fixup: create piper libdir also when not built

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

.env

@@ -66,4 +66,7 @@ MODELS_PATH=/models
 ### Python backends GRPC max workers
 ### Default number of workers for GRPC Python backends.
 ### This actually controls wether a backend can process multiple requests or not.
-# PYTHON_GRPC_MAX_WORKERS=1
+# PYTHON_GRPC_MAX_WORKERS=1
+
+### Define the number of parallel LLAMA.cpp workers (Defaults to 1)
+# LLAMACPP_PARALLEL=1

.github/workflows/bump_deps.yaml

@@ -12,6 +12,9 @@ jobs:
           - repository: "go-skynet/go-llama.cpp"
             variable: "GOLLAMA_VERSION"
             branch: "master"
+          - repository: "ggerganov/llama.cpp"
+            variable: "CPPLLAMA_VERSION"
+            branch: "master"
           - repository: "go-skynet/go-ggml-transformers.cpp"
             variable: "GOGGMLTRANSFORMERS_VERSION"
             branch: "master"
@@ -41,7 +44,7 @@ jobs:
             branch: "master"
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
       - name: Bump dependencies 🔧
         run: |
           bash .github/bump_deps.sh ${{ matrix.repository }} ${{ matrix.branch }} ${{ matrix.variable }}

.github/workflows/image.yml

@@ -14,15 +14,21 @@ concurrency:
   cancel-in-progress: true
 
 jobs:
-  docker:
+  image-build:
     strategy:
       matrix:
         include:
           - build-type: ''
-            platforms: 'linux/amd64,linux/arm64'
+            #platforms: 'linux/amd64,linux/arm64'
+            platforms: 'linux/amd64'
             tag-latest: 'auto'
             tag-suffix: ''
             ffmpeg: ''
+          - build-type: ''
+            platforms: 'linux/amd64'
+            tag-latest: 'false'
+            tag-suffix: '-ffmpeg'
+            ffmpeg: 'true'
           - build-type: 'cublas'
             cuda-major-version: 11
             cuda-minor-version: 7
@@ -37,11 +43,6 @@ jobs:
             tag-latest: 'false'
             tag-suffix: '-cublas-cuda12'
             ffmpeg: ''
-          - build-type: ''
-            platforms: 'linux/amd64,linux/arm64'
-            tag-latest: 'false'
-            tag-suffix: '-ffmpeg'
-            ffmpeg: 'true'
           - build-type: 'cublas'
             cuda-major-version: 11
             cuda-minor-version: 7
@@ -57,43 +58,54 @@ jobs:
             tag-suffix: '-cublas-cuda12-ffmpeg'
             ffmpeg: 'true'
 
-    runs-on: ubuntu-latest
+    runs-on: arc-runner-set 
     steps:
-      - name: Release space from worker
+      - name: Force Install GIT 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 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
+          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@v3
-
+        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

.github/workflows/release.yaml

@@ -19,7 +19,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Clone
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
         with:
           submodules: true
       - uses: actions/setup-go@v4
@@ -29,6 +29,12 @@ jobs:
         run: |
           sudo apt-get update
           sudo apt-get install build-essential ffmpeg
+
+          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 -j12 install
+
       - name: Build
         id: build
         env:
@@ -60,18 +66,26 @@ jobs:
     runs-on: macOS-latest
     steps:
       - name: Clone
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
         with:
           submodules: true
       - uses: actions/setup-go@v4
         with:
           go-version: '>=1.21.0'
+      - name: Dependencies
+        run: |
+          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
       - 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
           make dist
       - uses: actions/upload-artifact@v3
         with:

.github/workflows/test-gpu.yml

@@ -0,0 +1,63 @@
+---
+name: 'GPU tests'
+
+on:
+  pull_request:
+  push:
+    branches:
+      - master
+    tags:
+      - '*'
+
+concurrency:
+  group: ci-gpu-tests-${{ github.head_ref || github.ref }}-${{ github.repository }}
+  cancel-in-progress: true
+
+jobs:
+  ubuntu-latest:
+    runs-on: gpu
+    strategy:
+      matrix:
+        go-version: ['1.21.x']
+    steps:
+      - name: Clone
+        uses: actions/checkout@v4
+        with: 
+          submodules: true
+      - name: Setup Go ${{ matrix.go-version }}
+        uses: actions/setup-go@v4
+        with:
+          go-version: ${{ matrix.go-version }}
+      # You can test your matrix by printing the current Go version
+      - name: Display Go version
+        run: go version
+      - name: Dependencies
+        run: |
+          sudo apt-get update
+          sudo DEBIAN_FRONTEND=noninteractive apt-get install -y make wget
+      - name: Build
+        run: |
+          if [ ! -e /run/systemd/system ]; then
+            sudo mkdir /run/systemd/system
+          fi
+          sudo mkdir -p /host/tests/${{ github.head_ref || github.ref }}
+          sudo chmod -R 777 /host/tests/${{ github.head_ref || github.ref }}
+          make \
+            TEST_DIR="/host/tests/${{ github.head_ref || github.ref }}" \
+            BUILD_TYPE=cublas \
+            prepare-e2e run-e2e-image test-e2e
+      - name: Release space from worker ♻
+        if: always()
+        run: |
+          sudo rm -rf build || true
+          sudo rm -rf bin || true
+          sudo rm -rf dist || true
+          sudo docker logs $(sudo docker ps -q --filter ancestor=localai-tests) > logs.txt
+          sudo cat logs.txt || true
+          sudo rm -rf logs.txt
+          make clean || true
+          make \
+            TEST_DIR="/host/tests/${{ github.head_ref || github.ref }}" \
+            teardown-e2e || true
+          sudo rm -rf /host/tests/${{ github.head_ref || github.ref }} || true
+          docker system prune -f -a --volumes || true

.github/workflows/test.yml

@@ -14,7 +14,7 @@ concurrency:
   cancel-in-progress: true
 
 jobs:
-  ubuntu-latest:
+  tests-linux:
     runs-on: ubuntu-latest
     strategy:
       matrix:
@@ -53,7 +53,7 @@ jobs:
           sudo rm -rfv build || true
           df -h
       - name: Clone
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
         with: 
           submodules: true
       - name: Setup Go ${{ matrix.go-version }}
@@ -67,38 +67,43 @@ jobs:
         run: |
           sudo apt-get update
           sudo apt-get install build-essential ffmpeg
-          
+          curl https://repo.anaconda.com/pkgs/misc/gpgkeys/anaconda.asc | gpg --dearmor > conda.gpg && \
+             sudo install -o root -g root -m 644 conda.gpg /usr/share/keyrings/conda-archive-keyring.gpg && \
+              gpg --keyring /usr/share/keyrings/conda-archive-keyring.gpg --no-default-keyring --fingerprint 34161F5BF5EB1D4BFBBB8F0A8AEB4F8B29D82806 && \
+             sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" > /etc/apt/sources.list.d/conda.list' && \
+             sudo /bin/bash -c 'echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" | tee -a /etc/apt/sources.list.d/conda.list' && \
+             sudo apt-get update && \
+             sudo apt-get install -y conda
           sudo apt-get install -y ca-certificates cmake curl patch
           sudo apt-get install -y libopencv-dev && sudo ln -s /usr/include/opencv4/opencv2 /usr/include/opencv2
-          sudo pip install -r extra/requirements.txt
+          
+          sudo rm -rfv /usr/bin/conda || true
+          PATH=$PATH:/opt/conda/bin make -C extra/grpc/huggingface
 
-          sudo mkdir /build && sudo chmod -R 777 /build && cd /build && \
-          curl -L "https://github.com/gabime/spdlog/archive/refs/tags/v1.11.0.tar.gz" | \
-          tar -xzvf - && \
-          mkdir -p "spdlog-1.11.0/build" && \
-          cd "spdlog-1.11.0/build" && \
-          cmake ..  && \
-          make -j8 && \
-          sudo cmake --install . --prefix /usr && mkdir -p "lib/Linux-$(uname -m)" && \
-          cd /build && \
-          mkdir -p "lib/Linux-$(uname -m)/piper_phonemize" && \
-          curl -L "https://github.com/rhasspy/piper-phonemize/releases/download/v1.0.0/libpiper_phonemize-amd64.tar.gz" | \
-          tar -C "lib/Linux-$(uname -m)/piper_phonemize" -xzvf - && ls -liah /build/lib/Linux-$(uname -m)/piper_phonemize/ && \
-          sudo cp -rfv /build/lib/Linux-$(uname -m)/piper_phonemize/lib/. /usr/lib/ && \
-          sudo ln -s /usr/lib/libpiper_phonemize.so /usr/lib/libpiper_phonemize.so.1 && \
-          sudo cp -rfv /build/lib/Linux-$(uname -m)/piper_phonemize/include/. /usr/include/
+          # 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/ && \
+
+          # 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
+
+          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 -j12 install
       - name: Test
         run: |
-          ESPEAK_DATA="/build/lib/Linux-$(uname -m)/piper_phonemize/lib/espeak-ng-data" GO_TAGS="tts stablediffusion" make test
+          GO_TAGS="stablediffusion tts" make test
 
-  macOS-latest:
+  tests-apple:
     runs-on: macOS-latest
     strategy:
       matrix:
         go-version: ['1.21.x']
     steps:
       - name: Clone
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
         with: 
           submodules: true
       - name: Setup Go ${{ matrix.go-version }}
@@ -108,6 +113,14 @@ jobs:
       # You can test your matrix by printing the current Go version
       - name: Display Go version
         run: go version
+      - name: Dependencies
+        run: |
+          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
       - name: Test
         run: |
+          export C_INCLUDE_PATH=/usr/local/include
+          export CPLUS_INCLUDE_PATH=/usr/local/include
           CMAKE_ARGS="-DLLAMA_F16C=OFF -DLLAMA_AVX512=OFF -DLLAMA_AVX2=OFF -DLLAMA_FMA=OFF" make test

.gitignore

@@ -10,6 +10,8 @@ go-ggllm
 __pycache__/
 *.a
 get-sources
+/backend/cpp/llama/grpc-server
+/backend/cpp/llama/llama.cpp
 
 go-ggml-transformers
 go-gpt2

Dockerfile

@@ -1,22 +1,27 @@
 ARG GO_VERSION=1.21-bullseye
+ARG IMAGE_TYPE=extras
+# extras or core
 
-FROM golang:$GO_VERSION as requirements
+
+FROM golang:$GO_VERSION as requirements-core
 
 ARG BUILD_TYPE
 ARG CUDA_MAJOR_VERSION=11
 ARG CUDA_MINOR_VERSION=7
-ARG SPDLOG_VERSION="1.11.0"
-ARG PIPER_PHONEMIZE_VERSION='1.0.0'
 ARG TARGETARCH
 ARG TARGETVARIANT
 
 ENV BUILD_TYPE=${BUILD_TYPE}
-ENV EXTERNAL_GRPC_BACKENDS="huggingface-embeddings:/build/extra/grpc/huggingface/huggingface.py,autogptq:/build/extra/grpc/autogptq/autogptq.py,bark:/build/extra/grpc/bark/ttsbark.py,diffusers:/build/extra/grpc/diffusers/backend_diffusers.py,exllama:/build/extra/grpc/exllama/exllama.py,vall-e-x:/build/extra/grpc/vall-e-x/ttsvalle.py,vllm:/build/extra/grpc/vllm/backend_vllm.py"
+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 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"
 
 RUN apt-get update && \
-    apt-get install -y ca-certificates cmake curl patch pip
+    apt-get install -y ca-certificates curl patch pip cmake && apt-get clean
+
+
+COPY --chmod=644 custom-ca-certs/* /usr/local/share/ca-certificates/
+RUN update-ca-certificates
 
 # Use the variables in subsequent instructions
 RUN echo "Target Architecture: $TARGETARCH"
@@ -30,66 +35,62 @@ RUN if [ "${BUILD_TYPE}" = "cublas" ]; then \
     dpkg -i cuda-keyring_1.0-1_all.deb && \
     rm -f cuda-keyring_1.0-1_all.deb && \
     apt-get update && \
-    apt-get install -y cuda-nvcc-${CUDA_MAJOR_VERSION}-${CUDA_MINOR_VERSION} libcublas-dev-${CUDA_MAJOR_VERSION}-${CUDA_MINOR_VERSION} libcusparse-dev-${CUDA_MAJOR_VERSION}-${CUDA_MINOR_VERSION} libcusolver-dev-${CUDA_MAJOR_VERSION}-${CUDA_MINOR_VERSION} \
+    apt-get install -y cuda-nvcc-${CUDA_MAJOR_VERSION}-${CUDA_MINOR_VERSION} libcublas-dev-${CUDA_MAJOR_VERSION}-${CUDA_MINOR_VERSION} libcusparse-dev-${CUDA_MAJOR_VERSION}-${CUDA_MINOR_VERSION} libcusolver-dev-${CUDA_MAJOR_VERSION}-${CUDA_MINOR_VERSION}  && apt-get clean \
     ; fi
 ENV PATH /usr/local/cuda/bin:${PATH}
 
+# OpenBLAS requirements and stable diffusion
+RUN apt-get install -y \
+    libopenblas-dev \
+    libopencv-dev \ 
+    && apt-get clean
+
+# Set up OpenCV
+RUN ln -s /usr/include/opencv4/opencv2 /usr/include/opencv2
+
+WORKDIR /build
+
+RUN test -n "$TARGETARCH" \
+    || (echo 'warn: missing $TARGETARCH, either set this `ARG` manually, or run using `docker buildkit`')
+
 # Extras requirements
+FROM requirements-core as requirements-extras
+
+RUN curl https://repo.anaconda.com/pkgs/misc/gpgkeys/anaconda.asc | gpg --dearmor > conda.gpg && \
+    install -o root -g root -m 644 conda.gpg /usr/share/keyrings/conda-archive-keyring.gpg && \
+    gpg --keyring /usr/share/keyrings/conda-archive-keyring.gpg --no-default-keyring --fingerprint 34161F5BF5EB1D4BFBBB8F0A8AEB4F8B29D82806 && \
+    echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" > /etc/apt/sources.list.d/conda.list && \
+    echo "deb [arch=amd64 signed-by=/usr/share/keyrings/conda-archive-keyring.gpg] https://repo.anaconda.com/pkgs/misc/debrepo/conda stable main" | tee -a /etc/apt/sources.list.d/conda.list && \
+    apt-get update && \
+    apt-get install -y conda
+
 COPY extra/requirements.txt /build/extra/requirements.txt
 ENV PATH="/root/.cargo/bin:${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}${CUDA_MINOR_VERSION}-cp39-cp39-linux_x86_64.whl;\
-    fi
-RUN pip install -r /build/extra/requirements.txt && rm -rf /build/extra/requirements.txt
+#RUN if [ "${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
 
-WORKDIR /build
-
-# OpenBLAS requirements
-RUN apt-get install -y libopenblas-dev
-
-# Stable Diffusion requirements
-RUN apt-get install -y libopencv-dev && \
-    ln -s /usr/include/opencv4/opencv2 /usr/include/opencv2
-
-
-# piper requirements
-# Use pre-compiled Piper phonemization library (includes onnxruntime)
-#RUN if echo "${GO_TAGS}" | grep -q "tts"; then \
-RUN test -n "$TARGETARCH" \
-    || (echo 'warn: missing $TARGETARCH, either set this `ARG` manually, or run using `docker buildkit`')
-
-RUN curl -L "https://github.com/gabime/spdlog/archive/refs/tags/v${SPDLOG_VERSION}.tar.gz" | \
-    tar -xzvf - && \
-    mkdir -p "spdlog-${SPDLOG_VERSION}/build" && \
-    cd "spdlog-${SPDLOG_VERSION}/build" && \
-    cmake ..  && \
-    make -j8 && \
-    cmake --install . --prefix /usr && mkdir -p "lib/Linux-$(uname -m)" && \
-    cd /build && \
-    mkdir -p "lib/Linux-$(uname -m)/piper_phonemize" && \
-    curl -L "https://github.com/rhasspy/piper-phonemize/releases/download/v${PIPER_PHONEMIZE_VERSION}/libpiper_phonemize-${TARGETARCH:-$(go env GOARCH)}${TARGETVARIANT}.tar.gz" | \
-    tar -C "lib/Linux-$(uname -m)/piper_phonemize" -xzvf - && ls -liah /build/lib/Linux-$(uname -m)/piper_phonemize/ && \
-    cp -rfv /build/lib/Linux-$(uname -m)/piper_phonemize/lib/. /usr/lib/ && \
-    ln -s /usr/lib/libpiper_phonemize.so /usr/lib/libpiper_phonemize.so.1 && \
-    cp -rfv /build/lib/Linux-$(uname -m)/piper_phonemize/include/. /usr/include/
 # \
 #    ; fi
 
 ###################################
 ###################################
 
-FROM requirements as builder
+FROM requirements-${IMAGE_TYPE} as builder
 
 ARG GO_TAGS="stablediffusion tts"
-
+ARG GRPC_BACKENDS
+ARG BUILD_GRPC=true
+ENV GRPC_BACKENDS=${GRPC_BACKENDS}
 ENV GO_TAGS=${GO_TAGS}
 ENV NVIDIA_DRIVER_CAPABILITIES=compute,utility
 ENV NVIDIA_REQUIRE_CUDA="cuda>=${CUDA_MAJOR_VERSION}.0"
@@ -104,21 +105,43 @@ RUN make prepare
 COPY . .
 COPY .git .
 
