Hugging Face's logo

matthieulel
/
dinov2-base-finetuned-galaxy10-decals

Image Classification
Transformers
TensorBoard
Safetensors
dinov2
vision
Generated from Trainer
Model card Files Metrics Community
matthieulel commited on
Commit
d68f17f
·
verified ·
1 Parent(s): 9c315b2

Model save

Browse files
Files changed (2) hide show
  1. README.md +50 -34
  2. model.safetensors +1 -1
README.md CHANGED
@@ -5,20 +5,26 @@ tags:
5
  - generated_from_trainer
6
  metrics:
7
  - accuracy
 
 
 
8
  model-index:
9
- - name: dinov2-base-finetuned-galaxy
10
  results: []
11
  ---
12
 
13
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
14
  should probably proofread and complete it, then remove this comment. -->
15
 
16
- # dinov2-base-finetuned-galaxy
17
 
18
  This model is a fine-tuned version of [facebook/dinov2-base](https://huggingface.co/facebook/dinov2-base) on an unknown dataset.
19
  It achieves the following results on the evaluation set:
20
- - Loss: 0.5583
21
- - Accuracy: 0.8704
 
 
 
22
 
23
  ## Model description
24
 
@@ -38,45 +44,55 @@ More information needed
38
 
39
  The following hyperparameters were used during training:
40
  - learning_rate: 5e-05
41
- - train_batch_size: 32
42
- - eval_batch_size: 32
43
  - seed: 42
44
  - gradient_accumulation_steps: 4
45
- - total_train_batch_size: 128
46
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
47
  - lr_scheduler_type: linear
48
  - lr_scheduler_warmup_ratio: 0.1
49
- - num_epochs: 20
50
 
51
  ### Training results
52
 
53
- | Training Loss | Epoch | Step | Validation Loss | Accuracy |
54
- |:-------------:|:-------:|:----:|:---------------:|:--------:|
55
- | 0.8589 | 0.9978 | 112 | 0.7458 | 0.7508 |
56
- | 0.8465 | 1.9955 | 224 | 0.6569 | 0.7865 |
57
- | 0.732 | 2.9933 | 336 | 0.7021 | 0.7595 |
58
- | 0.7083 | 4.0 | 449 | 0.6094 | 0.7884 |
59
- | 0.662 | 4.9978 | 561 | 0.5464 | 0.8178 |
60
- | 0.6211 | 5.9955 | 673 | 0.5667 | 0.8134 |
61
- | 0.5374 | 6.9933 | 785 | 0.5135 | 0.8297 |
62
- | 0.5147 | 8.0 | 898 | 0.5184 | 0.8284 |
63
- | 0.5101 | 8.9978 | 1010 | 0.5775 | 0.8090 |
64
- | 0.4791 | 9.9955 | 1122 | 0.4965 | 0.8278 |
65
- | 0.4339 | 10.9933 | 1234 | 0.5054 | 0.8303 |
66
- | 0.4709 | 12.0 | 1347 | 0.6274 | 0.8034 |
67
- | 0.4092 | 12.9978 | 1459 | 0.4794 | 0.8485 |
68
- | 0.368 | 13.9955 | 1571 | 0.5032 | 0.8466 |
69
- | 0.3415 | 14.9933 | 1683 | 0.5258 | 0.8410 |
70
- | 0.2963 | 16.0 | 1796 | 0.5382 | 0.8416 |
71
- | 0.2614 | 16.9978 | 1908 | 0.5105 | 0.8560 |
72
- | 0.2542 | 17.9955 | 2020 | 0.5331 | 0.8535 |
73
- | 0.2436 | 18.9933 | 2132 | 0.5468 | 0.8610 |
74
- | 0.2275 | 19.9555 | 2240 | 0.5583 | 0.8704 |
 
 
 
 
 
 
 
 
 
 
75
 
76
 
77
  ### Framework versions
78
 
79
- - Transformers 4.40.0
80
- - Pytorch 2.2.1+cu121
81
- - Datasets 2.19.0
82
- - Tokenizers 0.19.1
 
5
  - generated_from_trainer
6
  metrics:
7
  - accuracy
8
+ - precision
9
+ - recall
10
+ - f1
11
  model-index:
12
+ - name: dinov2-base-finetuned-galaxy10-decals
13
  results: []
14
  ---
15
 
16
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
17
  should probably proofread and complete it, then remove this comment. -->
18
 
19
+ # dinov2-base-finetuned-galaxy10-decals
20
 
21
  This model is a fine-tuned version of [facebook/dinov2-base](https://huggingface.co/facebook/dinov2-base) on an unknown dataset.
22
  It achieves the following results on the evaluation set:
23
+ - Loss: 0.5435
24
+ - Accuracy: 0.8630
25
+ - Precision: 0.8609
26
+ - Recall: 0.8630
27
+ - F1: 0.8606
28
 
29
  ## Model description
30
 
 
44
 
45
  The following hyperparameters were used during training:
46
  - learning_rate: 5e-05
47
+ - train_batch_size: 128
48
+ - eval_batch_size: 128
49
  - seed: 42
50
  - gradient_accumulation_steps: 4
51
+ - total_train_batch_size: 512
52
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
53
  - lr_scheduler_type: linear
54
  - lr_scheduler_warmup_ratio: 0.1
55
+ - num_epochs: 30
56
 
