    request_deserializer=backend__pb2.HealthMessage.FromString,
+                    response_serializer=backend__pb2.StatusResponse.SerializeToString,
+            ),
+    }
+    generic_handler = grpc.method_handlers_generic_handler(
+            'backend.Backend', rpc_method_handlers)
+    server.add_generic_rpc_handlers((generic_handler,))
+
+
+ # This class is part of an EXPERIMENTAL API.
+class Backend(object):
+    """Missing associated documentation comment in .proto file."""
+
+    @staticmethod
+    def Health(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Health',
+            backend__pb2.HealthMessage.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Predict(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Predict',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def LoadModel(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/LoadModel',
+            backend__pb2.ModelOptions.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def PredictStream(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_stream(request, target, '/backend.Backend/PredictStream',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Embedding(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Embedding',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.EmbeddingResult.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def GenerateImage(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/GenerateImage',
+            backend__pb2.GenerateImageRequest.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def AudioTranscription(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/AudioTranscription',
+            backend__pb2.TranscriptRequest.SerializeToString,
+            backend__pb2.TranscriptResult.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def TTS(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/TTS',
+            backend__pb2.TTSRequest.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def TokenizeString(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/TokenizeString',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.TokenizationResponse.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Status(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Status',
+            backend__pb2.HealthMessage.SerializeToString,
+            backend__pb2.StatusResponse.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)

backend/python/mamba/Makefile

@@ -1,26 +1,16 @@
 .PHONY: mamba
-mamba: protogen
+mamba:
 	$(MAKE) -C ../common-env/transformers
 	bash install.sh
 
 .PHONY: run
-run: protogen
+run:
 	@echo "Running mamba..."
 	bash run.sh
 	@echo "mamba run."
 
 .PHONY: test
-test: protogen
+test:
 	@echo "Testing mamba..."
 	bash test.sh
-	@echo "mamba tested."
-
-.PHONY: protogen
-protogen: backend_pb2_grpc.py backend_pb2.py
-
-.PHONY: protogen-clean
-protogen-clean:
-	$(RM) backend_pb2_grpc.py backend_pb2.py
-
-backend_pb2_grpc.py backend_pb2.py:
-	python3 -m grpc_tools.protoc -I../.. --python_out=. --grpc_python_out=. backend.proto
+	@echo "mamba tested."

