From 6644af10c6c00edeea1844bb5d934fa72469942a Mon Sep 17 00:00:00 2001
From: Ettore Di Giacinto <mudler@users.noreply.github.com>
Date: Thu, 26 Jun 2025 18:41:38 +0200
Subject: [PATCH] feat: :warning:  reduce images size and stop bundling sources
 (#5721)

feat: reduce images size and stop bundling sources

Do not copy sources anymore, and reduce packages of the base images by
not using builder images.

If needed to rebuild, just build the container image from scratch by
following the docs. We will slowly try to migrate all backends to the
gallery to keep the core small.

This PR is a breaking change, it also sets the base folders to /models
and /backends instead of /build/models and /build/backends.

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
Signed-off-by: Ettore Di Giacinto <mudler@users.noreply.github.com>
---
 Dockerfile                                    | 203 ++++++++----------
 aio/entrypoint.sh                             |   2 +-
 .../content/docs/features/GPU-acceleration.md |   6 +-
 .../docs/getting-started/container-images.md  |  10 +-
 docs/content/docs/reference/nvidia-l4t.md     |   2 +-
 docs/static/install.sh                        |   8 +-
 entrypoint.sh                                 |  49 ++---
 7 files changed, 123 insertions(+), 157 deletions(-)

diff --git a/Dockerfile b/Dockerfile
index 3bfc4a388..a31ce0a8c 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -2,82 +2,17 @@ ARG BASE_IMAGE=ubuntu:22.04
 ARG GRPC_BASE_IMAGE=${BASE_IMAGE}
 ARG INTEL_BASE_IMAGE=${BASE_IMAGE}
 
-# The requirements-core target is common to all images.  It should not be placed in requirements-core unless every single build will use it.
 FROM ${BASE_IMAGE} AS requirements
 
-USER root
-
-ARG GO_VERSION=1.22.6
-ARG CMAKE_VERSION=3.26.4
-ARG CMAKE_FROM_SOURCE=false
-ARG TARGETARCH
-ARG TARGETVARIANT
-
 ENV DEBIAN_FRONTEND=noninteractive
 
 RUN apt-get update && \
     apt-get install -y --no-install-recommends \
-        build-essential \
-        ccache \
-        ca-certificates espeak-ng \
-        curl libssl-dev \
-        git \
-        git-lfs \
-        unzip upx-ucl python3 python-is-python3 && \
+        ca-certificates curl wget espeak-ng libgomp1 \
+        python3 python-is-python3 ffmpeg && \
     apt-get clean && \
     rm -rf /var/lib/apt/lists/*
 
-# Install CMake (the version in 22.04 is too old)
-RUN <<EOT bash
-    if [ "${CMAKE_FROM_SOURCE}}" = "true" ]; then
-        curl -L -s https://github.com/Kitware/CMake/releases/download/v${CMAKE_VERSION}/cmake-${CMAKE_VERSION}.tar.gz -o cmake.tar.gz && tar xvf cmake.tar.gz && cd cmake-${CMAKE_VERSION} && ./configure && make && make install
-    else
-        apt-get update && \
-        apt-get install -y \
-            cmake && \
-        apt-get clean && \
-        rm -rf /var/lib/apt/lists/*
-    fi
-EOT
-
-# Install Go
-RUN curl -L -s https://go.dev/dl/go${GO_VERSION}.linux-${TARGETARCH}.tar.gz | tar -C /usr/local -xz
-ENV PATH=$PATH:/root/go/bin:/usr/local/go/bin
-
-# Install grpc compilers and rice
-RUN go install google.golang.org/protobuf/cmd/protoc-gen-go@v1.34.2 && \
-    go install google.golang.org/grpc/cmd/protoc-gen-go-grpc@1958fcbe2ca8bd93af633f11e97d44e567e945af && \
-    go install github.com/GeertJohan/go.rice/rice@latest
-
-COPY --chmod=644 custom-ca-certs/* /usr/local/share/ca-certificates/
-RUN update-ca-certificates
-
-RUN test -n "$TARGETARCH" \
-    || (echo 'warn: missing $TARGETARCH, either set this `ARG` manually, or run using `docker buildkit`')
-
-# Use the variables in subsequent instructions
-RUN echo "Target Architecture: $TARGETARCH"
-RUN echo "Target Variant: $TARGETVARIANT"
-
-# Cuda
-ENV PATH=/usr/local/cuda/bin:${PATH}
-
-# HipBLAS requirements
-ENV PATH=/opt/rocm/bin:${PATH}
-
-# OpenBLAS requirements and stable diffusion
-RUN apt-get update && \
-    apt-get install -y --no-install-recommends \
-        libopenblas-dev && \
-    apt-get clean && \
-    rm -rf /var/lib/apt/lists/*
-
-WORKDIR /build
-
-
-###################################
-###################################
-
 # The requirements-drivers target is for BUILD_TYPE specific items.  If you need to install something specific to CUDA, or specific to ROCM, it goes here.
 FROM requirements AS requirements-drivers
 
