dream

docker-compose.override.yml
298 строк · 7.6 Кб
Перенос по словам
1
services:
2
  agent:
3
    command: sh -c 'bin/wait && python -m deeppavlov_agent.run agent.pipeline_config=assistant_dists/universal_prompted_assistant/pipeline_conf.json'
4
    environment:
5
      WAIT_HOSTS: "sentseg:8011, ranking-based-response-selector:8002, combined-classification:8087, 
6
        sentence-ranker:8128, 
7
        transformers-lm-gptjt:8161, openai-api-chatgpt:8145, openai-api-davinci3:8131,
8
        openai-api-gpt4:8159, openai-api-gpt4-32k:8160, openai-api-chatgpt-16k:8167,
9
        openai-api-gpt4-turbo:8180, gigachat-api:8187, dff-universal-prompted-skill:8147"
10
      WAIT_HOSTS_TIMEOUT: ${WAIT_TIMEOUT:-1000}
11

12
  sentseg:
13
    env_file: [ .env ]
14
    build:
15
      context: ./annotators/SentSeg/
16
    command: flask run -h 0.0.0.0 -p 8011
17
    environment:
18
      - FLASK_APP=server
19
    deploy:
20
      resources:
21
        limits:
22
          memory: 1.5G
23
        reservations:
24
          memory: 1.5G
25

26
  combined-classification:
27
    env_file: [ .env ]
28
    build:
29
      args:
30
        CONFIG: combined_classifier.json
31
        SERVICE_PORT: 8087
32
      context: .
33
      dockerfile: ./annotators/combined_classification/Dockerfile
34
    command: gunicorn --workers=1 server:app -b 0.0.0.0:8087 --timeout 600
35
    environment:
36
      - CUDA_VISIBLE_DEVICES=0
37
    deploy:
38
      resources:
39
        limits:
40
          memory: 2G
41
        reservations:
42
          memory: 2G
43

44
  ranking-based-response-selector:
45
    env_file: [ .env ]
46
    build:
47
      args:
48
        SERVICE_PORT: 8002
49
        SERVICE_NAME: response_selector
50
        LANGUAGE: EN
51
        SENTENCE_RANKER_ANNOTATION_NAME: sentence_ranker
52
        SENTENCE_RANKER_SERVICE_URL: http://sentence-ranker:8128/respond
53
        SENTENCE_RANKER_TIMEOUT: 3
54
        N_UTTERANCES_CONTEXT: 5
55
        FILTER_TOXIC_OR_BADLISTED: 1
56
      context: .
57
      dockerfile: ./response_selectors/ranking_based_response_selector/Dockerfile
58
    command: flask run -h 0.0.0.0 -p 8002
59
    environment:
60
      - FLASK_APP=server
61
    deploy:
62
      resources:
63
        limits:
64
          memory: 100M
65
        reservations:
66
          memory: 100M
67

68
  sentence-ranker:
69
    env_file: [ .env ]
70
    build:
71
      args:
72
        SERVICE_PORT: 8128
73
        SERVICE_NAME: sentence_ranker
74
        PRETRAINED_MODEL_NAME_OR_PATH: sentence-transformers/all-MiniLM-L6-v2
75
      context: ./services/sentence_ranker/
76
    command: flask run -h 0.0.0.0 -p 8128
77
    environment:
78
      - CUDA_VISIBLE_DEVICES=0
79
      - FLASK_APP=server
80
    deploy:
81
      resources:
82
        limits:
83
          memory: 3G
84
        reservations:
85
          memory: 3G
86

87
  transformers-lm-gptjt:
88
    env_file: [ .env ]
89
    build:
90
      args:
91
        SERVICE_PORT: 8161
92
        SERVICE_NAME: transformers_lm_gptjt
93
        PRETRAINED_MODEL_NAME_OR_PATH: togethercomputer/GPT-JT-6B-v1
94
        HALF_PRECISION: 1
95
      context: .
96
      dockerfile: ./services/transformers_lm/Dockerfile
97
    command: flask run -h 0.0.0.0 -p 8161
98
    environment:
99
      - CUDA_VISIBLE_DEVICES=0
100
      - FLASK_APP=server
101
    deploy:
102
      resources:
103
        limits:
104
          memory: 50G
105
        reservations:
106
          memory: 50G
107

108
  openai-api-chatgpt:
109
    env_file: [ .env ]
110
    build:
111
      args:
112
        SERVICE_PORT: 8145
113
        SERVICE_NAME: openai_api_chatgpt
114
        PRETRAINED_MODEL_NAME_OR_PATH: gpt-3.5-turbo
115
      context: .
116
      dockerfile: ./services/openai_api_lm/Dockerfile
117
    command: flask run -h 0.0.0.0 -p 8145
118
    environment:
119
      - FLASK_APP=server
120
    deploy:
121
      resources:
122
        limits:
123
          memory: 500M
124
        reservations:
125
          memory: 100M
126

127
  openai-api-davinci3:
128
    env_file: [ .env ]
129
    build:
130
      args:
131
        SERVICE_PORT: 8131
132
        SERVICE_NAME: openai_api_davinci3
133
        PRETRAINED_MODEL_NAME_OR_PATH: text-davinci-003
134
      context: .
135
      dockerfile: ./services/openai_api_lm/Dockerfile
136
    command: flask run -h 0.0.0.0 -p 8131
137
    environment:
138
      - FLASK_APP=server
139
    deploy:
140
      resources:
141
        limits:
142
          memory: 500M
143
        reservations:
144
          memory: 100M
145

