compose.yaml

services:
  laya:
    build:
      context: .
      args:
        TORCH_INDEX: "${LAYA_TORCH_INDEX:-cpu}"
        TORCH_VERSION: "${LAYA_TORCH_VERSION:-2.14.0}"
    environment:
      LAYA_DEVICE: "${LAYA_DEVICE:-cpu}"
      # The autocast dtype the runtime picks when it builds the model on the selected device
      # (`laya/agent.py`). An empty value means "use the checkpoint's own `amp_dtype`", which is
      # what these images have always served. README's threshold section measures what fp16 vs
      # bf16 decides differently.
      LAYA_CUDA_AMP: "${LAYA_CUDA_AMP:-}"
      LAYA_CPU_AMP: "${LAYA_CPU_AMP:-}"
      LAYA_MODEL: "${LAYA_MODEL:-auto}"
      LAYA_MODEL_PATH: "${LAYA_MODEL_PATH:-}"
      LAYA_REQUEST_FILE: "${LAYA_REQUEST_FILE:-/opt/laya/examples/request.json}"
      # The Hub commit, branch or tag every checkpoint download resolves to, or `reviewed` for the
      # reviewed SHAs in `laya/revisions.py`. Empty keeps the Hub default, which is what these
      # images have always done. This service builds its Router with no `revision=` argument
      # (`examples/docker/quickstart.py`), so this variable is the only way to pin from Compose.
      LAYA_REVISION: "${LAYA_REVISION:-}"
      HF_HUB_OFFLINE: "${HF_HUB_OFFLINE:-0}"
      HF_TOKEN: "${HF_TOKEN:-}"
      HF_TOKEN_FILE: "${HF_TOKEN_FILE:-}"
      OMP_NUM_THREADS: "${OMP_NUM_THREADS:-4}"
    volumes:
      - model-cache:/home/laya/.cache/huggingface
    init: true

volumes:
  model-cache:
    name: "${LAYA_CACHE_VOLUME:-${COMPOSE_PROJECT_NAME}_model-cache}"