@@ -85,7 +20,8 @@ ARG BUILD_TYPE
 ARG CUDA_MAJOR_VERSION=12
 ARG CUDA_MINOR_VERSION=0
 ARG SKIP_DRIVERS=false
-
+ARG TARGETARCH
+ARG TARGETVARIANT
 ENV BUILD_TYPE=${BUILD_TYPE}
 
 # Vulkan requirements
@@ -152,6 +88,83 @@ RUN if [ "${BUILD_TYPE}" = "hipblas" ] && [ "${SKIP_DRIVERS}" = "false" ]; then
         ldconfig \
     ; fi
 
+# Cuda
+ENV PATH=/usr/local/cuda/bin:${PATH}
+
+# HipBLAS requirements
+ENV PATH=/opt/rocm/bin:${PATH}
+
+###################################
+###################################
+
+# The requirements-core target is common to all images.  It should not be placed in requirements-core unless every single build will use it.
+FROM requirements-drivers AS build-requirements
+
+ARG GO_VERSION=1.22.6
+ARG CMAKE_VERSION=3.26.4
+ARG CMAKE_FROM_SOURCE=false
+ARG TARGETARCH
+ARG TARGETVARIANT
+
+
+RUN apt-get update && \
+    apt-get install -y --no-install-recommends \
+        build-essential \
+        ccache \
+        ca-certificates espeak-ng \
+        curl libssl-dev \
+        git \
+        git-lfs \
+        unzip upx-ucl python3 python-is-python3 && \
+    apt-get clean && \
+    rm -rf /var/lib/apt/lists/*
+
+# Install CMake (the version in 22.04 is too old)
+RUN <<EOT bash
+    if [ "${CMAKE_FROM_SOURCE}}" = "true" ]; then
+        curl -L -s https://github.com/Kitware/CMake/releases/download/v${CMAKE_VERSION}/cmake-${CMAKE_VERSION}.tar.gz -o cmake.tar.gz && tar xvf cmake.tar.gz && cd cmake-${CMAKE_VERSION} && ./configure && make && make install
+    else
+        apt-get update && \
+        apt-get install -y \
+            cmake && \
+        apt-get clean && \
+        rm -rf /var/lib/apt/lists/*
+    fi
+EOT
+
+# Install Go
+RUN curl -L -s https://go.dev/dl/go${GO_VERSION}.linux-${TARGETARCH}.tar.gz | tar -C /usr/local -xz
+ENV PATH=$PATH:/root/go/bin:/usr/local/go/bin
+
+# Install grpc compilers and rice
+RUN go install google.golang.org/protobuf/cmd/protoc-gen-go@v1.34.2 && \
+    go install google.golang.org/grpc/cmd/protoc-gen-go-grpc@1958fcbe2ca8bd93af633f11e97d44e567e945af && \
+    go install github.com/GeertJohan/go.rice/rice@latest
+
+COPY --chmod=644 custom-ca-certs/* /usr/local/share/ca-certificates/
+RUN update-ca-certificates
+
+
+# OpenBLAS requirements and stable diffusion
+RUN apt-get update && \
+    apt-get install -y --no-install-recommends \
+        libopenblas-dev && \
+    apt-get clean && \
+    rm -rf /var/lib/apt/lists/*
+
+RUN test -n "$TARGETARCH" \
+    || (echo 'warn: missing $TARGETARCH, either set this `ARG` manually, or run using `docker buildkit`')
+
+# Use the variables in subsequent instructions
+RUN echo "Target Architecture: $TARGETARCH"
+RUN echo "Target Variant: $TARGETVARIANT"
+
+
+
+
+WORKDIR /build
+
+
 ###################################
 ###################################
 
@@ -218,13 +231,14 @@ RUN git clone --recurse-submodules --jobs 4 -b ${GRPC_VERSION} --depth 1 --shall
 
 # The builder-base target has the arguments, variables, and copies shared between full builder images and the uncompiled devcontainer
 
-FROM requirements-drivers AS builder-base
+FROM build-requirements AS builder-base
 
