fix(vall-e-x): correctly install reqs in environment (#1377)

Signed-off-by: GitHub <noreply@github.com>
Co-authored-by: mudler <mudler@users.noreply.github.com>

.env

@@ -69,4 +69,21 @@ MODELS_PATH=/models
 # PYTHON_GRPC_MAX_WORKERS=1
 
 ### Define the number of parallel LLAMA.cpp workers (Defaults to 1)
-# LLAMACPP_PARALLEL=1
+# LLAMACPP_PARALLEL=1
+
+### Enable to run parallel requests
+# PARALLEL_REQUESTS=true
+
+### Watchdog settings
+###
+# Enables watchdog to kill backends that are inactive for too much time
+# WATCHDOG_IDLE=true
+#
+# Enables watchdog to kill backends that are busy for too much time
+# 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
+# WATCHDOG_BUSY_TIMEOUT=5m

.github/workflows/image.yml

@@ -14,8 +14,25 @@ concurrency:
   cancel-in-progress: true
 
 jobs:
-  image-build:
+  extras-image-build:
+    uses: ./.github/workflows/image_build.yml
+    with:
+      tag-latest: ${{ matrix.tag-latest }}
+      tag-suffix: ${{ matrix.tag-suffix }}
+      ffmpeg: ${{ matrix.ffmpeg }}
+      image-type: ${{ matrix.image-type }}
+      build-type: ${{ matrix.build-type }}
+      cuda-major-version: ${{ matrix.cuda-major-version }}
+      cuda-minor-version: ${{ matrix.cuda-minor-version }}
+      platforms: ${{ matrix.platforms }}
+      runs-on: ${{ matrix.runs-on }}
+    secrets:
+      dockerUsername: ${{ secrets.LOCALAI_REGISTRY_USERNAME }}
+      dockerPassword: ${{ secrets.LOCALAI_REGISTRY_PASSWORD }}
     strategy:
+      # Pushing with all jobs in parallel
+      # eats the bandwidth of all the nodes
+      max-parallel: ${{ github.event_name != 'pull_request' && 2 || 4 }}
       matrix:
         include:
           - build-type: ''
@@ -24,130 +41,117 @@ jobs:
             tag-latest: 'auto'
             tag-suffix: ''
             ffmpeg: ''
+            image-type: 'extras'
+            runs-on: 'arc-runner-set'
           - build-type: ''
             platforms: 'linux/amd64'
             tag-latest: 'false'
             tag-suffix: '-ffmpeg'
             ffmpeg: 'true'
+            image-type: 'extras'
+            runs-on: 'arc-runner-set'
           - build-type: 'cublas'
-            cuda-major-version: 11
-            cuda-minor-version: 7
+            cuda-major-version: "11"
+            cuda-minor-version: "7"
             platforms: 'linux/amd64'
             tag-latest: 'false'
             tag-suffix: '-cublas-cuda11'
             ffmpeg: ''
+            image-type: 'extras'
+            runs-on: 'arc-runner-set'
           - build-type: 'cublas'
-            cuda-major-version: 12
-            cuda-minor-version: 1
+            cuda-major-version: "12"
+            cuda-minor-version: "1"
             platforms: 'linux/amd64'
             tag-latest: 'false'
             tag-suffix: '-cublas-cuda12'
             ffmpeg: ''
+            image-type: 'extras'
+            runs-on: 'arc-runner-set'
           - build-type: 'cublas'
-            cuda-major-version: 11
-            cuda-minor-version: 7
+            cuda-major-version: "11"
+            cuda-minor-version: "7"
             platforms: 'linux/amd64'
             tag-latest: 'false'
             tag-suffix: '-cublas-cuda11-ffmpeg'
             ffmpeg: 'true'
+            image-type: 'extras'
+            runs-on: 'arc-runner-set'
           - build-type: 'cublas'
-            cuda-major-version: 12
-            cuda-minor-version: 1
+            cuda-major-version: "12"
+            cuda-minor-version: "1"
             platforms: 'linux/amd64'
             tag-latest: 'false'
             tag-suffix: '-cublas-cuda12-ffmpeg'
             ffmpeg: 'true'
-
-    runs-on: arc-runner-set 
-    steps:
-      - name: Force Install GIT latest
-        run: |
-          sudo apt-get update \
-          && sudo apt-get install -y software-properties-common \
-          && sudo apt-get update \
-          && sudo add-apt-repository -y ppa:git-core/ppa \
-          && sudo apt-get update \
-          && sudo apt-get install -y git
-      - name: Checkout
-        uses: actions/checkout@v4
-      # - 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 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
-      #     df -h
-      - name: Docker meta
-        id: meta
-        uses: docker/metadata-action@v5
-        with:
-          images: quay.io/go-skynet/local-ai
-          tags: |
-            type=ref,event=branch
-            type=semver,pattern={{raw}}
-            type=sha
-          flavor: |
-            latest=${{ matrix.tag-latest }}
-            suffix=${{ matrix.tag-suffix }}
-
-      - 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: Login to DockerHub
-        if: github.event_name != 'pull_request'
-        uses: docker/login-action@v3
-        with:
-          registry: quay.io
-          username: ${{ secrets.LOCALAI_REGISTRY_USERNAME }}
-          password: ${{ secrets.LOCALAI_REGISTRY_PASSWORD }}
-
-      - name: Build and push
-        uses: docker/build-push-action@v5
-        with:
-          builder: ${{ steps.buildx.outputs.name }}
-          build-args: |
-            BUILD_TYPE=${{ matrix.build-type }}
-            CUDA_MAJOR_VERSION=${{ matrix.cuda-major-version }}
-            CUDA_MINOR_VERSION=${{ matrix.cuda-minor-version }}
-            FFMPEG=${{ matrix.ffmpeg }}
-          context: .
-          file: ./Dockerfile
-          platforms: ${{ matrix.platforms }}
-          push: ${{ github.event_name != 'pull_request' }}
-          tags: ${{ steps.meta.outputs.tags }}
-          labels: ${{ steps.meta.outputs.labels }}
+            image-type: 'extras'
+            runs-on: 'arc-runner-set'
+          - build-type: ''
+            #platforms: 'linux/amd64,linux/arm64'
+            platforms: 'linux/amd64'
+            tag-latest: 'auto'
+            tag-suffix: ''
+            ffmpeg: ''
+            image-type: 'extras'
+            runs-on: 'arc-runner-set'
+  core-image-build:
+    uses: ./.github/workflows/image_build.yml
+    with:
+      tag-latest: ${{ matrix.tag-latest }}
+      tag-suffix: ${{ matrix.tag-suffix }}
+      ffmpeg: ${{ matrix.ffmpeg }}
+      image-type: ${{ matrix.image-type }}
+      build-type: ${{ matrix.build-type }}
+      cuda-major-version: ${{ matrix.cuda-major-version }}
+      cuda-minor-version: ${{ matrix.cuda-minor-version }}
+      platforms: ${{ matrix.platforms }}
+      runs-on: ${{ matrix.runs-on }}
+    secrets:
+      dockerUsername: ${{ secrets.LOCALAI_REGISTRY_USERNAME }}
+      dockerPassword: ${{ secrets.LOCALAI_REGISTRY_PASSWORD }}
+    strategy:
+      matrix:
+        include:
+          - build-type: ''
+            platforms: 'linux/amd64'
+            tag-latest: 'false'
+            tag-suffix: '-ffmpeg-core'
+            ffmpeg: 'true'
+            image-type: 'core'
+            runs-on: 'ubuntu-latest'
+          - build-type: 'cublas'
+            cuda-major-version: "11"
+            cuda-minor-version: "7"
+            platforms: 'linux/amd64'
+            tag-latest: 'false'
+            tag-suffix: '-cublas-cuda11-core'
+            ffmpeg: ''
+            image-type: 'core'
+            runs-on: 'ubuntu-latest'
+          - build-type: 'cublas'
+            cuda-major-version: "12"
+            cuda-minor-version: "1"
+            platforms: 'linux/amd64'
+            tag-latest: 'false'
+            tag-suffix: '-cublas-cuda12-core'
+            ffmpeg: ''
+            image-type: 'core'
+            runs-on: 'ubuntu-latest'
+          - build-type: 'cublas'
+            cuda-major-version: "11"
+            cuda-minor-version: "7"
+            platforms: 'linux/amd64'
+            tag-latest: 'false'
+            tag-suffix: '-cublas-cuda11-ffmpeg-core'
+            ffmpeg: 'true'
+            image-type: 'core'
+            runs-on: 'ubuntu-latest'
+          - build-type: 'cublas'
+            cuda-major-version: "12"
+            cuda-minor-version: "1"
+            platforms: 'linux/amd64'
+            tag-latest: 'false'
+            tag-suffix: '-cublas-cuda12-ffmpeg-core'
+            ffmpeg: 'true'
+            image-type: 'core'
+            runs-on: 'ubuntu-latest'

.github/workflows/image_build.yml

@@ -0,0 +1,147 @@
+---
+name: 'build container images (reusable)'
+
+on:
+  workflow_call:
+    inputs:
+      build-type:
+        description: 'Build type'
+        default: ''
+        type: string
+      cuda-major-version:
+        description: 'CUDA major version'
+        default: "11"
+        type: string
+      cuda-minor-version:
+        description: 'CUDA minor version'
+        default: "7"
+        type: string
+      platforms:
+        description: 'Platforms'
+        default: ''
+        type: string
+      tag-latest:
+        description: 'Tag latest'
+        default: ''
+        type: string
+      tag-suffix:
+        description: 'Tag suffix'
+        default: ''
+        type: string
+      ffmpeg:
+        description: 'FFMPEG'
+        default: ''
+        type: string
+      image-type:
+        description: 'Image type'
+        default: ''
+        type: string
+      runs-on:
+        description: 'Runs on'
+        required: true
+        default: ''
+        type: string
+    secrets:
+      dockerUsername:
+        required: true
+      dockerPassword:
+        required: true
+jobs:
+  reusable_image-build:
+    runs-on: ${{ inputs.runs-on }}
+    steps:
+      - name: Force Install GIT latest
+        run: |
+          sudo apt-get update \
+          && sudo apt-get install -y software-properties-common \
+          && sudo apt-get update \
+          && sudo add-apt-repository -y ppa:git-core/ppa \
+          && sudo apt-get update \
+          && sudo apt-get install -y git
+      - name: Checkout
+        uses: actions/checkout@v4
+      # - name: Release space from worker
+      #   run: |
+      #     echo "Listing top largest packages"
+      #     pkgs=$(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 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
+      #     df -h
+      - name: Docker meta
+        id: meta
+        uses: docker/metadata-action@v5
+        with:
+          images: quay.io/go-skynet/local-ai
+          tags: |
+            type=ref,event=branch
+            type=semver,pattern={{raw}}
+            type=sha
+          flavor: |
+            latest=${{ inputs.tag-latest }}
+            suffix=${{ inputs.tag-suffix }}
+
+      - 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: Login to DockerHub
+        if: github.event_name != 'pull_request'
+        uses: docker/login-action@v3
+        with:
+          registry: quay.io
+          username: ${{ secrets.dockerUsername }}
+          password: ${{ secrets.dockerPassword }}
+
+      - name: Build and push
+        uses: docker/build-push-action@v5
+        with:
+          builder: ${{ steps.buildx.outputs.name }}
+          build-args: |
+            BUILD_TYPE=${{ inputs.build-type }}
+            CUDA_MAJOR_VERSION=${{ inputs.cuda-major-version }}
+            CUDA_MINOR_VERSION=${{ inputs.cuda-minor-version }}
+            FFMPEG=${{ inputs.ffmpeg }}
+            IMAGE_TYPE=${{ inputs.image-type }}
+          context: .
+          file: ./Dockerfile
+          platforms: ${{ inputs.platforms }}
+          push: ${{ github.event_name != 'pull_request' }}
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}
+      - name: job summary
+        run: |
+          echo "Built image: ${{ steps.meta.outputs.labels }}" >> $GITHUB_STEP_SUMMARY

