version: '3.3'

services:
  server:
    image: tabbyml/tabby
    environment:
      - MODEL_BACKEND=triton
      - TRITON_TOKENIZER_NAME=/tokenizer
    volumes:
      - ../testdata/tiny-70M/tokenizer:/tokenizer
    links:
      - triton

  admin:
    links:
      - triton



  triton:
    image: tabbyml/fastertransformer_backend
    container_name: tabby-triton
    command: mpirun -n 1 --allow-run-as-root /opt/tritonserver/bin/tritonserver --model-repository=/model
    shm_size: 1gb
    volumes:
      - ../testdata/tiny-70M/models:/model
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: all
              capabilities: [gpu]