training_extensions/models/object_detection/model_templates/person-vehicle-bike-detection/person-vehicle-bike-detection-2001/template.yaml at develop · e-ddykim/training_extensions · 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
name: person-vehicle-bike-detection-2001
domain: Object Detection
problem: Person Vehicle Bike Detection
framework: OTEDetection v2.9.1
summary: Person Vehicle Bike Detection based on MobileNetV2 (SSD).
annotation_format: COCO
initial_weights: snapshot.pth
dependencies:
- sha256: 9c4190208d9ff7b7b860821c48264a50ee529e84f23d0d2f6947eceb64c2346a
  size: 14937205
  source: https://download.01.org/opencv/openvino_training_extensions/models/object_detection/v2/vehicle-person-bike-detection-2001-1.pth
  destination: snapshot.pth
- source: ../../../../../ote/tools/train.py
  destination: train.py
- source: ../../../../../ote/tools/eval.py
  destination: eval.py
- source: ../../../../../ote/tools/export.py
  destination: export.py
- source: ../../../../../ote/tools/compress.py
  destination: compress.py
- source: ../../../../../ote
  destination: packages/ote
- source: ../../requirements.txt
  destination: requirements.txt
dataset_requirements:
  classes:
  - vehicle
  - person
  - non-vehicle
max_nodes: 1
training_target:
- CPU
- GPU
inference_target:
- CPU
- iGPU
- VPU
hyper_parameters:
  basic:
    batch_size: 54
    base_learning_rate: 0.05
    epochs: 20
output_format:
  onnx:
    default: true
  openvino:
    default: true
    input_format: BGR
optimisations:
  nncf_quantization:
    config: compression_config.json
    default: false
metrics:
- display_name: Size
  key: size
  unit: Mp
  value: 1.84
- display_name: Complexity
  key: complexity
  unit: GFLOPs
  value: 1.86
- display_name: mAP @ [IoU=0.50:0.95]
  key: map
  unit: '%'
  value: 22.6
gpu_num: 4
tensorboard: true
config: model.py