.github/workflows/test.yml

@@ -78,13 +78,12 @@ jobs:
           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=$PATH:/opt/conda/bin make -C extra/grpc/huggingface
+          PATH=$PATH:/opt/conda/bin make -C backend/python/sentencetransformers
 
           # Pre-build piper before we start tests in order to have shared libraries in place
-          make go-piper && \
-          GO_TAGS="tts" make -C go-piper piper.o && \
-          sudo cp -rfv go-piper/piper/build/pi/lib/. /usr/lib/ && \
-
+          make sources/go-piper && \
+          GO_TAGS="tts" make -C sources/go-piper piper.o && \
+          sudo cp -rfv sources/go-piper/piper/build/pi/lib/. /usr/lib/ && \
           # Pre-build stable diffusion before we install a newer version of abseil (not compatible with stablediffusion-ncn)
           GO_TAGS="stablediffusion tts" GRPC_BACKENDS=backend-assets/grpc/stablediffusion make build
 

.gitignore

@@ -1,12 +1,5 @@
 # go-llama build artifacts
-go-llama
-go-llama-stable
-/gpt4all
-go-stable-diffusion
-go-piper
-/go-bert
-go-ggllm
-/piper
+/sources/
 __pycache__/
 *.a
 get-sources

.gitmodules

@@ -0,0 +1,3 @@
+[submodule "docs/themes/hugo-theme-relearn"]
+	path = docs/themes/hugo-theme-relearn
+	url = https://github.com/McShelby/hugo-theme-relearn.git

Dockerfile

@@ -12,7 +12,7 @@ ARG TARGETARCH
 ARG TARGETVARIANT
 
 ENV BUILD_TYPE=${BUILD_TYPE}
-ENV EXTERNAL_GRPC_BACKENDS="huggingface-embeddings:/build/extra/grpc/huggingface/run.sh,autogptq:/build/extra/grpc/autogptq/run.sh,bark:/build/extra/grpc/bark/run.sh,diffusers:/build/extra/grpc/diffusers/run.sh,exllama:/build/extra/grpc/exllama/run.sh,vall-e-x:/build/extra/grpc/vall-e-x/run.sh,vllm:/build/extra/grpc/vllm/run.sh"
+ENV EXTERNAL_GRPC_BACKENDS="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"
 ENV GALLERIES='[{"name":"model-gallery", "url":"github:go-skynet/model-gallery/index.yaml"}, {"url": "github:go-skynet/model-gallery/huggingface.yaml","name":"huggingface"}]'
 ARG GO_TAGS="stablediffusion tts"
 
@@ -64,20 +64,10 @@ RUN curl https://repo.anaconda.com/pkgs/misc/gpgkeys/anaconda.asc | gpg --dearmo
     apt-get update && \
     apt-get install -y conda
 
-COPY extra/requirements.txt /build/extra/requirements.txt
 ENV PATH="/root/.cargo/bin:${PATH}"
 RUN pip install --upgrade pip
 RUN curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y
-#RUN if [ "${TARGETARCH}" = "amd64" ]; then \
-#        pip install git+https://github.com/suno-ai/bark.git diffusers invisible_watermark transformers accelerate safetensors;\
-#    fi
-#RUN if [ "${BUILD_TYPE}" = "cublas" ] && [ "${TARGETARCH}" = "amd64" ]; then \
-#        pip install torch vllm && pip install auto-gptq https://github.com/jllllll/exllama/releases/download/0.0.10/exllama-0.0.10+cu${CUDA_MAJOR_VERSION}${CUDA_MINOR_VERSION}-cp39-cp39-linux_x86_64.whl;\
- #   fi
-#RUN pip install -r /build/extra/requirements.txt && rm -rf /build/extra/requirements.txt
 
-# Vall-e-X
-RUN git clone https://github.com/Plachtaa/VALL-E-X.git /usr/lib/vall-e-x && cd /usr/lib/vall-e-x && pip install -r requirements.txt
 
 # \
 #    ; fi
@@ -98,12 +88,9 @@ ENV NVIDIA_VISIBLE_DEVICES=all
 
 WORKDIR /build
 
-COPY Makefile .
-RUN make get-sources
-COPY go.mod .
-RUN make prepare
 COPY . .
 COPY .git .
+RUN make prepare
 
 # stablediffusion does not tolerate a newer version of abseil, build it first
 RUN GRPC_BACKENDS=backend-assets/grpc/stablediffusion make build
@@ -112,12 +99,17 @@ 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 && rm -rf grpc \
+       ../.. && make -j12 install \
     ; fi
 
 # Rebuild with defaults backends
 RUN make build
 
+RUN if [ ! -d "/build/sources/go-piper/piper/build/pi/lib/" ]; then \
+    mkdir -p /build/sources/go-piper/piper/build/pi/lib/ \
+    touch /build/sources/go-piper/piper/build/pi/lib/keep \
+    ; fi
+
 ###################################
 ###################################
 
@@ -149,48 +141,54 @@ WORKDIR /build
 # see https://github.com/go-skynet/LocalAI/pull/658#discussion_r1241971626 and
 # https://github.com/go-skynet/LocalAI/pull/434
 COPY . .
-RUN make prepare-sources
+
+COPY --from=builder /build/sources ./sources/
+COPY --from=builder /build/grpc ./grpc/
+
+RUN make prepare-sources && cd /build/grpc/cmake/build && make install && rm -rf grpc
 
 # Copy the binary
 COPY --from=builder /build/local-ai ./
 
 # Copy shared libraries for piper
-COPY --from=builder /build/go-piper/piper/build/pi/lib/* /usr/lib/
+COPY --from=builder /build/sources/go-piper/piper/build/pi/lib/* /usr/lib/
 
 # do not let stablediffusion rebuild (requires an older version of absl)
 COPY --from=builder /build/backend-assets/grpc/stablediffusion ./backend-assets/grpc/stablediffusion
 
 ## Duplicated from Makefile to avoid having a big layer that's hard to push
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-	PATH=$PATH:/opt/conda/bin make -C extra/grpc/autogptq \
+	PATH=$PATH:/opt/conda/bin make -C backend/python/autogptq \
     ; fi
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-	PATH=$PATH:/opt/conda/bin make -C extra/grpc/bark \
+	PATH=$PATH:/opt/conda/bin make -C backend/python/bark \
     ; fi
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-	PATH=$PATH:/opt/conda/bin make -C extra/grpc/diffusers \
+	PATH=$PATH:/opt/conda/bin make -C backend/python/diffusers \
     ; fi
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-	PATH=$PATH:/opt/conda/bin make -C extra/grpc/vllm \
+	PATH=$PATH:/opt/conda/bin make -C backend/python/vllm \
     ; fi
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-	PATH=$PATH:/opt/conda/bin make -C extra/grpc/huggingface \
+	PATH=$PATH:/opt/conda/bin make -C backend/python/sentencetransformers \
     ; fi
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-	PATH=$PATH:/opt/conda/bin make -C extra/grpc/vall-e-x \
+	PATH=$PATH:/opt/conda/bin make -C backend/python/transformers \
     ; fi
 RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
-	PATH=$PATH:/opt/conda/bin make -C extra/grpc/exllama \
+	PATH=$PATH:/opt/conda/bin make -C backend/python/vall-e-x \
+    ; fi
+RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
+	PATH=$PATH:/opt/conda/bin make -C backend/python/exllama \
+    ; fi
+RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
+	PATH=$PATH:/opt/conda/bin make -C backend/python/petals \
     ; fi
-
-# Copy VALLE-X as it's not a real "lib"
-RUN if [ -d /usr/lib/vall-e-x ]; then \
-    cp -rfv /usr/lib/vall-e-x/* ./ ; \ 
-    fi
 
 # we also copy exllama libs over to resolve exllama import error
+# TODO: check if this is still needed
 RUN if [ -d /usr/local/lib/python3.9/dist-packages/exllama ]; then \
-        cp -rfv /usr/local/lib/python3.9/dist-packages/exllama extra/grpc/exllama/;\
+        cp -rfv /usr/local/lib/python3.9/dist-packages/exllama backend/python/exllama/;\
     fi
 
 # Define the health check command

Entitlements.plist

@@ -0,0 +1,10 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
+<plist version="1.0">
+<dict>
+    <key>com.apple.security.network.client</key>
+    <true/>
+    <key>com.apple.security.network.server</key>
+    <true/>
+</dict>
+</plist>

Makefile

@@ -8,7 +8,7 @@ GOLLAMA_VERSION?=aeba71ee842819da681ea537e78846dc75949ac0
 
 GOLLAMA_STABLE_VERSION?=50cee7712066d9e38306eccadcfbb44ea87df4b7
 
-CPPLLAMA_VERSION?=a75fa576abba9d37f463580c379e4bbf1e1ad03c
+CPPLLAMA_VERSION?=5a7d3125e7c24f223659b7f0b7aa7736986e92c0
 
 # gpt4all version
 GPT4ALL_REPO?=https://github.com/nomic-ai/gpt4all
@@ -22,16 +22,16 @@ RWKV_REPO?=https://github.com/donomii/go-rwkv.cpp
 RWKV_VERSION?=c898cd0f62df8f2a7830e53d1d513bef4f6f792b
 
 # whisper.cpp version
-WHISPER_CPP_VERSION?=85ed71aaec8e0612a84c0b67804bde75aa75a273
+WHISPER_CPP_VERSION?=e369243ebd24c8a14201f6b4280bccbb7b6a7df3
 