-RUN ESPEAK_DATA=/build/lib/Linux-$(uname -m)/piper_phonemize/lib/espeak-ng-data make build
+# stablediffusion does not tolerate a newer version of abseil, build it first
+RUN GRPC_BACKENDS=backend-assets/grpc/stablediffusion make build
+
+RUN if [ "${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 \
+    ; fi
+
+# Rebuild with defaults backends
+RUN make build
+
+RUN if [ ! -d "/build/go-piper/piper/build/pi/lib/" ]; then \
+    mkdir -p /build/go-piper/piper/build/pi/lib/ \
+    touch /build/go-piper/piper/build/pi/lib/keep \
+    ; fi
 
 ###################################
 ###################################
 
-FROM requirements
+FROM requirements-${IMAGE_TYPE}
 
 ARG FFMPEG
 ARG BUILD_TYPE
 ARG TARGETARCH
+ARG IMAGE_TYPE=extras
 
 ENV BUILD_TYPE=${BUILD_TYPE}
 ENV REBUILD=false
 ENV HEALTHCHECK_ENDPOINT=http://localhost:8080/readyz
 
+ARG CUDA_MAJOR_VERSION=11
+ENV NVIDIA_DRIVER_CAPABILITIES=compute,utility
+ENV NVIDIA_REQUIRE_CUDA="cuda>=${CUDA_MAJOR_VERSION}.0"
+ENV NVIDIA_VISIBLE_DEVICES=all
+
 # Add FFmpeg
 RUN if [ "${FFMPEG}" = "true" ]; then \
     apt-get install -y ffmpeg \
@@ -132,15 +155,49 @@ WORKDIR /build
 # https://github.com/go-skynet/LocalAI/pull/434
 COPY . .
 RUN make prepare-sources
+
+# Copy the binary
 COPY --from=builder /build/local-ai ./
 
-# Copy VALLE-X as it's not a real "lib"
-RUN cp -rfv /usr/lib/vall-e-x/* ./
+# Copy shared libraries for piper
+COPY --from=builder /build/go-piper/piper/build/pi/lib/* /usr/lib/
 
-# To resolve exllama import error
-RUN if [ "${BUILD_TYPE}" = "cublas" ] && [ "${TARGETARCH:-$(go env GOARCH)}" = "amd64" ]; then \
+# do not let stablediffusion rebuild (requires an older version of absl)
+COPY --from=builder /build/backend-assets/grpc/stablediffusion ./backend-assets/grpc/stablediffusion
+
+## Duplicated from Makefile to avoid having a big layer that's hard to push
+RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
+	PATH=$PATH:/opt/conda/bin make -C extra/grpc/autogptq \
+    ; fi
+RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
+	PATH=$PATH:/opt/conda/bin make -C extra/grpc/bark \
+    ; fi
+RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
+	PATH=$PATH:/opt/conda/bin make -C extra/grpc/diffusers \
+    ; fi
+RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
+	PATH=$PATH:/opt/conda/bin make -C extra/grpc/vllm \
+    ; fi
+RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
+	PATH=$PATH:/opt/conda/bin make -C extra/grpc/huggingface \
+    ; fi
+RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
+	PATH=$PATH:/opt/conda/bin make -C extra/grpc/vall-e-x \
+    ; fi
+RUN if [ "${IMAGE_TYPE}" = "extras" ]; then \
+	PATH=$PATH:/opt/conda/bin make -C extra/grpc/exllama \
+    ; 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
+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/;\
     fi
+
 # Define the health check command
 HEALTHCHECK --interval=1m --timeout=10m --retries=10 \
   CMD curl -f $HEALTHCHECK_ENDPOINT || exit 1

Makefile

@@ -4,10 +4,12 @@ GOVET=$(GOCMD) vet
 BINARY_NAME=local-ai
 
 # llama.cpp versions
-GOLLAMA_VERSION?=1676dcd7a139b6cdfbaea5fd67f46dc25d9d8bcf
+GOLLAMA_VERSION?=aeba71ee842819da681ea537e78846dc75949ac0
 
 GOLLAMA_STABLE_VERSION?=50cee7712066d9e38306eccadcfbb44ea87df4b7
 
+CPPLLAMA_VERSION?=a75fa576abba9d37f463580c379e4bbf1e1ad03c
+
 # gpt4all version
 GPT4ALL_REPO?=https://github.com/nomic-ai/gpt4all
 GPT4ALL_VERSION?=27a8b020c36b0df8f8b82a252d261cda47cf44b8
@@ -26,17 +28,11 @@ WHISPER_CPP_VERSION?=85ed71aaec8e0612a84c0b67804bde75aa75a273
 BERT_VERSION?=6abe312cded14042f6b7c3cd8edf082713334a4d
 
 # go-piper version
-PIPER_VERSION?=56b8a81b4760a6fbee1a82e62f007ae7e8f010a7
-
-# go-bloomz version
-BLOOMZ_VERSION?=1834e77b83faafe912ad4092ccf7f77937349e2f
+PIPER_VERSION?=736f6fb639ab8e3397356e48eeb6bdcb9da88a78
 
 # stablediffusion version
 STABLEDIFFUSION_VERSION?=d89260f598afb809279bc72aa0107b4292587632
 
-# Go-ggllm
-GOGGLLM_VERSION?=862477d16eefb0805261c19c9b0d053e3b2b684b
-
 export BUILD_TYPE?=
 export STABLE_BUILD_TYPE?=$(BUILD_TYPE)
 export CMAKE_ARGS?=
@@ -45,6 +41,10 @@ CUDA_LIBPATH?=/usr/local/cuda/lib64/
 GO_TAGS?=
 BUILD_ID?=git
 
+TEST_DIR=/tmp/test
+
+RANDOM := $(shell bash -c 'echo $$RANDOM')
+
 VERSION?=$(shell git describe --always --tags || echo "dev" )
 # go tool nm ./local-ai | grep Commit
 LD_FLAGS?=
@@ -52,7 +52,6 @@ override LD_FLAGS += -X "github.com/go-skynet/LocalAI/internal.Version=$(VERSION
 override LD_FLAGS += -X "github.com/go-skynet/LocalAI/internal.Commit=$(shell git rev-parse HEAD)"
 
 OPTIONAL_TARGETS?=
-ESPEAK_DATA?=
 
 OS := $(shell uname -s)
 ARCH := $(shell uname -m)
@@ -62,6 +61,9 @@ WHITE  := $(shell tput -Txterm setaf 7)
 CYAN   := $(shell tput -Txterm setaf 6)
 RESET  := $(shell tput -Txterm sgr0)
 
+# Default Docker bridge IP
+E2E_BRIDGE_IP?=172.17.0.1
+
 ifndef UNAME_S
 UNAME_S := $(shell uname -s)
 endif
@@ -117,10 +119,18 @@ endif
 ifeq ($(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
 	OPTIONAL_GRPC+=backend-assets/grpc/piper
 endif
 
-GRPC_BACKENDS?=backend-assets/grpc/langchain-huggingface backend-assets/grpc/falcon-ggml backend-assets/grpc/bert-embeddings backend-assets/grpc/falcon backend-assets/grpc/bloomz backend-assets/grpc/llama 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-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 $(OPTIONAL_GRPC)
+GRPC_BACKENDS?=$(ALL_GRPC_BACKENDS) $(OPTIONAL_GRPC)
+
+# If empty, then we build all
+ifeq ($(GRPC_BACKENDS),)
+	GRPC_BACKENDS=$(ALL_GRPC_BACKENDS)
+endif
 
 .PHONY: all test build vendor
 
@@ -131,14 +141,6 @@ gpt4all:
 	git clone --recurse-submodules $(GPT4ALL_REPO) gpt4all
 	cd gpt4all && git checkout -b build $(GPT4ALL_VERSION) && git submodule update --init --recursive --depth 1
 
-## go-ggllm
-go-ggllm:
-	git clone --recurse-submodules https://github.com/mudler/go-ggllm.cpp go-ggllm
-	cd go-ggllm && git checkout -b build $(GOGGLLM_VERSION) && git submodule update --init --recursive --depth 1
-
-go-ggllm/libggllm.a: go-ggllm
-	$(MAKE) -C go-ggllm BUILD_TYPE=$(BUILD_TYPE) libggllm.a
-
 ## go-piper
 go-piper:
 	git clone --recurse-submodules https://github.com/mudler/go-piper go-piper
@@ -165,14 +167,6 @@ go-rwkv:
 go-rwkv/librwkv.a: go-rwkv
 	cd go-rwkv && cd rwkv.cpp &&	cmake . -DRWKV_BUILD_SHARED_LIBRARY=OFF &&	cmake --build . && 	cp librwkv.a ..
 
-## bloomz
-bloomz:
-	git clone --recurse-submodules https://github.com/go-skynet/bloomz.cpp bloomz
-	cd bloomz && git checkout -b build $(BLOOMZ_VERSION) && git submodule update --init --recursive --depth 1
-
-bloomz/libbloomz.a: bloomz
-	cd bloomz && make libbloomz.a
-
 go-bert/libgobert.a: go-bert
 	$(MAKE) -C go-bert libgobert.a
 
@@ -182,14 +176,10 @@ backend-assets/gpt4all: gpt4all/gpt4all-bindings/golang/libgpt4all.a
 	@cp gpt4all/gpt4all-bindings/golang/buildllm/*.dylib backend-assets/gpt4all/ || true
 	@cp gpt4all/gpt4all-bindings/golang/buildllm/*.dll backend-assets/gpt4all/ || true
 
-backend-assets/espeak-ng-data:
+backend-assets/espeak-ng-data: go-piper
 	mkdir -p backend-assets/espeak-ng-data
-ifdef ESPEAK_DATA
-	@cp -rf $(ESPEAK_DATA)/. backend-assets/espeak-ng-data
-else
-	@echo "ESPEAK_DATA not set, skipping tts. Note that this will break the tts functionality."
-	@touch backend-assets/espeak-ng-data/keep
-endif
+	$(MAKE) -C go-piper piper.o
+	@cp -rf 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
@@ -223,10 +213,10 @@ go-llama/libbinding.a: go-llama
 go-llama-stable/libbinding.a: go-llama-stable
 	$(MAKE) -C go-llama-stable BUILD_TYPE=$(STABLE_BUILD_TYPE) libbinding.a
 
-go-piper/libpiper_binding.a:
+go-piper/libpiper_binding.a: go-piper
 	$(MAKE) -C go-piper libpiper_binding.a example/main
 
-get-sources: go-llama go-llama-stable go-ggllm go-ggml-transformers gpt4all go-piper go-rwkv whisper.cpp go-bert bloomz go-stable-diffusion
+get-sources: go-llama go-llama-stable go-ggml-transformers gpt4all go-piper go-rwkv whisper.cpp go-bert go-stable-diffusion
 	touch $@
 
 replace:
@@ -235,10 +225,8 @@ replace:
 	$(GOCMD) mod edit -replace github.com/donomii/go-rwkv.cpp=$(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
-	$(GOCMD) mod edit -replace github.com/go-skynet/bloomz.cpp=$(shell pwd)/bloomz
 	$(GOCMD) mod edit -replace github.com/mudler/go-stable-diffusion=$(shell pwd)/go-stable-diffusion
 	$(GOCMD) mod edit -replace github.com/mudler/go-piper=$(shell pwd)/go-piper
-	$(GOCMD) mod edit -replace github.com/mudler/go-ggllm.cpp=$(shell pwd)/go-ggllm
 
 prepare-sources: get-sources replace
 	$(GOCMD) mod download
@@ -254,9 +242,7 @@ rebuild: ## Rebuilds the project
 	$(MAKE) -C whisper.cpp clean
 	$(MAKE) -C go-stable-diffusion clean
 	$(MAKE) -C go-bert clean
-	$(MAKE) -C bloomz clean
 	$(MAKE) -C go-piper clean
-	$(MAKE) -C go-ggllm clean
 	$(MAKE) build
 
 prepare: prepare-sources $(OPTIONAL_TARGETS)
@@ -274,12 +260,14 @@ clean: ## Remove build related file
 	rm -rf ./backend-assets
 	rm -rf ./go-rwkv
 	rm -rf ./go-bert
-	rm -rf ./bloomz
 	rm -rf ./whisper.cpp
 	rm -rf ./go-piper
-	rm -rf ./go-ggllm
 	rm -rf $(BINARY_NAME)
 	rm -rf release/
+	rm -rf ./backend/cpp/grpc/grpc_repo
+	rm -rf ./backend/cpp/grpc/build
+	rm -rf ./backend/cpp/grpc/installed_packages
+	$(MAKE) -C backend/cpp/llama clean
 
 ## Build:
 
@@ -302,12 +290,12 @@ run: prepare ## run local-ai
 test-models/testmodel:
 	mkdir test-models
 	mkdir test-dir
-	wget https://huggingface.co/nnakasato/ggml-model-test/resolve/main/ggml-model-q4.bin -O test-models/testmodel
-	wget https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-base.en.bin -O test-models/whisper-en
-	wget https://huggingface.co/skeskinen/ggml/resolve/main/all-MiniLM-L6-v2/ggml-model-q4_0.bin -O test-models/bert
-	wget https://cdn.openai.com/whisper/draft-20220913a/micro-machines.wav -O test-dir/audio.wav
-	wget https://huggingface.co/mudler/rwkv-4-raven-1.5B-ggml/resolve/main/RWKV-4-Raven-1B5-v11-Eng99%2525-Other1%2525-20230425-ctx4096_Q4_0.bin -O test-models/rwkv
-	wget https://raw.githubusercontent.com/saharNooby/rwkv.cpp/5eb8f09c146ea8124633ab041d9ea0b1f1db4459/rwkv/20B_tokenizer.json -O test-models/rwkv.tokenizer.json
+	wget -q https://huggingface.co/nnakasato/ggml-model-test/resolve/main/ggml-model-q4.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
+	wget -q https://huggingface.co/mudler/rwkv-4-raven-1.5B-ggml/resolve/main/RWKV-4-Raven-1B5-v11-Eng99%2525-Other1%2525-20230425-ctx4096_Q4_0.bin -O test-models/rwkv
+	wget -q https://raw.githubusercontent.com/saharNooby/rwkv.cpp/5eb8f09c146ea8124633ab041d9ea0b1f1db4459/rwkv/20B_tokenizer.json -O test-models/rwkv.tokenizer.json
 	cp tests/models_fixtures/* test-models
 
 prepare-test: grpcs
@@ -318,14 +306,34 @@ test: prepare test-models/testmodel grpcs
 	@echo 'Running tests'
 	export GO_TAGS="tts stablediffusion"
 	$(MAKE) prepare-test
-	HUGGINGFACE_GRPC=$(abspath ./)/extra/grpc/huggingface/huggingface.py TEST_DIR=$(abspath ./)/test-dir/ FIXTURES=$(abspath ./)/tests/fixtures CONFIG_FILE=$(abspath ./)/test-models/config.yaml MODELS_PATH=$(abspath ./)/test-models \
-	$(GOCMD) run github.com/onsi/ginkgo/v2/ginkgo --label-filter="!gpt4all && !llama && !llama-gguf" --flake-attempts 5 -v -r ./api ./pkg
+	HUGGINGFACE_GRPC=$(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 \
+	$(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
 	$(MAKE) test-llama-gguf
 	$(MAKE) test-tts
 	$(MAKE) test-stablediffusion
 
+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 $(TEST_DIR)/ggllm-test-model.bin
+	docker build --build-arg BUILD_GRPC=true --build-arg GRPC_BACKENDS="$(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
+
+test-e2e:
+	@echo 'Running e2e tests'
+	BUILD_TYPE=$(BUILD_TYPE) \
+	LOCALAI_API=http://$(E2E_BRIDGE_IP):5390/v1 \
+	$(GOCMD) run github.com/onsi/ginkgo/v2/ginkgo --flake-attempts 5 -v -r ./tests/e2e
+
+teardown-e2e:
+	rm -rf $(TEST_DIR) || true
+	docker stop $$(docker ps -q --filter ancestor=localai-tests)
+
 test-gpt4all: 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="gpt4all" --flake-attempts 5 -v -r ./api ./pkg
@@ -378,14 +386,20 @@ protogen-python:
 	python3 -m grpc_tools.protoc -Ipkg/grpc/proto/ --python_out=extra/grpc/vllm/ --grpc_python_out=extra/grpc/vllm/ pkg/grpc/proto/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
+
 
 backend-assets/grpc:
 	mkdir -p backend-assets/grpc
 
-backend-assets/grpc/falcon: backend-assets/grpc go-ggllm/libggllm.a
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/go-ggllm LIBRARY_PATH=$(shell pwd)/go-ggllm \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/falcon ./cmd/grpc/falcon/
-
 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=$(shell pwd)/go-llama \
@@ -395,6 +409,37 @@ ifeq ($(BUILD_TYPE),metal)
 	cp go-llama/build/bin/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
+
+backend/cpp/llama/grpc-server:
+ifdef BUILD_GRPC_FOR_BACKEND_LLAMA
+	backend/cpp/grpc/script/build_grpc.sh ${INSTALLED_PACKAGES}
+	export _PROTOBUF_PROTOC=${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) $(MAKE) -C backend/cpp/llama grpc-server 
+else
+	echo "BUILD_GRPC_FOR_BACKEND_LLAMA is not defined."
+	LLAMA_VERSION=$(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/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=$(shell pwd)/go-llama \
@@ -440,10 +485,6 @@ backend-assets/grpc/rwkv: backend-assets/grpc go-rwkv/librwkv.a
 	CGO_LDFLAGS="$(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/bloomz: backend-assets/grpc bloomz/libbloomz.a
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(shell pwd)/bloomz LIBRARY_PATH=$(shell pwd)/bloomz \
-	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/bloomz ./cmd/grpc/bloomz/
-
 backend-assets/grpc/bert-embeddings: backend-assets/grpc go-bert/libgobert.a
 	CGO_LDFLAGS="$(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/
@@ -451,12 +492,15 @@ backend-assets/grpc/bert-embeddings: backend-assets/grpc go-bert/libgobert.a
 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/
 
-backend-assets/grpc/stablediffusion: backend-assets/grpc 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/
+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/; \
+	fi
 
 backend-assets/grpc/piper: backend-assets/grpc backend-assets/espeak-ng-data go-piper/libpiper_binding.a
-	CGO_LDFLAGS="$(CGO_LDFLAGS)" LIBRARY_PATH=$(shell pwd)/go-piper \
+	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/whisper: backend-assets/grpc whisper.cpp/libwhisper.a

api/api.go

@@ -11,6 +11,7 @@ import (
 	"github.com/go-skynet/LocalAI/api/options"
 	"github.com/go-skynet/LocalAI/api/schema"
 	"github.com/go-skynet/LocalAI/internal"
+	"github.com/go-skynet/LocalAI/metrics"
 	"github.com/go-skynet/LocalAI/pkg/assets"
 
 	"github.com/gofiber/fiber/v2"
@@ -120,6 +121,9 @@ func App(opts ...options.AppOption) (*fiber.App, error) {
 
 	// Default middleware config
 	app.Use(recover.New())
+	if options.Metrics != nil {
+		app.Use(metrics.APIMiddleware(options.Metrics))
+	}
 
 	// Auth middleware checking if API key is valid. If no API key is set, no auth is required.
 	auth := func(c *fiber.Ctx) error {
@@ -229,5 +233,7 @@ func App(opts ...options.AppOption) (*fiber.App, error) {
 	app.Get("/v1/models", auth, openai.ListModelsEndpoint(options.Loader, cl))
 	app.Get("/models", auth, openai.ListModelsEndpoint(options.Loader, cl))
 
+	app.Get("/metrics", metrics.MetricsHandler())
+
 	return app, nil
 }

api/api_test.go

@@ -15,6 +15,7 @@ import (
 
 	. "github.com/go-skynet/LocalAI/api"
 	"github.com/go-skynet/LocalAI/api/options"
+	"github.com/go-skynet/LocalAI/metrics"
 	"github.com/go-skynet/LocalAI/pkg/gallery"
 	"github.com/go-skynet/LocalAI/pkg/model"
 	"github.com/go-skynet/LocalAI/pkg/utils"
@@ -162,8 +163,12 @@ var _ = Describe("API test", func() {
 				},
 			}
 
+			metricsService, err := metrics.SetupMetrics()
+			Expect(err).ToNot(HaveOccurred())
+
 			app, err = App(
 				append(commonOpts,
+					options.WithMetrics(metricsService),
 					options.WithContext(c),
 					options.WithGalleries(galleries),
 					options.WithModelLoader(modelLoader), options.WithBackendAssets(backendAssets), options.WithBackendAssetsOutput(tmpdir))...)
@@ -452,7 +457,7 @@ var _ = Describe("API test", func() {
 				Eventually(func() bool {
 					response := getModelStatus("http://127.0.0.1:9090/models/jobs/" + uuid)
 					return response["processed"].(bool)
-				}, "360s", "10s").Should(Equal(true))
+				}, "960s", "10s").Should(Equal(true))
 
 				resp, err := client.CreateChatCompletion(context.TODO(), openai.ChatCompletionRequest{Model: "gpt4all-j", Messages: []openai.ChatCompletionMessage{openai.ChatCompletionMessage{Role: "user", Content: "How are you?"}}})
 				Expect(err).ToNot(HaveOccurred())
@@ -479,9 +484,13 @@ var _ = Describe("API test", func() {
 				},
 			}
 
+			metricsService, err := metrics.SetupMetrics()
+			Expect(err).ToNot(HaveOccurred())
+
 			app, err = App(
 				append(commonOpts,
 					options.WithContext(c),
+					options.WithMetrics(metricsService),
 					options.WithAudioDir(tmpdir),
 					options.WithImageDir(tmpdir),
 					options.WithGalleries(galleries),
@@ -583,12 +592,15 @@ var _ = Describe("API test", func() {
 			modelLoader = model.NewModelLoader(os.Getenv("MODELS_PATH"))
 			c, cancel = context.WithCancel(context.Background())
 
-			var err error
+			metricsService, err := metrics.SetupMetrics()
+			Expect(err).ToNot(HaveOccurred())
+
 			app, err = App(
 				append(commonOpts,
 					options.WithExternalBackend("huggingface", os.Getenv("HUGGINGFACE_GRPC")),
 					options.WithContext(c),
 					options.WithModelLoader(modelLoader),
+					options.WithMetrics(metricsService),
 				)...)
 			Expect(err).ToNot(HaveOccurred())
 			go app.Listen("127.0.0.1:9090")
@@ -675,7 +687,7 @@ var _ = Describe("API test", func() {
 					Input: []string{"sun", "cat"},
 				},
 			)
-			Expect(err).ToNot(HaveOccurred())
+			Expect(err).ToNot(HaveOccurred(), err)
 			Expect(len(resp.Data[0].Embedding)).To(BeNumerically("==", 384))
 			Expect(len(resp.Data[1].Embedding)).To(BeNumerically("==", 384))
 