backend/python/mamba/backend_pb2_grpc.py

@@ -0,0 +1,363 @@
+# Generated by the gRPC Python protocol compiler plugin. DO NOT EDIT!
+"""Client and server classes corresponding to protobuf-defined services."""
+import grpc
+
+import backend_pb2 as backend__pb2
+
+
+class BackendStub(object):
+    """Missing associated documentation comment in .proto file."""
+
+    def __init__(self, channel):
+        """Constructor.
+
+        Args:
+            channel: A grpc.Channel.
+        """
+        self.Health = channel.unary_unary(
+                '/backend.Backend/Health',
+                request_serializer=backend__pb2.HealthMessage.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.Predict = channel.unary_unary(
+                '/backend.Backend/Predict',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.LoadModel = channel.unary_unary(
+                '/backend.Backend/LoadModel',
+                request_serializer=backend__pb2.ModelOptions.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.PredictStream = channel.unary_stream(
+                '/backend.Backend/PredictStream',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.Embedding = channel.unary_unary(
+                '/backend.Backend/Embedding',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.EmbeddingResult.FromString,
+                )
+        self.GenerateImage = channel.unary_unary(
+                '/backend.Backend/GenerateImage',
+                request_serializer=backend__pb2.GenerateImageRequest.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.AudioTranscription = channel.unary_unary(
+                '/backend.Backend/AudioTranscription',
+                request_serializer=backend__pb2.TranscriptRequest.SerializeToString,
+                response_deserializer=backend__pb2.TranscriptResult.FromString,
+                )
+        self.TTS = channel.unary_unary(
+                '/backend.Backend/TTS',
+                request_serializer=backend__pb2.TTSRequest.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.TokenizeString = channel.unary_unary(
+                '/backend.Backend/TokenizeString',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.TokenizationResponse.FromString,
+                )
+        self.Status = channel.unary_unary(
+                '/backend.Backend/Status',
+                request_serializer=backend__pb2.HealthMessage.SerializeToString,
+                response_deserializer=backend__pb2.StatusResponse.FromString,
+                )
+
+
+class BackendServicer(object):
+    """Missing associated documentation comment in .proto file."""
+
+    def Health(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Predict(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def LoadModel(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def PredictStream(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Embedding(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def GenerateImage(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def AudioTranscription(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def TTS(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def TokenizeString(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Status(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+
+def add_BackendServicer_to_server(servicer, server):
+    rpc_method_handlers = {
+            'Health': grpc.unary_unary_rpc_method_handler(
+                    servicer.Health,
+                    request_deserializer=backend__pb2.HealthMessage.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'Predict': grpc.unary_unary_rpc_method_handler(
+                    servicer.Predict,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'LoadModel': grpc.unary_unary_rpc_method_handler(
+                    servicer.LoadModel,
+                    request_deserializer=backend__pb2.ModelOptions.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'PredictStream': grpc.unary_stream_rpc_method_handler(
+                    servicer.PredictStream,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'Embedding': grpc.unary_unary_rpc_method_handler(
+                    servicer.Embedding,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.EmbeddingResult.SerializeToString,
+            ),
+            'GenerateImage': grpc.unary_unary_rpc_method_handler(
+                    servicer.GenerateImage,
+                    request_deserializer=backend__pb2.GenerateImageRequest.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'AudioTranscription': grpc.unary_unary_rpc_method_handler(
+                    servicer.AudioTranscription,
+                    request_deserializer=backend__pb2.TranscriptRequest.FromString,
+                    response_serializer=backend__pb2.TranscriptResult.SerializeToString,
+            ),
+            'TTS': grpc.unary_unary_rpc_method_handler(
+                    servicer.TTS,
+                    request_deserializer=backend__pb2.TTSRequest.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'TokenizeString': grpc.unary_unary_rpc_method_handler(
+                    servicer.TokenizeString,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.TokenizationResponse.SerializeToString,
+            ),
+            'Status': grpc.unary_unary_rpc_method_handler(
+                    servicer.Status,
+                    request_deserializer=backend__pb2.HealthMessage.FromString,
+                    response_serializer=backend__pb2.StatusResponse.SerializeToString,
+            ),
+    }
+    generic_handler = grpc.method_handlers_generic_handler(
+            'backend.Backend', rpc_method_handlers)
+    server.add_generic_rpc_handlers((generic_handler,))
+
+
+ # This class is part of an EXPERIMENTAL API.
+class Backend(object):
+    """Missing associated documentation comment in .proto file."""
+
+    @staticmethod
+    def Health(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Health',
+            backend__pb2.HealthMessage.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Predict(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Predict',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def LoadModel(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/LoadModel',
+            backend__pb2.ModelOptions.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def PredictStream(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_stream(request, target, '/backend.Backend/PredictStream',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Embedding(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Embedding',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.EmbeddingResult.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def GenerateImage(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/GenerateImage',
+            backend__pb2.GenerateImageRequest.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def AudioTranscription(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/AudioTranscription',
+            backend__pb2.TranscriptRequest.SerializeToString,
+            backend__pb2.TranscriptResult.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def TTS(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/TTS',
+            backend__pb2.TTSRequest.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def TokenizeString(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/TokenizeString',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.TokenizationResponse.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Status(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Status',
+            backend__pb2.HealthMessage.SerializeToString,
+            backend__pb2.StatusResponse.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)

backend/python/parler-tts/Makefile

@@ -1,39 +0,0 @@
-export CONDA_ENV_PATH = "parler.yml"
-SKIP_CONDA?=0
-ifeq ($(BUILD_TYPE), cublas)
-export CONDA_ENV_PATH = "parler-nvidia.yml"
-endif
-
-# Intel GPU are supposed to have dependencies installed in the main python
-# environment, so we skip conda installation for SYCL builds.
-# https://github.com/intel/intel-extension-for-pytorch/issues/538
-ifneq (,$(findstring sycl,$(BUILD_TYPE)))
-export SKIP_CONDA=1
-endif
-
-.PHONY: parler-tts
-parler-tts: protogen
-	@echo "Installing $(CONDA_ENV_PATH)..."
-	bash install.sh $(CONDA_ENV_PATH)
-
-.PHONY: run
-run: protogen
-	@echo "Running transformers..."
-	bash run.sh
-	@echo "transformers run."
-
-.PHONY: test
-test: protogen
-	@echo "Testing transformers..."
-	bash test.sh
-	@echo "transformers tested."
-
-.PHONY: protogen
-protogen: backend_pb2_grpc.py backend_pb2.py
-
-.PHONY: protogen-clean
-protogen-clean:
-	$(RM) backend_pb2_grpc.py backend_pb2.py
-
-backend_pb2_grpc.py backend_pb2.py:
-	python3 -m grpc_tools.protoc -I../.. --python_out=. --grpc_python_out=. backend.proto

