Added new localai:custom docker container, built from a customized version of localai that supports their memory addition 'localrecall' and their 'localagi' along wigh GPU support. Will be re-writing all of this into docker-compose methods in the future.
This commit is contained in:
parent
13028b8d2b
commit
e80f6c995a
27
Dockerfile
Normal file
27
Dockerfile
Normal file
@ -0,0 +1,27 @@
|
||||
FROM nvidia/cuda:12.2.0-runtime-ubuntu22.04
|
||||
|
||||
ENV DEBIAN_FRONTEND=noninteractive
|
||||
WORKDIR /app
|
||||
|
||||
RUN apt-get update && apt-get install -y --no-install-recommends \
|
||||
libgomp1 libgl1 ffmpeg curl python3 ca-certificates && \
|
||||
rm -rf /var/lib/apt/lists/*
|
||||
|
||||
COPY local-ai /usr/local/bin/local-ai
|
||||
|
||||
ENV MODEL_PATH=/models
|
||||
ENV CONFIG_PATH=/config
|
||||
ENV ENABLE_BACKENDS=llama-cuda,whispercpp,stablediffusion,ollama
|
||||
ENV AUTOLOAD_MODELS=true
|
||||
ENV INCLUDE_DEFAULT_MODELS=true
|
||||
|
||||
VOLUME /models
|
||||
VOLUME /config
|
||||
|
||||
HEALTHCHECK --interval=30s --timeout=5s --start-period=10s --retries=3 \
|
||||
CMD curl -f http://localhost:8080/readyz || exit 1
|
||||
|
||||
EXPOSE 8080
|
||||
|
||||
CMD ["/usr/local/bin/local-ai"]
|
||||
|
||||
@ -44,30 +44,55 @@ if [[ "$1" == "--update" ]]; then
|
||||
echo "⚠️ JupyterLab is not running, skipping update."
|
||||
fi
|
||||
|
||||
# --- LocalAI custom build and update ---
|
||||
LOCALAI_REPO=~/git-repos/LocalAI
|
||||
DOCKERFILE_URL="https://your-domain.com/path/to/Dockerfile"
|
||||
|
||||
# 🔁 Updating LocalAI (GPU build)
|
||||
if check_container_running "localai"; then
|
||||
echo "🧠 Updating LocalAI..."
|
||||
docker stop localai
|
||||
docker rm localai
|
||||
else
|
||||
echo "➕ Installing LocalAI..."
|
||||
fi
|
||||
echo "🔁 Rebuilding LocalAI (custom)..."
|
||||
|
||||
docker pull localai/localai:latest-aio-gpu-nvidia-cuda-12
|
||||
# Backup and remove existing repo
|
||||
if [ -d "$LOCALAI_REPO" ]; then
|
||||
echo "📦 Backing up existing LocalAI repo..."
|
||||
mv "$LOCALAI_REPO" "${LOCALAI_REPO}_backup_$(date +%s)"
|
||||
fi
|
||||
|
||||
# Ensure default models.yaml exists
|
||||
MODELS_YAML=~/ai-stack/localai/config/models.yaml
|
||||
if [ ! -f "$MODELS_YAML" ]; then
|
||||
echo "📄 Creating default models.yaml..."
|
||||
cat > "$MODELS_YAML" <<EOF
|
||||
- name: phi
|
||||
backend: llama-cuda
|
||||
parameters:
|
||||
model: /models/phi-2.gguf
|
||||
EOF
|
||||
fi
|
||||
# Clone latest LocalAI
|
||||
git clone https://github.com/mudler/LocalAI.git "$LOCALAI_REPO"
|
||||
|
||||
# Remove default Docker-related files
|
||||
cd "$LOCALAI_REPO"
|
||||
rm -f Dockerfile Dockerfile.aio docker-compose.yaml
|
||||
|
||||
# Prepare and build LocalAI binary
|
||||
# NOTE - This portion required quite a few custom additions in order to build correctly... if you are running this from a machine that hasn't built it already, just better probably to await the implementation of the Docker-Compose methods
|
||||
# Docker-compose methods will be coming soon.
|
||||
echo "🔧 Running make prepare..."
|
||||
make prepare
|
||||
|
||||
echo "🧠 Building LocalAI binary with GPU, localrecall, and localagi..."
|
||||
make build \
|
||||
TAGS="localrecall,localagi,gpu" \
|
||||
BACKENDS="llama-cuda whispercpp stablediffusion ollama" \
|
||||
GPU_SUPPORT="cuda"
|
||||
|
||||
|
||||
# Download your custom Dockerfile
|
||||
echo "📥 Downloading custom Dockerfile..."
|
||||
curl -fsSL "$DOCKERFILE_URL" -o Dockerfile
|
||||
|
||||
# Build the container
|
||||
echo "🔨 Building custom LocalAI Docker image..."
|
||||
docker build -t localai:custom .