 # bert.cpp version
 BERT_VERSION?=6abe312cded14042f6b7c3cd8edf082713334a4d
 
 # go-piper version
-PIPER_VERSION?=736f6fb639ab8e3397356e48eeb6bdcb9da88a78
+PIPER_VERSION?=5a4c9e28c84bac09ab6baa9f88457d852cb46bb2
 
 # stablediffusion version
-STABLEDIFFUSION_VERSION?=d89260f598afb809279bc72aa0107b4292587632
+STABLEDIFFUSION_VERSION?=902db5f066fd137697e3b69d0fa10d4782bd2c2f
 
 export BUILD_TYPE?=
 export STABLE_BUILD_TYPE?=$(BUILD_TYPE)
@@ -68,29 +68,39 @@ ifndef UNAME_S
 UNAME_S := $(shell uname -s)
 endif
 
-ifeq ($(UNAME_S),Darwin)
+ifeq ($(OS),Darwin)
 	CGO_LDFLAGS += -lcblas -framework Accelerate
-ifneq ($(BUILD_TYPE),metal)
-    # explicit disable metal if on Darwin and metal is disabled
-	CMAKE_ARGS+=-DLLAMA_METAL=OFF
-endif
+	ifeq ($(OSX_SIGNING_IDENTITY),)
+		OSX_SIGNING_IDENTITY := $(shell security find-identity -v -p codesigning | grep '"' | head -n 1 | sed -E 's/.*"(.*)"/\1/')
+	endif
+
+	# on OSX, if BUILD_TYPE is blank, we should default to use Metal
+	ifeq ($(BUILD_TYPE),)
+		BUILD_TYPE=metal
+	# disable metal if on Darwin and any other value is explicitly passed.
+	else ifneq ($(BUILD_TYPE),metal)
+		CMAKE_ARGS+=-DLLAMA_METAL=OFF
+	endif
 endif
 
 ifeq ($(BUILD_TYPE),openblas)
 	CGO_LDFLAGS+=-lopenblas
+	export WHISPER_OPENBLAS=1
 endif
 
 ifeq ($(BUILD_TYPE),cublas)
 	CGO_LDFLAGS+=-lcublas -lcudart -L$(CUDA_LIBPATH)
 	export LLAMA_CUBLAS=1
+	export WHISPER_CUBLAS=1
 endif
 
 ifeq ($(BUILD_TYPE),hipblas)
 	ROCM_HOME ?= /opt/rocm
 	export CXX=$(ROCM_HOME)/llvm/bin/clang++
 	export CC=$(ROCM_HOME)/llvm/bin/clang
-	# Llama-stable has no hipblas support, so override it here.
+	# llama-ggml has no hipblas support, so override it here.
 	export STABLE_BUILD_TYPE=
+	export WHISPER_HIPBLAS=1
 	GPU_TARGETS ?= gfx900,gfx90a,gfx1030,gfx1031,gfx1100
 	AMDGPU_TARGETS ?= "$(GPU_TARGETS)"
 	CMAKE_ARGS+=-DLLAMA_HIPBLAS=ON -DAMDGPU_TARGETS="$(AMDGPU_TARGETS)" -DGPU_TARGETS="$(GPU_TARGETS)"
@@ -100,10 +110,12 @@ endif
 ifeq ($(BUILD_TYPE),metal)
 	CGO_LDFLAGS+=-framework Foundation -framework Metal -framework MetalKit -framework MetalPerformanceShaders
 	export LLAMA_METAL=1
+	export WHISPER_METAL=1
 endif
 
 ifeq ($(BUILD_TYPE),clblas)
 	CGO_LDFLAGS+=-lOpenCL -lclblast
+	export WHISPER_CLBLAST=1
 endif
 
 # glibc-static or glibc-devel-static required
@@ -119,12 +131,12 @@ endif
 ifeq ($(findstring tts,$(GO_TAGS)),tts)
 #	OPTIONAL_TARGETS+=go-piper/libpiper_binding.a
 #	OPTIONAL_TARGETS+=backend-assets/espeak-ng-data
-	PIPER_CGO_CXXFLAGS+=-I$(shell pwd)/go-piper/piper/src/cpp -I$(shell pwd)/go-piper/piper/build/fi/include -I$(shell pwd)/go-piper/piper/build/pi/include -I$(shell pwd)/go-piper/piper/build/si/include
- 	PIPER_CGO_LDFLAGS+=-L$(shell pwd)/go-piper/piper/build/fi/lib -L$(shell pwd)/go-piper/piper/build/pi/lib -L$(shell pwd)/go-piper/piper/build/si/lib -lfmt -lspdlog
+	PIPER_CGO_CXXFLAGS+=-I$(shell pwd)/sources/go-piper/piper/src/cpp -I$(shell pwd)/sources/go-piper/piper/build/fi/include -I$(shell pwd)/sources/go-piper/piper/build/pi/include -I$(shell pwd)/sources/go-piper/piper/build/si/include
+ 	PIPER_CGO_LDFLAGS+=-L$(shell pwd)/sources/go-piper/piper/build/fi/lib -L$(shell pwd)/sources/go-piper/piper/build/pi/lib -L$(shell pwd)/sources/go-piper/piper/build/si/lib -lfmt -lspdlog -lucd
 	OPTIONAL_GRPC+=backend-assets/grpc/piper
 endif
 
-ALL_GRPC_BACKENDS=backend-assets/grpc/langchain-huggingface backend-assets/grpc/falcon-ggml backend-assets/grpc/bert-embeddings backend-assets/grpc/llama backend-assets/grpc/llama-cpp backend-assets/grpc/llama-stable backend-assets/grpc/gpt4all backend-assets/grpc/dolly backend-assets/grpc/gpt2 backend-assets/grpc/gptj backend-assets/grpc/gptneox backend-assets/grpc/mpt backend-assets/grpc/replit backend-assets/grpc/starcoder backend-assets/grpc/rwkv backend-assets/grpc/whisper $(OPTIONAL_GRPC)
+ALL_GRPC_BACKENDS=backend-assets/grpc/langchain-huggingface backend-assets/grpc/falcon-ggml backend-assets/grpc/bert-embeddings backend-assets/grpc/llama backend-assets/grpc/llama-cpp backend-assets/grpc/llama-ggml backend-assets/grpc/gpt4all backend-assets/grpc/dolly backend-assets/grpc/gpt2 backend-assets/grpc/gptj backend-assets/grpc/gptneox backend-assets/grpc/mpt backend-assets/grpc/replit backend-assets/grpc/starcoder backend-assets/grpc/rwkv backend-assets/grpc/whisper $(OPTIONAL_GRPC)
 GRPC_BACKENDS?=$(ALL_GRPC_BACKENDS) $(OPTIONAL_GRPC)
 
 # If empty, then we build all
@@ -137,112 +149,117 @@ endif
 all: help
 
 ## GPT4ALL
-gpt4all:
-	git clone --recurse-submodules $(GPT4ALL_REPO) gpt4all
-	cd gpt4all && git checkout -b build $(GPT4ALL_VERSION) && git submodule update --init --recursive --depth 1
+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
 
 ## go-piper
-go-piper:
-	git clone --recurse-submodules https://github.com/mudler/go-piper go-piper
-	cd go-piper && git checkout -b build $(PIPER_VERSION) && git submodule update --init --recursive --depth 1
+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
 
 ## BERT embeddings
-go-bert:
-	git clone --recurse-submodules https://github.com/go-skynet/go-bert.cpp go-bert
-	cd go-bert && git checkout -b build $(BERT_VERSION) && git submodule update --init --recursive --depth 1
+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
-go-stable-diffusion:
-	git clone --recurse-submodules https://github.com/mudler/go-stable-diffusion go-stable-diffusion
-	cd go-stable-diffusion && git checkout -b build $(STABLEDIFFUSION_VERSION) && git submodule update --init --recursive --depth 1
+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
 
-go-stable-diffusion/libstablediffusion.a:
-	$(MAKE) -C go-stable-diffusion libstablediffusion.a
+sources/go-stable-diffusion/libstablediffusion.a:
+	$(MAKE) -C sources/go-stable-diffusion libstablediffusion.a
 
 ## RWKV
-go-rwkv:
-	git clone --recurse-submodules $(RWKV_REPO) go-rwkv
-	cd go-rwkv && git checkout -b build $(RWKV_VERSION) && git submodule update --init --recursive --depth 1
+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
 
-go-rwkv/librwkv.a: go-rwkv
-	cd go-rwkv && cd rwkv.cpp &&	cmake . -DRWKV_BUILD_SHARED_LIBRARY=OFF &&	cmake --build . && 	cp librwkv.a ..
+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 ..
 
-go-bert/libgobert.a: go-bert
-	$(MAKE) -C go-bert libgobert.a
+sources/go-bert/libgobert.a: sources/go-bert
+	$(MAKE) -C sources/go-bert libgobert.a
 
-backend-assets/gpt4all: gpt4all/gpt4all-bindings/golang/libgpt4all.a
+backend-assets/gpt4all: sources/gpt4all/gpt4all-bindings/golang/libgpt4all.a
 	mkdir -p backend-assets/gpt4all
-	@cp gpt4all/gpt4all-bindings/golang/buildllm/*.so backend-assets/gpt4all/ || true
-	@cp gpt4all/gpt4all-bindings/golang/buildllm/*.dylib backend-assets/gpt4all/ || true
-	@cp gpt4all/gpt4all-bindings/golang/buildllm/*.dll backend-assets/gpt4all/ || true
+	@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: go-piper
+backend-assets/espeak-ng-data: sources/go-piper
 	mkdir -p backend-assets/espeak-ng-data
-	$(MAKE) -C go-piper piper.o
-	@cp -rf go-piper/piper/build/pi/share/espeak-ng-data/. backend-assets/espeak-ng-data
+	$(MAKE) -C sources/go-piper piper.o
+	@cp -rf sources/go-piper/piper/build/pi/share/espeak-ng-data/. backend-assets/espeak-ng-data
 
-gpt4all/gpt4all-bindings/golang/libgpt4all.a: gpt4all
-	$(MAKE) -C gpt4all/gpt4all-bindings/golang/ libgpt4all.a
+sources/gpt4all/gpt4all-bindings/golang/libgpt4all.a: sources/gpt4all
+	$(MAKE) -C sources/gpt4all/gpt4all-bindings/golang/ libgpt4all.a
 
 ## CEREBRAS GPT
-go-ggml-transformers:
-	git clone --recurse-submodules https://github.com/go-skynet/go-ggml-transformers.cpp go-ggml-transformers
-	cd go-ggml-transformers && git checkout -b build $(GOGPT2_VERSION) && git submodule update --init --recursive --depth 1
+sources/go-ggml-transformers:
+	git clone --recurse-submodules https://github.com/go-skynet/go-ggml-transformers.cpp sources/go-ggml-transformers
+	cd sources/go-ggml-transformers && git checkout -b build $(GOGPT2_VERSION) && git submodule update --init --recursive --depth 1
 