backend/python/parler-tts/install.sh

@@ -1,39 +0,0 @@
-#!/bin/bash
-set -ex
-
-SKIP_CONDA=${SKIP_CONDA:-0}
-
-# Check if environment exist
-conda_env_exists(){
-    ! conda list --name "${@}" >/dev/null 2>/dev/null
-}
-
-if [ $SKIP_CONDA -eq 1 ]; then
-    echo "Skipping conda environment installation"
-else
-    export PATH=$PATH:/opt/conda/bin
-    if conda_env_exists "parler" ; then
-        echo "Creating virtual environment..."
-        conda env create --name parler --file $1
-        echo "Virtual environment created."
-    else 
-        echo "Virtual environment already exists."
-    fi
-fi
-
-if [ $SKIP_CONDA -ne 1 ]; then
-    # Activate conda environment
-    source activate parler
-    # https://github.com/descriptinc/audiotools/issues/101
-    # incompatible protobuf versions.
-    curl -L https://raw.githubusercontent.com/protocolbuffers/protobuf/main/python/google/protobuf/internal/builder.py -o $CONDA_PREFIX/lib/python3.11/site-packages/google/protobuf/internal/builder.py
-fi
-
-if [ "$PIP_CACHE_PURGE" = true ] ; then
-    if [ $SKIP_CONDA -ne 1 ]; then
-        # Activate conda environment
-        source activate parler
-    fi
-
-    pip cache purge
-fi

backend/python/parler-tts/parler-nvidia.yml

@@ -1,48 +0,0 @@
-name: parler
-channels:
-  - defaults
-dependencies:
-  - _libgcc_mutex=0.1=main
-  - _openmp_mutex=5.1=1_gnu
-  - bzip2=1.0.8=h7b6447c_0
-  - ca-certificates=2023.08.22=h06a4308_0
-  - ld_impl_linux-64=2.38=h1181459_1
-  - libffi=3.4.4=h6a678d5_0
-  - libgcc-ng=11.2.0=h1234567_1
-  - libgomp=11.2.0=h1234567_1
-  - libstdcxx-ng=11.2.0=h1234567_1
-  - libuuid=1.41.5=h5eee18b_0
-  - ncurses=6.4=h6a678d5_0
-  - openssl=3.0.11=h7f8727e_2
-  - pip=23.2.1=py311h06a4308_0
-  - python=3.11.5=h955ad1f_0
-  - readline=8.2=h5eee18b_0
-  - setuptools=68.0.0=py311h06a4308_0
-  - sqlite=3.41.2=h5eee18b_0
-  - tk=8.6.12=h1ccaba5_0
-  - tzdata=2023c=h04d1e81_0
-  - wheel=0.41.2=py311h06a4308_0
-  - xz=5.4.2=h5eee18b_0
-  - zlib=1.2.13=h5eee18b_0
-  - pip:
-      - accelerate>=0.11.0
-      - grpcio==1.59.0
-      - numpy==1.26.0
-      - nvidia-cublas-cu12==12.1.3.1
-      - nvidia-cuda-cupti-cu12==12.1.105
-      - nvidia-cuda-nvrtc-cu12==12.1.105
-      - nvidia-cuda-runtime-cu12==12.1.105
-      - nvidia-cudnn-cu12==8.9.2.26
-      - nvidia-cufft-cu12==11.0.2.54
-      - nvidia-curand-cu12==10.3.2.106
-      - nvidia-cusolver-cu12==11.4.5.107
-      - nvidia-cusparse-cu12==12.1.0.106
-      - nvidia-nccl-cu12==2.18.1
-      - nvidia-nvjitlink-cu12==12.2.140
-      - nvidia-nvtx-cu12==12.1.105
-      - torch==2.1.0
-      - transformers>=4.34.0
-      - descript-audio-codec
-      - sentencepiece
-      - git+https://github.com/huggingface/parler-tts.git@10016fb0300c0dc31a0fb70e26f3affee7b62f16
-prefix: /opt/conda/envs/diffusers

