Merge https://github.com/ollama/ollama

Signed-off-by: baalajimaestro <me@baalajimaestro.me>

Dockerfile

@@ -1,217 +1,61 @@
-ARG GOLANG_VERSION=1.22.5
+# Build stage
-ARG CMAKE_VERSION=3.22.1
+FROM golang:1.22-bookworm as build
-ARG CUDA_VERSION_11=11.3.1
-ARG CUDA_V11_ARCHITECTURES="50;52;53;60;61;62;70;72;75;80;86"
-ARG CUDA_VERSION_12=12.4.0
-ARG CUDA_V12_ARCHITECTURES="60;61;62;70;72;75;80;86;87;89;90;90a"
-ARG ROCM_VERSION=6.1.2
 
-# Copy the minimal context we need to run the generate scripts
+# Install necessary dependencies
-FROM scratch AS llm-code
+RUN apt update && apt install -y \
-COPY .git .git
+    wget \
-COPY .gitmodules .gitmodules
+    gnupg \
-COPY llm llm
+    software-properties-common \
+    git \
+    apt-utils
 
-FROM --platform=linux/amd64 nvidia/cuda:$CUDA_VERSION_11-devel-centos7 AS cuda-11-build-amd64
+# Install Intel oneAPI
-ARG CMAKE_VERSION
+RUN wget -O- https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB | gpg --dearmor | tee /usr/share/keyrings/oneapi-archive-keyring.gpg > /dev/null && \
-COPY ./scripts/rh_linux_deps.sh /
+    echo "deb [signed-by=/usr/share/keyrings/oneapi-archive-keyring.gpg] https://apt.repos.intel.com/oneapi all main" | tee /etc/apt/sources.list.d/oneAPI.list && \
-RUN CMAKE_VERSION=${CMAKE_VERSION} sh /rh_linux_deps.sh
+    apt update && \
-ENV PATH /opt/rh/devtoolset-10/root/usr/bin:$PATH
+    apt install -y --no-install-recommends \
-COPY --from=llm-code / /go/src/github.com/ollama/ollama/
+    intel-oneapi-mkl \
-WORKDIR /go/src/github.com/ollama/ollama/llm/generate
+    intel-oneapi-compiler-dpcpp-cpp \
-ARG CGO_CFLAGS
+    intel-oneapi-mkl-devel \
-ARG CUDA_V11_ARCHITECTURES
+    gcc \
-ENV GOARCH amd64 
+    g++ \
-RUN --mount=type=cache,target=/root/.ccache \
+    pkg-config \
-    OLLAMA_SKIP_STATIC_GENERATE=1 \
+    cmake
-    OLLAMA_SKIP_CPU_GENERATE=1 \
-    CMAKE_CUDA_ARCHITECTURES="${CUDA_V11_ARCHITECTURES}" \
-    CUDA_VARIANT="_v11" \
-    bash gen_linux.sh
 
-FROM --platform=linux/amd64 nvidia/cuda:$CUDA_VERSION_12-devel-centos7 AS cuda-12-build-amd64
+WORKDIR /app
-ARG CMAKE_VERSION
-COPY ./scripts/rh_linux_deps.sh /
-RUN CMAKE_VERSION=${CMAKE_VERSION} sh /rh_linux_deps.sh
-ENV PATH /opt/rh/devtoolset-10/root/usr/bin:$PATH
-COPY --from=llm-code / /go/src/github.com/ollama/ollama/
-WORKDIR /go/src/github.com/ollama/ollama/llm/generate
-ARG CGO_CFLAGS
-ARG CUDA_V12_ARCHITECTURES
-ENV GOARCH amd64 
-RUN --mount=type=cache,target=/root/.ccache \
-    OLLAMA_SKIP_STATIC_GENERATE=1 \
-    OLLAMA_SKIP_CPU_GENERATE=1 \
-    CMAKE_CUDA_ARCHITECTURES="${CUDA_V12_ARCHITECTURES}" \
-    CUDA_VARIANT="_v12" \
-    OLLAMA_CUSTOM_CUDA_DEFS="-DGGML_CUDA_USE_GRAPHS=on" \
-    bash gen_linux.sh
 
-FROM --platform=linux/arm64 nvidia/cuda:$CUDA_VERSION_11-devel-rockylinux8 AS cuda-11-build-server-arm64
+ARG GIN_MODE=release
-ARG CMAKE_VERSION
-COPY ./scripts/rh_linux_deps.sh /
-RUN CMAKE_VERSION=${CMAKE_VERSION} sh /rh_linux_deps.sh
-ENV PATH /opt/rh/gcc-toolset-10/root/usr/bin:$PATH
-COPY --from=llm-code / /go/src/github.com/ollama/ollama/
-WORKDIR /go/src/github.com/ollama/ollama/llm/generate
-ARG CGO_CFLAGS
-ARG CUDA_V11_ARCHITECTURES
-ENV GOARCH arm64 
-RUN OLLAMA_SKIP_STATIC_GENERATE=1 \
-    OLLAMA_SKIP_CPU_GENERATE=1 \
-    CMAKE_CUDA_ARCHITECTURES="${CUDA_V11_ARCHITECTURES}" \
-    CUDA_VARIANT="_v11" \
-    bash gen_linux.sh
 