-go-ggml-transformers/libtransformers.a: go-ggml-transformers
-	$(MAKE) -C go-ggml-transformers BUILD_TYPE=$(BUILD_TYPE) libtransformers.a
+sources/go-ggml-transformers/libtransformers.a: sources/go-ggml-transformers
+	$(MAKE) -C sources/go-ggml-transformers BUILD_TYPE=$(BUILD_TYPE) libtransformers.a
 
-whisper.cpp:
-	git clone https://github.com/ggerganov/whisper.cpp.git
-	cd whisper.cpp && git checkout -b build $(WHISPER_CPP_VERSION) && git submodule update --init --recursive --depth 1
+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
 
-whisper.cpp/libwhisper.a: whisper.cpp
-	cd whisper.cpp && make libwhisper.a
+sources/whisper.cpp/libwhisper.a: sources/whisper.cpp
+	cd sources/whisper.cpp && make libwhisper.a
 
-go-llama:
-	git clone --recurse-submodules https://github.com/go-skynet/go-llama.cpp go-llama
-	cd go-llama && git checkout -b build $(GOLLAMA_VERSION) && git submodule update --init --recursive --depth 1
+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
 
-go-llama-stable:
-	git clone --recurse-submodules https://github.com/go-skynet/go-llama.cpp go-llama-stable
-	cd go-llama-stable && git checkout -b build $(GOLLAMA_STABLE_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
 
-go-llama/libbinding.a: go-llama
-	$(MAKE) -C go-llama BUILD_TYPE=$(BUILD_TYPE) libbinding.a
+sources/go-llama/libbinding.a: sources/go-llama
+	$(MAKE) -C sources/go-llama BUILD_TYPE=$(BUILD_TYPE) libbinding.a
 
-go-llama-stable/libbinding.a: go-llama-stable
-	$(MAKE) -C go-llama-stable BUILD_TYPE=$(STABLE_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
 
-go-piper/libpiper_binding.a: go-piper
-	$(MAKE) -C go-piper libpiper_binding.a example/main
+sources/go-piper/libpiper_binding.a: sources/go-piper
+	$(MAKE) -C sources/go-piper libpiper_binding.a example/main
 
-get-sources: go-llama go-llama-stable go-ggml-transformers gpt4all go-piper go-rwkv whisper.cpp go-bert go-stable-diffusion
+backend/cpp/llama/llama.cpp:
+	$(MAKE) -C backend/cpp/llama llama.cpp	
+
+get-sources: backend/cpp/llama/llama.cpp sources/go-llama sources/go-llama-ggml sources/go-ggml-transformers sources/gpt4all sources/go-piper sources/go-rwkv sources/whisper.cpp sources/go-bert sources/go-stable-diffusion
 	touch $@
 
 replace:
-	$(GOCMD) mod edit -replace github.com/nomic-ai/gpt4all/gpt4all-bindings/golang=$(shell pwd)/gpt4all/gpt4all-bindings/golang
-	$(GOCMD) mod edit -replace github.com/go-skynet/go-ggml-transformers.cpp=$(shell pwd)/go-ggml-transformers
-	$(GOCMD) mod edit -replace github.com/donomii/go-rwkv.cpp=$(shell pwd)/go-rwkv
-	$(GOCMD) mod edit -replace github.com/ggerganov/whisper.cpp=$(shell pwd)/whisper.cpp
-	$(GOCMD) mod edit -replace github.com/go-skynet/go-bert.cpp=$(shell pwd)/go-bert
-	$(GOCMD) mod edit -replace github.com/mudler/go-stable-diffusion=$(shell pwd)/go-stable-diffusion
-	$(GOCMD) mod edit -replace github.com/mudler/go-piper=$(shell pwd)/go-piper
+	$(GOCMD) mod edit -replace github.com/nomic-ai/gpt4all/gpt4all-bindings/golang=$(shell pwd)/sources/gpt4all/gpt4all-bindings/golang
+	$(GOCMD) mod edit -replace github.com/go-skynet/go-ggml-transformers.cpp=$(shell pwd)/sources/go-ggml-transformers
+	$(GOCMD) mod edit -replace github.com/donomii/go-rwkv.cpp=$(shell pwd)/sources/go-rwkv
+	$(GOCMD) mod edit -replace github.com/ggerganov/whisper.cpp=$(shell pwd)/sources/whisper.cpp
+	$(GOCMD) mod edit -replace github.com/ggerganov/whisper.cpp/bindings/go=$(shell pwd)/sources/whisper.cpp/bindings/go
+	$(GOCMD) mod edit -replace github.com/go-skynet/go-bert.cpp=$(shell pwd)/sources/go-bert
+	$(GOCMD) mod edit -replace github.com/mudler/go-stable-diffusion=$(shell pwd)/sources/go-stable-diffusion
+	$(GOCMD) mod edit -replace github.com/mudler/go-piper=$(shell pwd)/sources/go-piper
 
 prepare-sources: get-sources replace
 	$(GOCMD) mod download
+	touch $@
 
 ## GENERIC
 rebuild: ## Rebuilds the project
 	$(GOCMD) clean -cache
-	$(MAKE) -C go-llama clean
-	$(MAKE) -C go-llama-stable clean
-	$(MAKE) -C gpt4all/gpt4all-bindings/golang/ clean
-	$(MAKE) -C go-ggml-transformers clean
-	$(MAKE) -C go-rwkv clean
-	$(MAKE) -C whisper.cpp clean
-	$(MAKE) -C go-stable-diffusion clean
-	$(MAKE) -C go-bert clean
-	$(MAKE) -C go-piper 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-ggml-transformers 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 clean
+	$(MAKE) -C sources/go-piper clean
 	$(MAKE) build
 
 prepare: prepare-sources $(OPTIONAL_TARGETS)
@@ -251,17 +268,7 @@ prepare: prepare-sources $(OPTIONAL_TARGETS)
 clean: ## Remove build related file
 	$(GOCMD) clean -cache
 	rm -f prepare
-	rm -rf ./go-llama
-	rm -rf ./gpt4all
-	rm -rf ./go-llama-stable
-	rm -rf ./go-gpt2
-	rm -rf ./go-stable-diffusion
-	rm -rf ./go-ggml-transformers
-	rm -rf ./backend-assets
-	rm -rf ./go-rwkv
-	rm -rf ./go-bert
-	rm -rf ./whisper.cpp
-	rm -rf ./go-piper
+	rm -rf ./sources
 	rm -rf $(BINARY_NAME)
 	rm -rf release/
 	rm -rf ./backend/cpp/grpc/grpc_repo
@@ -283,6 +290,9 @@ dist: build
 	mkdir -p release
 	cp $(BINARY_NAME) release/$(BINARY_NAME)-$(BUILD_ID)-$(OS)-$(ARCH)
 
+osx-signed: build
+	codesign --deep --force --sign "$(OSX_SIGNING_IDENTITY)" --entitlements "./Entitlements.plist" "./$(BINARY_NAME)"
+
 ## Run
 run: prepare ## run local-ai
 	CGO_LDFLAGS="$(CGO_LDFLAGS)" $(GOCMD) run ./
@@ -306,7 +316,7 @@ test: prepare test-models/testmodel grpcs
 	@echo 'Running tests'
 	export GO_TAGS="tts stablediffusion"
 	$(MAKE) prepare-test
-	HUGGINGFACE_GRPC=$(abspath ./)/extra/grpc/huggingface/run.sh TEST_DIR=$(abspath ./)/test-dir/ FIXTURES=$(abspath ./)/tests/fixtures CONFIG_FILE=$(abspath ./)/test-models/config.yaml MODELS_PATH=$(abspath ./)/test-models \
+	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 5 --fail-fast -v -r ./api ./pkg
 	$(MAKE) test-gpt4all
 	$(MAKE) test-llama
@@ -374,39 +384,43 @@ protogen: protogen-go protogen-python
 
 protogen-go:
 	protoc --go_out=. --go_opt=paths=source_relative --go-grpc_out=. --go-grpc_opt=paths=source_relative \
-    pkg/grpc/proto/backend.proto
+    backend/backend.proto
 
 protogen-python:
-	python3 -m grpc_tools.protoc -Ipkg/grpc/proto/ --python_out=extra/grpc/huggingface/ --grpc_python_out=extra/grpc/huggingface/ pkg/grpc/proto/backend.proto
-	python3 -m grpc_tools.protoc -Ipkg/grpc/proto/ --python_out=extra/grpc/autogptq/ --grpc_python_out=extra/grpc/autogptq/ pkg/grpc/proto/backend.proto
-	python3 -m grpc_tools.protoc -Ipkg/grpc/proto/ --python_out=extra/grpc/exllama/ --grpc_python_out=extra/grpc/exllama/ pkg/grpc/proto/backend.proto
-	python3 -m grpc_tools.protoc -Ipkg/grpc/proto/ --python_out=extra/grpc/bark/ --grpc_python_out=extra/grpc/bark/ pkg/grpc/proto/backend.proto
-	python3 -m grpc_tools.protoc -Ipkg/grpc/proto/ --python_out=extra/grpc/diffusers/ --grpc_python_out=extra/grpc/diffusers/ pkg/grpc/proto/backend.proto
-	python3 -m grpc_tools.protoc -Ipkg/grpc/proto/ --python_out=extra/grpc/vall-e-x/ --grpc_python_out=extra/grpc/vall-e-x/ pkg/grpc/proto/backend.proto
-	python3 -m grpc_tools.protoc -Ipkg/grpc/proto/ --python_out=extra/grpc/vllm/ --grpc_python_out=extra/grpc/vllm/ pkg/grpc/proto/backend.proto
+	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/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/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
 
 ## GRPC
 # Note: it is duplicated in the Dockerfile
 prepare-extra-conda-environments:
-	$(MAKE) -C extra/grpc/autogptq
-	$(MAKE) -C extra/grpc/bark
-	$(MAKE) -C extra/grpc/diffusers
-	$(MAKE) -C extra/grpc/vllm
-	$(MAKE) -C extra/grpc/huggingface
-	$(MAKE) -C extra/grpc/vall-e-x
-	$(MAKE) -C extra/grpc/exllama
+	$(MAKE) -C backend/python/autogptq
+	$(MAKE) -C backend/python/bark
+	$(MAKE) -C backend/python/diffusers
+	$(MAKE) -C backend/python/vllm
+	$(MAKE) -C backend/python/sentencetransformers
+	$(MAKE) -C backend/python/transformers
+	$(MAKE) -C backend/python/vall-e-x
+	$(MAKE) -C backend/python/exllama
+	$(MAKE) -C backend/python/petals
 
 
 backend-assets/grpc:
 	mkdir -p backend-assets/grpc
 
-backend-assets/grpc/llama: backend-assets/grpc go-llama/libbinding.a
-	$(GOCMD) mod edit -replace github.com/go-skynet/go-llama.cpp=$(shell pwd)/go-llama
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/go-llama LIBRARY_PATH=$(shell pwd)/go-llama \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/llama ./cmd/grpc/llama/
-# TODO: every binary should have its own folder instead, so can have different metal implementations
+backend-assets/grpc/llama: backend-assets/grpc sources/go-llama/libbinding.a
+	$(GOCMD) mod edit -replace github.com/go-skynet/go-llama.cpp=$(shell pwd)/sources/go-llama
+	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/sources/go-llama LIBRARY_PATH=$(shell pwd)/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 go-llama/build/bin/ggml-metal.metal backend-assets/grpc/
+	cp backend/cpp/llama/llama.cpp/ggml-metal.metal backend-assets/grpc/
 endif
 