backend/python/parler-tts/parler.yml

@@ -1,36 +0,0 @@
-name: parler
-channels:
-  - defaults
-dependencies:
-  - _libgcc_mutex=0.1=main
-  - _openmp_mutex=5.1=1_gnu
-  - bzip2=1.0.8=h7b6447c_0
-  - ca-certificates=2023.08.22=h06a4308_0
-  - ld_impl_linux-64=2.38=h1181459_1
-  - libffi=3.4.4=h6a678d5_0
-  - libgcc-ng=11.2.0=h1234567_1
-  - libgomp=11.2.0=h1234567_1
-  - libstdcxx-ng=11.2.0=h1234567_1
-  - libuuid=1.41.5=h5eee18b_0
-  - ncurses=6.4=h6a678d5_0
-  - openssl=3.0.11=h7f8727e_2
-  - pip=23.2.1=py311h06a4308_0
-  - python=3.11.5=h955ad1f_0
-  - readline=8.2=h5eee18b_0
-  - setuptools=68.0.0=py311h06a4308_0
-  - sqlite=3.41.2=h5eee18b_0
-  - tk=8.6.12=h1ccaba5_0
-  - tzdata=2023c=h04d1e81_0
-  - wheel=0.41.2=py311h06a4308_0
-  - xz=5.4.2=h5eee18b_0
-  - zlib=1.2.13=h5eee18b_0
-  - pip:
-      - accelerate>=0.11.0
-      - numpy==1.26.0
-      - grpcio==1.59.0
-      - torch==2.1.0
-      - transformers>=4.34.0
-      - descript-audio-codec
-      - sentencepiece
-      - git+https://github.com/huggingface/parler-tts.git@10016fb0300c0dc31a0fb70e26f3affee7b62f16
-prefix: /opt/conda/envs/parler