|
||||
|
||||
# Remove existing container
|
||||
docker stop localai 2>/dev/null || true
|
||||
docker rm localai 2>/dev/null || true
|
||||
|
||||
# Create model and config directories if needed
|
||||
mkdir -p ~/ai-stack/localai/models
|
||||
mkdir -p ~/ai-stack/localai/config
|
||||
|
||||
# Run new container
|
||||
docker run -d \
|
||||
--name localai \
|
||||
--gpus all \
|
||||
@ -79,9 +104,48 @@ docker run -d \
|
||||
-e AUTOLOAD_MODELS=true \
|
||||
-e MODEL_PATH=/models \
|
||||
--restart unless-stopped \
|
||||
localai/localai:latest-aio-gpu-nvidia-cuda-12
|
||||
localai:custom
|
||||
|
||||
echo "✅ LocalAI running at http://localhost:8080"
|
||||
echo "✅ LocalAI (custom) is now running at http://localhost:8080"
|
||||
|
||||
|
||||
# # 🔁 Updating LocalAI (GPU build)
|
||||
# if check_container_running "localai"; then
|
||||
# echo "🧠 Updating LocalAI..."
|
||||
# docker stop localai
|
||||
# docker rm localai
|
||||
# else
|
||||
# echo "➕ Installing LocalAI..."
|
||||
# fi
|
||||
|
||||
# docker pull localai/localai:latest-aio-gpu-nvidia-cuda-12
|
||||
|
||||
# # Ensure default models.yaml exists
|
||||
# MODELS_YAML=~/ai-stack/localai/config/models.yaml
|
||||
# if [ ! -f "$MODELS_YAML" ]; then
|
||||
# echo "📄 Creating default models.yaml..."
|
||||
# cat > "$MODELS_YAML" <<EOF
|
||||
# - name: phi
|
||||
# backend: llama-cuda
|
||||
# parameters:
|
||||
# model: /models/phi-2.gguf
|
||||
# EOF
|
||||
# fi
|
||||
|
||||
# docker run -d \
|
||||
# --name localai \
|
||||
# --gpus all \
|
||||
# -p 8080:8080 \
|
||||
# -v ~/ai-stack/localai/models:/models \
|
||||
# -v ~/ai-stack/localai/config:/config \
|
||||
# -e ENABLE_BACKENDS=llama-cuda,ollama \
|
||||
# -e INCLUDE_DEFAULT_MODELS=true \
|
||||
# -e AUTOLOAD_MODELS=true \
|
||||
# -e MODEL_PATH=/models \
|
||||
# --restart unless-stopped \
|
||||
# localai/localai:latest-aio-gpu-nvidia-cuda-12
|
||||
|
||||
# echo "✅ LocalAI running at http://localhost:8080"
|
||||
|
||||
|
||||
|
||||
@ -456,6 +520,60 @@ EOF
|
||||
fi
|
||||
fi
|
||||
|
||||
if [[ "$1" != "--update" ]]; then
|
||||
if check_container_running "localai"; then
|
||||
echo "✅ Skipping LocalAI install (already running)."
|
||||
else
|
||||
read -p "➕ Install LocalAI (GPU-accelerated)? [y/N]: " localai_prompt
|
||||
if [[ "$localai_prompt" =~ ^[Yy]$ ]]; then
|
||||
echo "🧠 Installing LocalAI (NVIDIA CUDA 12)..."
|
||||
|
||||
mkdir -p ~/ai-stack/localai/models
|
||||
mkdir -p ~/ai-stack/localai/config
|
||||
|
||||
echo "📄 Creating default models.yaml..."
|
||||
cat > ~/ai-stack/localai/config/models.yaml <<EOF
|
||||
- name: phi
|
||||
backend: llama-cuda
|
||||
parameters:
|
||||
model: /models/phi-2.gguf
|
||||
EOF
|
||||
|
||||
echo "📦 Using custom built LocalAI image (localai:custom)"
|
||||
|
||||
docker run -d \
|
||||
--name localai \
|
||||
--gpus all \
|
||||
--network ai-stack-net \
|
||||
-p 8080:8080 \
|
||||
-v ~/ai-stack/localai/models:/models \
|
||||
-v ~/ai-stack/localai/config:/config \
|
||||
-e ENABLE_BACKENDS=llama-cuda,whispercpp,stablediffusion,ollama \
|
||||
-e INCLUDE_DEFAULT_MODELS=true \
|
||||
-e AUTOLOAD_MODELS=true \
|
||||
-e MODEL_PATH=/models \
|
||||
--restart unless-stopped \
|
||||
localai:custom
|
||||
|
||||
echo "🌐 LocalAI is now running at http://localhost:8080"
|
||||
fi
|
||||
fi
|
||||
fi
|
||||
|
||||
|
||||
docker run -d \
|
||||
--name localai \
|
||||
--gpus all \
|
||||
--network ai-stack-net \
|
||||
-p 8080:8080 \
|
||||
-v ~/ai-stack/localai/models:/models \
|
||||
-v ~/ai-stack/localai/config:/config \
|
||||
-e ENABLE_BACKENDS=llama-cuda,whispercpp,stablediffusion,ollama \
|
||||
-e INCLUDE_DEFAULT_MODELS=true \
|
||||
-e AUTOLOAD_MODELS=true \
|
||||
-e MODEL_PATH=/models \
|
||||
--restart unless-stopped \
|
||||
localai:custom
|
||||
|
||||
|
||||
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user