Hugging Face's logo

ChingizA
/
detr

Object Detection
Transformers
TensorBoard
Safetensors
detr
Generated from Trainer
Model card Files
xet
Metrics Community
ChingizA commited on
Commit
5a3c58b
·
verified ·
1 Parent(s): 89dfc79

End of training

Browse files
Files changed (1) hide show
  1. README.md +54 -29
README.md CHANGED
@@ -15,7 +15,7 @@ should probably proofread and complete it, then remove this comment. -->
15
 
16
  This model is a fine-tuned version of [facebook/detr-resnet-50](https://huggingface.co/facebook/detr-resnet-50) on an unknown dataset.
17
  It achieves the following results on the evaluation set:
18
- - Loss: 0.7342
19
 
20
  ## Model description
21
 
@@ -40,38 +40,63 @@ The following hyperparameters were used during training:
40
  - seed: 42
41
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
42
  - lr_scheduler_type: linear
43
- - num_epochs: 2
44
  - mixed_precision_training: Native AMP
45
 
46
  ### Training results
47
 
48
- | Training Loss | Epoch | Step | Validation Loss |
49
- |:-------------:|:-----:|:----:|:---------------:|
50
- | 1.1153 | 0.08 | 200 | 1.0413 |
51
- | 1.0375 | 0.16 | 400 | 0.9910 |
52
- | 1.0557 | 0.24 | 600 | 0.9525 |
53
- | 0.9463 | 0.32 | 800 | 0.9226 |
54
- | 0.9683 | 0.4 | 1000 | 0.9185 |
55
- | 0.9448 | 0.48 | 1200 | 0.8853 |
56
- | 0.9343 | 0.56 | 1400 | 0.9085 |
57
- | 0.9047 | 0.64 | 1600 | 0.8637 |
58
- | 0.9429 | 0.72 | 1800 | 0.8998 |
59
- | 0.846 | 0.8 | 2000 | 0.8494 |
60
- | 0.9111 | 0.88 | 2200 | 0.8380 |
61
- | 0.8987 | 0.96 | 2400 | 0.8217 |
62
- | 0.8348 | 1.04 | 2600 | 0.7994 |
63
- | 0.8578 | 1.12 | 2800 | 0.8096 |
64
- | 0.8063 | 1.2 | 3000 | 0.7920 |
65
- | 0.7922 | 1.28 | 3200 | 0.7797 |
66
- | 0.805 | 1.36 | 3400 | 0.7697 |
67
- | 0.8121 | 1.44 | 3600 | 0.7722 |
68
- | 0.798 | 1.52 | 3800 | 0.7664 |
69
- | 0.79 | 1.6 | 4000 | 0.7601 |
70
- | 0.7406 | 1.68 | 4200 | 0.7507 |
71
- | 0.7323 | 1.76 | 4400 | 0.7330 |
72
- | 0.8103 | 1.84 | 4600 | 0.7295 |
73
- | 0.7509 | 1.92 | 4800 | 0.7334 |
74
- | 0.7615 | 2.0 | 5000 | 0.7342 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
75
 
76
 
77
  ### Framework versions
 
15
 
16
  This model is a fine-tuned version of [facebook/detr-resnet-50](https://huggingface.co/facebook/detr-resnet-50) on an unknown dataset.
17
  It achieves the following results on the evaluation set:
18
+ - Loss: 0.7117
19
 
20
  ## Model description
21
 
 
40
  - seed: 42
41
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
42
  - lr_scheduler_type: linear
43
+ - num_epochs: 4
44
  - mixed_precision_training: Native AMP
45
 
46
  ### Training results
47
 
48
+ | Training Loss | Epoch | Step | Validation Loss |
49
+ |:-------------:|:-----:|:-----:|:---------------:|
50
+ | 2.0345 | 0.08 | 200 | 1.8447 |
51
+ | 1.5511 | 0.16 | 400 | 1.4217 |
52
+ | 1.444 | 0.24 | 600 | 1.3814 |
53
+ | 1.3746 | 0.32 | 800 | 1.3241 |
54
+ | 1.2361 | 0.4 | 1000 | 1.2589 |
55
+ | 1.3506 | 0.48 | 1200 | 1.2441 |
56
+ | 1.2833 | 0.56 | 1400 | 1.2052 |
57
+ | 1.1051 | 0.64 | 1600 | 1.0607 |
58
+ | 1.1091 | 0.72 | 1800 | 1.0610 |
59
+ | 1.0295 | 0.8 | 2000 | 1.0241 |
60
+ | 1.1376 | 0.88 | 2200 | 1.0846 |
61
+ | 1.1172 | 0.96 | 2400 | 1.1095 |
62
+ | 1.0186 | 1.04 | 2600 | 0.9978 |
63
+ | 1.0775 | 1.12 | 2800 | 1.0225 |
64
+ | 0.9973 | 1.2 | 3000 | 0.9934 |
65
+ | 1.006 | 1.28 | 3200 | 0.9886 |
66
+ | 0.9814 | 1.36 | 3400 | 0.9256 |
67
+ | 1.0253 | 1.44 | 3600 | 0.9209 |
68
+ | 0.9932 | 1.52 | 3800 | 0.9159 |
69
+ | 0.9307 | 1.6 | 4000 | 0.9058 |
70
+ | 0.9103 | 1.68 | 4200 | 0.9049 |
71
+ | 0.9034 | 1.76 | 4400 | 0.8643 |
72
+ | 0.9544 | 1.84 | 4600 | 0.9114 |
73
+ | 0.889 | 1.92 | 4800 | 0.8880 |
74
+ | 0.8888 | 2.0 | 5000 | 0.8515 |
75
+ | 0.8877 | 2.08 | 5200 | 0.8707 |
76
+ | 0.8799 | 2.16 | 5400 | 0.8458 |
77
+ | 0.8398 | 2.24 | 5600 | 0.8292 |
78
+ | 0.8181 | 2.32 | 5800 | 0.8226 |
79
+ | 0.8876 | 2.4 | 6000 | 0.8021 |
80
+ | 0.8893 | 2.48 | 6200 | 0.8173 |
81
+ | 0.8497 | 2.56 | 6400 | 0.7870 |
82
+ | 0.8369 | 2.64 | 6600 | 0.7719 |
83
+ | 0.8213 | 2.72 | 6800 | 0.7877 |
84
+ | 0.8044 | 2.8 | 7000 | 0.7763 |
85
+ | 0.8087 | 2.88 | 7200 | 0.7702 |
86
+ | 0.7616 | 2.96 | 7400 | 0.7570 |
87
+ | 0.7901 | 3.04 | 7600 | 0.7451 |
88
+ | 0.8454 | 3.12 | 7800 | 0.7560 |
89
+ | 0.7428 | 3.2 | 8000 | 0.7455 |
90
+ | 0.822 | 3.28 | 8200 | 0.7390 |
91
+ | 0.8293 | 3.36 | 8400 | 0.7324 |
92
+ | 0.7196 | 3.44 | 8600 | 0.7270 |
93
+ | 0.7508 | 3.52 | 8800 | 0.7357 |
94
+ | 0.783 | 3.6 | 9000 | 0.7293 |
95
+ | 0.7094 | 3.68 | 9200 | 0.7276 |
96
+ | 0.7811 | 3.76 | 9400 | 0.7178 |
97
+ | 0.7765 | 3.84 | 9600 | 0.7129 |
98
+ | 0.7542 | 3.92 | 9800 | 0.7165 |
99
+ | 0.756 | 4.0 | 10000 | 0.7117 |
100
 
101
 
102
  ### Framework versions