Hugging Face's logo

Gelmi
/
rtdetr-r50-ball-handler-finetune

Object Detection
Transformers
TensorBoard
Safetensors
rt_detr
Generated from Trainer
Model card Files Metrics Community
Gelmi commited on
Commit
8ba2078
·
verified ·
1 Parent(s): a9cdc02

End of training

Browse files
Files changed (6) hide show
  1. README.md +31 -43
  2. config.json +4 -17
  3. model.safetensors +2 -2
  4. runs/Feb20_15-02-20_f8a56163e644/events.out.tfevents.1740063750.f8a56163e644.432.0 +3 -0
  5. runs/Feb20_15-02-20_f8a56163e644/events.out.tfevents.1740067951.f8a56163e644.432.1 +3 -0
  6. training_args.bin +1 -1
README.md CHANGED
@@ -16,35 +16,23 @@ should probably proofread and complete it, then remove this comment. -->
16
 
17
  This model is a fine-tuned version of [PekingU/rtdetr_r50vd_coco_o365](https://huggingface.co/PekingU/rtdetr_r50vd_coco_o365) on an unknown dataset.
18
  It achieves the following results on the evaluation set:
19
- - Loss: 12.6727
20
- - Map: 0.2908
21
- - Map 50: 0.4405
22
- - Map 75: 0.2916
23
- - Map Small: 0.0665
24
- - Map Medium: 0.3114
25
- - Map Large: 0.3807
26
- - Mar 1: 0.2993
27
- - Mar 10: 0.5615
28
- - Mar 100: 0.583
29
- - Mar Small: 0.2
30
- - Mar Medium: 0.7166
31
- - Mar Large: 0.7854
32
- - Map Player: 0.4295
33
- - Mar 100 Player: 0.6432
34
- - Map Ball-handler: 0.3115
35
- - Mar 100 Ball-handler: 0.6877
36
- - Map Ball-handler-floor: 0.04
37
- - Mar 100 Ball-handler-floor: 0.85
38
- - Map Basketball: 0.3535
39
- - Mar 100 Basketball: 0.4
40
- - Map Head: -1.0
41
- - Mar 100 Head: -1.0
42
- - Map People: 0.7525
43
- - Mar 100 People: 0.75
44
- - Map Player-floor: 0.1487
45
- - Mar 100 Player-floor: 0.75
46
- - Map Rim: 0.0
47
- - Mar 100 Rim: 0.0
48
 