 ARG GO_TAGS="tts p2p"
 ARG GRPC_BACKENDS
 ARG MAKEFLAGS
 ARG LD_FLAGS="-s -w"
-
+ARG TARGETARCH
+ARG TARGETVARIANT
 ENV GRPC_BACKENDS=${GRPC_BACKENDS}
 ENV GO_TAGS=${GO_TAGS}
 ENV MAKEFLAGS=${MAKEFLAGS}
@@ -259,6 +273,8 @@ EOT
 
 # Compile backends first in a separate stage
 FROM builder-base AS builder-backends
+ARG TARGETARCH
+ARG TARGETVARIANT
 
 COPY --from=grpc /opt/grpc /usr/local
 
@@ -314,24 +330,13 @@ RUN if [ ! -d "/build/sources/go-piper/piper-phonemize/pi/lib/" ]; then \
 
 FROM builder-base AS devcontainer
 
-ARG FFMPEG
-
 COPY --from=grpc /opt/grpc /usr/local
 
 COPY .devcontainer-scripts /.devcontainer-scripts
 
-# Add FFmpeg
-RUN if [ "${FFMPEG}" = "true" ]; then \
-        apt-get update && \
-        apt-get install -y --no-install-recommends \
-            ffmpeg && \
-        apt-get clean && \
-        rm -rf /var/lib/apt/lists/* \
-    ; fi
-
 RUN apt-get update && \
     apt-get install -y --no-install-recommends \
-        ssh less wget
+        ssh less
 # For the devcontainer, leave apt functional in case additional devtools are needed at runtime.
 
 RUN go install github.com/go-delve/delve/cmd/dlv@latest
@@ -345,40 +350,16 @@ RUN go install github.com/mikefarah/yq/v4@latest
 # If you cannot find a more suitable place for an addition, this layer is a suitable place for it.
 FROM requirements-drivers
 
-ARG FFMPEG
-ARG BUILD_TYPE
-ARG TARGETARCH
-ARG MAKEFLAGS
-
-ENV BUILD_TYPE=${BUILD_TYPE}
-ENV REBUILD=false
 ENV HEALTHCHECK_ENDPOINT=http://localhost:8080/readyz
-ENV MAKEFLAGS=${MAKEFLAGS}
 
 ARG CUDA_MAJOR_VERSION=12
 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 update && \
-        apt-get install -y --no-install-recommends \
-            ffmpeg && \
-        apt-get clean && \
-        rm -rf /var/lib/apt/lists/* \
-    ; fi
+WORKDIR /
 
-WORKDIR /build
-
-# we start fresh & re-copy all assets because `make build` does not clean up nicely after itself
-# so when `entrypoint.sh` runs `make build` again (which it does by default), the build would fail
-# see https://github.com/go-skynet/LocalAI/pull/658#discussion_r1241971626 and
-# https://github.com/go-skynet/LocalAI/pull/434
-COPY . .
-
-COPY --from=builder /build/sources ./sources/
-COPY --from=grpc /opt/grpc /usr/local
+COPY ./entrypoint.sh .
 
 # Copy the binary
 COPY --from=builder /build/local-ai ./
@@ -387,12 +368,12 @@ COPY --from=builder /build/local-ai ./
 COPY --from=builder /build/sources/go-piper/piper-phonemize/pi/lib/* /usr/lib/
 
 # Make sure the models directory exists
-RUN mkdir -p /build/models /build/backends
+RUN mkdir -p /models /backends
 
 # Define the health check command
 HEALTHCHECK --interval=1m --timeout=10m --retries=10 \
   CMD curl -f ${HEALTHCHECK_ENDPOINT} || exit 1
 
-VOLUME /build/models /build/backends
+VOLUME /models /backends
 EXPOSE 8080
-ENTRYPOINT [ "/build/entrypoint.sh" ]
+ENTRYPOINT [ "/entrypoint.sh" ]
diff --git a/aio/entrypoint.sh b/aio/entrypoint.sh
index e24434ff9..a4b83a9da 100755
--- a/aio/entrypoint.sh
+++ b/aio/entrypoint.sh
@@ -135,4 +135,4 @@ check_vars
 
 echo "===> Starting LocalAI[$PROFILE] with the following models: $MODELS"
 
