Commit
•
12fa37c
1
Parent(s):
b7bb1b9
Upload README.md
Browse files
README.md
CHANGED
|
@@ -1,135 +1,192 @@
|
|
|
|
|
| 1 |
---
|
| 2 |
-
|
| 3 |
-
|
|
|
|
| 4 |
tags:
|
|
|
|
|
|
|
| 5 |
- generated_from_trainer
|
| 6 |
-
|
| 7 |
-
- accuracy
|
| 8 |
model-index:
|
| 9 |
- name: DinoVdeau_Aina-large-2024_06_12-batch-size32_epochs150_freeze
|
| 10 |
results: []
|
| 11 |
---
|
| 12 |
|
| 13 |
-
|
| 14 |
-
should probably proofread and complete it, then remove this comment. -->
|
| 15 |
|
| 16 |
-
# DinoVdeau_Aina-large-2024_06_12-batch-size32_epochs150_freeze
|
| 17 |
-
|
| 18 |
-
This model is a fine-tuned version of [facebook/dinov2-large](https://huggingface.co/facebook/dinov2-large) on the None dataset.
|
| 19 |
-
It achieves the following results on the evaluation set:
|
| 20 |
- Loss: 0.1378
|
| 21 |
- F1 Micro: 0.8118
|
| 22 |
- F1 Macro: 0.5888
|
| 23 |
- Roc Auc: 0.8738
|
| 24 |
- Accuracy: 0.5906
|
| 25 |
-
- Learning Rate: 0.0000
|
| 26 |
|
| 27 |
-
|
|
|
|
|
|
|
|
|
|
| 28 |
|
| 29 |
-
|
| 30 |
|
| 31 |
-
|
| 32 |
|
| 33 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 34 |
|
| 35 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 36 |
|
| 37 |
-
|
| 38 |
|
| 39 |
-
|
| 40 |
|
| 41 |
-
|
| 42 |
|
| 43 |
The following hyperparameters were used during training:
|
| 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 |
-
| 0.
|
| 78 |
-
| 0.
|
| 79 |
-
| 0.
|
| 80 |
-
| 0.
|
| 81 |
-
| 0.
|
| 82 |
-
| 0.
|
| 83 |
-
| 0.
|
| 84 |
-
| 0.
|
| 85 |
-
| 0.
|
| 86 |
-
| 0.
|
| 87 |
-
| 0.
|
| 88 |
-
| 0.
|
| 89 |
-
| 0.
|
| 90 |
-
| 0.
|
| 91 |
-
| 0.
|
| 92 |
-
| 0.
|
| 93 |
-
| 0.
|
| 94 |
-
| 0.
|
| 95 |
-
| 0.
|
| 96 |
-
| 0.
|
| 97 |
-
| 0.
|
| 98 |
-
| 0.
|
| 99 |
-
| 0.
|
| 100 |
-
| 0.
|
| 101 |
-
| 0.
|
| 102 |
-
| 0.
|
| 103 |
-
| 0.
|
| 104 |
-
| 0.
|
| 105 |
-
| 0.
|
| 106 |
-
| 0.
|
| 107 |
-
| 0.
|
| 108 |
-
| 0.
|
| 109 |
-
| 0.
|
| 110 |
-
| 0.
|
| 111 |
-
| 0.
|
| 112 |
-
| 0.
|
| 113 |
-
| 0.
|
| 114 |
-
| 0.
|
| 115 |
-
| 0.
|
| 116 |
-
| 0.
|
| 117 |
-
| 0.
|
| 118 |
-
| 0.
|
| 119 |
-
| 0.
|
| 120 |
-
| 0.
|
| 121 |
-
| 0.
|
| 122 |
-
| 0.
|
| 123 |
-
| 0.
|
| 124 |
-
| 0.
|
| 125 |
-
| 0.
|
| 126 |
-
| 0.
|
| 127 |
-
| 0.