 ## BACKEND CPP LLAMA START
@@ -425,7 +439,7 @@ ifdef BUILD_GRPC_FOR_BACKEND_LLAMA
 	export _PROTOBUF_PROTOC=${INSTALLED_PACKAGES}/bin/proto && \
 	export _GRPC_CPP_PLUGIN_EXECUTABLE=${INSTALLED_PACKAGES}/bin/grpc_cpp_plugin && \
 	export PATH=${PATH}:${INSTALLED_PACKAGES}/bin && \
-	CMAKE_ARGS="${ADDED_CMAKE_ARGS}" LLAMA_VERSION=$(CPPLLAMA_VERSION) $(MAKE) -C backend/cpp/llama grpc-server 
+	CMAKE_ARGS="${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) $(MAKE) -C backend/cpp/llama grpc-server			
@@ -440,71 +454,71 @@ ifeq ($(BUILD_TYPE),metal)
 	cp backend/cpp/llama/llama.cpp/build/bin/ggml-metal.metal backend-assets/grpc/
 endif
 
-backend-assets/grpc/llama-stable: backend-assets/grpc go-llama-stable/libbinding.a
-	$(GOCMD) mod edit -replace github.com/go-skynet/go-llama.cpp=$(shell pwd)/go-llama-stable
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/go-llama-stable LIBRARY_PATH=$(shell pwd)/go-llama \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/llama-stable ./cmd/grpc/llama-stable/
+backend-assets/grpc/llama-ggml: backend-assets/grpc sources/go-llama-ggml/libbinding.a
+	$(GOCMD) mod edit -replace github.com/go-skynet/go-llama.cpp=$(shell pwd)/sources/go-llama-ggml
+	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/sources/go-llama-ggml LIBRARY_PATH=$(shell pwd)/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/gpt4all: backend-assets/grpc backend-assets/gpt4all gpt4all/gpt4all-bindings/golang/libgpt4all.a
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/gpt4all/gpt4all-bindings/golang/ LIBRARY_PATH=$(shell pwd)/gpt4all/gpt4all-bindings/golang/ \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/gpt4all ./cmd/grpc/gpt4all/
+backend-assets/grpc/gpt4all: backend-assets/grpc backend-assets/gpt4all sources/gpt4all/gpt4all-bindings/golang/libgpt4all.a
+	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/sources/gpt4all/gpt4all-bindings/golang/ LIBRARY_PATH=$(shell pwd)/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/dolly: backend-assets/grpc go-ggml-transformers/libtransformers.a
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/go-ggml-transformers LIBRARY_PATH=$(shell pwd)/go-ggml-transformers \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/dolly ./cmd/grpc/dolly/
+backend-assets/grpc/dolly: backend-assets/grpc sources/go-ggml-transformers/libtransformers.a
+	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/sources/go-ggml-transformers LIBRARY_PATH=$(shell pwd)/sources/go-ggml-transformers \
+	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/dolly ./backend/go/llm/dolly/
 
-backend-assets/grpc/gpt2: backend-assets/grpc go-ggml-transformers/libtransformers.a
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/go-ggml-transformers LIBRARY_PATH=$(shell pwd)/go-ggml-transformers \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/gpt2 ./cmd/grpc/gpt2/
+backend-assets/grpc/gpt2: backend-assets/grpc sources/go-ggml-transformers/libtransformers.a
+	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/sources/go-ggml-transformers LIBRARY_PATH=$(shell pwd)/sources/go-ggml-transformers \
+	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/gpt2 ./backend/go/llm/gpt2/
 
-backend-assets/grpc/gptj: backend-assets/grpc go-ggml-transformers/libtransformers.a
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/go-ggml-transformers LIBRARY_PATH=$(shell pwd)/go-ggml-transformers \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/gptj ./cmd/grpc/gptj/
+backend-assets/grpc/gptj: backend-assets/grpc sources/go-ggml-transformers/libtransformers.a
+	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/sources/go-ggml-transformers LIBRARY_PATH=$(shell pwd)/sources/go-ggml-transformers \
+	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/gptj ./backend/go/llm/gptj/
 
-backend-assets/grpc/gptneox: backend-assets/grpc go-ggml-transformers/libtransformers.a
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/go-ggml-transformers LIBRARY_PATH=$(shell pwd)/go-ggml-transformers \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/gptneox ./cmd/grpc/gptneox/
+backend-assets/grpc/gptneox: backend-assets/grpc sources/go-ggml-transformers/libtransformers.a
+	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/sources/go-ggml-transformers LIBRARY_PATH=$(shell pwd)/sources/go-ggml-transformers \
+	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/gptneox ./backend/go/llm/gptneox/
 
-backend-assets/grpc/mpt: backend-assets/grpc go-ggml-transformers/libtransformers.a
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/go-ggml-transformers LIBRARY_PATH=$(shell pwd)/go-ggml-transformers \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/mpt ./cmd/grpc/mpt/
+backend-assets/grpc/mpt: backend-assets/grpc sources/go-ggml-transformers/libtransformers.a
+	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/sources/go-ggml-transformers LIBRARY_PATH=$(shell pwd)/sources/go-ggml-transformers \
+	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/mpt ./backend/go/llm/mpt/
 
-backend-assets/grpc/replit: backend-assets/grpc go-ggml-transformers/libtransformers.a
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/go-ggml-transformers LIBRARY_PATH=$(shell pwd)/go-ggml-transformers \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/replit ./cmd/grpc/replit/
+backend-assets/grpc/replit: backend-assets/grpc sources/go-ggml-transformers/libtransformers.a
+	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/sources/go-ggml-transformers LIBRARY_PATH=$(shell pwd)/sources/go-ggml-transformers \
+	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/replit ./backend/go/llm/replit/
 
-backend-assets/grpc/falcon-ggml: backend-assets/grpc go-ggml-transformers/libtransformers.a
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/go-ggml-transformers LIBRARY_PATH=$(shell pwd)/go-ggml-transformers \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/falcon-ggml ./cmd/grpc/falcon-ggml/
+backend-assets/grpc/falcon-ggml: backend-assets/grpc sources/go-ggml-transformers/libtransformers.a
+	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/sources/go-ggml-transformers LIBRARY_PATH=$(shell pwd)/sources/go-ggml-transformers \
+	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/falcon-ggml ./backend/go/llm/falcon-ggml/
 
-backend-assets/grpc/starcoder: backend-assets/grpc go-ggml-transformers/libtransformers.a
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/go-ggml-transformers LIBRARY_PATH=$(shell pwd)/go-ggml-transformers \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/starcoder ./cmd/grpc/starcoder/
+backend-assets/grpc/starcoder: backend-assets/grpc sources/go-ggml-transformers/libtransformers.a
+	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/sources/go-ggml-transformers LIBRARY_PATH=$(shell pwd)/sources/go-ggml-transformers \
+	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/starcoder ./backend/go/llm/starcoder/
 
-backend-assets/grpc/rwkv: backend-assets/grpc go-rwkv/librwkv.a
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/go-rwkv LIBRARY_PATH=$(shell pwd)/go-rwkv \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/rwkv ./cmd/grpc/rwkv/
+backend-assets/grpc/rwkv: backend-assets/grpc sources/go-rwkv/librwkv.a
+	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/sources/go-rwkv LIBRARY_PATH=$(shell pwd)/sources/go-rwkv \
+	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/rwkv ./backend/go/llm/rwkv
 
-backend-assets/grpc/bert-embeddings: backend-assets/grpc go-bert/libgobert.a
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/go-bert LIBRARY_PATH=$(shell pwd)/go-bert \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/bert-embeddings ./cmd/grpc/bert-embeddings/
+backend-assets/grpc/bert-embeddings: backend-assets/grpc sources/go-bert/libgobert.a
+	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/sources/go-bert LIBRARY_PATH=$(shell pwd)/sources/go-bert \
+	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/bert-embeddings ./backend/go/llm/bert/
 
 backend-assets/grpc/langchain-huggingface: backend-assets/grpc
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/langchain-huggingface ./cmd/grpc/langchain-huggingface/
+	$(GOCMD) build -ldflags "$(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 \
-		$(MAKE) go-stable-diffusion/libstablediffusion.a; \
-		CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/go-stable-diffusion/ LIBRARY_PATH=$(shell pwd)/go-stable-diffusion/ \
-		$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/stablediffusion ./cmd/grpc/stablediffusion/; \
+		$(MAKE) sources/go-stable-diffusion/libstablediffusion.a; \
+		CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/sources/go-stable-diffusion/ LIBRARY_PATH=$(shell pwd)/sources/go-stable-diffusion/ \
+		$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/stablediffusion ./backend/go/image/; \
 	fi
 
-backend-assets/grpc/piper: backend-assets/grpc backend-assets/espeak-ng-data go-piper/libpiper_binding.a
-	CGO_CXXFLAGS="$(PIPER_CGO_CXXFLAGS)" CGO_LDFLAGS="$(PIPER_CGO_LDFLAGS)" LIBRARY_PATH=$(shell pwd)/go-piper \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/piper ./cmd/grpc/piper/
+backend-assets/grpc/piper: backend-assets/grpc backend-assets/espeak-ng-data sources/go-piper/libpiper_binding.a
+	CGO_CXXFLAGS="$(PIPER_CGO_CXXFLAGS)" CGO_LDFLAGS="$(PIPER_CGO_LDFLAGS)" LIBRARY_PATH=$(shell pwd)/sources/go-piper \
+	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/piper ./backend/go/tts/
 
-backend-assets/grpc/whisper: backend-assets/grpc whisper.cpp/libwhisper.a
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/whisper.cpp LIBRARY_PATH=$(shell pwd)/whisper.cpp \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/whisper ./cmd/grpc/whisper/
+backend-assets/grpc/whisper: backend-assets/grpc sources/whisper.cpp/libwhisper.a
+	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/sources/whisper.cpp LIBRARY_PATH=$(shell pwd)/sources/whisper.cpp \
+	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/whisper ./backend/go/transcribe/
 
 grpcs: prepare $(GRPC_BACKENDS)

README.md

@@ -22,14 +22,13 @@
 
 > :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/)
-
+> [💻 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 a drop-in replacement REST API that's compatible with OpenAI API specifications for local inferencing. It allows you to run LLMs (and not only) locally or on-prem with consumer grade hardware, supporting multiple model families that are compatible with the ggml format, pytorch and more. Does not require GPU.
+**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.
 