49
  ## Model description
50
 
@@ -74,23 +62,23 @@ The following hyperparameters were used during training:
74
 
75
  ### Training results
76
 
77
- | Training Loss | Epoch | Step | Validation Loss | Map | Map 50 | Map 75 | Map Small | Map Medium | Map Large | Mar 1 | Mar 10 | Mar 100 | Mar Small | Mar Medium | Mar Large | Map Player | Mar 100 Player | Map Ball-handler | Mar 100 Ball-handler | Map Ball-handler-floor | Mar 100 Ball-handler-floor | Map Basketball | Mar 100 Basketball | Map Head | Mar 100 Head | Map People | Mar 100 People | Map Player-floor | Mar 100 Player-floor | Map Rim | Mar 100 Rim |
78
- |:-------------:|:-----:|:----:|:---------------:|:------:|:------:|:------:|:---------:|:----------:|:---------:|:------:|:------:|:-------:|:---------:|:----------:|:---------:|:----------:|:--------------:|:----------------:|:--------------------:|:----------------------:|:--------------------------:|:--------------:|:------------------:|:--------:|:------------:|:----------:|:--------------:|:----------------:|:--------------------:|:-------:|:-----------:|
79
- | No log | 1.0 | 232 | 12.1637 | 0.1613 | 0.2423 | 0.1719 | -1.0 | 0.1295 | 0.1419 | 0.2973 | 0.6614 | 0.7375 | -1.0 | 0.7151 | 0.7664 | 0.379 | 0.6798 | 0.2057 | 0.7063 | 0.0366 | 0.7893 | 0.0064 | 0.825 | -1.0 | -1.0 | 0.3062 | 0.84 | 0.0338 | 0.5848 | -1.0 | -1.0 |
80
- | No log | 2.0 | 464 | 11.9739 | 0.1493 | 0.2363 | 0.1613 | -1.0 | 0.1421 | 0.288 | 0.4387 | 0.6973 | 0.7511 | -1.0 | 0.6618 | 0.785 | 0.2426 | 0.6848 | 0.2934 | 0.7139 | 0.1291 | 0.796 | 0.1274 | 0.825 | -1.0 | -1.0 | 0.046 | 0.84 | 0.057 | 0.6469 | -1.0 | -1.0 |
81
- | 16.3591 | 3.0 | 696 | 12.2254 | 0.2604 | 0.3704 | 0.2989 | -1.0 | 0.1998 | 0.2663 | 0.4591 | 0.7305 | 0.7747 | -1.0 | 0.7052 | 0.799 | 0.3183 | 0.682 | 0.421 | 0.7366 | 0.0823 | 0.772 | 0.1393 | 0.85 | -1.0 | -1.0 | 0.5425 | 0.92 | 0.059 | 0.6877 | -1.0 | -1.0 |
82
- | 16.3591 | 4.0 | 928 | 12.6746 | 0.2151 | 0.3231 | 0.2456 | -1.0 | 0.2055 | 0.3285 | 0.4947 | 0.7412 | 0.7692 | -1.0 | 0.7274 | 0.8024 | 0.4005 | 0.6848 | 0.4143 | 0.7277 | 0.1024 | 0.7893 | 0.1678 | 0.925 | -1.0 | -1.0 | 0.1233 | 0.8 | 0.082 | 0.6881 | -1.0 | -1.0 |
83
- | 12.4571 | 5.0 | 1160 | 12.7896 | 0.2785 | 0.4006 | 0.3031 | -1.0 | 0.2273 | 0.3351 | 0.5048 | 0.761 | 0.7904 | -1.0 | 0.7003 | 0.8202 | 0.4383 | 0.6886 | 0.3893 | 0.7286 | 0.1583 | 0.8027 | 0.3371 | 0.925 | -1.0 | -1.0 | 0.2581 | 0.88 | 0.0899 | 0.7177 | -1.0 | -1.0 |
84
- | 12.4571 | 6.0 | 1392 | 12.8732 | 0.3974 | 0.5323 | 0.4427 | -1.0 | 0.3075 | 0.4263 | 0.5162 | 0.7599 | 0.7825 | -1.0 | 0.7101 | 0.819 | 0.4679 | 0.6912 | 0.4184 | 0.721 | 0.1791 | 0.8 | 0.8403 | 0.925 | -1.0 | -1.0 | 0.3377 | 0.86 | 0.1411 | 0.6979 | -1.0 | -1.0 |
85
- | 11.0844 | 7.0 | 1624 | 12.7970 | 0.3535 | 0.5039 | 0.3956 | -1.0 | 0.2993 | 0.4295 | 0.4771 | 0.7622 | 0.7779 | -1.0 | 0.7082 | 0.8071 | 0.4828 | 0.6966 | 0.4018 | 0.713 | 0.1785 | 0.78 | 0.6242 | 0.95 | -1.0 | -1.0 | 0.2852 | 0.82 | 0.1487 | 0.7078 | -1.0 | -1.0 |
86
- | 11.0844 | 8.0 | 1856 | 12.9335 | 0.3639 | 0.5035 | 0.4121 | -1.0 | 0.3213 | 0.411 | 0.4975 | 0.7731 | 0.788 | -1.0 | 0.7143 | 0.8226 | 0.4617 | 0.6905 | 0.3972 | 0.7109 | 0.183 | 0.7773 | 0.5394 | 0.925 | -1.0 | -1.0 | 0.427 | 0.9 | 0.175 | 0.7243 | -1.0 | -1.0 |
87
- | 10.1323 | 9.0 | 2088 | 13.1395 | 0.3272 | 0.4701 | 0.3811 | -1.0 | 0.3284 | 0.3028 | 0.4761 | 0.7724 | 0.7919 | -1.0 | 0.7067 | 0.8194 | 0.474 | 0.6904 | 0.3917 | 0.7265 | 0.1156 | 0.796 | 0.4345 | 0.95 | -1.0 | -1.0 | 0.3848 | 0.88 | 0.1628 | 0.7086 | -1.0 | -1.0 |
88
- | 10.1323 | 10.0 | 2320 | 13.0695 | 0.3422 | 0.4834 | 0.3906 | -1.0 | 0.3255 | 0.3584 | 0.499 | 0.7576 | 0.7741 | -1.0 | 0.6972 | 0.8022 | 0.4799 | 0.6907 | 0.3888 | 0.7239 | 0.1122 | 0.7973 | 0.5208 | 0.9 | -1.0 | -1.0 | 0.3781 | 0.82 | 0.1734 | 0.7128 | -1.0 | -1.0 |
89
 