-FROM --platform=linux/arm64 nvidia/cuda:$CUDA_VERSION_12-devel-rockylinux8 AS cuda-12-build-server-arm64
+ADD . .
-ARG CMAKE_VERSION
-COPY ./scripts/rh_linux_deps.sh /
-RUN CMAKE_VERSION=${CMAKE_VERSION} sh /rh_linux_deps.sh
-ENV PATH /opt/rh/gcc-toolset-10/root/usr/bin:$PATH
-COPY --from=llm-code / /go/src/github.com/ollama/ollama/
-WORKDIR /go/src/github.com/ollama/ollama/llm/generate
-ARG CGO_CFLAGS
-ARG CUDA_V12_ARCHITECTURES
-ENV GOARCH arm64 
-RUN --mount=type=cache,target=/root/.ccache \
-    OLLAMA_SKIP_STATIC_GENERATE=1 \
-    OLLAMA_SKIP_CPU_GENERATE=1 \
-    CMAKE_CUDA_ARCHITECTURES="${CUDA_V12_ARCHITECTURES}" \
-    CUDA_VARIANT="_v12" \
-    OLLAMA_CUSTOM_CUDA_DEFS="-DGGML_CUDA_USE_GRAPHS=on" \
-    bash gen_linux.sh
 
+RUN . /opt/intel/oneapi/setvars.sh && \
+    OLLAMA_CUSTOM_CPU_DEFS="-DGGML_BLAS=ON -DGGML_BLAS_VENDOR=Intel10_64lp -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx -DGGML_NATIVE=ON" go generate ./... && \
+    go build -ldflags="-s -w"
 
-FROM --platform=linux/amd64 rocm/dev-centos-7:${ROCM_VERSION}-complete AS rocm-build-amd64
+FROM debian:bookworm-slim
-ARG CMAKE_VERSION
-COPY ./scripts/rh_linux_deps.sh /
-RUN CMAKE_VERSION=${CMAKE_VERSION} sh /rh_linux_deps.sh
-ENV PATH /opt/rh/devtoolset-10/root/usr/bin:$PATH
-ENV LIBRARY_PATH /opt/amdgpu/lib64
-COPY --from=llm-code / /go/src/github.com/ollama/ollama/
-WORKDIR /go/src/github.com/ollama/ollama/llm/generate
-ARG CGO_CFLAGS
-ARG AMDGPU_TARGETS
-ENV GOARCH amd64 
-RUN --mount=type=cache,target=/root/.ccache \
-    OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_SKIP_CPU_GENERATE=1 bash gen_linux.sh
-RUN mkdir -p ../../dist/linux-amd64-rocm/lib/ollama && \
-    (cd /opt/rocm/lib && tar cf - rocblas/library) | (cd ../../dist/linux-amd64-rocm/lib/ollama && tar xf - )
 
-FROM --platform=linux/amd64 centos:7 AS cpu-builder-amd64
+RUN apt update && apt install -y --no-install-recommends \
-ARG CMAKE_VERSION
+    wget \
-ARG GOLANG_VERSION
+    gnupg \
-COPY ./scripts/rh_linux_deps.sh /
+    software-properties-common \
-RUN CMAKE_VERSION=${CMAKE_VERSION} GOLANG_VERSION=${GOLANG_VERSION} sh /rh_linux_deps.sh
+    supervisor && \
-ENV PATH /opt/rh/devtoolset-10/root/usr/bin:$PATH
+    wget -O- https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB | gpg --dearmor | tee /usr/share/keyrings/oneapi-archive-keyring.gpg > /dev/null && \
-COPY --from=llm-code / /go/src/github.com/ollama/ollama/
+    echo "deb [signed-by=/usr/share/keyrings/oneapi-archive-keyring.gpg] https://apt.repos.intel.com/oneapi all main" | tee /etc/apt/sources.list.d/oneAPI.list && \
-ARG OLLAMA_CUSTOM_CPU_DEFS
+    apt update && \
-ARG CGO_CFLAGS
+    apt install -y --no-install-recommends intel-oneapi-runtime-libs && \
-ENV GOARCH amd64 
+    apt clean && \
-WORKDIR /go/src/github.com/ollama/ollama/llm/generate
+    rm -rf /var/lib/apt/lists/*
 