@@ -792,10 +804,13 @@ var _ = Describe("API test", func() {
 			modelLoader = model.NewModelLoader(os.Getenv("MODELS_PATH"))
 			c, cancel = context.WithCancel(context.Background())
 
-			var err error
+			metricsService, err := metrics.SetupMetrics()
+			Expect(err).ToNot(HaveOccurred())
+
 			app, err = App(
 				append(commonOpts,
 					options.WithContext(c),
+					options.WithMetrics(metricsService),
 					options.WithModelLoader(modelLoader),
 					options.WithConfigFile(os.Getenv("CONFIG_FILE")))...,
 			)

api/backend/image.go

@@ -21,6 +21,7 @@ func ImageGeneration(height, width, mode, step, seed int, positive_prompt, negat
 			PipelineType:  c.Diffusers.PipelineType,
 			CFGScale:      c.Diffusers.CFGScale,
 			LoraAdapter:   c.LoraAdapter,
+			LoraScale:     c.LoraScale,
 			LoraBase:      c.LoraBase,
 			IMG2IMG:       c.Diffusers.IMG2IMG,
 			CLIPModel:     c.Diffusers.ClipModel,

api/backend/llm.go

@@ -26,7 +26,7 @@ type TokenUsage struct {
 	Completion int
 }
 
-func ModelInference(ctx context.Context, s string, loader *model.ModelLoader, c config.Config, o *options.Option, tokenCallback func(string, TokenUsage) bool) (func() (LLMResponse, error), error) {
+func ModelInference(ctx context.Context, s string, images []string, loader *model.ModelLoader, c config.Config, o *options.Option, tokenCallback func(string, TokenUsage) bool) (func() (LLMResponse, error), error) {
 	modelFile := c.Model
 
 	grpcOpts := gRPCModelOpts(c)
@@ -72,6 +72,7 @@ func ModelInference(ctx context.Context, s string, loader *model.ModelLoader, c
 	fn := func() (LLMResponse, error) {
 		opts := gRPCPredictOpts(c, loader.ModelPath)
 		opts.Prompt = s
+		opts.Images = images
 
 		tokenUsage := TokenUsage{}
 

api/backend/options.go

@@ -38,29 +38,35 @@ func gRPCModelOpts(c config.Config) *pb.ModelOptions {
 	}
 
 	return &pb.ModelOptions{
-		ContextSize:   int32(c.ContextSize),
-		Seed:          int32(c.Seed),
-		NBatch:        int32(b),
-		NoMulMatQ:     c.NoMulMatQ,
-		DraftModel:    c.DraftModel,
-		AudioPath:     c.VallE.AudioPath,
-		Quantization:  c.Quantization,
-		LoraAdapter:   c.LoraAdapter,
-		LoraBase:      c.LoraBase,
-		NGQA:          c.NGQA,
-		RMSNormEps:    c.RMSNormEps,
-		F16Memory:     c.F16,
-		MLock:         c.MMlock,
-		RopeFreqBase:  c.RopeFreqBase,
-		RopeFreqScale: c.RopeFreqScale,
-		NUMA:          c.NUMA,
-		Embeddings:    c.Embeddings,
-		LowVRAM:       c.LowVRAM,
-		NGPULayers:    int32(c.NGPULayers),
-		MMap:          c.MMap,
-		MainGPU:       c.MainGPU,
-		Threads:       int32(c.Threads),
-		TensorSplit:   c.TensorSplit,
+		ContextSize:    int32(c.ContextSize),
+		Seed:           int32(c.Seed),
+		NBatch:         int32(b),
+		NoMulMatQ:      c.NoMulMatQ,
+		DraftModel:     c.DraftModel,
+		AudioPath:      c.VallE.AudioPath,
+		Quantization:   c.Quantization,
+		MMProj:         c.MMProj,
+		YarnExtFactor:  c.YarnExtFactor,
+		YarnAttnFactor: c.YarnAttnFactor,
+		YarnBetaFast:   c.YarnBetaFast,
+		YarnBetaSlow:   c.YarnBetaSlow,
+		LoraAdapter:    c.LoraAdapter,
+		LoraBase:       c.LoraBase,
+		LoraScale:      c.LoraScale,
+		NGQA:           c.NGQA,
+		RMSNormEps:     c.RMSNormEps,
+		F16Memory:      c.F16,
+		MLock:          c.MMlock,
+		RopeFreqBase:   c.RopeFreqBase,
+		RopeFreqScale:  c.RopeFreqScale,
+		NUMA:           c.NUMA,
+		Embeddings:     c.Embeddings,
+		LowVRAM:        c.LowVRAM,
+		NGPULayers:     int32(c.NGPULayers),
+		MMap:           c.MMap,
+		MainGPU:        c.MainGPU,
+		Threads:        int32(c.Threads),
+		TensorSplit:    c.TensorSplit,
 		// AutoGPTQ
 		ModelBaseName:    c.AutoGPTQ.ModelBaseName,
 		Device:           c.AutoGPTQ.Device,

api/config/config.go

@@ -100,10 +100,18 @@ type LLMConfig struct {
 	NUMA            bool     `yaml:"numa"`
 	LoraAdapter     string   `yaml:"lora_adapter"`
 	LoraBase        string   `yaml:"lora_base"`
+	LoraScale       float32  `yaml:"lora_scale"`
 	NoMulMatQ       bool     `yaml:"no_mulmatq"`
 	DraftModel      string   `yaml:"draft_model"`
 	NDraft          int32    `yaml:"n_draft"`
 	Quantization    string   `yaml:"quantization"`
+	MMProj          string   `yaml:"mmproj"`
+
+	RopeScaling    string  `yaml:"rope_scaling"`
+	YarnExtFactor  float32 `yaml:"yarn_ext_factor"`
+	YarnAttnFactor float32 `yaml:"yarn_attn_factor"`
+	YarnBetaFast   float32 `yaml:"yarn_beta_fast"`
+	YarnBetaSlow   float32 `yaml:"yarn_beta_slow"`
 }
 
 type AutoGPTQ struct {

api/openai/chat.go

@@ -81,6 +81,10 @@ func ChatEndpoint(cm *config.ConfigLoader, o *options.Option) func(c *fiber.Ctx)
 			noActionDescription = config.FunctionsConfig.NoActionDescriptionName
 		}
 
+		if input.ResponseFormat == "json_object" {
+			input.Grammar = grammar.JSONBNF
+		}
+
 		// process functions if we have any defined or if we have a function call string
 		if len(input.Functions) > 0 && config.ShouldUseFunctions() {
 			log.Debug().Msgf("Response needs to process functions")
@@ -140,14 +144,14 @@ func ChatEndpoint(cm *config.ConfigLoader, o *options.Option) func(c *fiber.Ctx)
 				}
 			}
 			r := config.Roles[role]
-			contentExists := i.Content != nil && *i.Content != ""
+			contentExists := i.Content != nil && i.StringContent != ""
 			// First attempt to populate content via a chat message specific template
 			if config.TemplateConfig.ChatMessage != "" {
 				chatMessageData := model.ChatMessageTemplateData{
 					SystemPrompt: config.SystemPrompt,
 					Role:         r,
 					RoleName:     role,
-					Content:      *i.Content,
+					Content:      i.StringContent,
 					MessageIndex: messageIndex,
 				}
 				templatedChatMessage, err := o.Loader.EvaluateTemplateForChatMessage(config.TemplateConfig.ChatMessage, chatMessageData)
@@ -166,7 +170,7 @@ func ChatEndpoint(cm *config.ConfigLoader, o *options.Option) func(c *fiber.Ctx)
 			if content == "" {
 				if r != "" {
 					if contentExists {
-						content = fmt.Sprint(r, " ", *i.Content)
+						content = fmt.Sprint(r, i.StringContent)
 					}
 					if i.FunctionCall != nil {
 						j, err := json.Marshal(i.FunctionCall)
@@ -180,7 +184,7 @@ func ChatEndpoint(cm *config.ConfigLoader, o *options.Option) func(c *fiber.Ctx)
 					}
 				} else {
 					if contentExists {
-						content = fmt.Sprint(*i.Content)
+						content = fmt.Sprint(i.StringContent)
 					}
 					if i.FunctionCall != nil {
 						j, err := json.Marshal(i.FunctionCall)
@@ -334,7 +338,11 @@ func ChatEndpoint(cm *config.ConfigLoader, o *options.Option) func(c *fiber.Ctx)
 					// Otherwise ask the LLM to understand the JSON output and the context, and return a message
 					// Note: This costs (in term of CPU) another computation
 					config.Grammar = ""
-					predFunc, err := backend.ModelInference(input.Context, predInput, o.Loader, *config, o, nil)
+					images := []string{}
+					for _, m := range input.Messages {
+						images = append(images, m.StringImages...)
+					}
+					predFunc, err := backend.ModelInference(input.Context, predInput, images, o.Loader, *config, o, nil)
 					if err != nil {
 						log.Error().Msgf("inference error: %s", err.Error())
 						return

api/openai/completion.go

@@ -12,6 +12,7 @@ import (
 	config "github.com/go-skynet/LocalAI/api/config"
 	"github.com/go-skynet/LocalAI/api/options"
 	"github.com/go-skynet/LocalAI/api/schema"
+	"github.com/go-skynet/LocalAI/pkg/grammar"
 	model "github.com/go-skynet/LocalAI/pkg/model"
 	"github.com/gofiber/fiber/v2"
 	"github.com/google/uuid"
@@ -64,6 +65,10 @@ 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" {
+			input.Grammar = grammar.JSONBNF
+		}
+
 		log.Debug().Msgf("Parameter Config: %+v", config)
 
 		if input.Stream {

api/openai/inference.go

@@ -23,8 +23,13 @@ func ComputeChoices(
 		n = 1
 	}
 
+	images := []string{}
+	for _, m := range req.Messages {
+		images = append(images, m.StringImages...)
+	}
+
 	// get the model function to call for the result
-	predFunc, err := backend.ModelInference(req.Context, predInput, loader, *config, o, tokenCallback)
+	predFunc, err := backend.ModelInference(req.Context, predInput, images, loader, *config, o, tokenCallback)
 	if err != nil {
 		return result, backend.TokenUsage{}, err
 	}

api/openai/request.go

@@ -2,8 +2,11 @@ package openai
 
 import (
 	"context"
+	"encoding/base64"
 	"encoding/json"
 	"fmt"
+	"io/ioutil"
+	"net/http"
 	"os"
 	"path/filepath"
 	"strings"
@@ -24,7 +27,7 @@ func readInput(c *fiber.Ctx, o *options.Option, randomModel bool) (string, *sche
 	input.Cancel = cancel
 	// Get input data from the request body
 	if err := c.BodyParser(input); err != nil {
-		return "", nil, err
+		return "", nil, fmt.Errorf("failed parsing request body: %w", err)
 	}
 
 	modelFile := input.Model
@@ -61,6 +64,37 @@ func readInput(c *fiber.Ctx, o *options.Option, randomModel bool) (string, *sche
 	return modelFile, input, nil
 }
 
+// this function check if the string is an URL, if it's an URL downloads the image in memory
+// encodes it in base64 and returns the base64 string
+func getBase64Image(s string) (string, error) {
+	if strings.HasPrefix(s, "http") {
+		// download the image
+		resp, err := http.Get(s)
+		if err != nil {
+			return "", err
+		}
+		defer resp.Body.Close()
+
+		// read the image data into memory
+		data, err := ioutil.ReadAll(resp.Body)
+		if err != nil {
+			return "", err
+		}
+
+		// encode the image data in base64
+		encoded := base64.StdEncoding.EncodeToString(data)
+
+		// return the base64 string
+		return encoded, nil
+	}
+
+	// if the string instead is prefixed with "data:image/jpeg;base64,", drop it
+	if strings.HasPrefix(s, "data:image/jpeg;base64,") {
+		return strings.ReplaceAll(s, "data:image/jpeg;base64,", ""), nil
+	}
+	return "", fmt.Errorf("not valid string")
+}
+
 func updateConfig(config *config.Config, input *schema.OpenAIRequest) {
 	if input.Echo {
 		config.Echo = input.Echo
@@ -129,6 +163,35 @@ func updateConfig(config *config.Config, input *schema.OpenAIRequest) {
 		}
 	}
 
+	// Decode each request's message content
+	index := 0
+	for i, m := range input.Messages {
+		switch content := m.Content.(type) {
+		case string:
+			input.Messages[i].StringContent = content
+		case []interface{}:
+			dat, _ := json.Marshal(content)
+			c := []schema.Content{}
+			json.Unmarshal(dat, &c)
+			for _, pp := range c {
+				if pp.Type == "text" {
+					input.Messages[i].StringContent = pp.Text
+				} else if pp.Type == "image_url" {
+					// Detect if pp.ImageURL is an URL, if it is download the image and encode it in base64:
+					base64, err := getBase64Image(pp.ImageURL.URL)
+					if err == nil {
+						input.Messages[i].StringImages = append(input.Messages[i].StringImages, base64) // TODO: make sure that we only return base64 stuff
+						// set a placeholder for each image
+						input.Messages[i].StringContent = fmt.Sprintf("[img-%d]", index) + input.Messages[i].StringContent
+						index++
+					} else {
+						fmt.Print("Failed encoding image", err)
+					}
+				}
+			}
+		}
+	}
+
 	if input.RepeatPenalty != 0 {
 		config.RepeatPenalty = input.RepeatPenalty
 	}

api/options/options.go

@@ -7,6 +7,7 @@ import (
 
 	"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"
 )
 
@@ -24,6 +25,7 @@ type Option struct {
 	PreloadModelsFromPath               string
 	CORSAllowOrigins                    string
 	ApiKeys                             []string
+	Metrics                             *metrics.Metrics
 
 	Galleries []gallery.Gallery
 
@@ -198,3 +200,9 @@ func WithApiKeys(apiKeys []string) AppOption {
 		o.ApiKeys = apiKeys
 	}
 }
+
+func WithMetrics(meter *metrics.Metrics) AppOption {
+	return func(o *Option) {
+		o.Metrics = meter
+	}
+}

api/schema/openai.go

@@ -55,11 +55,25 @@ type Choice struct {
 	Text         string   `json:"text,omitempty"`
 }
 
+type Content struct {
+	Type     string     `json:"type" yaml:"type"`
+	Text     string     `json:"text" yaml:"text"`
+	ImageURL ContentURL `json:"image_url" yaml:"image_url"`
+}
+
+type ContentURL struct {
+	URL string `json:"url" yaml:"url"`
+}
+
 type Message struct {
 	// The message role
 	Role string `json:"role,omitempty" yaml:"role"`
 	// The message content
-	Content *string `json:"content" yaml:"content"`
+	Content interface{} `json:"content" yaml:"content"`
+
+	StringContent string   `json:"string_content,omitempty" yaml:"string_content,omitempty"`
+	StringImages  []string `json:"string_images,omitempty" yaml:"string_images,omitempty"`
+
 	// A result of a function call
 	FunctionCall interface{} `json:"function_call,omitempty" yaml:"function_call,omitempty"`
 }

backend/cpp/grpc/.gitignore

@@ -0,0 +1,3 @@
+installed_packages/
+grpc_build/
+grpc_repo/

backend/cpp/grpc/script/build_grpc.sh

@@ -0,0 +1,81 @@
+#!/bin/bash
+
+# Builds locally from sources the packages needed by the llama cpp backend.
+
+# Makes sure a few base packages exist.
+# sudo apt-get --no-upgrade -y install g++ gcc binutils cmake git build-essential autoconf libtool pkg-config 
+
+SCRIPT_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" &> /dev/null && pwd )"
+echo "Script directory: $SCRIPT_DIR"
+
+CPP_INSTALLED_PACKAGES_DIR=$1
+if [ -z ${CPP_INSTALLED_PACKAGES_DIR} ]; then 
+    echo "CPP_INSTALLED_PACKAGES_DIR env variable not set. Don't know where to install: failed."; 
+    echo
+    exit -1
+fi
+
+if [ -d "${CPP_INSTALLED_PACKAGES_DIR}" ]; then
+  echo "gRPC installation directory already exists. Nothing to do."
+  exit 0
+fi
+
+# The depth when cloning a git repo. 1 speeds up the clone when the repo history is not needed.
+GIT_CLONE_DEPTH=1
+
+NUM_BUILD_THREADS=$(nproc --ignore=1)
+
+# Google gRPC --------------------------------------------------------------------------------------
+TAG_LIB_GRPC="v1.59.0"
+GIT_REPO_LIB_GRPC="https://github.com/grpc/grpc.git"
+GRPC_REPO_DIR="${SCRIPT_DIR}/../grpc_repo"
+GRPC_BUILD_DIR="${SCRIPT_DIR}/../grpc_build"
+SRC_DIR_LIB_GRPC="${GRPC_REPO_DIR}/grpc"
+
+echo "SRC_DIR_LIB_GRPC: ${SRC_DIR_LIB_GRPC}"
+echo "GRPC_REPO_DIR: ${GRPC_REPO_DIR}"
+echo "GRPC_BUILD_DIR: ${GRPC_BUILD_DIR}"
+
+mkdir -pv ${GRPC_REPO_DIR}
+
+rm   -rf ${GRPC_BUILD_DIR}
+mkdir -pv ${GRPC_BUILD_DIR}
+
+mkdir -pv ${CPP_INSTALLED_PACKAGES_DIR}
+	
+if [ -d "${SRC_DIR_LIB_GRPC}" ]; then
+  echo "gRPC source already exists locally. Not cloned again."
+else  
+  ( cd ${GRPC_REPO_DIR} && \
+    git clone --depth ${GIT_CLONE_DEPTH} -b ${TAG_LIB_GRPC} ${GIT_REPO_LIB_GRPC} && \
+    cd ${SRC_DIR_LIB_GRPC} && \
+    git submodule update --init --recursive --depth ${GIT_CLONE_DEPTH} 
+  )    
+fi
+
+( cd ${GRPC_BUILD_DIR} && \
+  cmake -G "Unix Makefiles" \
+     -DCMAKE_BUILD_TYPE=Release \
+     -DgRPC_INSTALL=ON \
+     -DEXECUTABLE_OUTPUT_PATH=${CPP_INSTALLED_PACKAGES_DIR}/grpc/bin \
+     -DLIBRARY_OUTPUT_PATH=${CPP_INSTALLED_PACKAGES_DIR}/grpc/lib \
+     -DgRPC_BUILD_TESTS=OFF \
+     -DgRPC_BUILD_CSHARP_EXT=OFF \
+     -DgRPC_BUILD_GRPC_CPP_PLUGIN=ON \
+     -DgRPC_BUILD_GRPC_CSHARP_PLUGIN=OFF \
+     -DgRPC_BUILD_GRPC_NODE_PLUGIN=OFF \
+     -DgRPC_BUILD_GRPC_OBJECTIVE_C_PLUGIN=OFF \
+     -DgRPC_BUILD_GRPC_PHP_PLUGIN=OFF \
+     -DgRPC_BUILD_GRPC_PYTHON_PLUGIN=ON \
+     -DgRPC_BUILD_GRPC_RUBY_PLUGIN=OFF \
+     -Dprotobuf_WITH_ZLIB=ON \
+     -DRE2_BUILD_TESTING=OFF \
+     -DCMAKE_INSTALL_PREFIX=${CPP_INSTALLED_PACKAGES_DIR}/ \
+     ${SRC_DIR_LIB_GRPC}  && \
+  cmake --build .  -- -j ${NUM_BUILD_THREADS} && \
+  cmake --build .  --target install -- -j ${NUM_BUILD_THREADS} 
+)
+
+rm -rf ${GRPC_BUILD_DIR}
+rm -rf ${GRPC_REPO_DIR}
+