-<p align="center"><b>Follow LocalAI </b></p>
+<p  align="center"><b>Follow LocalAI </b></p>
 
 <p align="center">
 <a href="https://twitter.com/LocalAI_API" target="blank">
@@ -39,7 +38,7 @@
 <img src="https://dcbadge.vercel.app/api/server/uJAeKSAGDy?style=flat-square&theme=default-inverted" alt="Join LocalAI Discord Community"/>
 </a>
 
-<p align="center"><b>Connect with the Creator </b></p>
+<p  align="center"><b>Connect with the Creator </b></p>
 
 <p align="center">
 <a href="https://twitter.com/mudler_it" target="blank">
@@ -50,7 +49,7 @@
 </a>
 </p>
 
-<p align="center"><b>Share LocalAI Repository</b></p>
+<p  align="center"><b>Share LocalAI Repository</b></p>
 
 <p align="center">
 
@@ -64,6 +63,22 @@
 
 </p>
 
+## 💻 [Getting started](https://localai.io/basics/getting_started/index.html)
+
+## 🔥🔥 Hot topics / Roadmap
+
+[Roadmap](https://github.com/mudler/LocalAI/issues?q=is%3Aissue+is%3Aopen+label%3Aroadmap)
+
+🆕 New! [LLM finetuning guide](https://localai.io/advanced/fine-tuning/)
+
+Hot topics (looking for contributors):
+- Backends v2: https://github.com/mudler/LocalAI/issues/1126
+- Improving UX v2: https://github.com/mudler/LocalAI/issues/1373
+
+If you want to help and contribute, issues up for grabs: https://github.com/mudler/LocalAI/issues?q=is%3Aissue+is%3Aopen+label%3A%22up+for+grabs%22
+
+
+
 <hr>
 
 In a nutshell:
@@ -79,8 +94,6 @@ LocalAI was created by [Ettore Di Giacinto](https://github.com/mudler/) and is a
 
 Note that this started just as a [fun weekend project](https://localai.io/#backstory) in order to try to create the necessary pieces for a full AI assistant like `ChatGPT`: the community is growing fast and we are working hard to make it better and more stable. If you want to help, please consider contributing (see below)!
 
-## 🔥🔥 [Hot topics / Roadmap](https://localai.io/#-hot-topics--roadmap)
-
 ## 🚀 [Features](https://localai.io/features/)
 
 - 📖 [Text generation with GPTs](https://localai.io/features/text-generation/) (`llama.cpp`, `gpt4all.cpp`, ... [:book: and more](https://localai.io/model-compatibility/index.html#model-compatibility-table))
@@ -91,8 +104,32 @@ Note that this started just as a [fun weekend project](https://localai.io/#backs
 - 🧠 [Embeddings generation for vector databases](https://localai.io/features/embeddings/)
 - ✍️ [Constrained grammars](https://localai.io/features/constrained_grammars/)
 - 🖼️ [Download Models directly from Huggingface ](https://localai.io/models/)
+- 🆕 [Vision API](https://localai.io/features/gpt-vision/)
 
+## 💻 Usage
 
+Check out the [Getting started](https://localai.io/basics/getting_started/index.html) section in our documentation.
+
+### 🔗 Community and integrations
+
+WebUIs:
+- https://github.com/Jirubizu/localai-admin
+- https://github.com/go-skynet/LocalAI-frontend
+
+Model galleries
+- https://github.com/go-skynet/model-gallery
+
+Other:
+- Helm chart https://github.com/go-skynet/helm-charts
+
+### 🔗 Resources
+
+- 🆕 New! [LLM finetuning guide](https://localai.io/advanced/fine-tuning/)
+- [How to build locally](https://localai.io/basics/build/index.html)
+- [How to install in Kubernetes](https://localai.io/basics/getting_started/index.html#run-localai-in-kubernetes)
+- [Projects integrating LocalAI](https://localai.io/integrations/)
+- [How tos section](https://localai.io/howtos/) (curated by our community)
+  
 ## :book: 🎥 [Media, Blogs, Social](https://localai.io/basics/news/#media-blogs-social)
 
 - [Create a slackbot for teams and OSS projects that answer to documentation](https://mudler.pm/posts/smart-slackbot-for-teams/)
@@ -100,21 +137,6 @@ Note that this started just as a [fun weekend project](https://localai.io/#backs
 - [Question Answering on Documents locally with LangChain, LocalAI, Chroma, and GPT4All](https://mudler.pm/posts/localai-question-answering/)
 - [Tutorial to use k8sgpt with LocalAI](https://medium.com/@tyler_97636/k8sgpt-localai-unlock-kubernetes-superpowers-for-free-584790de9b65)
 
-## 💻 Usage
-
-Check out the [Getting started](https://localai.io/basics/getting_started/index.html) section in our documentation.
-
-### 💡 Example: Use Luna-AI Llama model
-
-See the [documentation](https://localai.io/basics/getting_started)
-
-### 🔗 Resources
-
-- [How to build locally](https://localai.io/basics/build/index.html)
-- [How to install in Kubernetes](https://localai.io/basics/getting_started/index.html#run-localai-in-kubernetes)
-- [Projects integrating LocalAI](https://localai.io/integrations/)
-- [How tos section](https://localai.io/howtos/) (curated by our community)
-  
 ## Citation
 
 If you utilize this repository, data in a downstream project, please consider citing it with:

api/api.go

@@ -13,6 +13,7 @@ import (
 	"github.com/go-skynet/LocalAI/internal"
 	"github.com/go-skynet/LocalAI/metrics"
 	"github.com/go-skynet/LocalAI/pkg/assets"
+	"github.com/go-skynet/LocalAI/pkg/model"
 
 	"github.com/gofiber/fiber/v2"
 	"github.com/gofiber/fiber/v2/middleware/cors"
@@ -79,6 +80,22 @@ func Startup(opts ...options.AppOption) (*options.Option, *config.ConfigLoader,
 		options.Loader.StopAllGRPC()
 	}()
 
+	if options.WatchDog {
+		wd := model.NewWatchDog(
+			options.Loader,
+			options.WatchDogBusyTimeout,
+			options.WatchDogIdleTimeout,
+			options.WatchDogBusy,
+			options.WatchDogIdle)
+		options.Loader.SetWatchDog(wd)
+		go wd.Run()
+		go func() {
+			<-options.Context.Done()
+			log.Debug().Msgf("Context canceled, shutting down")
+			wd.Shutdown()
+		}()
+	}
+
 	return options, cl, nil
 }
 

api/api_test.go

@@ -301,7 +301,7 @@ var _ = Describe("API test", func() {
 				response := postModelApplyRequest("http://127.0.0.1:9090/models/apply", modelApplyRequest{
 					URL:       "github:go-skynet/model-gallery/openllama_3b.yaml",
 					Name:      "openllama_3b",
-					Overrides: map[string]interface{}{"backend": "llama-stable", "mmap": true, "f16": true, "context_size": 128},
+					Overrides: map[string]interface{}{"backend": "llama-ggml", "mmap": true, "f16": true, "context_size": 128},
 				})
 
 				Expect(response["uuid"]).ToNot(BeEmpty(), fmt.Sprint(response))
@@ -704,7 +704,7 @@ var _ = Describe("API test", func() {
 		})
 
 		Context("External gRPC calls", func() {
-			It("calculate embeddings with huggingface", func() {
+			It("calculate embeddings with sentencetransformers", func() {
 				if runtime.GOOS != "linux" {
 					Skip("test supported only on linux")
 				}

api/backend/options.go

@@ -16,6 +16,10 @@ func modelOpts(c config.Config, o *options.Option, opts []model.Option) []model.
 		opts = append(opts, model.WithSingleActiveBackend())
 	}
 
+	if o.ParallelBackendRequests {
+		opts = append(opts, model.EnableParallelRequests)
+	}
+
 	if c.GRPC.Attempts != 0 {
 		opts = append(opts, model.WithGRPCAttempts(c.GRPC.Attempts))
 	}

api/config/config.go

@@ -277,7 +277,7 @@ func (cm *ConfigLoader) LoadConfigs(path string) error {
 	}
 	for _, file := range files {
 		// Skip templates, YAML and .keep files
-		if !strings.Contains(file.Name(), ".yaml") {
+		if !strings.Contains(file.Name(), ".yaml") && !strings.Contains(file.Name(), ".yml") {
 			continue
 		}
 		c, err := ReadConfig(filepath.Join(path, file.Name()))

api/localai/backend_monitor.go

@@ -123,13 +123,12 @@ func BackendMonitorEndpoint(bm BackendMonitor) func(c *fiber.Ctx) error {
 			return err
 		}
 
-		client := bm.options.Loader.CheckIsLoaded(backendId)
-
-		if client == nil {
+		model := bm.options.Loader.CheckIsLoaded(backendId)
+		if model == "" {
 			return fmt.Errorf("backend %s is not currently loaded", backendId)
 		}
 
-		status, rpcErr := client.Status(context.TODO())
+		status, rpcErr := model.GRPC(false, nil).Status(context.TODO())
 		if rpcErr != nil {
 			log.Warn().Msgf("backend %s experienced an error retrieving status info: %s", backendId, rpcErr.Error())
 			val, slbErr := bm.SampleLocalBackendProcess(backendId)

api/openai/chat.go

@@ -81,7 +81,7 @@ func ChatEndpoint(cm *config.ConfigLoader, o *options.Option) func(c *fiber.Ctx)
 			noActionDescription = config.FunctionsConfig.NoActionDescriptionName
 		}
 
-		if input.ResponseFormat == "json_object" {
+		if input.ResponseFormat.Type == "json_object" {
 			input.Grammar = grammar.JSONBNF
 		}
 

api/openai/completion.go

@@ -65,7 +65,7 @@ func CompletionEndpoint(cm *config.ConfigLoader, o *options.Option) func(c *fibe
 			return fmt.Errorf("failed reading parameters from request:%w", err)
 		}
 
-		if input.ResponseFormat == "json_object" {
+		if input.ResponseFormat.Type == "json_object" {
 			input.Grammar = grammar.JSONBNF
 		}
 

api/openai/image.go

@@ -100,7 +100,7 @@ func ImageEndpoint(cm *config.ConfigLoader, o *options.Option) func(c *fiber.Ctx
 		}
 
 		b64JSON := false
-		if input.ResponseFormat == "b64_json" {
+		if input.ResponseFormat.Type == "b64_json" {
 			b64JSON = true
 		}
 		// src and clip_skip

api/options/options.go

@@ -4,10 +4,11 @@ import (
 	"context"
 	"embed"
 	"encoding/json"
+	"time"
 