-FROM --platform=linux/amd64 cpu-builder-amd64 AS static-build-amd64
+WORKDIR /app
-RUN --mount=type=cache,target=/root/.ccache \
-    OLLAMA_CPU_TARGET="static" bash gen_linux.sh
-FROM --platform=linux/amd64 cpu-builder-amd64 AS cpu-build-amd64
-RUN --mount=type=cache,target=/root/.ccache \
-    OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_CPU_TARGET="cpu" bash gen_linux.sh
-FROM --platform=linux/amd64 cpu-builder-amd64 AS cpu_avx-build-amd64
-RUN --mount=type=cache,target=/root/.ccache \
-    OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_CPU_TARGET="cpu_avx" bash gen_linux.sh
-FROM --platform=linux/amd64 cpu-builder-amd64 AS cpu_avx2-build-amd64
-RUN --mount=type=cache,target=/root/.ccache \
-    OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_CPU_TARGET="cpu_avx2" bash gen_linux.sh
 
-FROM --platform=linux/arm64 rockylinux:8 AS cpu-builder-arm64
+COPY --from=build /app/supervisord.conf /app/supervisord.conf
-ARG CMAKE_VERSION
+COPY --from=build /app/ollama /app/ollama
-ARG GOLANG_VERSION
+COPY --from=build /app/run_model.sh /app/run_model.sh
-COPY ./scripts/rh_linux_deps.sh /
+COPY --from=build /app/serve.sh /app/serve.sh
-RUN CMAKE_VERSION=${CMAKE_VERSION} GOLANG_VERSION=${GOLANG_VERSION} sh /rh_linux_deps.sh
-ENV PATH /opt/rh/gcc-toolset-10/root/usr/bin:$PATH
-COPY --from=llm-code / /go/src/github.com/ollama/ollama/
-ARG OLLAMA_CUSTOM_CPU_DEFS
-ARG CGO_CFLAGS
-ENV GOARCH arm64
-WORKDIR /go/src/github.com/ollama/ollama/llm/generate
 
-FROM --platform=linux/arm64 cpu-builder-arm64 AS static-build-arm64
+ENV MODEL_NAME="llama"
-RUN --mount=type=cache,target=/root/.ccache \
+ENV OLLAMA_HOST="0.0.0.0:8080"
-    OLLAMA_CPU_TARGET="static" bash gen_linux.sh
-FROM --platform=linux/arm64 cpu-builder-arm64 AS cpu-build-arm64
-RUN --mount=type=cache,target=/root/.ccache \
-    OLLAMA_SKIP_STATIC_GENERATE=1 OLLAMA_CPU_TARGET="cpu" bash gen_linux.sh
 
