project_terrier/anima_module.yaml at main · RobotFlow-Labs/project_terrier · 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
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
apiVersion: anima.robotflow.dev/v1
kind: Module
metadata:
  name: project_terrier
  codename: TERRIER
  wave: 10
  theme: WARDOG
  tier: T3
  defense_score: 38

spec:
  paper:
    title: "Progressive Representation Learning for Real-Time UAV Tracking"
    arxiv: "https://arxiv.org/abs/2409.16652"
    doi: "https://doi.org/10.1109/IROS58592.2024.10803050"
    authors:
      - "Changhong Fu"
      - "Xiang Lei"
      - "Haobo Zuo"
      - "Liangliang Yao"
      - "Guangze Zheng"
      - "Jia Pan"

  inputs:
    - name: rgb_frame
      kind: image
      shape: [H, W, 3]
      description: "BGR frame from ZED 2i or standard camera"
    - name: initial_box
      kind: box_xywh
      description: "Target box in first-frame pixel coordinates"

  outputs:
    - name: tracked_box
      kind: box_xywh
      description: "Predicted box in image coordinates"
    - name: track_score
      kind: scalar
      description: "Confidence of the current prediction"

  runtime:
    backends:
      - cpu
      - cuda
      - mlx
    entrypoint: "anima_terrier.cli:main"
    serve:
      dockerfile: Dockerfile.serve
      compose: docker-compose.serve.yml
      endpoint: "/update"
      port: 8401

  training:
    recipe: configs/default.toml
    epochs: 70
    optimizer: sgd_momentum
    lr_schedule: warmup_log_decay
    checkpoint_dir: /mnt/artifacts-datai/checkpoints/project_terrier
    log_dir: /mnt/artifacts-datai/logs/project_terrier

  datasets:
    training:
      - coco
      - got10k
      - lasot
    adaptation:
      - mega_uav_1_8m
      - visdrone
      - uavdt
      - dronevehicle
      - seadronessee
    evaluation:
      - uavtrack112
      - uavtrack112_L
      - uav123

  integration:
    ros2:
      node: anima_terrier_tracker
      topics:
        in: /camera/rgb
        out: /terrier/track
    detector_handoff: yolo26m