backend/python/parler-tts/parler_tts_server.py

@@ -1,125 +0,0 @@
-#!/usr/bin/env python3
-"""
-Extra gRPC server for MusicgenForConditionalGeneration models.
-"""
-from concurrent import futures
-
-import argparse
-import signal
-import sys
-import os
-
-import time
-import backend_pb2
-import backend_pb2_grpc
-
-import grpc
-
-from scipy.io.wavfile import write as write_wav
-
-from parler_tts import ParlerTTSForConditionalGeneration
-from transformers import AutoTokenizer
-import soundfile as sf  
-import torch
-
-_ONE_DAY_IN_SECONDS = 60 * 60 * 24
-
-# If MAX_WORKERS are specified in the environment use it, otherwise default to 1
-MAX_WORKERS = int(os.environ.get('PYTHON_GRPC_MAX_WORKERS', '1'))
-
-# Implement the BackendServicer class with the service methods
-class BackendServicer(backend_pb2_grpc.BackendServicer):
-    """
-    A gRPC servicer for the backend service.
-
-    This class implements the gRPC methods for the backend service, including Health, LoadModel, and Embedding.
-    """
-    def Health(self, request, context):
-        """
-        A gRPC method that returns the health status of the backend service.
-
-        Args:
-            request: A HealthRequest object that contains the request parameters.
-            context: A grpc.ServicerContext object that provides information about the RPC.
-
-        Returns:
-            A Reply object that contains the health status of the backend service.
-        """
-        return backend_pb2.Reply(message=bytes("OK", 'utf-8'))
-
-    def LoadModel(self, request, context):
-        """
-        A gRPC method that loads a model into memory.
-
-        Args:
-            request: A LoadModelRequest object that contains the request parameters.
-            context: A grpc.ServicerContext object that provides information about the RPC.
-
-        Returns:
-            A Result object that contains the result of the LoadModel operation.
-        """
-        model_name = request.Model
-        device = "cuda:0" if torch.cuda.is_available() else "cpu"
-        try:
-            self.model = ParlerTTSForConditionalGeneration.from_pretrained(model_name).to(device)
-            self.tokenizer = AutoTokenizer.from_pretrained(model_name)
-        except Exception as err:
-            return backend_pb2.Result(success=False, message=f"Unexpected {err=}, {type(err)=}")
-
-        return backend_pb2.Result(message="Model loaded successfully", success=True)
-
-    def TTS(self, request, context):
-        model_name = request.model
-        voice = request.voice
-        if voice == "":
-            voice = "A female speaker with a slightly low-pitched voice delivers her words quite expressively, in a very confined sounding environment with clear audio quality. She speaks very fast."
-        if model_name == "":
-            return backend_pb2.Result(success=False, message="request.model is required")
-        try:
-            device = "cuda:0" if torch.cuda.is_available() else "cpu"
-            input_ids = self.tokenizer(voice, return_tensors="pt").input_ids.to(device)
-            prompt_input_ids = self.tokenizer(request.text, return_tensors="pt").input_ids.to(device)
-           
-            generation = self.model.generate(input_ids=input_ids, prompt_input_ids=prompt_input_ids)
-            audio_arr = generation.cpu().numpy().squeeze()
-            print("[parler-tts] TTS generated!", file=sys.stderr)
-            sf.write(request.dst, audio_arr, self.model.config.sampling_rate)
-            print("[parler-tts] TTS saved to", request.dst, file=sys.stderr)
-            print("[parler-tts] TTS for", file=sys.stderr)
-            print(request, file=sys.stderr)
-        except Exception as err:
-            return backend_pb2.Result(success=False, message=f"Unexpected {err=}, {type(err)=}")
-        return backend_pb2.Result(success=True)
-
-
-def serve(address):
-    server = grpc.server(futures.ThreadPoolExecutor(max_workers=MAX_WORKERS))
-    backend_pb2_grpc.add_BackendServicer_to_server(BackendServicer(), server)
-    server.add_insecure_port(address)
-    server.start()
-    print("[parler-tts] Server started. Listening on: " + address, file=sys.stderr)
-
-    # Define the signal handler function
-    def signal_handler(sig, frame):
-        print("[parler-tts] Received termination signal. Shutting down...")
-        server.stop(0)
-        sys.exit(0)
-
-    # Set the signal handlers for SIGINT and SIGTERM
-    signal.signal(signal.SIGINT, signal_handler)
-    signal.signal(signal.SIGTERM, signal_handler)
-
-    try:
-        while True:
-            time.sleep(_ONE_DAY_IN_SECONDS)
-    except KeyboardInterrupt:
-        server.stop(0)
-
-if __name__ == "__main__":
-    parser = argparse.ArgumentParser(description="Run the gRPC server.")
-    parser.add_argument(
-        "--addr", default="localhost:50051", help="The address to bind the server to."
-    )
-    args = parser.parse_args()
-    print(f"[parler-tts] startup: {args}", file=sys.stderr)
-    serve(args.addr)

backend/python/parler-tts/run.sh

@@ -1,16 +0,0 @@
-#!/bin/bash
-
-##
-## A bash script wrapper that runs the parler-tts server with conda
-
-echo "Launching gRPC server for parler-tts"
-
-export PATH=$PATH:/opt/conda/bin
-
-# Activate conda environment
-source activate parler
-
-# get the directory where the bash script is located
-DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" >/dev/null 2>&1 && pwd )"
-
-python $DIR/parler_tts_server.py $@

backend/python/parler-tts/test.sh

@@ -1,11 +0,0 @@
-#!/bin/bash
-##
-## A bash script wrapper that runs the transformers server with conda
-
-# Activate conda environment
-source activate parler
-
-# get the directory where the bash script is located
-DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" >/dev/null 2>&1 && pwd )"
-
-python -m unittest $DIR/test_parler.py