backend/cpp/llama/CMakeLists.txt

@@ -0,0 +1,74 @@
+
+## XXX: In some versions of CMake clip wasn't being built before llama.
+## This is an hack for now, but it should be fixed in the future.
+set(TARGET myclip)
+add_library(${TARGET} clip.cpp clip.h)
+install(TARGETS ${TARGET} LIBRARY)
+target_link_libraries(${TARGET} PRIVATE common ggml ${CMAKE_THREAD_LIBS_INIT})
+target_compile_features(${TARGET} PRIVATE cxx_std_11)
+if (NOT MSVC)
+    target_compile_options(${TARGET} PRIVATE -Wno-cast-qual) # stb_image.h
+endif()
+
+set(TARGET grpc-server)
+# END CLIP hack
+set(CMAKE_CXX_STANDARD 17)
+cmake_minimum_required(VERSION 3.15)
+set(TARGET grpc-server)
+set(_PROTOBUF_LIBPROTOBUF libprotobuf)
+set(_REFLECTION grpc++_reflection)
+if (${CMAKE_SYSTEM_NAME} MATCHES "Darwin")
+    link_directories("/opt/homebrew/lib")
+    include_directories("/opt/homebrew/include")
+endif()
+
+find_package(absl CONFIG REQUIRED)
+find_package(Protobuf CONFIG REQUIRED)
+find_package(gRPC CONFIG REQUIRED)
+
+find_program(_PROTOBUF_PROTOC protoc)
+set(_GRPC_GRPCPP grpc++)
+find_program(_GRPC_CPP_PLUGIN_EXECUTABLE grpc_cpp_plugin)
+
+include_directories(${CMAKE_CURRENT_BINARY_DIR})
+include_directories(${Protobuf_INCLUDE_DIRS})
+
+message(STATUS "Using protobuf version ${Protobuf_VERSION} | Protobuf_INCLUDE_DIRS: ${Protobuf_INCLUDE_DIRS} | CMAKE_CURRENT_BINARY_DIR: ${CMAKE_CURRENT_BINARY_DIR}")
+
+# Proto file
+get_filename_component(hw_proto "../../../../../../pkg/grpc/proto/backend.proto" ABSOLUTE)
+get_filename_component(hw_proto_path "${hw_proto}" PATH)
+
+# Generated sources
+set(hw_proto_srcs "${CMAKE_CURRENT_BINARY_DIR}/backend.pb.cc")
+set(hw_proto_hdrs "${CMAKE_CURRENT_BINARY_DIR}/backend.pb.h")
+set(hw_grpc_srcs "${CMAKE_CURRENT_BINARY_DIR}/backend.grpc.pb.cc")
+set(hw_grpc_hdrs "${CMAKE_CURRENT_BINARY_DIR}/backend.grpc.pb.h")
+
+add_custom_command(
+      OUTPUT "${hw_proto_srcs}" "${hw_proto_hdrs}" "${hw_grpc_srcs}" "${hw_grpc_hdrs}"
+      COMMAND ${_PROTOBUF_PROTOC}
+      ARGS --grpc_out "${CMAKE_CURRENT_BINARY_DIR}"
+        --cpp_out "${CMAKE_CURRENT_BINARY_DIR}"
+        -I "${hw_proto_path}"
+        --plugin=protoc-gen-grpc="${_GRPC_CPP_PLUGIN_EXECUTABLE}"
+        "${hw_proto}"
+      DEPENDS "${hw_proto}")
+
+# hw_grpc_proto
+add_library(hw_grpc_proto
+  ${hw_grpc_srcs}
+  ${hw_grpc_hdrs}
+  ${hw_proto_srcs}
+  ${hw_proto_hdrs} )
+
+add_executable(${TARGET} grpc-server.cpp json.hpp )
+target_link_libraries(${TARGET} PRIVATE common llama myclip ${CMAKE_THREAD_LIBS_INIT} absl::flags hw_grpc_proto
+  absl::flags_parse
+  gRPC::${_REFLECTION}
+  gRPC::${_GRPC_GRPCPP}
+  protobuf::${_PROTOBUF_LIBPROTOBUF})
+target_compile_features(${TARGET} PRIVATE cxx_std_11)
+if(TARGET BUILD_INFO)
+  add_dependencies(${TARGET} BUILD_INFO)
+endif()

backend/cpp/llama/Makefile

@@ -0,0 +1,50 @@
+
+LLAMA_VERSION?=d9b33fe95bd257b36c84ee5769cc048230067d6f
+
+CMAKE_ARGS?=
+BUILD_TYPE?=
+
+# If build type is cublas, then we set -DLLAMA_CUBLAS=ON to CMAKE_ARGS automatically
+ifeq ($(BUILD_TYPE),cublas)
+	CMAKE_ARGS+=-DLLAMA_CUBLAS=ON
+# If build type is openblas then we set -DLLAMA_BLAS=ON -DLLAMA_BLAS_VENDOR=OpenBLAS
+# to CMAKE_ARGS automatically
+else ifeq ($(BUILD_TYPE),openblas)
+	CMAKE_ARGS+=-DLLAMA_BLAS=ON -DLLAMA_BLAS_VENDOR=OpenBLAS
+# If build type is clblast (openCL) we set -DLLAMA_CLBLAST=ON -DCLBlast_DIR=/some/path
+else ifeq ($(BUILD_TYPE),clblast)
+	CMAKE_ARGS+=-DLLAMA_CLBLAST=ON -DCLBlast_DIR=/some/path
+# If it's hipblas we do have also to set CC=/opt/rocm/llvm/bin/clang CXX=/opt/rocm/llvm/bin/clang++ 
+else ifeq ($(BUILD_TYPE),hipblas)
+	CMAKE_ARGS+=-DLLAMA_HIPBLAS=ON
+endif
+
+llama.cpp:
+	git clone --recurse-submodules https://github.com/ggerganov/llama.cpp llama.cpp
+	cd llama.cpp && git checkout -b build $(LLAMA_VERSION) && git submodule update --init --recursive --depth 1
+
+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/
+	cp -rfv $(abspath ./)/json.hpp llama.cpp/examples/grpc-server/
+	echo "add_subdirectory(grpc-server)" >> llama.cpp/examples/CMakeLists.txt
+## XXX: In some versions of CMake clip wasn't being built before llama.
+## This is an hack for now, but it should be fixed in the future.
+	cp -rfv llama.cpp/examples/llava/clip.h llama.cpp/examples/grpc-server/clip.h
+	cp -rfv llama.cpp/examples/llava/clip.cpp llama.cpp/examples/grpc-server/clip.cpp
+
+rebuild:
+	cp -rfv $(abspath ./)/CMakeLists.txt llama.cpp/examples/grpc-server/
+	cp -rfv $(abspath ./)/grpc-server.cpp llama.cpp/examples/grpc-server/
+	cp -rfv $(abspath ./)/json.hpp llama.cpp/examples/grpc-server/
+	rm -rf grpc-server
+	$(MAKE) grpc-server
+
+clean:
+	rm -rf llama.cpp
+	rm -rf grpc-server
+
+grpc-server: llama.cpp llama.cpp/examples/grpc-server
+	cd llama.cpp && mkdir -p build && cd build && cmake .. $(CMAKE_ARGS) && cmake --build . --config Release
+	cp llama.cpp/build/bin/grpc-server .

cmd/grpc/bloomz/main.go

@@ -1,23 +0,0 @@
-package main
-
-// Note: this is started internally by LocalAI and a server is allocated for each model
-
-import (
-	"flag"
-
-	bloomz "github.com/go-skynet/LocalAI/pkg/backend/llm/bloomz"
-
-	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, &bloomz.LLM{}); err != nil {
-		panic(err)
-	}
-}

cmd/grpc/falcon/main.go

@@ -1,25 +0,0 @@
-package main
-
-// GRPC Falcon server
-
-// Note: this is started internally by LocalAI and a server is allocated for each model
-
-import (
-	"flag"
-
-	falcon "github.com/go-skynet/LocalAI/pkg/backend/llm/falcon"
-
-	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, &falcon.LLM{}); err != nil {
-		panic(err)
-	}
-}

entrypoint.sh

@@ -5,7 +5,7 @@ cd /build
 
 if [ "$REBUILD" != "false" ]; then
 	rm -rf ./local-ai
-	ESPEAK_DATA=/build/lib/Linux-$(uname -m)/piper_phonemize/lib/espeak-ng-data make build -j${BUILD_PARALLELISM:-1}
+	make build -j${BUILD_PARALLELISM:-1}
 else
 	echo "@@@@@"
 	echo "Skipping rebuild"

examples/bruno/LocalAI Test Requests/image generation/Generate image.bru

@@ -0,0 +1,25 @@
+meta {
+  name: Generate image
+  type: http
+  seq: 1
+}
+
+post {
+  url: {{PROTOCOL}}{{HOST}}:{{PORT}}/v1/images/generations
+  body: json
+  auth: none
+}
+
+headers {
+  Content-Type: application/json
+}
+
+body:json {
+  {
+    "prompt": "<positive prompt>|<negative prompt>",
+    "model": "model-name",
+    "step": 51,
+    "size": "1024x1024",
+    "image": ""
+  }
+}

examples/bruno/LocalAI Test Requests/llm text/chat/chat completion -simple- 1 message-.bru

@@ -15,10 +15,16 @@ headers {
 }
 
 body:json {
-  {
-       "model": "{{DEFAULT_MODEL}}",
-       "messages": [{"role": "user", "content": "How could one use friction to cook an egg?"}],
-       "max_tokens": 256,
-       "temperature": 0.2
+  {
+    "model": "{{DEFAULT_MODEL}}",
+    "messages": [
+      {
+        "role": "user",
+        "content": "How could one use friction to cook an egg?"
+      }
+    ],
+    "max_tokens": 256,
+    "temperature": 0.2,
+    "grammar": ""
   }
 }

examples/chainlit/Dockerfile

@@ -0,0 +1,16 @@
+# Use an official Python runtime as a parent image
+FROM harbor.home.sfxworks.net/docker/library/python:3.9-slim
+
+# Set the working directory in the container
+WORKDIR /app
+
+# Copy the current directory contents into the container at /app
+COPY requirements.txt /app
+
+# Install any needed packages specified in requirements.txt
+RUN pip install -r requirements.txt
+
+COPY . /app
+
+# Run app.py when the container launches
+CMD ["chainlit", "run", "-h", "--host", "0.0.0.0", "main.py" ]

examples/chainlit/README.md

@@ -0,0 +1,25 @@
+# LocalAI Demonstration with Embeddings and Chainlit
+
+This demonstration shows you how to use embeddings with existing data in `LocalAI`, and how to integrate it with Chainlit for an interactive querying experience. We are using the `llama_index` library to facilitate the embedding and querying processes, and `chainlit` to provide an interactive interface. The `Weaviate` client is used as the embedding source.
+
+## Prerequisites
+
+Before proceeding, make sure you have the following installed:
+- Weaviate client
+- LocalAI and its dependencies
+- Chainlit and its dependencies
+
+## Getting Started
+
+1. Clone this repository:
+2. Navigate to the project directory:
+3. Run the example: `chainlit run main.py`
+
+# Highlight on `llama_index` and `chainlit`
+
+`llama_index` is the key library that facilitates the process of embedding and querying data in LocalAI. It provides a seamless interface to integrate various components, such as `WeaviateVectorStore`, `LocalAI`, `ServiceContext`, and more, for a smooth querying experience.
+
+`chainlit` is used to provide an interactive interface for users to query the data and see the results in real-time. It integrates with llama_index to handle the querying process and display the results to the user.
+
+In this example, `llama_index` is used to set up the `VectorStoreIndex` and `QueryEngine`, and `chainlit` is used to handle the user interactions with `LocalAI` and display the results.
+

examples/chainlit/config.yaml

@@ -0,0 +1,16 @@
+localAI:
+  temperature: 0
+  modelName: gpt-3.5-turbo
+  apiBase: http://local-ai.default
+  apiKey: stub
+  streaming: True
+weviate:
+  url: http://weviate.local
+  index: AIChroma
+query:
+  mode: hybrid
+  topK: 1
+  alpha: 0.0
+  chunkSize: 1024
+embedding:
+  model: BAAI/bge-small-en-v1.5

examples/chainlit/main.py

@@ -0,0 +1,82 @@
+import os
+
+import weaviate
+from llama_index.storage.storage_context import StorageContext
+from llama_index.vector_stores import WeaviateVectorStore
+
+from llama_index.query_engine.retriever_query_engine import RetrieverQueryEngine
+from llama_index.callbacks.base import CallbackManager
+from llama_index import (
+    LLMPredictor,
+    ServiceContext,
+    StorageContext,
+    VectorStoreIndex,
+)
+import chainlit as cl
+
+from llama_index.llms import LocalAI
+from llama_index.embeddings import HuggingFaceEmbedding
+import yaml
+
+# Load the configuration file
+with open("config.yaml", "r") as ymlfile:
+    cfg = yaml.safe_load(ymlfile)
+
+# Get the values from the configuration file or set the default values
+temperature = cfg['localAI'].get('temperature', 0)
+model_name = cfg['localAI'].get('modelName', "gpt-3.5-turbo")
+api_base = cfg['localAI'].get('apiBase', "http://local-ai.default")
+api_key = cfg['localAI'].get('apiKey', "stub")
+streaming = cfg['localAI'].get('streaming', True)
+weaviate_url = cfg['weviate'].get('url', "http://weviate.default")
+index_name = cfg['weviate'].get('index', "AIChroma")
+query_mode = cfg['query'].get('mode', "hybrid")
+topK = cfg['query'].get('topK', 1)
+alpha = cfg['query'].get('alpha', 0.0)
+embed_model_name = cfg['embedding'].get('model', "BAAI/bge-small-en-v1.5")
+chunk_size = cfg['query'].get('chunkSize', 1024)
+
+
+embed_model = HuggingFaceEmbedding(model_name=embed_model_name)
+
+
+llm = LocalAI(temperature=temperature, model_name=model_name, api_base=api_base, api_key=api_key, streaming=streaming)
+llm.globally_use_chat_completions = True;
+client = weaviate.Client(weaviate_url)
+vector_store = WeaviateVectorStore(weaviate_client=client, index_name=index_name)
+storage_context = StorageContext.from_defaults(vector_store=vector_store)
+
+@cl.on_chat_start
+async def factory():
+
+    llm_predictor = LLMPredictor(
+        llm=llm
+    )
+    
+    service_context = ServiceContext.from_defaults(embed_model=embed_model, callback_manager=CallbackManager([cl.LlamaIndexCallbackHandler()]), llm_predictor=llm_predictor, chunk_size=chunk_size)
+
+    index = VectorStoreIndex.from_vector_store(
+        vector_store,
+        storage_context=storage_context,
+        service_context=service_context
+    )
+
+    query_engine = index.as_query_engine(vector_store_query_mode=query_mode, similarity_top_k=topK, alpha=alpha, streaming=True)
+
+    cl.user_session.set("query_engine", query_engine)
+
+
+@cl.on_message
+async def main(message: cl.Message):
+    query_engine = cl.user_session.get("query_engine")
+    response = await cl.make_async(query_engine.query)(message.content)
+
+    response_message = cl.Message(content="")
+
+    for token in response.response_gen:
+        await response_message.stream_token(token=token)
+
+    if response.response_txt:
+        response_message.content = response.response_txt
+
+    await response_message.send()

examples/chainlit/requirements.txt

@@ -0,0 +1,7 @@
+llama_hub==0.0.41
+llama_index==0.8.55
+Requests==2.31.0
+weaviate_client==3.25.1
+transformers
+torch
+chainlit

examples/configurations/README.md

@@ -0,0 +1,42 @@
+## Advanced configuration
+
+This section contains examples on how to install models manually with config files.
+
+### Prerequisites
+
+First clone LocalAI:
+
+```bash
+git clone https://github.com/go-skynet/LocalAI
+
+cd LocalAI
+```
+
+Setup the model you prefer from the examples below and then start LocalAI:
+
+```bash
+docker compose up -d --pull always
+```
+
+If LocalAI is already started, you can restart it with 
+
+```bash
+docker compose restart
+```
+
+See also the getting started: https://localai.io/basics/getting_started/
+
+### Mistral
+
+To setup mistral copy the files inside `mistral` in the `models` folder:
+
+```bash
+cp -r examples/configurations/mistral/* models/
+```
+
+Now download the model:
+
+```bash
+wget https://huggingface.co/TheBloke/Mistral-7B-OpenOrca-GGUF/resolve/main/mistral-7b-openorca.Q6_K.gguf -O models/mistral-7b-openorca.Q6_K.gguf
+```
+

examples/configurations/llava/README.md

@@ -0,0 +1,18 @@
+![llava](https://github.com/mudler/LocalAI/assets/2420543/cb0a0897-3b58-4350-af66-e6f4387b58d3)
+
+## Setup
+
+```
+mkdir models
+wget https://huggingface.co/mys/ggml_bakllava-1/resolve/main/ggml-model-q4_k.gguf -O models/ggml-model-q4_k.gguf
+wget https://huggingface.co/mys/ggml_bakllava-1/resolve/main/mmproj-model-f16.gguf -O models/mmproj-model-f16.gguf
+docker run -p 8080:8080 -v $PWD/models:/models -ti --rm quay.io/go-skynet/local-ai:master --models-path /models --threads 4
+```
+
+## Try it out
+
+```
+curl http://localhost:8080/v1/chat/completions -H "Content-Type: application/json" -d '{
+     "model": "llava",
+     "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}]}'
+```

examples/configurations/llava/chat-simple.tmpl

@@ -0,0 +1,3 @@
+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:

examples/configurations/llava/llava.yaml

@@ -0,0 +1,20 @@
+
+context_size: 4096
+f16: true
+threads: 11
+gpu_layers: 90
+name: llava
+mmap: true
+backend: llama-cpp
+roles:
+  user: "USER:"
+  assistant: "ASSISTANT:"
+  system: "SYSTEM:"
+parameters:
+  model: ggml-model-q4_k.gguf
+  temperature: 0.2
+  top_k: 40
+  top_p: 0.95
+template:
+  chat: chat-simple
+mmproj: mmproj-model-f16.gguf

examples/configurations/mistral/chatml-block.tmpl

@@ -0,0 +1,3 @@
+{{.Input}}
+<|im_start|>assistant
+

examples/configurations/mistral/chatml.tmpl

@@ -0,0 +1,3 @@
+<|im_start|>{{if eq .RoleName "assistant"}}assistant{{else if eq .RoleName "system"}}system{{else if eq .RoleName "user"}}user{{end}}
+{{if .Content}}{{.Content}}{{end}}
+<|im_end|>

examples/configurations/mistral/completion.tmpl

@@ -0,0 +1 @@
+{{.Input}}

examples/configurations/mistral/mistral.yaml

@@ -0,0 +1,16 @@
+name: mistral
+mmap: true
+parameters:
+  model: mistral-7b-openorca.Q6_K.gguf
+  temperature: 0.2
+  top_k: 40
+  top_p: 0.95
+template:
+  chat_message: chatml
+  chat: chatml-block
+  completion: completion
+context_size: 4096
+f16: true
+stopwords:
+- <|im_end|>
+threads: 4

examples/llamaindex/README.md

@@ -0,0 +1,30 @@
+# LocalAI Demonstration with Embeddings
+
+This demonstration shows you how to use embeddings with existing data in LocalAI. We are using the `llama_index` library to facilitate the embedding and querying processes. The `Weaviate` client is used as the embedding source.
+
+## Prerequisites
+
+Before proceeding, make sure you have the following installed:
+- Weaviate client
+- LocalAI and its dependencies
+- llama_index and its dependencies 
+
+## Getting Started
+
+1. Clone this repository:
+
+2. Navigate to the project directory:
+
+3. Run the example:
+
+`python main.py`
+
+```
+Downloading (…)lve/main/config.json: 100%|███████████████████████████| 684/684 [00:00<00:00, 6.01MB/s]
+Downloading model.safetensors: 100%|███████████████████████████████| 133M/133M [00:03<00:00, 39.5MB/s]
+Downloading (…)okenizer_config.json: 100%|███████████████████████████| 366/366 [00:00<00:00, 2.79MB/s]
+Downloading (…)solve/main/vocab.txt: 100%|█████████████████████████| 232k/232k [00:00<00:00, 6.00MB/s]
+Downloading (…)/main/tokenizer.json: 100%|█████████████████████████| 711k/711k [00:00<00:00, 18.8MB/s]
+Downloading (…)cial_tokens_map.json: 100%|███████████████████████████| 125/125 [00:00<00:00, 1.18MB/s]
+LocalAI is a community-driven project that aims to make AI accessible to everyone. It was created by Ettore Di Giacinto and is focused on providing various AI-related features such as text generation with GPTs, text to audio, audio to text, image generation, and more. The project is constantly growing and evolving, with a roadmap for future improvements. Anyone is welcome to contribute, provide feedback, and submit pull requests to help make LocalAI better.
+```

examples/llamaindex/main.py

@@ -0,0 +1,38 @@
+import os
+
+import weaviate
+
+from llama_index import ServiceContext, VectorStoreIndex, StorageContext
+from llama_index.llms import LocalAI
+from llama_index.vector_stores import WeaviateVectorStore
+from llama_index.storage.storage_context import StorageContext
+
+# Weaviate client setup
+client = weaviate.Client("http://weviate.default")
+
+# Weaviate vector store setup
+vector_store = WeaviateVectorStore(weaviate_client=client, index_name="AIChroma")
+
+# Storage context setup
+storage_context = StorageContext.from_defaults(vector_store=vector_store)
+
+# LocalAI setup
+llm = LocalAI(temperature=0, model_name="gpt-3.5-turbo", api_base="http://local-ai.default", api_key="stub")
+llm.globally_use_chat_completions = True;
+
+# Service context setup
+service_context = ServiceContext.from_defaults(llm=llm, embed_model="local")
+
+# Load index from stored vectors
+index = VectorStoreIndex.from_vector_store(
+    vector_store,
+    storage_context=storage_context,
+    service_context=service_context
+)
+
+# Query engine setup
+query_engine = index.as_query_engine(similarity_top_k=1, vector_store_query_mode="hybrid")
+
+# Query example
+response = query_engine.query("What is LocalAI?")
+print(response)

extra/grpc/README.md

@@ -0,0 +1,38 @@
+# Common commands about conda environment
+
+## Create a new empty conda environment
+
+```
+conda create --name <env-name> python=<your version> -y
+
+conda create --name autogptq python=3.11 -y
+```
+
+## To activate the environment
+
+As of conda 4.4
+```
+conda activate autogptq
+```
+
+The conda version older than 4.4
+
+```
+source activate autogptq
+```
+
+## Install the packages to your environment
+
+Sometimes you need to install the packages from the conda-forge channel
+
+By using `conda`
+```
+conda install <your-package-name>
+
+conda install -c conda-forge <your package-name>
+```
+
+Or by using `pip`
+```
+pip install <your-package-name>
+```

extra/grpc/autogptq/Makefile

@@ -0,0 +1,5 @@
+.PONY: autogptq
+autogptq:
+	@echo "Creating virtual environment..."
+	@conda env create --name autogptq --file autogptq.yml
+	@echo "Virtual environment created."

