-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathdocker-compose.yml
More file actions
83 lines (68 loc) · 1.76 KB
/
Copy pathdocker-compose.yml
File metadata and controls
83 lines (68 loc) · 1.76 KB
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
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
version: '3.8'
services:
atlas-train:
build:
context: .
dockerfile: Dockerfile
image: atlas:latest
container_name: atlas-training
# GPU support
runtime: nvidia
environment:
- NVIDIA_VISIBLE_DEVICES=all
- NVIDIA_DRIVER_CAPABILITIES=compute,utility
- CUDA_VISIBLE_DEVICES=0
# Mount volumes for data and outputs
volumes:
# Your dataset (adjust path as needed)
- ./data:/app/data:ro
# Output directories
- ./outputs:/app/outputs
- ./checkpoints:/app/checkpoints
- ./logs:/app/logs
# Weights & Biases cache
- ~/.wandb:/root/.wandb
# Resource limits (adjust based on your system)
deploy:
resources:
reservations:
devices:
- driver: nvidia
count: 1
capabilities: [gpu]
# Logging
logging:
driver: "json-file"
options:
max-size: "10m"
max-file: "3"
# Keep container running
stdin_open: true
tty: true
# Optional: Jupyter notebook service for experimentation
atlas-jupyter:
build:
context: .
dockerfile: Dockerfile
image: atlas:latest
container_name: atlas-jupyter
runtime: nvidia
environment:
- NVIDIA_VISIBLE_DEVICES=all
- NVIDIA_DRIVER_CAPABILITIES=compute,utility
- CUDA_VISIBLE_DEVICES=0
volumes:
- ./:/app
- ~/.wandb:/root/.wandb
ports:
- "8888:8888"
deploy:
resources:
reservations:
devices:
- driver: nvidia
count: 1
capabilities: [gpu]
command: jupyter notebook --ip=0.0.0.0 --port=8888 --no-browser --allow-root
stdin_open: true
tty: true