+	"github.com/go-skynet/LocalAI/metrics"
 	"github.com/go-skynet/LocalAI/pkg/gallery"
 	model "github.com/go-skynet/LocalAI/pkg/model"
-	"github.com/go-skynet/LocalAI/metrics"
 	"github.com/rs/zerolog/log"
 )
 
@@ -36,7 +37,13 @@ type Option struct {
 
 	AutoloadGalleries bool
 
-	SingleBackend bool
+	SingleBackend           bool
+	ParallelBackendRequests bool
+
+	WatchDogIdle                             bool
+	WatchDogBusy                             bool
+	WatchDog                                 bool
+	WatchDogBusyTimeout, WatchDogIdleTimeout time.Duration
 }
 
 type AppOption func(*Option)
@@ -62,10 +69,40 @@ func WithCors(b bool) AppOption {
 	}
 }
 
+var EnableWatchDog = func(o *Option) {
+	o.WatchDog = true
+}
+
+var EnableWatchDogIdleCheck = func(o *Option) {
+	o.WatchDog = true
+	o.WatchDogIdle = true
+}
+
+var EnableWatchDogBusyCheck = func(o *Option) {
+	o.WatchDog = true
+	o.WatchDogBusy = true
+}
+
+func SetWatchDogBusyTimeout(t time.Duration) AppOption {
+	return func(o *Option) {
+		o.WatchDogBusyTimeout = t
+	}
+}
+
+func SetWatchDogIdleTimeout(t time.Duration) AppOption {
+	return func(o *Option) {
+		o.WatchDogIdleTimeout = t
+	}
+}
+
 var EnableSingleBackend = func(o *Option) {
 	o.SingleBackend = true
 }
 
+var EnableParallelBackendRequests = func(o *Option) {
+	o.ParallelBackendRequests = true
+}
+
 var EnableGalleriesAutoload = func(o *Option) {
 	o.AutoloadGalleries = true
 }

api/schema/openai.go

@@ -83,6 +83,12 @@ type OpenAIModel struct {
 	Object string `json:"object"`
 }
 
+type ChatCompletionResponseFormatType string
+
+type ChatCompletionResponseFormat struct {
+	Type ChatCompletionResponseFormatType `json:"type,omitempty"`
+}
+
 type OpenAIRequest struct {
 	config.PredictionOptions
 
@@ -92,7 +98,7 @@ type OpenAIRequest struct {
 	// whisper
 	File string `json:"file" validate:"required"`
 	//whisper/image
-	ResponseFormat string `json:"response_format"`
+	ResponseFormat ChatCompletionResponseFormat `json:"response_format"`
 	// image
 	Size string `json:"size"`
 	// Prompt is read only by completion/image API calls

backend/cpp/llama/CMakeLists.txt

@@ -36,7 +36,7 @@ include_directories(${Protobuf_INCLUDE_DIRS})
 message(STATUS "Using protobuf version ${Protobuf_VERSION} | Protobuf_INCLUDE_DIRS: ${Protobuf_INCLUDE_DIRS} | CMAKE_CURRENT_BINARY_DIR: ${CMAKE_CURRENT_BINARY_DIR}")
 
 # Proto file
-get_filename_component(hw_proto "../../../../../../pkg/grpc/proto/backend.proto" ABSOLUTE)
+get_filename_component(hw_proto "../../../../../../backend/backend.proto" ABSOLUTE)
 get_filename_component(hw_proto_path "${hw_proto}" PATH)
 
 # Generated sources

backend/go/image/main.go

@@ -5,7 +5,6 @@ package main
 import (
 	"flag"
 
-	bert "github.com/go-skynet/LocalAI/pkg/backend/llm/bert"
 	grpc "github.com/go-skynet/LocalAI/pkg/grpc"
 )
 
@@ -16,7 +15,7 @@ var (
 func main() {
 	flag.Parse()
 
-	if err := grpc.StartServer(*addr, &bert.Embeddings{}); err != nil {
+	if err := grpc.StartServer(*addr, &StableDiffusion{}); err != nil {
 		panic(err)
 	}
 }

backend/go/image/stablediffusion.go

@@ -1,4 +1,4 @@
-package image
+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)

backend/go/llm/bert/bert.go

@@ -1,4 +1,4 @@
-package bert
+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)

backend/go/llm/bert/main.go

@@ -5,8 +5,6 @@ package main
 import (
 	"flag"
 
-	rwkv "github.com/go-skynet/LocalAI/pkg/backend/llm/rwkv"
-
 	grpc "github.com/go-skynet/LocalAI/pkg/grpc"
 )
 
@@ -17,7 +15,7 @@ var (
 func main() {
 	flag.Parse()
 
-	if err := grpc.StartServer(*addr, &rwkv.LLM{}); err != nil {
+	if err := grpc.StartServer(*addr, &Embeddings{}); err != nil {
 		panic(err)
 	}
 }

backend/go/llm/dolly/main.go

@@ -5,7 +5,7 @@ package main
 import (
 	"flag"
 
-	transformers "github.com/go-skynet/LocalAI/pkg/backend/llm/transformers"
+	transformers "github.com/go-skynet/LocalAI/backend/go/llm/transformers"
 
 	grpc "github.com/go-skynet/LocalAI/pkg/grpc"
 )

backend/go/llm/falcon-ggml/main.go

@@ -5,7 +5,7 @@ package main
 import (
 	"flag"
 
-	transformers "github.com/go-skynet/LocalAI/pkg/backend/llm/transformers"
+	transformers "github.com/go-skynet/LocalAI/backend/go/llm/transformers"
 
 	grpc "github.com/go-skynet/LocalAI/pkg/grpc"
 )

backend/go/llm/gpt2/main.go

@@ -5,7 +5,7 @@ package main
 import (
 	"flag"
 
-	transformers "github.com/go-skynet/LocalAI/pkg/backend/llm/transformers"
+	transformers "github.com/go-skynet/LocalAI/backend/go/llm/transformers"
 
 	grpc "github.com/go-skynet/LocalAI/pkg/grpc"
 )

backend/go/llm/gpt4all/gpt4all.go

@@ -1,4 +1,4 @@
-package gpt4all
+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)

backend/go/llm/gpt4all/main.go

@@ -5,8 +5,6 @@ package main
 import (
 	"flag"
 
-	gpt4all "github.com/go-skynet/LocalAI/pkg/backend/llm/gpt4all"
-
 	grpc "github.com/go-skynet/LocalAI/pkg/grpc"
 )
 
@@ -17,7 +15,7 @@ var (
 func main() {
 	flag.Parse()
 
-	if err := grpc.StartServer(*addr, &gpt4all.LLM{}); err != nil {
+	if err := grpc.StartServer(*addr, &LLM{}); err != nil {
 		panic(err)
 	}
 }

backend/go/llm/gptj/main.go

@@ -5,7 +5,7 @@ package main
 import (
 	"flag"
 
-	transformers "github.com/go-skynet/LocalAI/pkg/backend/llm/transformers"
+	transformers "github.com/go-skynet/LocalAI/backend/go/llm/transformers"
 
 	grpc "github.com/go-skynet/LocalAI/pkg/grpc"
 )

backend/go/llm/gptneox/main.go

@@ -5,7 +5,7 @@ package main
 import (
 	"flag"
 
-	transformers "github.com/go-skynet/LocalAI/pkg/backend/llm/transformers"
+	transformers "github.com/go-skynet/LocalAI/backend/go/llm/transformers"
 
 	grpc "github.com/go-skynet/LocalAI/pkg/grpc"
 )

backend/go/llm/langchain/langchain.go

@@ -1,4 +1,4 @@
-package langchain
+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)

backend/go/llm/langchain/main.go

@@ -0,0 +1,21 @@
+package main
+
+// Note: this is started internally by LocalAI and a server is allocated for each model
+
+import (
+	"flag"
+
+	grpc "github.com/go-skynet/LocalAI/pkg/grpc"
+)
+
+var (
+	addr = flag.String("addr", "localhost:50051", "the address to connect to")
+)
+
+func main() {
+	flag.Parse()
+
+	if err := grpc.StartServer(*addr, &LLM{}); err != nil {
+		panic(err)
+	}
+}

backend/go/llm/llama-ggml/llama.go

@@ -1,4 +1,4 @@
-package llama
+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)

backend/go/llm/llama-ggml/main.go

@@ -3,8 +3,6 @@ package main
 import (
 	"flag"
 
-	llama "github.com/go-skynet/LocalAI/pkg/backend/llm/llama-stable"
-
 	grpc "github.com/go-skynet/LocalAI/pkg/grpc"
 )
 
@@ -15,7 +13,7 @@ var (
 func main() {
 	flag.Parse()
 
-	if err := grpc.StartServer(*addr, &llama.LLM{}); err != nil {
+	if err := grpc.StartServer(*addr, &LLM{}); err != nil {
 		panic(err)
 	}
 }

backend/go/llm/llama/llama.go

@@ -1,4 +1,4 @@
-package llama
+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)

backend/go/llm/llama/main.go

@@ -1,12 +1,12 @@
 package main
 
+// GRPC Falcon server
+
 // Note: this is started internally by LocalAI and a server is allocated for each model
 
 import (
 	"flag"
 
-	tts "github.com/go-skynet/LocalAI/pkg/backend/tts"
-
 	grpc "github.com/go-skynet/LocalAI/pkg/grpc"
 )
 
@@ -17,7 +17,7 @@ var (
 func main() {
 	flag.Parse()
 
-	if err := grpc.StartServer(*addr, &tts.Piper{}); err != nil {
+	if err := grpc.StartServer(*addr, &LLM{}); err != nil {
 		panic(err)
 	}
 }

backend/go/llm/mpt/main.go

@@ -5,7 +5,7 @@ package main
 import (
 	"flag"
 
-	transformers "github.com/go-skynet/LocalAI/pkg/backend/llm/transformers"
+	transformers "github.com/go-skynet/LocalAI/backend/go/llm/transformers"
 
 	grpc "github.com/go-skynet/LocalAI/pkg/grpc"
 )

backend/go/llm/replit/main.go

@@ -5,7 +5,7 @@ package main
 import (
 	"flag"
 