90
 
91
  ### Framework versions
92
 
93
- - Transformers 4.49.0.dev0
94
  - Pytorch 2.5.1+cu124
95
- - Datasets 3.2.0
96
  - Tokenizers 0.21.0
 
16
 
17
  This model is a fine-tuned version of [PekingU/rtdetr_r50vd_coco_o365](https://huggingface.co/PekingU/rtdetr_r50vd_coco_o365) on an unknown dataset.
18
  It achieves the following results on the evaluation set:
19
+ - Loss: 12.7110
20
+ - Map: 0.2479
21
+ - Map 50: 0.4308
22
+ - Map 75: 0.2793
23
+ - Map Small: -1.0
24
+ - Map Medium: 0.2466
25
+ - Map Large: 0.3394
26
+ - Mar 1: 0.3356
27
+ - Mar 10: 0.6144
28
+ - Mar 100: 0.6671
29
+ - Mar Small: -1.0
30
+ - Mar Medium: 0.6408
31
+ - Mar Large: 0.7885
32
+ - Map Player-krm2: -1.0
33
+ - Mar 100 Player-krm2: -1.0
34
+ - Map Ball-handler: 0.2479
35
+ - Mar 100 Ball-handler: 0.6671
 
 
 
 
 
 
 
 
 
 
 
 
36
 
37
  ## Model description
38
 
 
62
 
63
  ### Training results
64
 
65
+ | Training Loss | Epoch | Step | Validation Loss | Map | Map 50 | Map 75 | Map Small | Map Medium | Map Large | Mar 1 | Mar 10 | Mar 100 | Mar Small | Mar Medium | Mar Large | Map Player-krm2 | Mar 100 Player-krm2 | Map Ball-handler | Mar 100 Ball-handler |
66
+ |:-------------:|:-----:|:----:|:---------------:|:------:|:------:|:------:|:---------:|:----------:|:---------:|:------:|:------:|:-------:|:---------:|:----------:|:---------:|:---------------:|:-------------------:|:----------------:|:--------------------:|
67
+ | No log | 1.0 | 232 | 41.5447 | 0.0261 | 0.0559 | 0.0222 | -1.0 | 0.0217 | 0.0328 | 0.0931 | 0.3912 | 0.569 | -1.0 | 0.4766 | 0.646 | -1.0 | -1.0 | 0.0261 | 0.569 |
68
+ | No log | 2.0 | 464 | 15.2104 | 0.0949 | 0.1776 | 0.0947 | -1.0 | 0.0633 | 0.1464 | 0.1577 | 0.5781 | 0.6903 | -1.0 | 0.629 | 0.7414 | -1.0 | -1.0 | 0.0949 | 0.6903 |
69
+ | 147.7731 | 3.0 | 696 | 14.0317 | 0.1437 | 0.2467 | 0.1575 | -1.0 | 0.1602 | 0.1901 | 0.2461 | 0.6799 | 0.7476 | -1.0 | 0.6772 | 0.8063 | -1.0 | -1.0 | 0.1437 | 0.7476 |
70
+ | 147.7731 | 4.0 | 928 | 12.8751 | 0.2289 | 0.3855 | 0.2488 | -1.0 | 0.2196 | 0.3006 | 0.3486 | 0.6956 | 0.7476 | -1.0 | 0.6759 | 0.8075 | -1.0 | -1.0 | 0.2289 | 0.7476 |
71
+ | 13.1994 | 5.0 | 1160 | 11.8435 | 0.3042 | 0.4872 | 0.3603 | -1.0 | 0.2955 | 0.349 | 0.442 | 0.6981 | 0.7455 | -1.0 | 0.6772 | 0.8023 | -1.0 | -1.0 | 0.3042 | 0.7455 |
72
+ | 13.1994 | 6.0 | 1392 | 12.0752 | 0.2591 | 0.4108 | 0.2987 | -1.0 | 0.2905 | 0.3361 | 0.4295 | 0.7019 | 0.7558 | -1.0 | 0.6966 | 0.8052 | -1.0 | -1.0 | 0.2591 | 0.7558 |
73
+ | 10.1568 | 7.0 | 1624 | 11.7779 | 0.2518 | 0.4233 | 0.2686 | -1.0 | 0.2971 | 0.2998 | 0.4389 | 0.6975 | 0.7408 | -1.0 | 0.6772 | 0.7937 | -1.0 | -1.0 | 0.2518 | 0.7408 |
74
+ | 10.1568 | 8.0 | 1856 | 11.6643 | 0.3024 | 0.4817 | 0.344 | -1.0 | 0.3419 | 0.3257 | 0.4251 | 0.6906 | 0.7414 | -1.0 | 0.6828 | 0.7902 | -1.0 | -1.0 | 0.3024 | 0.7414 |
75
+ | 8.5355 | 9.0 | 2088 | 11.6303 | 0.311 | 0.5091 | 0.3684 | -1.0 | 0.3569 | 0.3358 | 0.4915 | 0.6975 | 0.7389 | -1.0 | 0.6697 | 0.7966 | -1.0 | -1.0 | 0.311 | 0.7389 |
76
+ | 8.5355 | 10.0 | 2320 | 11.6071 | 0.3094 | 0.5061 | 0.345 | -1.0 | 0.3546 | 0.3265 | 0.4661 | 0.6903 | 0.731 | -1.0 | 0.6579 | 0.792 | -1.0 | -1.0 | 0.3094 | 0.731 |
77
 