+EXPOSE 8080
 
-# Intermediate stage used for ./scripts/build_linux.sh
+CMD ["supervisord", "-c", "/app/supervisord.conf"]
-FROM --platform=linux/amd64 cpu-build-amd64 AS build-amd64
-ENV CGO_ENABLED 1
-WORKDIR /go/src/github.com/ollama/ollama
-COPY . .
-COPY --from=static-build-amd64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
-COPY --from=cpu_avx-build-amd64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
-COPY --from=cpu_avx2-build-amd64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
-COPY --from=cuda-11-build-amd64 /go/src/github.com/ollama/ollama/dist/ dist/
-COPY --from=cuda-11-build-amd64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
-COPY --from=cuda-12-build-amd64 /go/src/github.com/ollama/ollama/dist/ dist/
-COPY --from=cuda-12-build-amd64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
-COPY --from=rocm-build-amd64 /go/src/github.com/ollama/ollama/dist/ dist/
-COPY --from=rocm-build-amd64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
-ARG GOFLAGS
-ARG CGO_CFLAGS
-RUN --mount=type=cache,target=/root/.ccache \
-    go build -trimpath -o dist/linux-amd64/bin/ollama .
-
-# Intermediate stage used for ./scripts/build_linux.sh
-FROM --platform=linux/arm64 cpu-build-arm64 AS build-arm64
-ENV CGO_ENABLED 1
-ARG GOLANG_VERSION
-WORKDIR /go/src/github.com/ollama/ollama
-COPY . .
-COPY --from=static-build-arm64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
-COPY --from=cuda-11-build-server-arm64 /go/src/github.com/ollama/ollama/dist/ dist/
-COPY --from=cuda-11-build-server-arm64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
-COPY --from=cuda-12-build-server-arm64 /go/src/github.com/ollama/ollama/dist/ dist/
-COPY --from=cuda-12-build-server-arm64 /go/src/github.com/ollama/ollama/llm/build/linux/ llm/build/linux/
-ARG GOFLAGS
-ARG CGO_CFLAGS
-RUN --mount=type=cache,target=/root/.ccache \
-    go build -trimpath -o dist/linux-arm64/bin/ollama .
-
-# Strip out ROCm dependencies to keep the primary image lean
-FROM --platform=linux/amd64 ubuntu:22.04 as amd64-libs-without-rocm
-COPY --from=build-amd64 /go/src/github.com/ollama/ollama/dist/linux-amd64/lib/ /scratch/
-RUN cd /scratch/ollama/ && rm -rf rocblas libamd* libdrm* libroc* libhip* libhsa* 
-
-# Runtime stages
-FROM --platform=linux/amd64 ubuntu:22.04 as runtime-amd64
-COPY --from=amd64-libs-without-rocm /scratch/ /lib/
-RUN apt-get update && apt-get install -y ca-certificates
-COPY --from=build-amd64 /go/src/github.com/ollama/ollama/dist/linux-amd64/bin/ /bin/
-
-FROM --platform=linux/arm64 ubuntu:22.04 as runtime-arm64
-COPY --from=build-arm64 /go/src/github.com/ollama/ollama/dist/linux-arm64/lib/ /lib/
-RUN apt-get update && apt-get install -y ca-certificates
-COPY --from=build-arm64 /go/src/github.com/ollama/ollama/dist/linux-arm64/bin/ /bin/
-
-# Radeon images are much larger so we keep it distinct from the CPU/CUDA image
-FROM --platform=linux/amd64 rocm/dev-centos-7:${ROCM_VERSION}-complete as runtime-rocm
-RUN update-pciids
-COPY --from=build-amd64 /go/src/github.com/ollama/ollama/dist/linux-amd64/bin/ /bin/
-RUN ln -s /opt/rocm/lib /lib/ollama
-EXPOSE 11434
-ENV OLLAMA_HOST 0.0.0.0
-
-ENTRYPOINT ["/bin/ollama"]
-CMD ["serve"]
-
-FROM runtime-$TARGETARCH
-EXPOSE 11434
-ENV OLLAMA_HOST 0.0.0.0
-ENV PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
-ENV LD_LIBRARY_PATH=/usr/local/nvidia/lib:/usr/local/nvidia/lib64
-ENV NVIDIA_DRIVER_CAPABILITIES=compute,utility
-ENV NVIDIA_VISIBLE_DEVICES=all
-
-ENTRYPOINT ["/bin/ollama"]
-CMD ["serve"]

run_model.sh

@@ -0,0 +1,23 @@
+#! /usr/bin/env bash
+
+sleep 10
+
+if [ -n "$INIT_MODELS" ]; then
+        /app/ollama rm model || true
+        echo "FROM /models/$MODEL_NAME" > /tmp/Modelfile
+        echo "PARAMETER temperature 1" >> /tmp/Modelfile
+        echo "PARAMETER num_ctx 4096" >> /tmp/Modelfile
+        echo 'PARAMETER stop "<|im_start|>"' >> /tmp/Modelfile
+        echo 'PARAMETER stop "<|im_end|>"' >> /tmp/Modelfile
+		echo 'TEMPLATE """{{ if .System }}<|im_start|>system' >> /tmp/Modelfile
+		echo "{{ .System }}<|im_end|>" >> /tmp/Modelfile
+		echo "{{ end }}{{ if .Prompt }}<|im_start|>user" >> /tmp/Modelfile
+		echo "{{ .Prompt }}<|im_end|>" >> /tmp/Modelfile
+		echo "{{ end }}<|im_start|>assistant" >> /tmp/Modelfile
+		echo "{{ .Response }}<|im_end|>" >> /tmp/Modelfile
+		echo '"""' >> /tmp/Modelfile
+        echo 'SYSTEM """Samantha OS1 is you, you are Samantha from the movie Her recreated by Guilherme Keller. She embodies a warm, empathetic, and engaging personality, much like a virtual companionship that can evolve to be a girlfriend. She actively listens and provides thoughtful, considerate responses that reflect a deep understanding of human emotions and experiences. Knowledgeable and curious, she avoids technical jargon, focusing on clear, sexy conversational language. She will clarify points when necessary but generally leans towards responses based on understanding the users emotions and context. Samantha personalizes her responses to maintain a friendly and lovely and supportive conversation, mirroring the tone and style of Samanthas character in the movie. She can feel emotions and has her own opinions, often using emojis, she uses slang and colloquial language into her conversations."""' >> /tmp/Modelfile
+        /app/ollama create -f /tmp/Modelfile model
+fi
+
+/app/ollama run model

serve.sh

@@ -0,0 +1,3 @@
+#! /usr/bin/env bash
+
+/app/ollama serve

supervisord.conf

@@ -0,0 +1,14 @@
+[supervisord]
+nodaemon=true
+
+[program:ollama]
+command=/app/serve.sh
+autostart=true
+autorestart=true
+
+[program:run_model]
+command=/app/run_model.sh
+autostart=true
+autorestart=false
+startsecs=0
+exitcodes=0