-	transformers "github.com/go-skynet/LocalAI/pkg/backend/llm/transformers"
+	transformers "github.com/go-skynet/LocalAI/backend/go/llm/transformers"
 
 	grpc "github.com/go-skynet/LocalAI/pkg/grpc"
 )

backend/go/llm/rwkv/main.go

@@ -0,0 +1,21 @@
+package main
+
+// Note: this is started internally by LocalAI and a server is allocated for each model
+
+import (
+	"flag"
+
+	grpc "github.com/go-skynet/LocalAI/pkg/grpc"
+)
+
+var (
+	addr = flag.String("addr", "localhost:50051", "the address to connect to")
+)
+
+func main() {
+	flag.Parse()
+
+	if err := grpc.StartServer(*addr, &LLM{}); err != nil {
+		panic(err)
+	}
+}

backend/go/llm/rwkv/rwkv.go

@@ -1,4 +1,4 @@
-package rwkv
+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)

backend/go/llm/starcoder/main.go

@@ -5,7 +5,7 @@ package main
 import (
 	"flag"
 
-	transformers "github.com/go-skynet/LocalAI/pkg/backend/llm/transformers"
+	transformers "github.com/go-skynet/LocalAI/backend/go/llm/transformers"
 
 	grpc "github.com/go-skynet/LocalAI/pkg/grpc"
 )

backend/go/transcribe/main.go

@@ -0,0 +1,21 @@
+package main
+
+// Note: this is started internally by LocalAI and a server is allocated for each model
+
+import (
+	"flag"
+
+	grpc "github.com/go-skynet/LocalAI/pkg/grpc"
+)
+
+var (
+	addr = flag.String("addr", "localhost:50051", "the address to connect to")
+)
+
+func main() {
+	flag.Parse()
+
+	if err := grpc.StartServer(*addr, &Whisper{}); err != nil {
+		panic(err)
+	}
+}

backend/go/transcribe/transcript.go

@@ -1,4 +1,4 @@
-package transcribe
+package main
 
 import (
 	"fmt"

backend/go/transcribe/whisper.go

@@ -1,4 +1,4 @@
-package transcribe
+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)

backend/go/tts/main.go

@@ -0,0 +1,21 @@
+package main
+
+// Note: this is started internally by LocalAI and a server is allocated for each model
+
+import (
+	"flag"
+
+	grpc "github.com/go-skynet/LocalAI/pkg/grpc"
+)
+
+var (
+	addr = flag.String("addr", "localhost:50051", "the address to connect to")
+)
+
+func main() {
+	flag.Parse()
+
+	if err := grpc.StartServer(*addr, &Piper{}); err != nil {
+		panic(err)
+	}
+}

backend/go/tts/piper.go

@@ -1,4 +1,4 @@
-package tts
+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)

backend/python/autogptq/Makefile

@@ -1,4 +1,4 @@
-.PONY: autogptq
+.PHONY: autogptq
 autogptq:
 	@echo "Creating virtual environment..."
 	@conda env create --name autogptq --file autogptq.yml

backend/python/bark/Makefile

@@ -1,5 +1,5 @@
-.PONY: ttsbark
+.PHONY: ttsbark
 ttsbark:
 	@echo "Creating virtual environment..."
 	@conda env create --name ttsbark --file ttsbark.yml
-	@echo "Virtual environment created."
+	@echo "Virtual environment created."

backend/python/bark/ttsbark.py

@@ -1,8 +1,7 @@
-"""
-This is the extra gRPC server of LocalAI
-"""
-
 #!/usr/bin/env python3
+"""
+This is an extra gRPC server of LocalAI for Bark TTS
+"""
 from concurrent import futures
 import time
 import argparse

backend/python/diffusers/Makefile

@@ -1,11 +1,11 @@
-.PONY: diffusers
+.PHONY: diffusers
 diffusers:
 	@echo "Creating virtual environment..."
 	@conda env create --name diffusers --file diffusers.yml
 	@echo "Virtual environment created."
 
-.PONY: run
+.PHONY: run
 run:
 	@echo "Running diffusers..."
 	bash run.sh
-	@echo "Diffusers run."
+	@echo "Diffusers run."

backend/python/exllama/Makefile

@@ -1,11 +1,11 @@
-.PONY: exllama
+.PHONY: exllama
 exllama:
 	@echo "Creating virtual environment..."
 	@conda env create --name exllama --file exllama.yml
 	@echo "Virtual environment created."
 
-.PONY: run
+.PHONY: run
 run:
 	@echo "Running exllama..."
 	bash run.sh
-	@echo "exllama run."
+	@echo "exllama run."

backend/python/exllama/exllama.py

@@ -63,6 +63,19 @@ class BackendServicer(backend_pb2_grpc.BackendServicer):
 
             config = ExLlamaConfig(model_config_path)               # create config from config.json
             config.model_path = model_path                          # supply path to model weights file
+            if (request.ContextSize):
+                config.max_seq_len = request.ContextSize            # override max sequence length
+                config.max_attention_size = request.ContextSize**2  # Should be set to context_size^2. 
+                # https://github.com/turboderp/exllama/issues/220#issuecomment-1720324163
+
+            # Set Rope scaling.
+            if (request.RopeFreqScale):
+                # Alpha value for Rope scaling. 
+                # Higher value increases context but adds perplexity.
+                # alpha_value and compress_pos_emb are mutually exclusive.
+                # https://github.com/turboderp/exllama/issues/115
+                config.alpha_value = request.RopeFreqScale
+                config.calculate_rotary_embedding_base()
 
             model = ExLlama(config)                                 # create ExLlama instance and load the weights
             tokenizer = ExLlamaTokenizer(tokenizer_path)            # create tokenizer from tokenizer model file

backend/python/petals/Makefile

@@ -0,0 +1,11 @@
+.PHONY: petals
+petals:
+	@echo "Creating virtual environment..."
+	@conda env create --name petals --file petals.yml
+	@echo "Virtual environment created."
+
+.PHONY: run
+run:
+	@echo "Running petals..."
+	bash run.sh
+	@echo "petals run."

backend/python/petals/backend_petals.py

@@ -0,0 +1,140 @@
+#!/usr/bin/env python3
+from concurrent import futures
+import time
+import argparse
+import signal
+import sys
+import os
+
+import backend_pb2
+import backend_pb2_grpc
+
+import grpc
+import torch
+from transformers import AutoTokenizer
+from petals import AutoDistributedModelForCausalLM
+
+_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 that implements the Backend service defined in backend.proto.
+    """
+    def Health(self, request, context):
+        """
+        Returns a health check message.
+
+        Args:
+            request: The health check request.
+            context: The gRPC context.
+
+        Returns:
+            backend_pb2.Reply: The health check reply.
+        """
+        return backend_pb2.Reply(message=bytes("OK", 'utf-8'))
+
+    def LoadModel(self, request, context):
+        """
+        Loads a language model.
+
+        Args:
+            request: The load model request.
+            context: The gRPC context.
+
+        Returns:
+            backend_pb2.Result: The load model result.
+        """
+        try:
+            self.tokenizer = AutoTokenizer.from_pretrained(request.Model, use_fast=False, add_bos_token=False)
+            self.model = AutoDistributedModelForCausalLM.from_pretrained(request.Model)
+            self.cuda = False
+            if request.CUDA:
+                self.model = self.model.cuda()
+                self.cuda = True
+
+        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 Predict(self, request, context):
+        """
+        Generates text based on the given prompt and sampling parameters.
+
+        Args:
+            request: The predict request.
+            context: The gRPC context.
+
+        Returns:
+            backend_pb2.Result: The predict result.
+        """
+
+        inputs = self.tokenizer(request.Prompt, return_tensors="pt")["input_ids"]
+        if self.cuda:
+            inputs = inputs.cuda()
+ 
+        if request.Tokens == 0:
+            # Max to max value if tokens are not specified
+            request.Tokens = 8192
+
+        # TODO: kwargs and map all parameters
+        outputs = self.model.generate(inputs, max_new_tokens=request.Tokens)
+
+        generated_text = self.tokenizer.decode(outputs[0])
+        # Remove prompt from response if present
+        if request.Prompt in generated_text:
+            generated_text = generated_text.replace(request.Prompt, "")
+
+        return backend_pb2.Result(message=bytes(generated_text, encoding='utf-8'))
+
+    def PredictStream(self, request, context):
+        """
+        Generates text based on the given prompt and sampling parameters, and streams the results.
+
+        Args:
+            request: The predict stream request.
+            context: The gRPC context.
+
+        Returns:
+            backend_pb2.Result: The predict stream result.
+        """
+        # Implement PredictStream RPC
+        #for reply in some_data_generator():
+        #    yield reply
+        # Not implemented yet
+        return self.Predict(request, context)
+
+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("Server started. Listening on: " + address, file=sys.stderr)
+
+    # Define the signal handler function
+    def signal_handler(sig, frame):
+        print("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()
+
+    serve(args.addr)

backend/python/petals/petals.yml

@@ -0,0 +1,29 @@
+name: petals
+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:
+      - git+https://github.com/bigscience-workshop/petals
+prefix: /opt/conda/envs/petals

backend/python/petals/run.sh

@@ -0,0 +1,21 @@
+#!/bin/bash
+
+##
+## A bash script wrapper that runs the exllama server with conda
+
+export PATH=$PATH:/opt/conda/bin
+
+# Activate conda environment
+# if source is available use it, or use conda
+#
+if [ -f /opt/conda/bin/activate ]; then
+    source activate petals
+else
+    eval "$(conda shell.bash hook)"
+    conda activate petals
+fi
+
+# get the directory where the bash script is located
+DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" >/dev/null 2>&1 && pwd )"
+
+python $DIR/backend_petals.py $@

backend/python/sentencetransformers/Makefile

@@ -0,0 +1,18 @@
+.PHONY: sentencetransformers
+sentencetransformers:
+	@echo "Creating virtual environment..."
+	@conda env create --name sentencetransformers --file sentencetransformers.yml
+	@echo "Virtual environment created."
+
+.PHONY: run
+run:
+	@echo "Running sentencetransformers..."
+	bash run.sh
+	@echo "sentencetransformers run."
+
+# It is not working well by using command line. It only6 works with IDE like VSCode.
+.PHONY: test
+test:
+	@echo "Testing sentencetransformers..."
+	bash test.sh
+	@echo "sentencetransformers tested."

backend/python/sentencetransformers/README.md

@@ -0,0 +1,5 @@
+# Creating a separate environment for the sentencetransformers project
+
+```
+make sentencetransformers
+```

backend/python/sentencetransformers/run.sh

@@ -0,0 +1,14 @@
+#!/bin/bash
+
+##
+## A bash script wrapper that runs the sentencetransformers server with conda
+
+export PATH=$PATH:/opt/conda/bin
+
+# Activate conda environment
+source activate sentencetransformers
+
+# get the directory where the bash script is located
+DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" >/dev/null 2>&1 && pwd )"
+
+python $DIR/sentencetransformers.py $@

backend/python/sentencetransformers/sentencetransformers.py

@@ -1,7 +1,7 @@
+#!/usr/bin/env python3
 """
 Extra gRPC server for HuggingFace SentenceTransformer models.
 """
-#!/usr/bin/env python3
 from concurrent import futures
 
 import argparse