57
  ### Training results
58
 
59
+ | Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 |
60
+ |:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|
61
+ | 1.1698 | 0.99 | 31 | 0.9105 | 0.6635 | 0.6861 | 0.6635 | 0.6488 |
62
+ | 0.7528 | 1.98 | 62 | 0.6494 | 0.7790 | 0.8147 | 0.7790 | 0.7613 |
63
+ | 0.6893 | 2.98 | 93 | 0.6538 | 0.7943 | 0.8020 | 0.7943 | 0.7912 |
64
+ | 0.6554 | 4.0 | 125 | 0.6333 | 0.7886 | 0.8071 | 0.7886 | 0.7773 |
65
+ | 0.6342 | 4.99 | 156 | 0.5532 | 0.8134 | 0.8198 | 0.8134 | 0.8138 |
66
+ | 0.5565 | 5.98 | 187 | 0.5324 | 0.8207 | 0.8343 | 0.8207 | 0.8180 |
67
+ | 0.5475 | 6.98 | 218 | 0.5030 | 0.8354 | 0.8387 | 0.8354 | 0.8343 |
68
+ | 0.5271 | 8.0 | 250 | 0.4834 | 0.8337 | 0.8444 | 0.8337 | 0.8325 |
69
+ | 0.5086 | 8.99 | 281 | 0.4611 | 0.8433 | 0.8441 | 0.8433 | 0.8382 |
70
+ | 0.4341 | 9.98 | 312 | 0.4437 | 0.8506 | 0.8539 | 0.8506 | 0.8509 |
71
+ | 0.4557 | 10.98 | 343 | 0.4559 | 0.8484 | 0.8529 | 0.8484 | 0.8495 |
72
+ | 0.4179 | 12.0 | 375 | 0.5942 | 0.8129 | 0.8257 | 0.8129 | 0.8133 |
73
+ | 0.4243 | 12.99 | 406 | 0.4599 | 0.8540 | 0.8537 | 0.8540 | 0.8518 |
74
+ | 0.372 | 13.98 | 437 | 0.4743 | 0.8410 | 0.8472 | 0.8410 | 0.8403 |
75
+ | 0.4003 | 14.98 | 468 | 0.4749 | 0.8478 | 0.8471 | 0.8478 | 0.8461 |
76
+ | 0.344 | 16.0 | 500 | 0.4678 | 0.8596 | 0.8575 | 0.8596 | 0.8572 |
77
+ | 0.3252 | 16.99 | 531 | 0.5024 | 0.8472 | 0.8470 | 0.8472 | 0.8459 |
78
+ | 0.3166 | 17.98 | 562 | 0.5038 | 0.8439 | 0.8442 | 0.8439 | 0.8418 |
79
+ | 0.2978 | 18.98 | 593 | 0.5240 | 0.8365 | 0.8351 | 0.8365 | 0.8349 |
80
+ | 0.2748 | 20.0 | 625 | 0.5176 | 0.8512 | 0.8497 | 0.8512 | 0.8472 |
81
+ | 0.2691 | 20.99 | 656 | 0.5529 | 0.8534 | 0.8514 | 0.8534 | 0.8506 |
82
+ | 0.2571 | 21.98 | 687 | 0.5441 | 0.8563 | 0.8573 | 0.8563 | 0.8535 |
83
+ | 0.2451 | 22.98 | 718 | 0.5440 | 0.8427 | 0.8427 | 0.8427 | 0.8412 |
84
+ | 0.2256 | 24.0 | 750 | 0.5489 | 0.8506 | 0.8467 | 0.8506 | 0.8474 |
85
+ | 0.2304 | 24.99 | 781 | 0.5695 | 0.8534 | 0.8492 | 0.8534 | 0.8498 |
86
+ | 0.2102 | 25.98 | 812 | 0.5347 | 0.8568 | 0.8533 | 0.8568 | 0.8540 |
87
+ | 0.2172 | 26.98 | 843 | 0.5399 | 0.8523 | 0.8525 | 0.8523 | 0.8518 |
88
+ | 0.1953 | 28.0 | 875 | 0.5699 | 0.8551 | 0.8547 | 0.8551 | 0.8522 |
89
+ | 0.2035 | 28.99 | 906 | 0.5538 | 0.8630 | 0.8630 | 0.8630 | 0.8607 |
90
+ | 0.1926 | 29.76 | 930 | 0.5435 | 0.8630 | 0.8609 | 0.8630 | 0.8606 |
91
 
92
 
93
  ### Framework versions
94
 
95
+ - Transformers 4.37.2
96
+ - Pytorch 2.3.0
97
+ - Datasets 2.19.1
98
+ - Tokenizers 0.15.1
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:36e626a28616290a31749d9cdb11d1b8997ec32b3069d711fb5935ff04aac7ec
3
  size 346409112
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:785f4ddef22804a541c3db231b81d1f7c9f3e66c0d7f36225fb3e360db280e79
3
  size 346409112