Hugging Face's logo

valadhi
/
swin-tiny-patch4-window7-224-finetuned-agrivision

Image Classification
Transformers
PyTorch
swin
Generated from Trainer
Eval Results
Inference Endpoints
Model card Files Community
valadhi commited on
Commit
c05a4a6
·
1 Parent(s): 3e588e7

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +34 -54
README.md CHANGED
@@ -21,7 +21,7 @@ model-index:
21
  metrics:
22
  - name: Accuracy
23
  type: accuracy
24
- value: 0.8410462776659959
25
  ---
26
 
27
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -31,8 +31,8 @@ should probably proofread and complete it, then remove this comment. -->
31
 
32
  This model is a fine-tuned version of [microsoft/swin-tiny-patch4-window7-224](https://huggingface.co/microsoft/swin-tiny-patch4-window7-224) on the imagefolder dataset.
33
  It achieves the following results on the evaluation set:
34
- - Loss: 0.6828
35
- - Accuracy: 0.8410
36
 
37
  ## Model description
38
 
@@ -60,62 +60,42 @@ The following hyperparameters were used during training:
60
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
61
  - lr_scheduler_type: linear
62
  - lr_scheduler_warmup_ratio: 0.1
63
- - num_epochs: 50
64
 
65
  ### Training results
66
 
67
  | Training Loss | Epoch | Step | Validation Loss | Accuracy |
68
  |:-------------:|:-----:|:----:|:---------------:|:--------:|
69
- | 0.9718 | 1.0 | 35 | 0.9797 | 0.6398 |
70
- | 0.2715 | 2.0 | 70 | 0.6352 | 0.7485 |
71
- | 0.1615 | 3.0 | 105 | 0.4674 | 0.8129 |
72
- | 0.0959 | 4.0 | 140 | 0.5570 | 0.7847 |
73
- | 0.0563 | 5.0 | 175 | 0.2903 | 0.8833 |
74
- | 0.0824 | 6.0 | 210 | 0.4108 | 0.8451 |
75
- | 0.0441 | 7.0 | 245 | 0.5063 | 0.8209 |
76
- | 0.0306 | 8.0 | 280 | 0.3053 | 0.8753 |
77
- | 0.0279 | 9.0 | 315 | 0.4467 | 0.8592 |
78
- | 0.0223 | 10.0 | 350 | 0.4209 | 0.8551 |
79
- | 0.026 | 11.0 | 385 | 0.5075 | 0.8531 |
80
- | 0.0344 | 12.0 | 420 | 0.4922 | 0.8551 |
81
- | 0.0079 | 13.0 | 455 | 0.7256 | 0.7807 |
82
- | 0.0156 | 14.0 | 490 | 0.5177 | 0.8370 |
83
- | 0.0184 | 15.0 | 525 | 0.5858 | 0.8350 |
84
- | 0.0121 | 16.0 | 560 | 0.4655 | 0.8652 |
85
- | 0.0089 | 17.0 | 595 | 0.9852 | 0.7586 |
86
- | 0.028 | 18.0 | 630 | 0.5087 | 0.8571 |
87
- | 0.003 | 19.0 | 665 | 0.5447 | 0.8491 |
88
- | 0.0015 | 20.0 | 700 | 0.5610 | 0.8390 |
89
- | 0.008 | 21.0 | 735 | 0.5702 | 0.8571 |
90
- | 0.0071 | 22.0 | 770 | 0.6043 | 0.8451 |
91
- | 0.0006 | 23.0 | 805 | 0.5951 | 0.8451 |
92
- | 0.0077 | 24.0 | 840 | 0.6436 | 0.8310 |
93
- | 0.0007 | 25.0 | 875 | 0.4439 | 0.8793 |
94
- | 0.0081 | 26.0 | 910 | 0.4689 | 0.8732 |
95
- | 0.0036 | 27.0 | 945 | 0.5058 | 0.8712 |
96
- | 0.0004 | 28.0 | 980 | 0.5426 | 0.8592 |
97
- | 0.0007 | 29.0 | 1015 | 0.4835 | 0.8632 |
98
- | 0.005 | 30.0 | 1050 | 0.4958 | 0.8652 |
99
- | 0.001 | 31.0 | 1085 | 0.7008 | 0.8390 |
100
- | 0.0022 | 32.0 | 1120 | 0.6210 | 0.8632 |
101
- | 0.0029 | 33.0 | 1155 | 0.6328 | 0.8471 |
102
- | 0.0001 | 34.0 | 1190 | 0.5887 | 0.8551 |
103
- | 0.0019 | 35.0 | 1225 | 0.6666 | 0.8390 |
104
- | 0.0028 | 36.0 | 1260 | 0.6372 | 0.8571 |
105
- | 0.0077 | 37.0 | 1295 | 0.5972 | 0.8632 |
106
- | 0.0019 | 38.0 | 1330 | 0.5053 | 0.8753 |
107
- | 0.0042 | 39.0 | 1365 | 0.8244 | 0.8249 |
108
- | 0.0022 | 40.0 | 1400 | 0.7626 | 0.8410 |
109
- | 0.0036 | 41.0 | 1435 | 0.6884 | 0.8410 |
110
- | 0.0016 | 42.0 | 1470 | 0.6704 | 0.8410 |
111
- | 0.0011 | 43.0 | 1505 | 0.5821 | 0.8531 |
112
- | 0.0001 | 44.0 | 1540 | 0.5815 | 0.8571 |
113
- | 0.0003 | 45.0 | 1575 | 0.6694 | 0.8431 |
114
- | 0.0007 | 46.0 | 1610 | 0.6877 | 0.8431 |
115
- | 0.0 | 47.0 | 1645 | 0.6863 | 0.8390 |
116
- | 0.0 | 48.0 | 1680 | 0.6967 | 0.8431 |
117
- | 0.0001 | 49.0 | 1715 | 0.6851 | 0.8410 |
118
- | 0.0 | 50.0 | 1750 | 0.6828 | 0.8410 |
119
 
