LocalAI

#  http://localhost:8080
# CPU: latest
# Nvidia: latest-gpu-nvidia-cuda-12, latest-gpu-nvidia-cuda-11, latest-nvidia-l4t-arm64
# AIO: latest-aio-gpu-nvidia-cuda-12
# AMD GPU ROCm: latest-gpu-hipblas
# 健康 http://localhost:8080/readyz
# 参考 https://localai.io/basics/container/
# 会直接下载一些模型 https://github.com/mudler/LocalAI/tree/master/aio/gpu-8g
# 内置模型名 whisper-1, stablediffusion
# 支持 https_proxy
docker run --rm -ti --gpus all \
  -p 8080:8080 \
  -v $PWD/localai/models:/models:cached \
  -v $PWD/localai/backends:/backends:cached \
  --name localai docker.m.daocloud.io/localai/localai:latest-gpu-nvidia-cuda-12

docker run --rm -ti --gpus all -p 8080:8080 -v $PWD/localai/models:/build/models:cached --name localai docker.m.daocloud.io/localai/localai:latest-gpu-nvidia-cuda-12

docker exec -it localai bash
./local-ai
./local-ai models list
./local-ai models install flux.1-krea-dev-ggml

# a cute baby sea otter|malformed
# POSITIVE: a cute baby sea otter
# NEGATIVE: malformed
# 1024x1024 ~3min
curl http://localhost:8080/v1/images/generations --json '{
  "prompt": "floating hair, portrait, ((loli)), ((one girl)), cute face, hidden hands, asymmetrical bangs, beautiful detailed eyes, eye shadow, hair ornament, ribbons, bowties, buttons, pleated skirt, (((masterpiece))), ((best quality)), colorful|((part of the head)), ((((mutated hands and fingers)))), deformed, blurry, bad anatomy, disfigured, poorly drawn face, mutation, mutated, extra limb, ugly, poorly drawn hands, missing limb, blurry, floating limbs, disconnected limbs, malformed hands, blur, out of focus, long neck, long body, Octane renderer, lowres, bad anatomy, bad hands, text",
  "model": "flux.1-krea-dev-ggml",
  "size": "1024x1024"
}'

version: '3.9'
services:
  api:
    image: localai/localai:localai/localai:latest-aio-gpu-nvidia-cuda-12
    healthcheck:
      test: ['CMD', 'curl', '-f', 'http://localhost:8080/readyz']
      interval: 1m
      timeout: 20m
      retries: 5
    ports:
      - 8080:8080
    environment:
      - DEBUG=true
    volumes:
      - ./models:/build/models:cached
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities: [gpu]

python -c "import torch; print(torch.__version__)"
python -c "import torchvision; print(torchvision.__version__)"