backend/python/parler-tts/test_parler.py

@@ -1,81 +0,0 @@
-"""
-A test script to test the gRPC service
-"""
-import unittest
-import subprocess
-import time
-import backend_pb2
-import backend_pb2_grpc
-
-import grpc
-
-
-class TestBackendServicer(unittest.TestCase):
-    """
-    TestBackendServicer is the class that tests the gRPC service
-    """
-    def setUp(self):
-        """
-        This method sets up the gRPC service by starting the server
-        """
-        self.service = subprocess.Popen(["python3", "parler_tts_server.py", "--addr", "localhost:50051"])
-        time.sleep(10)
-
-    def tearDown(self) -> None:
-        """
-        This method tears down the gRPC service by terminating the server
-        """
-        self.service.terminate()
-        self.service.wait()
-
-    def test_server_startup(self):
-        """
-        This method tests if the server starts up successfully
-        """
-        try:
-            self.setUp()
-            with grpc.insecure_channel("localhost:50051") as channel:
-                stub = backend_pb2_grpc.BackendStub(channel)
-                response = stub.Health(backend_pb2.HealthMessage())
-                self.assertEqual(response.message, b'OK')
-        except Exception as err:
-            print(err)
-            self.fail("Server failed to start")
-        finally:
-            self.tearDown()
-
-    def test_load_model(self):
-        """
-        This method tests if the model is loaded successfully
-        """
-        try:
-            self.setUp()
-            with grpc.insecure_channel("localhost:50051") as channel:
-                stub = backend_pb2_grpc.BackendStub(channel)
-                response = stub.LoadModel(backend_pb2.ModelOptions(Model="parler-tts/parler_tts_mini_v0.1"))
-                self.assertTrue(response.success)
-                self.assertEqual(response.message, "Model loaded successfully")
-        except Exception as err:
-            print(err)
-            self.fail("LoadModel service failed")
-        finally:
-            self.tearDown()
-
-    def test_tts(self):
-        """
-        This method tests if the embeddings are generated successfully
-        """
-        try:
-            self.setUp()
-            with grpc.insecure_channel("localhost:50051") as channel:
-                stub = backend_pb2_grpc.BackendStub(channel)
-                response = stub.LoadModel(backend_pb2.ModelOptions(Model="parler-tts/parler_tts_mini_v0.1"))
-                self.assertTrue(response.success)
-                tts_request = backend_pb2.TTSRequest(text="Hey, how are you doing today?")
-                tts_response = stub.TTS(tts_request)
-                self.assertIsNotNone(tts_response)
-        except Exception as err:
-            print(err)
-            self.fail("TTS service failed")
-        finally:
-            self.tearDown()

backend/python/petals/Makefile

@@ -1,27 +1,17 @@
 .PHONY: petals
-petals: protogen
+petals:
 	@echo "Creating virtual environment..."
 	bash install.sh "petals.yml"
 	@echo "Virtual environment created."
 
 .PHONY: run
-run: protogen
+run:
 	@echo "Running petals..."
 	bash run.sh
 	@echo "petals run."
 
 .PHONY: test
-test: protogen
+test:
 	@echo "Testing petals..."
 	bash test.sh
 	@echo "petals tested."
-
-.PHONY: protogen
-protogen: backend_pb2_grpc.py backend_pb2.py
-
-.PHONY: protogen-clean
-protogen-clean:
-	$(RM) backend_pb2_grpc.py backend_pb2.py
-
-backend_pb2_grpc.py backend_pb2.py:
-	python3 -m grpc_tools.protoc -I../.. --python_out=. --grpc_python_out=. backend.proto

