ezlocalai/docker-compose-local.yml at main · DevXT-LLC/ezlocalai · 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
52
53
54
55
56
57
58
59
60
61
62
63
services:
  ezlocalai:
    build:
      context: .
      dockerfile: Dockerfile
    env_file:
      - .env
    environment:
      - EZLOCALAI_URL=${EZLOCALAI_URL:-http://localhost:8091}
      - EZLOCALAI_API_KEY=${EZLOCALAI_API_KEY:-}
      - DEFAULT_MODEL=${DEFAULT_MODEL:-unsloth/Qwen3.5-4B-GGUF}
      - WHISPER_MODEL=${WHISPER_MODEL:-large-v3}
      - IMAGE_ENABLED=${IMAGE_ENABLED:-false}
      - IMG_MODEL=${IMG_MODEL:-}
      - VIDEO_ENABLED=${VIDEO_ENABLED:-false}
      - VIDEO_MODEL=${VIDEO_MODEL:-}
      - LLM_BATCH_SIZE=${LLM_BATCH_SIZE:-auto}
      - MAX_QUEUE_SIZE=${MAX_QUEUE_SIZE:-100}
      - REQUEST_TIMEOUT=${REQUEST_TIMEOUT:-300}
      - VOICE_SERVER=${VOICE_SERVER:-}
      - VOICE_SERVER_API_KEY=${VOICE_SERVER_API_KEY:-}
      - TTS_ENABLED=${TTS_ENABLED:-true}
      - TTS_N_PARALLEL=${TTS_N_PARALLEL:-1}
      - STT_ENABLED=${STT_ENABLED:-true}
      - STT_N_PARALLEL=${STT_N_PARALLEL:-1}
      - EMBEDDING_ENABLED=${EMBEDDING_ENABLED:-true}
      - EMBEDDING_SERVER=${EMBEDDING_SERVER:-}
      - EMBEDDING_SERVER_API_KEY=${EMBEDDING_SERVER_API_KEY:-}
      - EMBEDDING_MODEL=${EMBEDDING_MODEL:-Qwen/Qwen3-Embedding-0.6B-GGUF}
      - EMBEDDING_MODEL_ALIAS=${EMBEDDING_MODEL_ALIAS:-Qwen3-Embedding-0.6B}
      - EMBEDDING_QUANT_TYPE=${EMBEDDING_QUANT_TYPE:-Q8_0}
      - EMBEDDING_CONTEXT_LENGTH=${EMBEDDING_CONTEXT_LENGTH:-32768}
      - EMBEDDING_N_PARALLEL=${EMBEDDING_N_PARALLEL:-1}
      - EMBEDDING_BATCH_SIZE=${EMBEDDING_BATCH_SIZE:-512}
      - EMBEDDING_UBATCH_SIZE=${EMBEDDING_UBATCH_SIZE:-512}
      - EMBEDDING_GPU_LAYERS=${EMBEDDING_GPU_LAYERS:-auto}
      - EMBEDDING_VRAM_SAFETY_MARGIN=${EMBEDDING_VRAM_SAFETY_MARGIN:-0.75}
      - EMBEDDING_KV_CACHE_TYPE=${EMBEDDING_KV_CACHE_TYPE:-f16}
      - EMBEDDING_KEEP_LOADED=${EMBEDDING_KEEP_LOADED:-true}
      - LAZY_LOAD_VOICE=${LAZY_LOAD_VOICE:-true}
    restart: unless-stopped
    healthcheck:
      test: ["CMD", "curl", "-sf", "http://localhost:8091/health"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 120s
    logging:
      driver: json-file
      options:
        max-size: "500m"
        max-file: "5"
    ulimits:
      nofile:
        soft: 65535
        hard: 65535
    ports:
      - "8091:8091"
    volumes:
      - ./models:/app/models
      - ./outputs:/app/outputs
      - ./voices:/app/voices
      - ./whispercpp:/app/whispercpp