146
  openai-api-gpt4:
147
    env_file: [ .env ]
148
    build:
149
      args:
150
        SERVICE_PORT: 8159
151
        SERVICE_NAME: openai_api_gpt4
152
        PRETRAINED_MODEL_NAME_OR_PATH: gpt-4
153
      context: .
154
      dockerfile: ./services/openai_api_lm/Dockerfile
155
    command: flask run -h 0.0.0.0 -p 8159
156
    environment:
157
      - FLASK_APP=server
158
    deploy:
159
      resources:
160
        limits:
161
          memory: 500M
162
        reservations:
163
          memory: 100M
164

165
  openai-api-gpt4-32k:
166
    env_file: [ .env ]
167
    build:
168
      args:
169
        SERVICE_PORT: 8160
170
        SERVICE_NAME: openai_api_gpt4_32k
171
        PRETRAINED_MODEL_NAME_OR_PATH: gpt-4-32k
172
      context: .
173
      dockerfile: ./services/openai_api_lm/Dockerfile
174
    command: flask run -h 0.0.0.0 -p 8160
175
    environment:
176
      - FLASK_APP=server
177
    deploy:
178
      resources:
179
        limits:
180
          memory: 500M
181
        reservations:
182
          memory: 100M
183

184
  openai-api-chatgpt-16k:
185
    env_file: [ .env ]
186
    build:
187
      args:
188
        SERVICE_PORT: 8167
189
        SERVICE_NAME: openai_api_chatgpt_16k
190
        PRETRAINED_MODEL_NAME_OR_PATH: gpt-3.5-turbo-16k
191
      context: .
192
      dockerfile: ./services/openai_api_lm/Dockerfile
193
    command: flask run -h 0.0.0.0 -p 8167
194
    environment:
195
      - FLASK_APP=server
196
    deploy:
197
      resources:
198
        limits:
199
          memory: 500M
200
        reservations:
201
          memory: 100M
202

203
  openai-api-gpt4-turbo:
204
    env_file: [ .env ]
205
    build:
206
      args:
207
        SERVICE_PORT: 8180
208
        SERVICE_NAME: openai_api_gpt4_turbo
209
        PRETRAINED_MODEL_NAME_OR_PATH: gpt-4-1106-preview
210
      context: .
211
      dockerfile: ./services/openai_api_lm/Dockerfile
212
    command: flask run -h 0.0.0.0 -p 8180
213
    environment:
214
      - FLASK_APP=server
215
    deploy:
216
      resources:
217
        limits:
218
          memory: 500M
219
        reservations:
220
          memory: 100M
221

222
  gigachat-api:
223
    env_file: [ .env ]
224
    build:
225
      args:
226
        SERVICE_PORT: 8187
227
        SERVICE_NAME: gigachat_api
228
        PRETRAINED_MODEL_NAME_OR_PATH: GigaChat:1.3.23.1
229
      context: .
230
      dockerfile: ./services/gigachat_api_lm/Dockerfile
231
    command: flask run -h 0.0.0.0 -p 8187
232
    environment:
233
      - FLASK_APP=server
234
    deploy:
235
      resources:
236
        limits:
237
          memory: 500M
238
        reservations:
239
          memory: 100M
240

241
  anthropic-api-claude-v1:
242
    env_file: [ .env ]
243
    build:
244
      args:
245
        SERVICE_PORT: 8164
246
        SERVICE_NAME: anthropic_api_claude_v1
247
        PRETRAINED_MODEL_NAME_OR_PATH: claude-1
248
      context: .
249
      dockerfile: ./services/anthropic_api_lm/Dockerfile
250
    command: flask run -h 0.0.0.0 -p 8164
251
    environment:
252
      - FLASK_APP=server
253
    deploy:
254
      resources:
255
        limits:
256
          memory: 100M
257
        reservations:
258
          memory: 100M
259

260
  anthropic-api-claude-instant-v1:
261
    env_file: [ .env ]
262
    build:
263
      args:
264
        SERVICE_PORT: 8163
265
        SERVICE_NAME: anthropic_api_claude_instant_v1
266
        PRETRAINED_MODEL_NAME_OR_PATH: claude-instant-1
267
      context: .
268
      dockerfile: ./services/anthropic_api_lm/Dockerfile
269
    command: flask run -h 0.0.0.0 -p 8163
270
    environment:
271
      - FLASK_APP=server
272
    deploy:
273
      resources:
274
        limits:
275
          memory: 100M
276
        reservations:
277
          memory: 100M
278

279
  dff-universal-prompted-skill:
280
    env_file: [ .env ]
281
    build:
282
      args:
283
        SERVICE_PORT: 8147
284
        SERVICE_NAME: dff_universal_prompted_skill
285
        GENERATIVE_TIMEOUT: 120
286
        N_UTTERANCES_CONTEXT: 7
287
        DEFAULT_LM_SERVICE_URL: http://transformers-lm-gptjt:8161/respond
288
        DEFAULT_LM_SERVICE_CONFIG: default_generative_config.json
289
      context: .
290
      dockerfile: ./skills/dff_universal_prompted_skill/Dockerfile
291
    deploy:
292
      resources:
293
        limits:
294
          memory: 128M
295
        reservations:
296
          memory: 128M
297

298
version: '3.7'
299
dream

Использование cookies