-exec /build/entrypoint.sh "$@"
+exec /entrypoint.sh "$@"
diff --git a/docs/content/docs/features/GPU-acceleration.md b/docs/content/docs/features/GPU-acceleration.md
index 4fd3c039a..550c013c1 100644
--- a/docs/content/docs/features/GPU-acceleration.md
+++ b/docs/content/docs/features/GPU-acceleration.md
@@ -268,7 +268,7 @@ The image list is on [quay](https://quay.io/repository/go-skynet/local-ai?tab=ta
 To run LocalAI with Docker and sycl starting `phi-2`, you can use the following command as an example:
 
 ```bash
-docker run -e DEBUG=true --privileged -ti -v $PWD/models:/build/models -p 8080:8080  -v /dev/dri:/dev/dri --rm quay.io/go-skynet/local-ai:master-gpu-intel-f32-ffmpeg-core phi-2
+docker run -e DEBUG=true --privileged -ti -v $PWD/models:/models -p 8080:8080  -v /dev/dri:/dev/dri --rm quay.io/go-skynet/local-ai:master-gpu-intel-f32-ffmpeg-core phi-2
 ```
 
 ### Notes
@@ -296,7 +296,7 @@ To use Vulkan, use the images with the `vulkan` tag, for example `{{< version >}
 To run LocalAI with Docker and Vulkan, you can use the following command as an example:
 
 ```bash
-docker run -p 8080:8080 -e DEBUG=true -v $PWD/models:/build/models localai/localai:latest-vulkan-ffmpeg-core
+docker run -p 8080:8080 -e DEBUG=true -v $PWD/models:/models localai/localai:latest-vulkan-ffmpeg-core
 ```
 
 ### Notes
@@ -308,7 +308,7 @@ These flags are the same as the sections above, depending on the hardware, for [
 If you have mixed hardware, you can pass flags for multiple GPUs, for example:
 
 ```bash
-docker run -p 8080:8080 -e DEBUG=true -v $PWD/models:/build/models \
+docker run -p 8080:8080 -e DEBUG=true -v $PWD/models:/models \
 --gpus=all \ # nvidia passthrough
 --device /dev/dri --device /dev/kfd \ # AMD/Intel passthrough
 localai/localai:latest-vulkan-ffmpeg-core
diff --git a/docs/content/docs/getting-started/container-images.md b/docs/content/docs/getting-started/container-images.md
index 4edbc6191..67b0c44ee 100644
--- a/docs/content/docs/getting-started/container-images.md
+++ b/docs/content/docs/getting-started/container-images.md
@@ -92,7 +92,7 @@ services:
       - DEBUG=true
       # ...
     volumes:
-      - ./models:/build/models:cached
+      - ./models:/models:cached
     # decomment the following piece if running with Nvidia GPUs
     # deploy:
     #   resources:
@@ -105,21 +105,21 @@ services:
 
 {{% alert icon="💡" %}}
 
-**Models caching**: The **AIO** image will download the needed models on the first run if not already present and store those in `/build/models` inside the container. The AIO models will be automatically updated with new versions of AIO images.
+**Models caching**: The **AIO** image will download the needed models on the first run if not already present and store those in `/models` inside the container. The AIO models will be automatically updated with new versions of AIO images.
 
 You can change the directory inside the container by specifying a `MODELS_PATH` environment variable (or `--models-path`). 
 
-If you want to use a named model or a local directory, you can mount it as a volume to `/build/models`:
+If you want to use a named model or a local directory, you can mount it as a volume to `/models`:
 
 ```bash
-docker run -p 8080:8080 --name local-ai -ti -v $PWD/models:/build/models localai/localai:latest-aio-cpu
+docker run -p 8080:8080 --name local-ai -ti -v $PWD/models:/models localai/localai:latest-aio-cpu
 ```
 
 or associate a volume:
 
 ```bash
 docker volume create localai-models
-docker run -p 8080:8080 --name local-ai -ti -v localai-models:/build/models localai/localai:latest-aio-cpu
+docker run -p 8080:8080 --name local-ai -ti -v localai-models:/models localai/localai:latest-aio-cpu
 ```
 
 {{% /alert %}}
diff --git a/docs/content/docs/reference/nvidia-l4t.md b/docs/content/docs/reference/nvidia-l4t.md
index ce0fd5e95..b019aa70c 100644
--- a/docs/content/docs/reference/nvidia-l4t.md
+++ b/docs/content/docs/reference/nvidia-l4t.md
@@ -35,7 +35,7 @@ docker pull quay.io/go-skynet/local-ai:master-nvidia-l4t-arm64-core
 Run the LocalAI container on Nvidia ARM64 devices using the following command, where `/data/models` is the directory containing the models:
 