extra/grpc/autogptq/README.md

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

extra/grpc/autogptq/autogptq.py

@@ -1,15 +1,15 @@
 #!/usr/bin/env python3
-import grpc
 from concurrent import futures
-import time
-import backend_pb2
-import backend_pb2_grpc
 import argparse
 import signal
 import sys
 import os
-from auto_gptq import AutoGPTQForCausalLM, BaseQuantizeConfig
-from pathlib import Path
+import time
+
+import grpc
+import backend_pb2
+import backend_pb2_grpc
+from auto_gptq import AutoGPTQForCausalLM
 from transformers import AutoTokenizer
 from transformers import TextGenerationPipeline
 

extra/grpc/autogptq/autogptq.yml

@@ -0,0 +1,86 @@
+name: autogptq
+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
+  - wheel=0.41.2=py311h06a4308_0
+  - xz=5.4.2=h5eee18b_0
+  - zlib=1.2.13=h5eee18b_0
+  - pip:
+      - accelerate==0.23.0
+      - aiohttp==3.8.5
+      - aiosignal==1.3.1
+      - async-timeout==4.0.3
+      - attrs==23.1.0
+      - auto-gptq==0.4.2
+      - certifi==2023.7.22
+      - charset-normalizer==3.3.0
+      - datasets==2.14.5
+      - dill==0.3.7
+      - filelock==3.12.4
+      - frozenlist==1.4.0
+      - fsspec==2023.6.0
+      - grpcio==1.59.0
+      - huggingface-hub==0.16.4
+      - idna==3.4
+      - jinja2==3.1.2
+      - markupsafe==2.1.3
+      - mpmath==1.3.0
+      - multidict==6.0.4
+      - multiprocess==0.70.15
+      - networkx==3.1
+      - 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
+      - packaging==23.2
+      - pandas==2.1.1
+      - peft==0.5.0
+      - protobuf==4.24.4
+      - psutil==5.9.5
+      - pyarrow==13.0.0
+      - python-dateutil==2.8.2
+      - pytz==2023.3.post1
+      - pyyaml==6.0.1
+      - regex==2023.10.3
+      - requests==2.31.0
+      - rouge==1.0.1
+      - safetensors==0.3.3
+      - six==1.16.0
+      - sympy==1.12
+      - tokenizers==0.14.0
+      - torch==2.1.0
+      - tqdm==4.66.1
+      - transformers==4.34.0
+      - triton==2.1.0
+      - typing-extensions==4.8.0
+      - tzdata==2023.3
+      - urllib3==2.0.6
+      - xxhash==3.4.1
+      - yarl==1.9.2

extra/grpc/autogptq/run.sh

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

extra/grpc/bark/Makefile

@@ -0,0 +1,5 @@
+.PONY: ttsbark
+ttsbark:
+	@echo "Creating virtual environment..."
+	@conda env create --name ttsbark --file ttsbark.yml
+	@echo "Virtual environment created."

extra/grpc/bark/README.md

@@ -0,0 +1,16 @@
+# Creating a separate environment for ttsbark project
+
+```
+make ttsbark
+```
+
+# Testing the gRPC server
+
+```
+<The path of your python interpreter> -m unittest test_ttsbark.py
+```
+
+For example
+```
+/opt/conda/envs/bark/bin/python -m unittest extra/grpc/bark/test_ttsbark.py
+``````

extra/grpc/bark/run.sh

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

extra/grpc/bark/test_ttsbark.py

@@ -0,0 +1,32 @@
+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):
+        self.service = subprocess.Popen(["python3", "ttsbark.py", "--addr", "localhost:50051"])
+
+    def tearDown(self) -> None:
+        self.service.terminate()
+        self.service.wait()
+
+    def test_server_startup(self):
+        time.sleep(2)
+        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()

extra/grpc/bark/ttsbark.py

@@ -1,18 +1,23 @@
+"""
+This is the extra gRPC server of LocalAI
+"""
+
 #!/usr/bin/env python3
-import grpc
 from concurrent import futures
 import time
-import backend_pb2
-import backend_pb2_grpc
 import argparse
 import signal
 import sys
 import os
-from auto_gptq import AutoGPTQForCausalLM, BaseQuantizeConfig
-from pathlib import Path
-from bark import SAMPLE_RATE, generate_audio, preload_models
 from scipy.io.wavfile import write as write_wav
 
+import backend_pb2
+import backend_pb2_grpc
+from bark import SAMPLE_RATE, generate_audio, preload_models
+
+import grpc
+
+
 _ONE_DAY_IN_SECONDS = 60 * 60 * 24
 
 # If MAX_WORKERS are specified in the environment use it, otherwise default to 1
@@ -20,6 +25,9 @@ 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):
+    """
+    BackendServicer is the class that implements the gRPC service
+    """
     def Health(self, request, context):
         return backend_pb2.Reply(message=bytes("OK", 'utf-8'))
     def LoadModel(self, request, context):
@@ -83,4 +91,4 @@ if __name__ == "__main__":
     )
     args = parser.parse_args()
 
-    serve(args.addr)
+    serve(args.addr)

extra/grpc/bark/ttsbark.yml

@@ -0,0 +1,96 @@
+name: bark
+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
+  - wheel=0.41.2=py311h06a4308_0
+  - xz=5.4.2=h5eee18b_0
+  - zlib=1.2.13=h5eee18b_0
+  - pip:
+      - accelerate==0.23.0
+      - aiohttp==3.8.5
+      - aiosignal==1.3.1
+      - 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
+      - charset-normalizer==3.3.0
+      - datasets==2.14.5
+      - dill==0.3.7
+      - einops==0.7.0
+      - encodec==0.1.1
+      - filelock==3.12.4
+      - frozenlist==1.4.0
+      - fsspec==2023.6.0
+      - funcy==2.0
+      - grpcio==1.59.0
+      - huggingface-hub==0.16.4
+      - idna==3.4
+      - jinja2==3.1.2
+      - jmespath==1.0.1
+      - markupsafe==2.1.3
+      - mpmath==1.3.0
+      - multidict==6.0.4
+      - multiprocess==0.70.15
+      - networkx==3.1
+      - 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
+      - packaging==23.2
+      - pandas==2.1.1
+      - peft==0.5.0
+      - protobuf==4.24.4
+      - psutil==5.9.5
+      - pyarrow==13.0.0
+      - python-dateutil==2.8.2
+      - pytz==2023.3.post1
+      - pyyaml==6.0.1
+      - regex==2023.10.3
+      - requests==2.31.0
+      - rouge==1.0.1
+      - s3transfer==0.7.0
+      - safetensors==0.3.3
+      - scipy==1.11.3
+      - six==1.16.0
+      - sympy==1.12
+      - tokenizers==0.14.0
+      - torch==2.1.0
+      - torchaudio==2.1.0
+      - tqdm==4.66.1
+      - transformers==4.34.0
+      - triton==2.1.0
+      - typing-extensions==4.8.0
+      - tzdata==2023.3
+      - urllib3==1.26.17
+      - xxhash==3.4.1
+      - yarl==1.9.2
+prefix: /opt/conda/envs/bark

extra/grpc/diffusers/Makefile

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

extra/grpc/diffusers/README.md

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

extra/grpc/diffusers/backend_diffusers.py

@@ -1,27 +1,32 @@
 #!/usr/bin/env python3
-import grpc
 from concurrent import futures
-import time
-import backend_pb2
-import backend_pb2_grpc
+
 import argparse
+from collections import defaultdict
+from enum import Enum
 import signal
 import sys
+import time
 import os
 
-# import diffusers
-import torch
-from torch import autocast
-from diffusers import StableDiffusionXLPipeline, StableDiffusionDepth2ImgPipeline, DPMSolverMultistepScheduler, StableDiffusionPipeline, DiffusionPipeline, EulerAncestralDiscreteScheduler
-from diffusers.pipelines.stable_diffusion import safety_checker
-from compel import Compel
 from PIL import Image
-from io import BytesIO
+import torch
+
+import backend_pb2
+import backend_pb2_grpc
+
+import grpc
+
+from diffusers import StableDiffusionXLPipeline, StableDiffusionDepth2ImgPipeline, DPMSolverMultistepScheduler, StableDiffusionPipeline, DiffusionPipeline, EulerAncestralDiscreteScheduler
 from diffusers import StableDiffusionImg2ImgPipeline
+from diffusers.pipelines.stable_diffusion import safety_checker
+
+from compel import Compel
+
 from transformers import CLIPTextModel
-from enum import Enum
-from collections import defaultdict
 from safetensors.torch import load_file
+
+
 _ONE_DAY_IN_SECONDS = 60 * 60 * 24
 COMPEL=os.environ.get("COMPEL", "1") == "1"
 CLIPSKIP=os.environ.get("CLIPSKIP", "1") == "1"

extra/grpc/diffusers/diffusers.yml

@@ -0,0 +1,74 @@
+name: diffusers
+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==0.23.0
+      - certifi==2023.7.22
+      - charset-normalizer==3.3.0
+      - compel==2.0.2
+      - diffusers==0.21.4
+      - filelock==3.12.4
+      - fsspec==2023.9.2
+      - grpcio==1.59.0
+      - huggingface-hub==0.17.3
+      - idna==3.4
+      - importlib-metadata==6.8.0
+      - jinja2==3.1.2
+      - markupsafe==2.1.3
+      - mpmath==1.3.0
+      - networkx==3.1
+      - 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
+      - packaging==23.2
+      - pillow==10.0.1
+      - protobuf==4.24.4
+      - psutil==5.9.5
+      - pyparsing==3.1.1
+      - pyyaml==6.0.1
+      - regex==2023.10.3
+      - requests==2.31.0
+      - safetensors==0.4.0
+      - sympy==1.12
+      - tokenizers==0.14.1
+      - torch==2.1.0
+      - tqdm==4.66.1
+      - transformers==4.34.0
+      - triton==2.1.0
+      - typing-extensions==4.8.0
+      - urllib3==2.0.6
+      - zipp==3.17.0
+prefix: /opt/conda/envs/diffusers

extra/grpc/diffusers/run.sh

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

extra/grpc/exllama/Makefile

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

extra/grpc/exllama/README.md

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

extra/grpc/exllama/exllama.yml

@@ -0,0 +1,55 @@
+name: exllama
+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:
+      - filelock==3.12.4
+      - fsspec==2023.9.2
+      - grpcio==1.59.0
+      - jinja2==3.1.2
+      - markupsafe==2.1.3
+      - mpmath==1.3.0
+      - networkx==3.1
+      - ninja==1.11.1
+      - 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
+      - protobuf==4.24.4
+      - safetensors==0.3.2
+      - sentencepiece==0.1.99
+      - sympy==1.12
+      - torch==2.1.0
+      - triton==2.1.0
+      - typing-extensions==4.8.0
+prefix: /opt/conda/envs/exllama

extra/grpc/exllama/run.sh

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

extra/grpc/huggingface/Makefile

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

extra/grpc/huggingface/README.md

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

extra/grpc/huggingface/huggingface.py

@@ -1,13 +1,20 @@
+"""
+Extra gRPC server for HuggingFace SentenceTransformer models.
+"""
 #!/usr/bin/env python3
-import grpc
 from concurrent import futures
-import time
-import backend_pb2
-import backend_pb2_grpc
+
 import argparse
 import signal
 import sys
 import os
+
+import time
+import backend_pb2
+import backend_pb2_grpc
+
+import grpc
+
 from sentence_transformers import SentenceTransformer
 
 _ONE_DAY_IN_SECONDS = 60 * 60 * 24
@@ -17,18 +24,56 @@ 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
         try:
             self.model = SentenceTransformer(model_name)
         except Exception as err:
             return backend_pb2.Result(success=False, message=f"Unexpected {err=}, {type(err)=}")
+
         # Implement your logic here for the LoadModel service
         # Replace this with your desired response
         return backend_pb2.Result(message="Model loaded successfully", success=True)
+
     def Embedding(self, request, context):
+        """
+        A gRPC method that calculates embeddings for a given sentence.
+
+        Args:
+            request: An EmbeddingRequest object that contains the request parameters.
+            context: A grpc.ServicerContext object that provides information about the RPC.
+
+        Returns:
+            An EmbeddingResult object that contains the calculated embeddings.
+        """
         # Implement your logic here for the Embedding service
         # Replace this with your desired response
         print("Calculated embeddings for: " + request.Embeddings, file=sys.stderr)
@@ -66,4 +111,4 @@ if __name__ == "__main__":
     )
     args = parser.parse_args()
 
-    serve(args.addr)
+    serve(args.addr)

extra/grpc/huggingface/huggingface.yml

@@ -0,0 +1,77 @@
+name: huggingface
+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:
+      - certifi==2023.7.22
+      - charset-normalizer==3.3.0
+      - click==8.1.7
+      - filelock==3.12.4
+      - fsspec==2023.9.2
+      - grpcio==1.59.0
+      - huggingface-hub==0.17.3
+      - idna==3.4
+      - install==1.3.5
+      - jinja2==3.1.2
+      - joblib==1.3.2
+      - markupsafe==2.1.3
+      - mpmath==1.3.0
+      - networkx==3.1
+      - nltk==3.8.1
+      - 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
+      - packaging==23.2
+      - pillow==10.0.1
+      - protobuf==4.24.4
+      - pyyaml==6.0.1
+      - regex==2023.10.3
+      - requests==2.31.0
+      - safetensors==0.4.0
+      - scikit-learn==1.3.1
+      - scipy==1.11.3
+      - sentence-transformers==2.2.2
+      - sentencepiece==0.1.99
+      - sympy==1.12
+      - threadpoolctl==3.2.0
+      - tokenizers==0.14.1
+      - torch==2.1.0
+      - torchvision==0.16.0
+      - tqdm==4.66.1
+      - transformers==4.34.0
+      - triton==2.1.0
+      - typing-extensions==4.8.0
+      - urllib3==2.0.6
+prefix: /opt/conda/envs/huggingface

extra/grpc/huggingface/run.sh

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

extra/grpc/huggingface/test.sh

@@ -0,0 +1,11 @@
+#!/bin/bash
+##
+## A bash script wrapper that runs the huggingface server with conda
+
+# Activate conda environment
+source activate huggingface
+
+# 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_huggingface.py

extra/grpc/huggingface/test_huggingface.py

@@ -0,0 +1,81 @@
+"""
+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", "huggingface.py", "--addr", "localhost:50051"])
+
+    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
+        """
+        time.sleep(2)
+        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="bert-base-nli-mean-tokens"))
+                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_embedding(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="bert-base-nli-mean-tokens"))
+                self.assertTrue(response.success)
+                embedding_request = backend_pb2.PredictOptions(Embeddings="This is a test sentence.")
+                embedding_response = stub.Embedding(embedding_request)
+                self.assertIsNotNone(embedding_response.embeddings)
+        except Exception as err:
+            print(err)
+            self.fail("Embedding service failed")
+        finally:
+            self.tearDown()

extra/grpc/vall-e-x/Makefile

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

extra/grpc/vall-e-x/README.md

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

extra/grpc/vall-e-x/run.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+
+##
+## A bash script wrapper that runs the ttsvalle server with conda
+export PATH=$PATH:/opt/conda/bin
+
+# Activate conda environment
+source activate ttsvalle
+
+# get the directory where the bash script is located
+DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" >/dev/null 2>&1 && pwd )"
+
+python $DIR/ttvalle.py $@

extra/grpc/vall-e-x/ttsvalle.py

@@ -1,14 +1,15 @@
 #!/usr/bin/env python3
-import grpc
+
 from concurrent import futures
-import time
-import backend_pb2
-import backend_pb2_grpc
 import argparse
 import signal
 import sys
 import os
-from pathlib import Path
+import time
+import backend_pb2
+import backend_pb2_grpc
+
+import grpc
 
 from utils.generation import SAMPLE_RATE, generate_audio, preload_models
 from scipy.io.wavfile import write as write_wav
@@ -21,9 +22,34 @@ 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):
+    """
+    gRPC servicer for backend services.
+    """
     def Health(self, request, context):
+        """
+        Health check service.
+
+        Args:
+            request: A backend_pb2.HealthRequest instance.
+            context: A grpc.ServicerContext instance.
+
+        Returns:
+            A backend_pb2.Reply instance with message "OK".
+        """
         return backend_pb2.Reply(message=bytes("OK", 'utf-8'))
+
     def LoadModel(self, request, context):
+        """
+        Load model service.
+
+        Args:
+            request: A backend_pb2.LoadModelRequest instance.
+            context: A grpc.ServicerContext instance.
+
+        Returns:
+            A backend_pb2.Result instance with message "Model loaded successfully" and success=True if successful.
+            A backend_pb2.Result instance with success=False and error message if unsuccessful.
+        """
         model_name = request.Model
         try:
             print("Preparing models, please wait", file=sys.stderr)
@@ -49,6 +75,17 @@ class BackendServicer(backend_pb2_grpc.BackendServicer):
         return backend_pb2.Result(message="Model loaded successfully", success=True)
 
     def TTS(self, request, context):
+        """
+        Text-to-speech service.
+
+        Args:
+            request: A backend_pb2.TTSRequest instance.
+            context: A grpc.ServicerContext instance.
+
+        Returns:
+            A backend_pb2.Result instance with success=True if successful.
+            A backend_pb2.Result instance with success=False and error message if unsuccessful.
+        """
         model = request.model
         print(request, file=sys.stderr)
         try:
@@ -97,4 +134,4 @@ if __name__ == "__main__":
     )
     args = parser.parse_args()
 