78
 
79
  ### Framework versions
80
 
81
+ - Transformers 4.50.0.dev0
82
  - Pytorch 2.5.1+cu124
83
+ - Datasets 3.3.1
84
  - Tokenizers 0.21.0
config.json CHANGED
@@ -63,28 +63,15 @@
63
  "freeze_backbone_batch_norms": true,
64
  "hidden_expansion": 1.0,
65
  "id2label": {
66
- "0": "player",
67
- "1": "ball-handler",
68
- "2": "ball-handler-floor",
69
- "3": "basketball",
70
- "4": "head",
71
- "5": "people",
72
- "6": "player",
73
- "7": "player-floor",
74
- "8": "rim"
75
  },
76
  "initializer_bias_prior_prob": null,
77
  "initializer_range": 0.01,
78
  "is_encoder_decoder": true,
79
  "label2id": {
80
  "ball-handler": 1,
81
- "ball-handler-floor": 2,
82
- "basketball": 3,
83
- "head": 4,
84
- "people": 5,
85
- "player": 6,
86
- "player-floor": 7,
87
- "rim": 8
88
  },
89
  "label_noise_ratio": 0.5,
90
  "layer_norm_eps": 1e-05,
@@ -101,7 +88,7 @@
101
  "num_queries": 300,
102
  "positional_encoding_temperature": 10000,
103
  "torch_dtype": "float32",
104
- "transformers_version": "4.49.0.dev0",
105
  "use_focal_loss": true,
106
  "use_pretrained_backbone": false,
107
  "use_timm_backbone": false,
 
63
  "freeze_backbone_batch_norms": true,
64
  "hidden_expansion": 1.0,
65
  "id2label": {
66
+ "0": "player-kRM2",
67
+ "1": "ball-handler"
 
 
 
 
 
 
 
68
  },
69
  "initializer_bias_prior_prob": null,
70
  "initializer_range": 0.01,
71
  "is_encoder_decoder": true,
72
  "label2id": {
73
  "ball-handler": 1,
74
+ "player-kRM2": 0
 
 
 
 
 
 
75
  },
76
  "label_noise_ratio": 0.5,
77
  "layer_norm_eps": 1e-05,
 
88
  "num_queries": 300,
89
  "positional_encoding_temperature": 10000,
90
  "torch_dtype": "float32",
91
+ "transformers_version": "4.50.0.dev0",
92
  "use_focal_loss": true,
93
  "use_pretrained_backbone": false,
94
  "use_timm_backbone": false,
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:75db8cdb1d376ccb3d575f354d426d4e1927b6f71e824078f24a55e85695a341
3
- size 171592228
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:956fdd49f1209c566693bf2db8d0dc8ed616a48cf5b43f1cc24b753e08ce2e9e
3
+ size 171534680
runs/Feb20_15-02-20_f8a56163e644/events.out.tfevents.1740063750.f8a56163e644.432.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5d685cad797da7805436ba3ed033f69cdcecb2b22ea67b49cbf1c692a824c7fd
3
+ size 19205
runs/Feb20_15-02-20_f8a56163e644/events.out.tfevents.1740067951.f8a56163e644.432.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:115f2cf7e696c8ef86edf205127b22f71a494b3c9b98ffff36c26aab1ae3f9c0
3
+ size 1222
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:32ce522aa53c0da6fff3d3eccde330bab740702d2687daad67e36a31b7156eb7
3
  size 5368
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:889ffd9886e1fde260b31f18700de7bde32e1a0a0a73c8b4c9526ec120aadad7
3
  size 5368