Register
Login
Resources
Docs Blog Datasets Glossary Case Studies Tutorials & Webinars
Product
Data Engine LLMs Platform Enterprise
Pricing Explore
Connect to our Discord channel
Resources  
Docs Blog Datasets Glossary Case Studies Tutorials & Webinars
Explore Pricing
Product  
Data Engine LLMs Platform Enterprise
Connect to Dagshub Discord channel
Register
Login

Deci-AI
/
super-gradients
connected to https://github.com/Deci-AI/super-gradients.git

Watch 4
Issues 62 Pull Requests 44 Discussions 0

#869 Add DagsHub Logger to Super Gradients

Merged
Ghost merged 1 commits into Deci-AI:master from timho102003:dagshub_logger
Conversation 0 Files changed 4 Commits 1
File Compare
super-gradients
/
src
/
super_gradients
/
recipes
/
dataset_params
/
roboflow_detection_dataset_params.yaml
View left file View right file
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
84
85
86 
  1. data_dir: /data/rf100
    2. 

  2. 

  3. dataset_name: # Set the name of the dataset you want to use (e.g. "digits-t2eg6")
    4. 

  4. 

  5. train_dataset_params:
    6. 

  6.   data_dir: ${..data_dir} # root path to Robflow datasets
    7. 

  7.   dataset_name: ${..dataset_name}
    8. 

  8.   split: train
    9. 

  9.   input_dim: [640, 640]
    10. 

  10.   cache_dir:
    11. 

  11.   cache: False
    12. 

  12.   transforms:
    13. 

  13.     - DetectionRandomAffine:
    14. 

  14.         degrees: 0.                  # rotation degrees, randomly sampled from [-degrees, degrees]
    15. 

  15.         translate: 0.1                # image translation fraction
    16. 

  16.         scales: [ 0.5, 1.5 ]              # random rescale range (keeps size by padding/cropping) after mosaic transform.
    17. 

  17.         shear: 0.0                    # shear degrees, randomly sampled from [-degrees, degrees]
    18. 

  18.         target_size: ${dataset_params.train_dataset_params.input_dim}
    19. 

  19.         filter_box_candidates: True   # whether to filter out transformed bboxes by edge size, area ratio, and aspect ratio.
    20. 

  20.         wh_thr: 2                     # edge size threshold when filter_box_candidates = True (pixels)
    21. 

  21.         area_thr: 0.1                 # threshold for area ratio between original image and the transformed one, when filter_box_candidates = True
    22. 

  22.         ar_thr: 20                    # aspect ratio threshold when filter_box_candidates = True
    23. 

  23.         border_value: 128
    24. 

  24.     - DetectionHSV:
    25. 

  25.         prob: 1.0                       # probability to apply HSV transform
    26. 

  26.         hgain: 5                        # HSV transform hue gain (randomly sampled from [-hgain, hgain])
    27. 

  27.         sgain: 30                       # HSV transform saturation gain (randomly sampled from [-sgain, sgain])
    28. 

  28.         vgain: 30                       # HSV transform value gain (randomly sampled from [-vgain, vgain])
    29. 

  29.     - DetectionHorizontalFlip:
    30. 

  30.         prob: 0.5                       # probability to apply horizontal flip
    31. 

  31.     - DetectionPaddedRescale:
    32. 

  32.         input_dim: ${dataset_params.train_dataset_params.input_dim}
    33. 

  33.         max_targets: 120
    34. 

  34.     - DetectionTargetsFormatTransform:
    35. 

  35.         input_dim: ${dataset_params.train_dataset_params.input_dim}
    36. 

  36.         output_format: LABEL_CXCYWH
    37. 

  37.   tight_box_rotation: False
    38. 

  38.   class_inclusion_list:
    39. 

  39.   max_num_samples:
    40. 

  40.   with_crowd: False
    41. 

  41.   verbose: 0
    42. 

  42. 

  43. train_dataloader_params:
    44. 

  44.   shuffle: True
    45. 

  45.   batch_size: 16
    46. 

  46.   num_workers: 0
    47. 

  47.   sampler:
    48. 

  48.   drop_last: False
    49. 

  49.   pin_memory: True
    50. 

  50.   worker_init_fn:
    51. 

  51.     _target_: super_gradients.training.utils.utils.load_func
    52. 

  52.     dotpath: super_gradients.training.datasets.datasets_utils.worker_init_reset_seed
    53. 

  53.   collate_fn: # collate function for trainset
    54. 

  54.     _target_: super_gradients.training.utils.detection_utils.DetectionCollateFN
    55. 

  55. 

  56. val_dataset_params:
    57. 

  57.   data_dir: ${..data_dir} # root path to Robflow datasets
    58. 

  58.   dataset_name: ${..dataset_name}
    59. 

  59.   split: valid
    60. 

  60.   input_dim: [640, 640]
    61. 

  61.   cache_dir:
    62. 

  62.   cache: False
    63. 

  63.   transforms:
    64. 

  64.   - DetectionPaddedRescale:
    65. 

  65.       input_dim: ${dataset_params.val_dataset_params.input_dim}
    66. 

  66.   - DetectionTargetsFormatTransform:
    67. 

  67.       max_targets: 50
    68. 

  68.       input_dim: ${dataset_params.val_dataset_params.input_dim}
    69. 

  69.       output_format: LABEL_CXCYWH
    70. 

  70.   tight_box_rotation: False
    71. 

  71.   class_inclusion_list:
    72. 

  72.   max_num_samples:
    73. 

  73.   with_crowd: True
    74. 

  74.   verbose: 0
    75. 

  75. 

  76. val_dataloader_params:
    77. 

  77.   batch_size: 64
    78. 

  78.   num_workers: 0
    79. 

  79.   sampler:
    80. 

  80.   drop_last: False
    81. 

  81.   pin_memory: True
    82. 

  82.   collate_fn: # collate function for valset
    83. 

  83.     _target_: super_gradients.training.utils.detection_utils.CrowdDetectionCollateFN
    84. 

  84. 

  85. 

  86. _convert_: all
    87.
Discard
Prev
Next
Tip!

Press p or to see the previous file or, n or to see the next file