-    serve(args.addr)
+    serve(args.addr)

extra/grpc/vall-e-x/ttsvalle.yml

@@ -0,0 +1,101 @@
+name: ttsvalle
+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=py310h06a4308_0
+  - python=3.10.13=h955ad1f_0
+  - readline=8.2=h5eee18b_0
+  - setuptools=68.0.0=py310h06a4308_0
+  - sqlite=3.41.2=h5eee18b_0
+  - tk=8.6.12=h1ccaba5_0
+  - tzdata=2023c=h04d1e81_0
+  - wheel=0.41.2=py310h06a4308_0
+  - xz=5.4.2=h5eee18b_0
+  - zlib=1.2.13=h5eee18b_0
+  - pip:
+      - aiofiles==23.2.1
+      - altair==5.1.2
+      - annotated-types==0.6.0
+      - anyio==3.7.1
+      - click==8.1.7
+      - cn2an==0.5.22
+      - cython==3.0.3
+      - einops==0.7.0
+      - encodec==0.1.1
+      - eng-to-ipa==0.0.2
+      - fastapi==0.103.2
+      - ffmpeg-python==0.2.0
+      - ffmpy==0.3.1
+      - fsspec==2023.9.2
+      - future==0.18.3
+      - gradio==3.47.1
+      - gradio-client==0.6.0
+      - grpcio==1.59.0
+      - h11==0.14.0
+      - httpcore==0.18.0
+      - httpx==0.25.0
+      - huggingface-hub==0.17.3
+      - importlib-resources==6.1.0
+      - inflect==7.0.0
+      - jieba==0.42.1
+      - langid==1.1.6
+      - llvmlite==0.41.0
+      - more-itertools==10.1.0
+      - nltk==3.8.1
+      - numba==0.58.0
+      - numpy==1.25.2
+      - 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
+      - openai-whisper==20230306
+      - orjson==3.9.7
+      - proces==0.1.7
+      - protobuf==4.24.4
+      - pydantic==2.4.2
+      - pydantic-core==2.10.1
+      - pydub==0.25.1
+      - pyopenjtalk-prebuilt==0.3.0
+      - pypinyin==0.49.0
+      - python-multipart==0.0.6
+      - regex==2023.10.3
+      - safetensors==0.4.0
+      - semantic-version==2.10.0
+      - soundfile==0.12.1
+      - starlette==0.27.0
+      - sudachidict-core==20230927
+      - sudachipy==0.6.7
+      - tokenizers==0.14.1
+      - toolz==0.12.0
+      - torch==2.1.0
+      - torchaudio==2.1.0
+      - torchvision==0.16.0
+      - tqdm==4.66.1
+      - transformers==4.34.0
+      - triton==2.1.0
+      - unidecode==1.3.7
+      - uvicorn==0.23.2
+      - vocos==0.0.3
+      - websockets==11.0.3
+      - wget==3.2
+prefix: /opt/conda/envs/ttsvalle

extra/grpc/vllm/Makefile

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

extra/grpc/vllm/README.md

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

extra/grpc/vllm/backend_vllm.py

@@ -1,15 +1,15 @@
 #!/usr/bin/env python3
-import grpc
 from concurrent import futures
 import time
-import backend_pb2
-import backend_pb2_grpc
 import argparse
 import signal
 import sys
-import os, glob
+import os
 
-from pathlib import Path
+import backend_pb2
+import backend_pb2_grpc
+
+import grpc
 from vllm import LLM, SamplingParams
 
 _ONE_DAY_IN_SECONDS = 60 * 60 * 24
@@ -19,7 +19,20 @@ 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 generate(self,prompt, max_new_tokens):
+        """
+        Generates text based on the given prompt and maximum number of new tokens.
+
+        Args:
+            prompt (str): The prompt to generate text from.
+            max_new_tokens (int): The maximum number of new tokens to generate.
+
+        Returns:
+            str: The generated text.
+        """
         self.generator.end_beam_search()
 
         # Tokenizing the input
@@ -41,9 +54,31 @@ class BackendServicer(backend_pb2_grpc.BackendServicer):
             if token.item() == self.generator.tokenizer.eos_token_id:
                 break
         return decoded_text
+
     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:
             if request.Quantization != "":
                 self.llm = LLM(model=request.Model, quantization=request.Quantization)
@@ -54,6 +89,16 @@ class BackendServicer(backend_pb2_grpc.BackendServicer):
         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.
+        """
         if request.TopP == 0:
             request.TopP = 0.9
 
@@ -68,6 +113,16 @@ class BackendServicer(backend_pb2_grpc.BackendServicer):
         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
@@ -104,4 +159,4 @@ if __name__ == "__main__":
     )
     args = parser.parse_args()
 
-    serve(args.addr)
+    serve(args.addr)

extra/grpc/vllm/run.sh

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

extra/grpc/vllm/test_backend_vllm.py

@@ -0,0 +1,41 @@
+import unittest
+import subprocess
+import time
+import backend_pb2
+import backend_pb2_grpc
+
+import grpc
+
+import unittest
+import subprocess
+import time
+import grpc
+import backend_pb2_grpc
+import backend_pb2
+
+class TestBackendServicer(unittest.TestCase):
+    """
+    TestBackendServicer is the class that tests the gRPC service.
+
+    This class contains methods to test the startup and shutdown of the gRPC service.
+    """
+    def setUp(self):
+        self.service = subprocess.Popen(["python", "backend_vllm.py", "--addr", "localhost:50051"])
+
+    def tearDown(self) -> None:
+        self.service.terminate()
+        self.service.wait()
+
+    def test_server_startup(self):
+        time.sleep(2)
+        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()

extra/grpc/vllm/vllm.yml

@@ -0,0 +1,99 @@
+name: vllm
+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
+  - wheel=0.41.2=py311h06a4308_0
+  - xz=5.4.2=h5eee18b_0
+  - zlib=1.2.13=h5eee18b_0
+  - pip:
+      - aiosignal==1.3.1
+      - anyio==3.7.1
+      - attrs==23.1.0
+      - certifi==2023.7.22
+      - charset-normalizer==3.3.0
+      - click==8.1.7
+      - cmake==3.27.6
+      - fastapi==0.103.2
+      - filelock==3.12.4
+      - frozenlist==1.4.0
+      - fsspec==2023.9.2
+      - grpcio==1.59.0
+      - h11==0.14.0
+      - httptools==0.6.0
+      - huggingface-hub==0.17.3
+      - idna==3.4
+      - jinja2==3.1.2
+      - jsonschema==4.19.1
+      - jsonschema-specifications==2023.7.1
+      - lit==17.0.2
+      - markupsafe==2.1.3
+      - mpmath==1.3.0
+      - msgpack==1.0.7
+      - networkx==3.1
+      - ninja==1.11.1
+      - numpy==1.26.0
+      - nvidia-cublas-cu11==11.10.3.66
+      - nvidia-cuda-cupti-cu11==11.7.101
+      - nvidia-cuda-nvrtc-cu11==11.7.99
+      - nvidia-cuda-runtime-cu11==11.7.99
+      - nvidia-cudnn-cu11==8.5.0.96
+      - nvidia-cufft-cu11==10.9.0.58
+      - nvidia-curand-cu11==10.2.10.91
+      - nvidia-cusolver-cu11==11.4.0.1
+      - nvidia-cusparse-cu11==11.7.4.91
+      - nvidia-nccl-cu11==2.14.3
+      - nvidia-nvtx-cu11==11.7.91
+      - packaging==23.2
+      - pandas==2.1.1
+      - protobuf==4.24.4
+      - psutil==5.9.5
+      - pyarrow==13.0.0
+      - pydantic==1.10.13
+      - python-dateutil==2.8.2
+      - python-dotenv==1.0.0
+      - pytz==2023.3.post1
+      - pyyaml==6.0.1
+      - ray==2.7.0
+      - referencing==0.30.2
+      - regex==2023.10.3
+      - requests==2.31.0
+      - rpds-py==0.10.4
+      - safetensors==0.4.0
+      - sentencepiece==0.1.99
+      - six==1.16.0
+      - sniffio==1.3.0
+      - starlette==0.27.0
+      - sympy==1.12
+      - tokenizers==0.14.1
+      - torch==2.0.1
+      - tqdm==4.66.1
+      - transformers==4.34.0
+      - triton==2.0.0
+      - typing-extensions==4.8.0
+      - tzdata==2023.3
+      - urllib3==2.0.6
+      - uvicorn==0.23.2
+      - uvloop==0.17.0
+      - vllm==0.2.0
+      - watchfiles==0.20.0
+      - websockets==11.0.3
+      - xformers==0.0.22
+prefix: /opt/conda/envs/vllm

go.mod

@@ -10,7 +10,7 @@ require (
 	github.com/go-skynet/go-bert.cpp v0.0.0-20230716133540-6abe312cded1
 	github.com/go-skynet/go-ggml-transformers.cpp v0.0.0-20230714203132-ffb09d7dd71e
 	github.com/go-skynet/go-llama.cpp v0.0.0-20231009155254-aeba71ee8428
-	github.com/gofiber/fiber/v2 v2.49.2
+	github.com/gofiber/fiber/v2 v2.50.0
 	github.com/google/uuid v1.3.1
 	github.com/hashicorp/go-multierror v1.1.1
 	github.com/hpcloud/tail v1.0.0
@@ -20,18 +20,23 @@ require (
 	github.com/mudler/go-ggllm.cpp v0.0.0-20230709223052-862477d16eef
 	github.com/mudler/go-processmanager v0.0.0-20230818213616-f204007f963c
 	github.com/mudler/go-stable-diffusion v0.0.0-20230605122230-d89260f598af
-	github.com/nomic-ai/gpt4all/gpt4all-bindings/golang v0.0.0-20231013181651-22de3c56bdd4
+	github.com/nomic-ai/gpt4all/gpt4all-bindings/golang v0.0.0-20231022042237-c25dc5193530
 	github.com/onsi/ginkgo/v2 v2.13.0
-	github.com/onsi/gomega v1.28.0
+	github.com/onsi/gomega v1.28.1
 	github.com/otiai10/openaigo v1.6.0
 	github.com/phayes/freeport v0.0.0-20220201140144-74d24b5ae9f5
+	github.com/prometheus/client_golang v1.17.0
 	github.com/rs/zerolog v1.31.0
 	github.com/sashabaranov/go-openai v1.16.0
 	github.com/schollz/progressbar/v3 v3.13.1
-	github.com/tmc/langchaingo v0.0.0-20230929160525-e16b77704b8d
+	github.com/tmc/langchaingo v0.0.0-20231019140956-c636b3da7701
 	github.com/urfave/cli/v2 v2.25.7
 	github.com/valyala/fasthttp v1.50.0
-	google.golang.org/grpc v1.58.3
+	go.opentelemetry.io/otel v1.19.0
+	go.opentelemetry.io/otel/exporters/prometheus v0.42.0
+	go.opentelemetry.io/otel/metric v1.19.0
+	go.opentelemetry.io/otel/sdk/metric v1.19.0
+	google.golang.org/grpc v1.59.0
 	google.golang.org/protobuf v1.31.0
 	gopkg.in/yaml.v2 v2.4.0
 	gopkg.in/yaml.v3 v3.0.1
@@ -49,23 +54,30 @@ require (
 )
 
 require (
+	github.com/beorn7/perks v1.0.1 // indirect
+	github.com/cespare/xxhash/v2 v2.2.0 // indirect
 	github.com/dlclark/regexp2 v1.8.1 // indirect
 	github.com/dsnet/compress v0.0.2-0.20210315054119-f66993602bf5 // indirect
+	github.com/go-logr/stdr v1.2.2 // indirect
 	github.com/golang/protobuf v1.5.3 // indirect
 	github.com/golang/snappy v0.0.2 // indirect
 	github.com/klauspost/pgzip v1.2.5 // indirect
-	github.com/kr/text v0.2.0 // indirect
+	github.com/matttproud/golang_protobuf_extensions v1.0.4 // indirect
 	github.com/mitchellh/colorstring v0.0.0-20190213212951-d06e56a500db // indirect
-	github.com/modern-go/concurrent v0.0.0-20180228061459-e0a39a4cb421 // indirect
+	github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd // indirect
 	github.com/modern-go/reflect2 v1.0.2 // indirect
 	github.com/nwaples/rardecode v1.1.0 // indirect
 	github.com/pierrec/lz4/v4 v4.1.2 // indirect
 	github.com/pkoukk/tiktoken-go v0.1.2 // indirect
+	github.com/prometheus/client_model v0.4.1-0.20230718164431-9a2bf3000d16 // indirect
+	github.com/prometheus/common v0.44.0 // indirect
+	github.com/prometheus/procfs v0.11.1 // indirect
 	github.com/ulikunitz/xz v0.5.9 // indirect
 	github.com/xi2/xz v0.0.0-20171230120015-48954b6210f8 // indirect
-	golang.org/x/term v0.11.0 // indirect
-	google.golang.org/genproto/googleapis/rpc v0.0.0-20230711160842-782d3b101e98 // indirect
-	gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c // indirect
+	go.opentelemetry.io/otel/sdk v1.19.0 // indirect
+	go.opentelemetry.io/otel/trace v1.19.0 // indirect
+	golang.org/x/term v0.13.0 // indirect
+	google.golang.org/genproto/googleapis/rpc v0.0.0-20230822172742-b8732ec3820d // indirect
 	gopkg.in/fsnotify.v1 v1.4.7 // indirect
 	gopkg.in/tomb.v1 v1.0.0-20141024135613-dd632973f1e7 // indirect
 )
@@ -77,7 +89,7 @@ require (
 	github.com/go-audio/riff v1.0.0 // indirect
 	github.com/go-logr/logr v1.2.4 // indirect
 	github.com/go-task/slim-sprig v0.0.0-20230315185526-52ccab3ef572 // indirect
-	github.com/google/go-cmp v0.5.9 // indirect
+	github.com/google/go-cmp v0.6.0 // indirect
 	github.com/google/pprof v0.0.0-20210407192527-94a9f03dee38 // indirect
 	github.com/hashicorp/errwrap v1.0.0 // indirect
 	github.com/klauspost/compress v1.16.7 // indirect
@@ -90,8 +102,8 @@ require (
 	github.com/valyala/bytebufferpool v1.0.0 // indirect
 	github.com/valyala/tcplisten v1.0.0 // indirect
 	github.com/xrash/smetrics v0.0.0-20201216005158-039620a65673 // indirect
-	golang.org/x/net v0.14.0 // indirect
-	golang.org/x/sys v0.12.0 // indirect
-	golang.org/x/text v0.12.0 // indirect
+	golang.org/x/net v0.17.0 // indirect
+	golang.org/x/sys v0.13.0 // indirect
+	golang.org/x/text v0.13.0 // indirect
 	golang.org/x/tools v0.12.0 // indirect
 )