120
 
121
  ### Framework versions
 
21
  metrics:
22
  - name: Accuracy
23
  type: accuracy
24
+ value: 0.9362186788154897
25
  ---
26
 
27
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
31
 
32
  This model is a fine-tuned version of [microsoft/swin-tiny-patch4-window7-224](https://huggingface.co/microsoft/swin-tiny-patch4-window7-224) on the imagefolder dataset.
33
  It achieves the following results on the evaluation set:
34
+ - Loss: 0.2783
35
+ - Accuracy: 0.9362
36
 
37
  ## Model description
38
 
 
60
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
61
  - lr_scheduler_type: linear
62
  - lr_scheduler_warmup_ratio: 0.1
63
+ - num_epochs: 30
64
 
65
  ### Training results
66
 
67
  | Training Loss | Epoch | Step | Validation Loss | Accuracy |
68
  |:-------------:|:-----:|:----:|:---------------:|:--------:|
69
+ | 0.5829 | 1.0 | 31 | 0.7480 | 0.7267 |
70
+ | 0.1199 | 2.0 | 62 | 0.4407 | 0.8246 |
71
+ | 0.1028 | 3.0 | 93 | 0.4477 | 0.8246 |
72
+ | 0.0533 | 4.0 | 124 | 0.4606 | 0.8292 |
73
+ | 0.0411 | 5.0 | 155 | 0.2470 | 0.9180 |
74
+ | 0.022 | 6.0 | 186 | 0.1568 | 0.9544 |
75
+ | 0.0206 | 7.0 | 217 | 0.4187 | 0.8793 |
76
+ | 0.0069 | 8.0 | 248 | 0.2498 | 0.9203 |
77
+ | 0.0053 | 9.0 | 279 | 0.2654 | 0.9226 |
78
+ | 0.0094 | 10.0 | 310 | 0.2343 | 0.9385 |
79
+ | 0.0152 | 11.0 | 341 | 0.3421 | 0.9021 |
80
+ | 0.0047 | 12.0 | 372 | 0.4494 | 0.8724 |
81
+ | 0.0128 | 13.0 | 403 | 0.5360 | 0.8679 |
82
+ | 0.0024 | 14.0 | 434 | 0.2775 | 0.9112 |
83
+ | 0.0127 | 15.0 | 465 | 0.2911 | 0.8975 |
84
+ | 0.0038 | 16.0 | 496 | 0.2337 | 0.9294 |
85
+ | 0.0001 | 17.0 | 527 | 0.2207 | 0.9408 |
86
+ | 0.0054 | 18.0 | 558 | 0.2506 | 0.9362 |
87
+ | 0.0011 | 19.0 | 589 | 0.3778 | 0.8952 |
88
+ | 0.0002 | 20.0 | 620 | 0.2316 | 0.9408 |
89
+ | 0.0003 | 21.0 | 651 | 0.2133 | 0.9431 |
90
+ | 0.0009 | 22.0 | 682 | 0.2519 | 0.9339 |
91
+ | 0.0004 | 23.0 | 713 | 0.2931 | 0.9203 |
92
+ | 0.0001 | 24.0 | 744 | 0.2847 | 0.9271 |
93
+ | 0.0003 | 25.0 | 775 | 0.2831 | 0.9317 |
94
+ | 0.0008 | 26.0 | 806 | 0.2919 | 0.9271 |
95
+ | 0.0003 | 27.0 | 837 | 0.2798 | 0.9362 |
96
+ | 0.0008 | 28.0 | 868 | 0.2857 | 0.9362 |
97
+ | 0.0008 | 29.0 | 899 | 0.2780 | 0.9362 |
98
+ | 0.0013 | 30.0 | 930 | 0.2783 | 0.9362 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
99
 
100
 
101
  ### Framework versions