backend/python/petals/backend_pb2_grpc.py

@@ -0,0 +1,363 @@
+# Generated by the gRPC Python protocol compiler plugin. DO NOT EDIT!
+"""Client and server classes corresponding to protobuf-defined services."""
+import grpc
+
+import backend_pb2 as backend__pb2
+
+
+class BackendStub(object):
+    """Missing associated documentation comment in .proto file."""
+
+    def __init__(self, channel):
+        """Constructor.
+
+        Args:
+            channel: A grpc.Channel.
+        """
+        self.Health = channel.unary_unary(
+                '/backend.Backend/Health',
+                request_serializer=backend__pb2.HealthMessage.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.Predict = channel.unary_unary(
+                '/backend.Backend/Predict',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.LoadModel = channel.unary_unary(
+                '/backend.Backend/LoadModel',
+                request_serializer=backend__pb2.ModelOptions.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.PredictStream = channel.unary_stream(
+                '/backend.Backend/PredictStream',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.Reply.FromString,
+                )
+        self.Embedding = channel.unary_unary(
+                '/backend.Backend/Embedding',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.EmbeddingResult.FromString,
+                )
+        self.GenerateImage = channel.unary_unary(
+                '/backend.Backend/GenerateImage',
+                request_serializer=backend__pb2.GenerateImageRequest.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.AudioTranscription = channel.unary_unary(
+                '/backend.Backend/AudioTranscription',
+                request_serializer=backend__pb2.TranscriptRequest.SerializeToString,
+                response_deserializer=backend__pb2.TranscriptResult.FromString,
+                )
+        self.TTS = channel.unary_unary(
+                '/backend.Backend/TTS',
+                request_serializer=backend__pb2.TTSRequest.SerializeToString,
+                response_deserializer=backend__pb2.Result.FromString,
+                )
+        self.TokenizeString = channel.unary_unary(
+                '/backend.Backend/TokenizeString',
+                request_serializer=backend__pb2.PredictOptions.SerializeToString,
+                response_deserializer=backend__pb2.TokenizationResponse.FromString,
+                )
+        self.Status = channel.unary_unary(
+                '/backend.Backend/Status',
+                request_serializer=backend__pb2.HealthMessage.SerializeToString,
+                response_deserializer=backend__pb2.StatusResponse.FromString,
+                )
+
+
+class BackendServicer(object):
+    """Missing associated documentation comment in .proto file."""
+
+    def Health(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Predict(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def LoadModel(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def PredictStream(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Embedding(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def GenerateImage(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def AudioTranscription(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def TTS(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def TokenizeString(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+    def Status(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+
+def add_BackendServicer_to_server(servicer, server):
+    rpc_method_handlers = {
+            'Health': grpc.unary_unary_rpc_method_handler(
+                    servicer.Health,
+                    request_deserializer=backend__pb2.HealthMessage.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'Predict': grpc.unary_unary_rpc_method_handler(
+                    servicer.Predict,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'LoadModel': grpc.unary_unary_rpc_method_handler(
+                    servicer.LoadModel,
+                    request_deserializer=backend__pb2.ModelOptions.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'PredictStream': grpc.unary_stream_rpc_method_handler(
+                    servicer.PredictStream,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.Reply.SerializeToString,
+            ),
+            'Embedding': grpc.unary_unary_rpc_method_handler(
+                    servicer.Embedding,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.EmbeddingResult.SerializeToString,
+            ),
+            'GenerateImage': grpc.unary_unary_rpc_method_handler(
+                    servicer.GenerateImage,
+                    request_deserializer=backend__pb2.GenerateImageRequest.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'AudioTranscription': grpc.unary_unary_rpc_method_handler(
+                    servicer.AudioTranscription,
+                    request_deserializer=backend__pb2.TranscriptRequest.FromString,
+                    response_serializer=backend__pb2.TranscriptResult.SerializeToString,
+            ),
+            'TTS': grpc.unary_unary_rpc_method_handler(
+                    servicer.TTS,
+                    request_deserializer=backend__pb2.TTSRequest.FromString,
+                    response_serializer=backend__pb2.Result.SerializeToString,
+            ),
+            'TokenizeString': grpc.unary_unary_rpc_method_handler(
+                    servicer.TokenizeString,
+                    request_deserializer=backend__pb2.PredictOptions.FromString,
+                    response_serializer=backend__pb2.TokenizationResponse.SerializeToString,
+            ),
+            'Status': grpc.unary_unary_rpc_method_handler(
+                    servicer.Status,
+                    request_deserializer=backend__pb2.HealthMessage.FromString,
+                    response_serializer=backend__pb2.StatusResponse.SerializeToString,
+            ),
+    }
+    generic_handler = grpc.method_handlers_generic_handler(
+            'backend.Backend', rpc_method_handlers)
+    server.add_generic_rpc_handlers((generic_handler,))
+
+
+ # This class is part of an EXPERIMENTAL API.
+class Backend(object):
+    """Missing associated documentation comment in .proto file."""
+
+    @staticmethod
+    def Health(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Health',
+            backend__pb2.HealthMessage.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Predict(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Predict',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def LoadModel(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/LoadModel',
+            backend__pb2.ModelOptions.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def PredictStream(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_stream(request, target, '/backend.Backend/PredictStream',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.Reply.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Embedding(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Embedding',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.EmbeddingResult.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def GenerateImage(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/GenerateImage',
+            backend__pb2.GenerateImageRequest.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def AudioTranscription(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/AudioTranscription',
+            backend__pb2.TranscriptRequest.SerializeToString,
+            backend__pb2.TranscriptResult.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def TTS(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/TTS',
+            backend__pb2.TTSRequest.SerializeToString,
+            backend__pb2.Result.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def TokenizeString(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/TokenizeString',
+            backend__pb2.PredictOptions.SerializeToString,
+            backend__pb2.TokenizationResponse.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+
+    @staticmethod
+    def Status(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(request, target, '/backend.Backend/Status',
+            backend__pb2.HealthMessage.SerializeToString,
+            backend__pb2.StatusResponse.FromString,
+            options, channel_credentials,
+            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)