 ```bash
-docker run -e DEBUG=true -p 8080:8080 -v /data/models:/build/models  -ti --restart=always --name local-ai --runtime nvidia --gpus all quay.io/go-skynet/local-ai:master-nvidia-l4t-arm64-core
+docker run -e DEBUG=true -p 8080:8080 -v /data/models:/models  -ti --restart=always --name local-ai --runtime nvidia --gpus all quay.io/go-skynet/local-ai:master-nvidia-l4t-arm64-core
 ```
 
 Note: `/data/models` is the directory containing the models. You can replace it with the directory containing your models.
diff --git a/docs/static/install.sh b/docs/static/install.sh
index d0c21b63a..357444387 100755
--- a/docs/static/install.sh
+++ b/docs/static/install.sh
@@ -663,7 +663,7 @@ install_docker() {
         IMAGE_TAG=${LOCALAI_VERSION}-vulkan
 
         info "Starting LocalAI Docker container..."
-        $SUDO docker run -v local-ai-data:/build/models \
+        $SUDO docker run -v local-ai-data:/models \
             --device /dev/dri \
             --restart=always \
             -e API_KEY=$API_KEY \
@@ -690,7 +690,7 @@ install_docker() {
         fi
 
         info "Starting LocalAI Docker container..."
-        $SUDO docker run -v local-ai-data:/build/models \
+        $SUDO docker run -v local-ai-data:/models \
             --gpus all \
             --restart=always \
             -e API_KEY=$API_KEY \
@@ -705,7 +705,7 @@ install_docker() {
         fi
 
         info "Starting LocalAI Docker container..."
-        $SUDO docker run -v local-ai-data:/build/models \
+        $SUDO docker run -v local-ai-data:/models \
             --device /dev/dri \
             --device /dev/kfd \
             --group-add=video \
@@ -723,7 +723,7 @@ install_docker() {
         fi
 
         info "Starting LocalAI Docker container..."
-        $SUDO docker run -v local-ai-data:/build/models \
+        $SUDO docker run -v local-ai-data:/models \
             --device /dev/dri \
             --restart=always \
             -e API_KEY=$API_KEY \
diff --git a/entrypoint.sh b/entrypoint.sh
index 389c846d5..fdaa92eba 100755
--- a/entrypoint.sh
+++ b/entrypoint.sh
@@ -1,7 +1,7 @@
 #!/bin/bash
 set -e
 
-cd /build
+cd /
 
 # If we have set EXTRA_BACKENDS, then we need to prepare the backends
 if [ -n "$EXTRA_BACKENDS" ]; then
@@ -13,38 +13,23 @@ if [ -n "$EXTRA_BACKENDS" ]; then
 	done
 fi
 
-if [ "$REBUILD" != "false" ]; then
-	rm -rf ./local-ai
-	make build -j${BUILD_PARALLELISM:-1}
+echo "CPU info:"
+grep -e "model\sname" /proc/cpuinfo | head -1
+grep -e "flags" /proc/cpuinfo | head -1
+if grep -q -e "\savx\s" /proc/cpuinfo ; then
+	echo "CPU:    AVX    found OK"
 else
-	echo "@@@@@"
-	echo "Skipping rebuild"
-	echo "@@@@@"
-	echo "If you are experiencing issues with the pre-compiled builds, try setting REBUILD=true"
-	echo "If you are still experiencing issues with the build, try setting CMAKE_ARGS and disable the instructions set as needed:"
-	echo 'CMAKE_ARGS="-DGGML_F16C=OFF -DGGML_AVX512=OFF -DGGML_AVX2=OFF -DGGML_FMA=OFF"'
-	echo "see the documentation at: https://localai.io/basics/build/index.html"
-	echo "Note: See also https://github.com/go-skynet/LocalAI/issues/288"
-	echo "@@@@@"
-	echo "CPU info:"
-	grep -e "model\sname" /proc/cpuinfo | head -1
-	grep -e "flags" /proc/cpuinfo | head -1
-	if grep -q -e "\savx\s" /proc/cpuinfo ; then
-		echo "CPU:    AVX    found OK"
-	else
-		echo "CPU: no AVX    found"
-	fi
-	if grep -q -e "\savx2\s" /proc/cpuinfo ; then
-		echo "CPU:    AVX2   found OK"
-	else
-		echo "CPU: no AVX2   found"
-	fi
-	if grep -q -e "\savx512" /proc/cpuinfo ; then
-		echo "CPU:    AVX512 found OK"
-	else
-		echo "CPU: no AVX512 found"
-	fi
-	echo "@@@@@"
+	echo "CPU: no AVX    found"
+fi
+if grep -q -e "\savx2\s" /proc/cpuinfo ; then
+	echo "CPU:    AVX2   found OK"
+else
+	echo "CPU: no AVX2   found"
+fi
+if grep -q -e "\savx512" /proc/cpuinfo ; then
+	echo "CPU:    AVX512 found OK"
+else
+	echo "CPU: no AVX512 found"
 fi
 
 exec ./local-ai "$@"