go.sum

@@ -1,13 +1,16 @@
 github.com/andybalholm/brotli v1.0.1/go.mod h1:loMXtMfwqflxFJPmdbJO0a3KNoPuLBgiu3qAvBg8x/Y=
 github.com/andybalholm/brotli v1.0.5 h1:8uQZIdzKmjc/iuPu7O2ioW48L81FgatrcpfFmiq/cCs=
 github.com/andybalholm/brotli v1.0.5/go.mod h1:fO7iG3H7G2nSZ7m0zPUDn85XEX2GTukHGRSepvi9Eig=
+github.com/beorn7/perks v1.0.1 h1:VlbKKnNfV8bJzeqoa4cOKqO6bYr3WgKZxO8Z16+hsOM=
+github.com/beorn7/perks v1.0.1/go.mod h1:G2ZrVWU2WbWT9wwq4/hrbKbnv/1ERSJQ0ibhJ6rlkpw=
+github.com/cespare/xxhash/v2 v2.2.0 h1:DC2CZ1Ep5Y4k3ZQ899DldepgrayRUGE6BBZ/cd9Cj44=
+github.com/cespare/xxhash/v2 v2.2.0/go.mod h1:VGX0DQ3Q6kWi7AoAeZDth3/j3BFtOZR5XLFGgcrjCOs=
 github.com/chzyer/logex v1.1.10/go.mod h1:+Ywpsq7O8HXn0nuIou7OrIPyXbp3wmkHB+jjWRnGsAI=
 github.com/chzyer/readline v0.0.0-20180603132655-2972be24d48e/go.mod h1:nSuG5e5PlCu98SY8svDHJxuZscDgtXS6KTTbou5AhLI=
 github.com/chzyer/test v0.0.0-20180213035817-a1ea475d72b1/go.mod h1:Q3SI9o4m/ZMnBNeIyt5eFwwo7qiLfzFZmjNmxjkiQlU=
 github.com/coreos/go-systemd/v22 v22.5.0/go.mod h1:Y58oyj3AT4RCenI/lSvhwexgC+NSVTIJ3seZv2GcEnc=
 github.com/cpuguy83/go-md2man/v2 v2.0.2 h1:p1EgwI/C7NhT0JmVkwCD2ZBK8j4aeHQX2pMHHBfMQ6w=
 github.com/cpuguy83/go-md2man/v2 v2.0.2/go.mod h1:tgQtvFlXSQOSOSIRvRPT7W67SCa46tRHOmNcaadrF8o=
-github.com/creack/pty v1.1.9/go.mod h1:oKZEueFk5CKHvIhNR5MUki03XCEU+Q6VDXinZuGJ33E=
 github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
@@ -29,8 +32,11 @@ github.com/go-audio/riff v1.0.0 h1:d8iCGbDvox9BfLagY94fBynxSPHO80LmZCaOsmKxokA=
 github.com/go-audio/riff v1.0.0/go.mod h1:l3cQwc85y79NQFCRB7TiPoNiaijp6q8Z0Uv38rVG498=
 github.com/go-audio/wav v1.1.0 h1:jQgLtbqBzY7G+BM8fXF7AHUk1uHUviWS4X39d5rsL2g=
 github.com/go-audio/wav v1.1.0/go.mod h1:mpe9qfwbScEbkd8uybLuIpTgHyrISw/OTuvjUW2iGtE=
+github.com/go-logr/logr v1.2.2/go.mod h1:jdQByPbusPIv2/zmleS9BjJVeZ6kBagPoEUsqbVz/1A=
 github.com/go-logr/logr v1.2.4 h1:g01GSCwiDw2xSZfjJ2/T9M+S6pFdcNtFYsp+Y43HYDQ=
 github.com/go-logr/logr v1.2.4/go.mod h1:jdQByPbusPIv2/zmleS9BjJVeZ6kBagPoEUsqbVz/1A=
+github.com/go-logr/stdr v1.2.2 h1:hSWxHoqTgW2S2qGc0LTAI563KZ5YKYRhT3MFKZMbjag=
+github.com/go-logr/stdr v1.2.2/go.mod h1:mMo/vtBO5dYbehREoey6XUKy/eSumjCCveDpRre4VKE=
 github.com/go-ole/go-ole v1.2.6 h1:/Fpf6oFPoeFik9ty7siob0G6Ke8QvQEuVcuChpwXzpY=
 github.com/go-ole/go-ole v1.2.6/go.mod h1:pprOEPIfldk/42T2oK7lQ4v4JSDwmV0As9GaiUsvbm0=
 github.com/go-skynet/bloomz.cpp v0.0.0-20230529155654-1834e77b83fa h1:gxr68r/6EWroay4iI81jxqGCDbKotY4+CiwdUkBz2NQ=
@@ -45,8 +51,8 @@ github.com/go-task/slim-sprig v0.0.0-20210107165309-348f09dbbbc0/go.mod h1:fyg78
 github.com/go-task/slim-sprig v0.0.0-20230315185526-52ccab3ef572 h1:tfuBGBXKqDEevZMzYi5KSi8KkcZtzBcTgAUUtapy0OI=
 github.com/go-task/slim-sprig v0.0.0-20230315185526-52ccab3ef572/go.mod h1:9Pwr4B2jHnOSGXyyzV8ROjYa2ojvAY6HCGYYfMoC3Ls=
 github.com/godbus/dbus/v5 v5.0.4/go.mod h1:xhWf0FNVPg57R7Z0UbKHbJfkEywrmjJnf7w5xrFpKfA=
-github.com/gofiber/fiber/v2 v2.49.2 h1:ONEN3/Vc+dUCxxDgZZwpqvhISgHqb+bu+isBiEyKEQs=
-github.com/gofiber/fiber/v2 v2.49.2/go.mod h1:gNsKnyrmfEWFpJxQAV0qvW6l70K1dZGno12oLtukcts=
+github.com/gofiber/fiber/v2 v2.50.0 h1:ia0JaB+uw3GpNSCR5nvC5dsaxXjRU5OEu36aytx+zGw=
+github.com/gofiber/fiber/v2 v2.50.0/go.mod h1:21eytvay9Is7S6z+OgPi7c7n4++tnClWmhpimVHMimw=
 github.com/golang/protobuf v1.2.0/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
 github.com/golang/protobuf v1.4.0-rc.1/go.mod h1:ceaxUfeHdC40wWswd/P6IGgMaK3YpKi5j83Wpe3EHw8=
 github.com/golang/protobuf v1.4.0-rc.1.0.20200221234624-67d41d38c208/go.mod h1:xKAWHe0F5eneWXFV3EuXVDTCmh+JuBKY0li0aMyXATA=
@@ -67,6 +73,8 @@ github.com/google/go-cmp v0.5.5/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/
 github.com/google/go-cmp v0.5.6/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
 github.com/google/go-cmp v0.5.9 h1:O2Tfq5qg4qc4AmwVlvv0oLiVAGB7enBSJ2x2DqQFi38=
 github.com/google/go-cmp v0.5.9/go.mod h1:17dUlkBOakJ0+DkrSSNjCkIjxS6bF9zb3elmeNGIjoY=
+github.com/google/go-cmp v0.6.0 h1:ofyhxvXcZhMsU5ulbFiLKl/XBFqE1GSq7atu8tAmTRI=
+github.com/google/go-cmp v0.6.0/go.mod h1:17dUlkBOakJ0+DkrSSNjCkIjxS6bF9zb3elmeNGIjoY=
 github.com/google/gofuzz v1.0.0/go.mod h1:dBl0BpW6vV/+mYPU4Po3pmUjxk6FQPldtuIdl/M65Eg=
 github.com/google/pprof v0.0.0-20210407192527-94a9f03dee38 h1:yAJXTCF9TqKcTiHJAE8dj7HMvPfh66eeA2JYW7eFpSE=
 github.com/google/pprof v0.0.0-20210407192527-94a9f03dee38/go.mod h1:kpwsk12EmLew5upagYY7GY0pfYCcupk39gWOCRROcvE=
@@ -91,10 +99,8 @@ github.com/klauspost/compress v1.16.7/go.mod h1:ntbaceVETuRiXiv4DpjP66DpAtAGkEQs
 github.com/klauspost/cpuid v1.2.0/go.mod h1:Pj4uuM528wm8OyEC2QMXAi2YiTZ96dNQPGgoMS4s3ek=
 github.com/klauspost/pgzip v1.2.5 h1:qnWYvvKqedOF2ulHpMG72XQol4ILEJ8k2wwRl/Km8oE=
 github.com/klauspost/pgzip v1.2.5/go.mod h1:Ch1tH69qFZu15pkjo5kYi6mth2Zzwzt50oCQKQE9RUs=
-github.com/kr/pretty v0.2.1 h1:Fmg33tUaq4/8ym9TJN1x7sLJnHVwhP33CNkpYV/7rwI=
-github.com/kr/pretty v0.2.1/go.mod h1:ipq/a2n7PKx3OHsz4KJII5eveXtPO4qwEXGdVfWzfnI=
-github.com/kr/pty v1.1.1/go.mod h1:pFQYn66WHrOpPYNljwOMqo10TkYh1fy3cYio2l3bCsQ=
-github.com/kr/text v0.1.0/go.mod h1:4Jbv+DJW3UT/LiOwJeYQe1efqtUx/iVham/4vfdArNI=
+github.com/kr/pretty v0.3.1 h1:flRD4NNwYAUpkphVc1HcthR4KEIFJ65n8Mw5qdRn3LE=
+github.com/kr/pretty v0.3.1/go.mod h1:hoEshYVHaxMs3cyo3Yncou5ZscifuDolrwPKZanG3xk=
 github.com/kr/text v0.2.0 h1:5Nx0Ya0ZqY2ygV366QzturHI13Jq95ApcVaJBhpS+AY=
 github.com/kr/text v0.2.0/go.mod h1:eLer722TekiGuMkidMxC/pM04lWEeraHUUmBw8l2grE=
 github.com/lufia/plan9stats v0.0.0-20211012122336-39d0f177ccd0 h1:6E+4a0GO5zZEnZ81pIr0yLvtUWk2if982qA3F3QD6H4=
@@ -108,12 +114,15 @@ github.com/mattn/go-isatty v0.0.19/go.mod h1:W+V8PltTTMOvKvAeJH7IuucS94S2C6jfK/D
 github.com/mattn/go-runewidth v0.0.14/go.mod h1:Jdepj2loyihRzMpdS35Xk/zdY8IAYHsh153qUoGf23w=
 github.com/mattn/go-runewidth v0.0.15 h1:UNAjwbU9l54TA3KzvqLGxwWjHmMgBUVhBiTjelZgg3U=
 github.com/mattn/go-runewidth v0.0.15/go.mod h1:Jdepj2loyihRzMpdS35Xk/zdY8IAYHsh153qUoGf23w=
+github.com/matttproud/golang_protobuf_extensions v1.0.4 h1:mmDVorXM7PCGKw94cs5zkfA9PSy5pEvNWRP0ET0TIVo=
+github.com/matttproud/golang_protobuf_extensions v1.0.4/go.mod h1:BSXmuO+STAnVfrANrmjBb36TMTDstsz7MSK+HVaYKv4=
 github.com/mholt/archiver/v3 v3.5.1 h1:rDjOBX9JSF5BvoJGvjqK479aL70qh9DIpZCl+k7Clwo=
 github.com/mholt/archiver/v3 v3.5.1/go.mod h1:e3dqJ7H78uzsRSEACH1joayhuSyhnonssnDhppzS1L4=
 github.com/mitchellh/colorstring v0.0.0-20190213212951-d06e56a500db h1:62I3jR2EmQ4l5rM/4FEfDWcRD+abF5XlKShorW5LRoQ=
 github.com/mitchellh/colorstring v0.0.0-20190213212951-d06e56a500db/go.mod h1:l0dey0ia/Uv7NcFFVbCLtqEBQbrT4OCwCSKTEv6enCw=
-github.com/modern-go/concurrent v0.0.0-20180228061459-e0a39a4cb421 h1:ZqeYNhU3OHLH3mGKHDcjJRFFRrJa6eAM5H+CtDdOsPc=
 github.com/modern-go/concurrent v0.0.0-20180228061459-e0a39a4cb421/go.mod h1:6dJC0mAP4ikYIbvyc7fijjWJddQyLn8Ig3JB5CqoB9Q=
+github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd h1:TRLaZ9cD/w8PVh93nsPXa1VrQ6jlwL5oN8l14QlcNfg=
+github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd/go.mod h1:6dJC0mAP4ikYIbvyc7fijjWJddQyLn8Ig3JB5CqoB9Q=
 github.com/modern-go/reflect2 v1.0.2 h1:xBagoLtFs94CBntxluKeaWgTMpvLxC4ur3nMaC9Gz0M=
 github.com/modern-go/reflect2 v1.0.2/go.mod h1:yWuevngMOJpCy52FWWMvUC8ws7m/LJsjYzDa0/r8luk=
 github.com/mudler/go-ggllm.cpp v0.0.0-20230709223052-862477d16eef h1:OJZtJ5vYhlkTJI0RHIl62kOkhiINQEhZgsXlwmmNDhM=
@@ -124,10 +133,12 @@ github.com/mudler/go-processmanager v0.0.0-20230818213616-f204007f963c h1:CI5uGw
 github.com/mudler/go-processmanager v0.0.0-20230818213616-f204007f963c/go.mod h1:gY3wyrhkRySJtmtI/JPt4a2mKv48h/M9pEZIW+SjeC0=
 github.com/mudler/go-stable-diffusion v0.0.0-20230605122230-d89260f598af h1:XFq6OUqsWQam0OrEr05okXsJK/TQur3zoZTHbiZD3Ks=
 github.com/mudler/go-stable-diffusion v0.0.0-20230605122230-d89260f598af/go.mod h1:8ufRkpz/S/9ahkaxzZ5i4WMgO9w4InEhuRoT7vK5Rnw=
-github.com/nomic-ai/gpt4all/gpt4all-bindings/golang v0.0.0-20231012185707-10f9b49313d0 h1:cvx7GSYYXMnmLLHey5qF9QjS0nPyOk8tsozPvhwLvgs=
-github.com/nomic-ai/gpt4all/gpt4all-bindings/golang v0.0.0-20231012185707-10f9b49313d0/go.mod h1:4T3CHXyrt+7FQHXaxULZfPjHbD8/99WuDDJa0YVZARI=
 github.com/nomic-ai/gpt4all/gpt4all-bindings/golang v0.0.0-20231013181651-22de3c56bdd4 h1:82J4t94Mmt0lva/OoxNlHkKrMSdSUZXkAjTFnlFFsow=
 github.com/nomic-ai/gpt4all/gpt4all-bindings/golang v0.0.0-20231013181651-22de3c56bdd4/go.mod h1:4T3CHXyrt+7FQHXaxULZfPjHbD8/99WuDDJa0YVZARI=
+github.com/nomic-ai/gpt4all/gpt4all-bindings/golang v0.0.0-20231016205817-9a19c740ee84 h1:AiFzd+M2Uxz67fdn4nCnKR70me5yf88rXhoqhvfRDak=
+github.com/nomic-ai/gpt4all/gpt4all-bindings/golang v0.0.0-20231016205817-9a19c740ee84/go.mod h1:4T3CHXyrt+7FQHXaxULZfPjHbD8/99WuDDJa0YVZARI=
+github.com/nomic-ai/gpt4all/gpt4all-bindings/golang v0.0.0-20231022042237-c25dc5193530 h1:YXMxHwHMB9jCBo2Yu5gz3mTB3T1TnZs/HmPLv15LUSA=
+github.com/nomic-ai/gpt4all/gpt4all-bindings/golang v0.0.0-20231022042237-c25dc5193530/go.mod h1:4T3CHXyrt+7FQHXaxULZfPjHbD8/99WuDDJa0YVZARI=
 github.com/nwaples/rardecode v1.1.0 h1:vSxaY8vQhOcVr4mm5e8XllHWTiM4JF507A0Katqw7MQ=
 github.com/nwaples/rardecode v1.1.0/go.mod h1:5DzqNKiOdpKKBH87u8VlvAnPZMXcGRhxWkRpHbbfGS0=
 github.com/nxadm/tail v1.4.4/go.mod h1:kenIhsEOeOJmVchQTgglprH7qJGnHDVpk1VPCcaMI8A=
@@ -144,6 +155,8 @@ github.com/onsi/gomega v1.10.1/go.mod h1:iN09h71vgCQne3DLsj+A5owkum+a2tYe+TOCB1y
 github.com/onsi/gomega v1.16.0/go.mod h1:HnhC7FXeEQY45zxNK3PPoIUhzk/80Xly9PcubAlGdZY=
 github.com/onsi/gomega v1.28.0 h1:i2rg/p9n/UqIDAMFUJ6qIUUMcsqOuUHgbpbu235Vr1c=
 github.com/onsi/gomega v1.28.0/go.mod h1:A1H2JE76sI14WIP57LMKj7FVfCHx3g3BcZVjJG8bjX8=
+github.com/onsi/gomega v1.28.1 h1:MijcGUbfYuznzK/5R4CPNoUP/9Xvuo20sXfEm6XxoTA=
+github.com/onsi/gomega v1.28.1/go.mod h1:9sxs+SwGrKI0+PWe4Fxa9tFQQBG5xSsSbMXOI8PPpoQ=
 github.com/otiai10/mint v1.6.1 h1:kgbTJmOpp/0ce7hk3H8jiSuR0MXmpwWRfqUdKww17qg=
 github.com/otiai10/mint v1.6.1/go.mod h1:MJm72SBthJjz8qhefc4z1PYEieWmy8Bku7CjcAqyUSM=
 github.com/otiai10/openaigo v1.6.0 h1:YTQEbtDSvawETOB/Kmb/6JvuHdHH/eIpSQfHVufiwY8=
@@ -159,15 +172,23 @@ github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZb
 github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
 github.com/power-devops/perfstat v0.0.0-20210106213030-5aafc221ea8c h1:ncq/mPwQF4JjgDlrVEn3C11VoGHZN7m8qihwgMEtzYw=
 github.com/power-devops/perfstat v0.0.0-20210106213030-5aafc221ea8c/go.mod h1:OmDBASR4679mdNQnz2pUhc2G8CO2JrUAVFDRBDP/hJE=
+github.com/prometheus/client_golang v1.17.0 h1:rl2sfwZMtSthVU752MqfjQozy7blglC+1SOtjMAMh+Q=
+github.com/prometheus/client_golang v1.17.0/go.mod h1:VeL+gMmOAxkS2IqfCq0ZmHSL+LjWfWDUmp1mBz9JgUY=
+github.com/prometheus/client_model v0.4.1-0.20230718164431-9a2bf3000d16 h1:v7DLqVdK4VrYkVD5diGdl4sxJurKJEMnODWRJlxV9oM=
+github.com/prometheus/client_model v0.4.1-0.20230718164431-9a2bf3000d16/go.mod h1:oMQmHW1/JoDwqLtg57MGgP/Fb1CJEYF2imWWhWtMkYU=
+github.com/prometheus/common v0.44.0 h1:+5BrQJwiBB9xsMygAB3TNvpQKOwlkc25LbISbrdOOfY=
+github.com/prometheus/common v0.44.0/go.mod h1:ofAIvZbQ1e/nugmZGz4/qCb9Ap1VoSTIO7x0VV9VvuY=
+github.com/prometheus/procfs v0.11.1 h1:xRC8Iq1yyca5ypa9n1EZnWZkt7dwcoRPQwX/5gwaUuI=
+github.com/prometheus/procfs v0.11.1/go.mod h1:eesXgaPo1q7lBpVMoMy0ZOFTth9hBn4W/y0/p/ScXhY=
 github.com/rivo/uniseg v0.2.0 h1:S1pD9weZBuJdFmowNwbpi7BJ8TNftyUImj/0WQi72jY=
 github.com/rivo/uniseg v0.2.0/go.mod h1:J6wj4VEh+S6ZtnVlnTBMWIodfgj8LQOQFoIToxlJtxc=
+github.com/rogpeppe/go-internal v1.11.0 h1:cWPaGQEPrBb5/AsnsZesgZZ9yb1OQ+GOISoDNXVBh4M=
+github.com/rogpeppe/go-internal v1.11.0/go.mod h1:ddIwULY96R17DhadqLgMfk9H9tvdUzkipdSkR5nkCZA=
 github.com/rs/xid v1.5.0/go.mod h1:trrq9SKmegXys3aeAKXMUTdJsYXVwGY3RLcfgqegfbg=
 github.com/rs/zerolog v1.31.0 h1:FcTR3NnLWW+NnTwwhFWiJSZr4ECLpqCm6QsEnyvbV4A=
 github.com/rs/zerolog v1.31.0/go.mod h1:/7mN4D5sKwJLZQ2b/znpjC3/GQWY/xaDXUM0kKWRHss=
 github.com/russross/blackfriday/v2 v2.1.0 h1:JIOH55/0cWyOuilr9/qlrm0BSXldqnqwMsf35Ld67mk=
 github.com/russross/blackfriday/v2 v2.1.0/go.mod h1:+Rmxgy9KzJVeS9/2gXHxylqXiyQDYRxCVz55jmeOWTM=
-github.com/sashabaranov/go-openai v1.15.4 h1:BXCR0Uxk5RipeY4yBC7g6pBVfcjh8jwrMNOYdie6yuk=
-github.com/sashabaranov/go-openai v1.15.4/go.mod h1:lj5b/K+zjTSFxVLijLSTDZuP7adOgerWeFyZLUhAKRg=
 github.com/sashabaranov/go-openai v1.16.0 h1:34W6WV84ey6OpW0p2UewZkdMu82AxGC+BzpU6iiauRw=
 github.com/sashabaranov/go-openai v1.16.0/go.mod h1:lj5b/K+zjTSFxVLijLSTDZuP7adOgerWeFyZLUhAKRg=
 github.com/schollz/progressbar/v3 v3.13.1 h1:o8rySDYiQ59Mwzy2FELeHY5ZARXZTVJC7iHD6PEFUiE=
@@ -195,8 +216,10 @@ github.com/tklauser/go-sysconf v0.3.12/go.mod h1:Ho14jnntGE1fpdOqQEEaiKRpvIavV0h
 github.com/tklauser/numcpus v0.6.0/go.mod h1:FEZLMke0lhOUG6w2JadTzp0a+Nl8PF/GFkQ5UVIcaL4=
 github.com/tklauser/numcpus v0.6.1 h1:ng9scYS7az0Bk4OZLvrNXNSAO2Pxr1XXRAPyjhIx+Fk=
 github.com/tklauser/numcpus v0.6.1/go.mod h1:1XfjsgE2zo8GVw7POkMbHENHzVg3GzmoZ9fESEdAacY=
-github.com/tmc/langchaingo v0.0.0-20230929160525-e16b77704b8d h1:i4+wYULVM2/3Yb/aDE7Z4s2v7vqtQERQWh5lopBEuig=
-github.com/tmc/langchaingo v0.0.0-20230929160525-e16b77704b8d/go.mod h1:R+a8fqt6nmKyYj7KSpr/m9oxqE6OJLbLyO9pxeHpjLU=
+github.com/tmc/langchaingo v0.0.0-20231016073620-a02d4fdc0f3a h1:BziGpoF5ZVWMDy6Z1adXnYndRye2fiYWZlmknUFksGA=
+github.com/tmc/langchaingo v0.0.0-20231016073620-a02d4fdc0f3a/go.mod h1:SiwyRS7sBSSi6f3NB4dKENw69X6br/wZ2WRkM+8pZWk=
+github.com/tmc/langchaingo v0.0.0-20231019140956-c636b3da7701 h1:LquLgmFiKf6eDXdwoUKCIGn5NsR34cLXC6ySYhiE6bA=
+github.com/tmc/langchaingo v0.0.0-20231019140956-c636b3da7701/go.mod h1:SiwyRS7sBSSi6f3NB4dKENw69X6br/wZ2WRkM+8pZWk=
 github.com/ulikunitz/xz v0.5.8/go.mod h1:nbz6k7qbPmH4IRqmfOplQw/tblSgqTqBwxkY0oWt/14=
 github.com/ulikunitz/xz v0.5.9 h1:RsKRIA2MO8x56wkkcd3LbtcE/uMszhb6DpRf+3uwa3I=
 github.com/ulikunitz/xz v0.5.9/go.mod h1:nbz6k7qbPmH4IRqmfOplQw/tblSgqTqBwxkY0oWt/14=
@@ -215,6 +238,18 @@ github.com/xrash/smetrics v0.0.0-20201216005158-039620a65673/go.mod h1:N3UwUGtsr
 github.com/yuin/goldmark v1.2.1/go.mod h1:3hX8gzYuyVAZsxl0MRgGTJEmQBFcNTphYh9decYSb74=
 github.com/yusufpapurcu/wmi v1.2.3 h1:E1ctvB7uKFMOJw3fdOW32DwGE9I7t++CRUEMKvFoFiw=
 github.com/yusufpapurcu/wmi v1.2.3/go.mod h1:SBZ9tNy3G9/m5Oi98Zks0QjeHVDvuK0qfxQmPyzfmi0=
+go.opentelemetry.io/otel v1.19.0 h1:MuS/TNf4/j4IXsZuJegVzI1cwut7Qc00344rgH7p8bs=
+go.opentelemetry.io/otel v1.19.0/go.mod h1:i0QyjOq3UPoTzff0PJB2N66fb4S0+rSbSB15/oyH9fY=
+go.opentelemetry.io/otel/exporters/prometheus v0.42.0 h1:jwV9iQdvp38fxXi8ZC+lNpxjK16MRcZlpDYvbuO1FiA=
+go.opentelemetry.io/otel/exporters/prometheus v0.42.0/go.mod h1:f3bYiqNqhoPxkvI2LrXqQVC546K7BuRDL/kKuxkujhA=
+go.opentelemetry.io/otel/metric v1.19.0 h1:aTzpGtV0ar9wlV4Sna9sdJyII5jTVJEvKETPiOKwvpE=
+go.opentelemetry.io/otel/metric v1.19.0/go.mod h1:L5rUsV9kM1IxCj1MmSdS+JQAcVm319EUrDVLrt7jqt8=
+go.opentelemetry.io/otel/sdk v1.19.0 h1:6USY6zH+L8uMH8L3t1enZPR3WFEmSTADlqldyHtJi3o=
+go.opentelemetry.io/otel/sdk v1.19.0/go.mod h1:NedEbbS4w3C6zElbLdPJKOpJQOrGUJ+GfzpjUvI0v1A=
+go.opentelemetry.io/otel/sdk/metric v1.19.0 h1:EJoTO5qysMsYCa+w4UghwFV/ptQgqSL/8Ni+hx+8i1k=
+go.opentelemetry.io/otel/sdk/metric v1.19.0/go.mod h1:XjG0jQyFJrv2PbMvwND7LwCEhsJzCzV5210euduKcKY=
+go.opentelemetry.io/otel/trace v1.19.0 h1:DFVQmlVbfVeOuBRrwdtaehRrWiL1JoVs9CPIQ1Dzxpg=
+go.opentelemetry.io/otel/trace v1.19.0/go.mod h1:mfaSyvGyEJEI0nyV2I4qhNQnbBOUUmYZpYojqMnX2vo=
 golang.org/x/crypto v0.0.0-20190308221718-c2843e01d9a2/go.mod h1:djNgcEr1/C05ACkg1iLfiJU5Ep61QUkGW8qpdssI0+w=
 golang.org/x/crypto v0.0.0-20191011191535-87dc89f01550/go.mod h1:yigFU9vqHzYiE8UmvKecakEJjdnWj3jj499lnFckfCI=
 golang.org/x/crypto v0.0.0-20200622213623-75b288015ac9/go.mod h1:LzIPMQfyMNhhGPhUkYOs5KpL4U8rLKemX1yGLhDgUto=
@@ -227,9 +262,10 @@ golang.org/x/net v0.0.0-20190620200207-3b0461eec859/go.mod h1:z5CRVTTTmAJ677TzLL
 golang.org/x/net v0.0.0-20200520004742-59133d7f0dd7/go.mod h1:qpuaurCH72eLCgpAm/N6yyVIVM9cpaDIP3A8BGJEC5A=
 golang.org/x/net v0.0.0-20201021035429-f5854403a974/go.mod h1:sp8m0HH+o8qH0wwXwYZr8TS3Oi6o0r6Gce1SSxlDquU=
 golang.org/x/net v0.0.0-20210428140749-89ef3d95e781/go.mod h1:OJAsFXCWl8Ukc7SiCT/9KSuxbyM7479/AVlXFRxuMCk=
-golang.org/x/net v0.14.0 h1:BONx9s002vGdD9umnlX1Po8vOZmrgH34qlHcD1MfK14=
-golang.org/x/net v0.14.0/go.mod h1:PpSgVXXLK0OxS0F31C1/tv6XNguvCrnXIDrFMspZIUI=
+golang.org/x/net v0.17.0 h1:pVaXccu2ozPjCXewfr1S7xza/zcXTity9cCdXQYSjIM=
+golang.org/x/net v0.17.0/go.mod h1:NxSsAGuq816PNPmqtQdLE42eU2Fs7NoRIZrHJAlaCOE=
 golang.org/x/sync v0.0.0-20180314180146-1d60e4601c6f/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20181221193216-37e7f081c4d4/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sync v0.0.0-20190423024810-112230192c58/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sync v0.0.0-20201020160332-67f06af15bc9/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sys v0.0.0-20180909124046-d0be0721c37e/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
@@ -252,17 +288,18 @@ golang.org/x/sys v0.6.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.8.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.10.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.11.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.12.0 h1:CM0HF96J0hcLAwsHPJZjfdNzs0gftsLfgKt57wWHJ0o=
 golang.org/x/sys v0.12.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/sys v0.13.0 h1:Af8nKPmuFypiUBjVoU9V20FiaFXOcuZI21p0ycVYYGE=
+golang.org/x/sys v0.13.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/term v0.0.0-20201126162022-7de9c90e9dd1/go.mod h1:bj7SfCRtBDWHUb9snDiAeCFNEtKQo2Wmx5Cou7ajbmo=
 golang.org/x/term v0.6.0/go.mod h1:m6U89DPEgQRMq3DNkDClhWw02AUbt2daBVO4cn4Hv9U=
-golang.org/x/term v0.11.0 h1:F9tnn/DA/Im8nCwm+fX+1/eBwi4qFjRT++MhtVC4ZX0=
-golang.org/x/term v0.11.0/go.mod h1:zC9APTIj3jG3FdV/Ons+XE1riIZXG4aZ4GTHiPZJPIU=
+golang.org/x/term v0.13.0 h1:bb+I9cTfFazGW51MZqBVmZy7+JEJMouUHTUSKVQLBek=
+golang.org/x/term v0.13.0/go.mod h1:LTmsnFJwVN6bCy1rVCoS+qHT1HhALEFxKncY3WNNh4U=
 golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
 golang.org/x/text v0.3.3/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
 golang.org/x/text v0.3.6/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
-golang.org/x/text v0.12.0 h1:k+n5B8goJNdU7hSvEtMUz3d1Q6D/XW4COJSJR6fN0mc=
-golang.org/x/text v0.12.0/go.mod h1:TvPlkZtksWOMsz7fbANvkp4WM8x/WCo/om8BMLbz+aE=
+golang.org/x/text v0.13.0 h1:ablQoSUd0tRdKxZewP80B+BaqeKJuVhuRxj/dkrun3k=
+golang.org/x/text v0.13.0/go.mod h1:TvPlkZtksWOMsz7fbANvkp4WM8x/WCo/om8BMLbz+aE=
 golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
 golang.org/x/tools v0.0.0-20191119224855-298f0cb1881e/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
 golang.org/x/tools v0.0.0-20201224043029-2b0845dc783e/go.mod h1:emZCQorbCU4vsT4fOWvOPXz4eW1wZW4PmDk9uLelYpA=
@@ -274,8 +311,12 @@ golang.org/x/xerrors v0.0.0-20191204190536-9bdfabe68543/go.mod h1:I/5z698sn9Ka8T
 golang.org/x/xerrors v0.0.0-20200804184101-5ec99f83aff1/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
 google.golang.org/genproto/googleapis/rpc v0.0.0-20230711160842-782d3b101e98 h1:bVf09lpb+OJbByTj913DRJioFFAjf/ZGxEz7MajTp2U=
 google.golang.org/genproto/googleapis/rpc v0.0.0-20230711160842-782d3b101e98/go.mod h1:TUfxEVdsvPg18p6AslUXFoLdpED4oBnGwyqk3dV1XzM=
+google.golang.org/genproto/googleapis/rpc v0.0.0-20230822172742-b8732ec3820d h1:uvYuEyMHKNt+lT4K3bN6fGswmK8qSvcreM3BwjDh+y4=
+google.golang.org/genproto/googleapis/rpc v0.0.0-20230822172742-b8732ec3820d/go.mod h1:+Bk1OCOj40wS2hwAMA+aCW9ypzm63QTBBHp6lQ3p+9M=
 google.golang.org/grpc v1.58.3 h1:BjnpXut1btbtgN/6sp+brB2Kbm2LjNXnidYujAVbSoQ=
 google.golang.org/grpc v1.58.3/go.mod h1:tgX3ZQDlNJGU96V6yHh1T/JeoBQ2TXdr43YbYSsCJk0=
+google.golang.org/grpc v1.59.0 h1:Z5Iec2pjwb+LEOqzpB2MR12/eKFhDPhuqW91O+4bwUk=
+google.golang.org/grpc v1.59.0/go.mod h1:aUPDwccQo6OTjy7Hct4AfBPD1GptF4fyUjIkQ9YtF98=
 google.golang.org/protobuf v0.0.0-20200109180630-ec00e32a8dfd/go.mod h1:DFci5gLYBciE7Vtevhsrf46CRTquxDuWsQurQQe4oz8=
 google.golang.org/protobuf v0.0.0-20200221191635-4d8936d0db64/go.mod h1:kwYJMbMJ01Woi6D6+Kah6886xMZcty6N08ah7+eCXa0=
 google.golang.org/protobuf v0.0.0-20200228230310-ab0ca4ff8a60/go.mod h1:cfTl7dwQJ+fmap5saPgwCLgHXTUD7jkjRqWcaiX5VyM=

