dream

1
name: transformers-lm-llama7bru
2
endpoints:
3
- respond
4
- generate_goals
5
compose:
6
  env_file:
7
  - .env
8
  build:
9
    args:
10
      SERVICE_PORT: 8149
11
      SERVICE_NAME: transformers_lm_llama7bru
12
      PRETRAINED_MODEL_NAME_OR_PATH: IlyaGusev/llama_7b_ru_turbo_alpaca_lora
13
      LANGUAGE: RU
14
      CUDA_VISIBLE_DEVICES: '0'
15
      FLASK_APP: server
16
    context: .
17
    dockerfile: ./services/transformers_peft_lm/Dockerfile
18
  command: flask run -h 0.0.0.0 -p 8149
19
  environment:
20
  - CUDA_VISIBLE_DEVICES=0
21
  - FLASK_APP=server
22
  deploy:
23
    resources:
24
      limits:
25
        memory: 50G
26
      reservations:
27
        memory: 50G
28
  volumes:
29
  - ./services/transformers_peft_lm:/src
30
  - ./common:/src/common
31
  - ~/.deeppavlov/cache:/root/.cache
32
  ports:
33
  - 8149:8149
34
proxy:
35
  command:
36
  - nginx
37
  - -g
38
  - daemon off;
39
  build:
40
    context: dp/proxy/
41
    dockerfile: Dockerfile
42
  environment:
43
  - PROXY_PASS=dream.deeppavlov.ai:8149
44
  - PORT=8149
45