Avano/docker-compose.yml at main · ma14ch/Avano · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51

services:
  ai-tts:
    build:
      context: .
      args:
        PRELOAD_MODELS: "false"
    container_name: ai-ASR
    gpus: all
    environment:
      - NVIDIA_VISIBLE_DEVICES=all
      - NVIDIA_DRIVER_CAPABILITIES=compute,utility
      - PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True
      - ASR_MODEL=${ASR_MODEL:-persian-v4}
    ports:
      - "5016:5016"
    volumes:
      - ./src:/app/src
      - ./models:/app/models
      - ./hf-cache:/app/hf-cache   # persist HF/transformers cache
    networks:
      - localnet

  ui:
    build:
      context: .
      args:
        PRELOAD_MODELS: "false"
    container_name: ai-ASR-ui
    gpus: all
    environment:
      - NVIDIA_VISIBLE_DEVICES=all
      - NVIDIA_DRIVER_CAPABILITIES=compute,utility
      - CUDA_VISIBLE_DEVICES=1
      - PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True
      - ASR_API_URL=http://ai-tts:5016/api/inference/
      - ASR_API_TIMEOUT=3600
    command: python3 -m ui.gradio_app
    ports:
      - "7860:7860"
    volumes:
      - ./src:/app/src
      - ./models:/app/models
      - ./ui:/app/ui
      - ./hf-cache:/app/hf-cache   # share the same cache with API
    networks:
      - localnet

networks:
  localnet:
    driver: bridge