main.go

@@ -18,6 +18,7 @@ import (
 	"github.com/go-skynet/LocalAI/internal"
 	"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"
 	"github.com/rs/zerolog/log"
 	progressbar "github.com/schollz/progressbar/v3"
@@ -214,6 +215,12 @@ For a list of compatible model, check out: https://localai.io/model-compatibilit
 				return err
 			}
 
+			metrics, err := metrics.SetupMetrics()
+			if err != nil {
+				return err
+			}
+			opts = append(opts, options.WithMetrics(metrics))
+
 			app, err := api.App(opts...)
 			if err != nil {
 				return err

metrics/metrics.go

@@ -0,0 +1,83 @@
+package metrics
+
+import (
+	"context"
+	"time"
+
+	"github.com/gofiber/fiber/v2"
+	"github.com/gofiber/fiber/v2/middleware/adaptor"
+	"github.com/prometheus/client_golang/prometheus/promhttp"
+	"go.opentelemetry.io/otel/attribute"
+	"go.opentelemetry.io/otel/exporters/prometheus"
+	api "go.opentelemetry.io/otel/metric"
+	"go.opentelemetry.io/otel/sdk/metric"
+)
+
+type Metrics struct {
+	meter         api.Meter
+	apiTimeMetric api.Float64Histogram
+}
+
+// setupOTelSDK bootstraps the OpenTelemetry pipeline.
+// If it does not return an error, make sure to call shutdown for proper cleanup.
+func SetupMetrics() (*Metrics, error) {
+	exporter, err := prometheus.New()
+	if err != nil {
+		return nil, err
+	}
+	provider := metric.NewMeterProvider(metric.WithReader(exporter))
+	meter := provider.Meter("github.com/go-skynet/LocalAI")
+
+	apiTimeMetric, err := meter.Float64Histogram("api_call", api.WithDescription("api calls"))
+	if err != nil {
+		return nil, err
+	}
+
+	return &Metrics{
+		meter:         meter,
+		apiTimeMetric: apiTimeMetric,
+	}, nil
+}
+
+func MetricsHandler() fiber.Handler {
+	return adaptor.HTTPHandler(promhttp.Handler())
+}
+
+type apiMiddlewareConfig struct {
+	Filter  func(c *fiber.Ctx) bool
+	metrics *Metrics
+}
+
+func APIMiddleware(metrics *Metrics) fiber.Handler {
+	cfg := apiMiddlewareConfig{
+		metrics: metrics,
+		Filter: func(c *fiber.Ctx) bool {
+			if c.Path() == "/metrics" {
+				return true
+			}
+			return false
+		},
+	}
+
+	return func(c *fiber.Ctx) error {
+		if cfg.Filter != nil && cfg.Filter(c) {
+			return c.Next()
+		}
+		path := c.Path()
+		method := c.Method()
+
+		start := time.Now()
+		err := c.Next()
+		elapsed := float64(time.Since(start)) / float64(time.Second)
+		cfg.metrics.ObserveAPICall(method, path, elapsed)
+		return err
+	}
+}
+
+func (m *Metrics) ObserveAPICall(method string, path string, duration float64) {
+	opts := api.WithAttributes(
+		attribute.String("method", method),
+		attribute.String("path", path),
+	)
+	m.apiTimeMetric.Record(context.Background(), duration, opts)
+}

pkg/backend/llm/bloomz/bloomz.go

@@ -1,59 +0,0 @@
-package bloomz
-
-// 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 (
-	"fmt"
-
-	"github.com/go-skynet/LocalAI/pkg/grpc/base"
-	pb "github.com/go-skynet/LocalAI/pkg/grpc/proto"
-
-	"github.com/go-skynet/bloomz.cpp"
-)
-
-type LLM struct {
-	base.SingleThread
-
-	bloomz *bloomz.Bloomz
-}
-
-func (llm *LLM) Load(opts *pb.ModelOptions) error {
-	model, err := bloomz.New(opts.ModelFile)
-	llm.bloomz = model
-	return err
-}
-
-func buildPredictOptions(opts *pb.PredictOptions) []bloomz.PredictOption {
-	predictOptions := []bloomz.PredictOption{
-		bloomz.SetTemperature(float64(opts.Temperature)),
-		bloomz.SetTopP(float64(opts.TopP)),
-		bloomz.SetTopK(int(opts.TopK)),
-		bloomz.SetTokens(int(opts.Tokens)),
-		bloomz.SetThreads(int(opts.Threads)),
-	}
-
-	if opts.Seed != 0 {
-		predictOptions = append(predictOptions, bloomz.SetSeed(int(opts.Seed)))
-	}
-
-	return predictOptions
-}
-
-func (llm *LLM) Predict(opts *pb.PredictOptions) (string, error) {
-	return llm.bloomz.Predict(opts.Prompt, buildPredictOptions(opts)...)
-}
-
-// fallback to Predict
-func (llm *LLM) PredictStream(opts *pb.PredictOptions, results chan string) error {
-	go func() {
-		res, err := llm.bloomz.Predict(opts.Prompt, buildPredictOptions(opts)...)
-
-		if err != nil {
-			fmt.Println("err: ", err)
-		}
-		results <- res
-		close(results)
-	}()
-
-	return nil
-}

pkg/backend/llm/falcon/falcon.go

@@ -1,145 +0,0 @@
-package falcon
-
-// 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 (
-	"fmt"
-
-	"github.com/go-skynet/LocalAI/pkg/grpc/base"
-	pb "github.com/go-skynet/LocalAI/pkg/grpc/proto"
-
-	ggllm "github.com/mudler/go-ggllm.cpp"
-)
-
-type LLM struct {
-	base.SingleThread
-
-	falcon *ggllm.Falcon
-}
-
-func (llm *LLM) Load(opts *pb.ModelOptions) error {
-	ggllmOpts := []ggllm.ModelOption{}
-	if opts.ContextSize != 0 {
-		ggllmOpts = append(ggllmOpts, ggllm.SetContext(int(opts.ContextSize)))
-	}
-	// F16 doesn't seem to produce good output at all!
-	//if c.F16 {
-	//	llamaOpts = append(llamaOpts, llama.EnableF16Memory)
-	//}
-
-	if opts.NGPULayers != 0 {
-		ggllmOpts = append(ggllmOpts, ggllm.SetGPULayers(int(opts.NGPULayers)))
-	}
-
-	ggllmOpts = append(ggllmOpts, ggllm.SetMMap(opts.MMap))
-	ggllmOpts = append(ggllmOpts, ggllm.SetMainGPU(opts.MainGPU))
-	ggllmOpts = append(ggllmOpts, ggllm.SetTensorSplit(opts.TensorSplit))
-	if opts.NBatch != 0 {
-		ggllmOpts = append(ggllmOpts, ggllm.SetNBatch(int(opts.NBatch)))
-	} else {
-		ggllmOpts = append(ggllmOpts, ggllm.SetNBatch(512))
-	}
-
-	model, err := ggllm.New(opts.ModelFile, ggllmOpts...)
-	llm.falcon = model
-	return err
-}
-
-func buildPredictOptions(opts *pb.PredictOptions) []ggllm.PredictOption {
-	predictOptions := []ggllm.PredictOption{
-		ggllm.SetTemperature(float64(opts.Temperature)),
-		ggllm.SetTopP(float64(opts.TopP)),
-		ggllm.SetTopK(int(opts.TopK)),
-		ggllm.SetTokens(int(opts.Tokens)),
-		ggllm.SetThreads(int(opts.Threads)),
-	}
-
-	if opts.PromptCacheAll {
-		predictOptions = append(predictOptions, ggllm.EnablePromptCacheAll)
-	}
-
-	if opts.PromptCacheRO {
-		predictOptions = append(predictOptions, ggllm.EnablePromptCacheRO)
-	}
-
-	// Expected absolute path
-	if opts.PromptCachePath != "" {
-		predictOptions = append(predictOptions, ggllm.SetPathPromptCache(opts.PromptCachePath))
-	}
-
-	if opts.Mirostat != 0 {
-		predictOptions = append(predictOptions, ggllm.SetMirostat(int(opts.Mirostat)))
-	}
-
-	if opts.MirostatETA != 0 {
-		predictOptions = append(predictOptions, ggllm.SetMirostatETA(float64(opts.MirostatETA)))
-	}
-
-	if opts.MirostatTAU != 0 {
-		predictOptions = append(predictOptions, ggllm.SetMirostatTAU(float64(opts.MirostatTAU)))
-	}
-
-	if opts.Debug {
-		predictOptions = append(predictOptions, ggllm.Debug)
-	}
-
-	predictOptions = append(predictOptions, ggllm.SetStopWords(opts.StopPrompts...))
-
-	if opts.PresencePenalty != 0 {
-		predictOptions = append(predictOptions, ggllm.SetPenalty(float64(opts.PresencePenalty)))
-	}
-
-	if opts.NKeep != 0 {
-		predictOptions = append(predictOptions, ggllm.SetNKeep(int(opts.NKeep)))
-	}
-
-	if opts.Batch != 0 {
-		predictOptions = append(predictOptions, ggllm.SetBatch(int(opts.Batch)))
-	}
-
-	if opts.IgnoreEOS {
-		predictOptions = append(predictOptions, ggllm.IgnoreEOS)
-	}
-
-	if opts.Seed != 0 {
-		predictOptions = append(predictOptions, ggllm.SetSeed(int(opts.Seed)))
-	}
-
-	//predictOptions = append(predictOptions, llama.SetLogitBias(c.Seed))
-
-	predictOptions = append(predictOptions, ggllm.SetFrequencyPenalty(float64(opts.FrequencyPenalty)))
-	predictOptions = append(predictOptions, ggllm.SetMlock(opts.MLock))
-	predictOptions = append(predictOptions, ggllm.SetMemoryMap(opts.MMap))
-	predictOptions = append(predictOptions, ggllm.SetPredictionMainGPU(opts.MainGPU))
-	predictOptions = append(predictOptions, ggllm.SetPredictionTensorSplit(opts.TensorSplit))
-	predictOptions = append(predictOptions, ggllm.SetTailFreeSamplingZ(float64(opts.TailFreeSamplingZ)))
-	predictOptions = append(predictOptions, ggllm.SetTypicalP(float64(opts.TypicalP)))
-	return predictOptions
-}
-
-func (llm *LLM) Predict(opts *pb.PredictOptions) (string, error) {
-	return llm.falcon.Predict(opts.Prompt, buildPredictOptions(opts)...)
-}
-
-func (llm *LLM) PredictStream(opts *pb.PredictOptions, results chan string) error {
-
-	predictOptions := buildPredictOptions(opts)
-
-	predictOptions = append(predictOptions, ggllm.SetTokenCallback(func(token string) bool {
-		if token == "<|endoftext|>" {
-			return true
-		}
-		results <- token
-		return true
-	}))
-
-	go func() {
-		_, err := llm.falcon.Predict(opts.Prompt, predictOptions...)
-		if err != nil {
-			fmt.Println("err: ", err)
-		}
-		close(results)
-	}()
-
-	return nil
-}