|
| 128 |
-
|
| 129 |
-
|
| 130 |
-
|
| 131 |
-
|
| 132 |
-
|
| 133 |
-
|
| 134 |
-
|
| 135 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
|
| 2 |
---
|
| 3 |
+
language:
|
| 4 |
+
- eng
|
| 5 |
+
license: wtfpl
|
| 6 |
tags:
|
| 7 |
+
- multilabel-image-classification
|
| 8 |
+
- multilabel
|
| 9 |
- generated_from_trainer
|
| 10 |
+
base_model: facebook/dinov2-large
|
|
|
|
| 11 |
model-index:
|
| 12 |
- name: DinoVdeau_Aina-large-2024_06_12-batch-size32_epochs150_freeze
|
| 13 |
results: []
|
| 14 |
---
|
| 15 |
|
| 16 |
+
DinoVd'eau is a fine-tuned version of [facebook/dinov2-large](https://huggingface.co/facebook/dinov2-large). It achieves the following results on the test set:
|
|
|
|
| 17 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
- Loss: 0.1378
|
| 19 |
- F1 Micro: 0.8118
|
| 20 |
- F1 Macro: 0.5888
|
| 21 |
- Roc Auc: 0.8738
|
| 22 |
- Accuracy: 0.5906
|
|
|
|
| 23 |
|
| 24 |
+
---
|
| 25 |
+
|
| 26 |
+
# Model description
|
| 27 |
+
DinoVd'eau is a model built on top of dinov2 model for underwater multilabel image classification.The classification head is a combination of linear, ReLU, batch normalization, and dropout layers.
|
| 28 |
|
| 29 |
+
The source code for training the model can be found in this [Git repository](https://github.com/SeatizenDOI/DinoVdeau).
|
| 30 |
|
| 31 |
+
- **Developed by:** [lombardata](https://huggingface.co/lombardata), credits to [César Leblanc](https://huggingface.co/CesarLeblanc) and [Victor Illien](https://huggingface.co/groderg)
|
| 32 |
|
| 33 |
+
---
|
| 34 |
+
|
| 35 |
+
# Intended uses & limitations
|
| 36 |
+
You can use the raw model for classify diverse marine species, encompassing coral morphotypes classes taken from the Global Coral Reef Monitoring Network (GCRMN), habitats classes and seagrass species.
|
| 37 |
+
|
| 38 |
+
---
|
| 39 |
|
| 40 |
+
# Training and evaluation data
|
| 41 |
+
Details on the number of images for each class are given in the following table:
|
| 42 |
+
| Class | train | val | test | Total |
|
| 43 |
+
|:--------|--------:|------:|-------:|--------:|
|
| 44 |
+
| Acr | 509 | 170 | 170 | 849 |
|
| 45 |
+
| Ech | 149 | 55 | 49 | 253 |
|
| 46 |
+
| Gal | 149 | 49 | 52 | 250 |
|
| 47 |
+
| Mtp | 278 | 93 | 92 | 463 |
|
| 48 |
+
| Poc | 166 | 54 | 60 | 280 |
|
| 49 |
+
| Por | 265 | 88 | 88 | 441 |
|
| 50 |
+
| ALGAE | 1221 | 407 | 407 | 2035 |
|
| 51 |
+
| RDC | 185 | 65 | 69 | 319 |
|
| 52 |
+
| SG | 1388 | 463 | 462 | 2313 |
|
| 53 |
+
| P | 198 | 66 | 66 | 330 |
|
| 54 |
+
| R | 1106 | 368 | 369 | 1843 |
|
| 55 |
+
| S | 2178 | 726 | 726 | 3630 |
|
| 56 |
+
| UNK | 132 | 44 | 44 | 220 |
|
| 57 |
|
| 58 |
+
---
|
| 59 |
|
| 60 |
+
# Training procedure
|
| 61 |
|
| 62 |
+
## Training hyperparameters
|
| 63 |
|
| 64 |
The following hyperparameters were used during training:
|
| 65 |
+
|
| 66 |
+
- **Number of Epochs**: 150
|
| 67 |
+
- **Learning Rate**: 0.001
|
| 68 |
+
- **Train Batch Size**: 32
|
| 69 |
+
- **Eval Batch Size**: 32
|
| 70 |
+
- **Optimizer**: Adam
|
| 71 |
+
- **LR Scheduler Type**: ReduceLROnPlateau with a patience of 5 epochs and a factor of 0.1
|
| 72 |
+
- **Freeze Encoder**: Yes
|
| 73 |
+
- **Data Augmentation**: Yes
|
| 74 |
+
|
| 75 |
+
|
| 76 |
+
## Data Augmentation
|
| 77 |
+
Data were augmented using the following transformations :
|
| 78 |
+
|
| 79 |
+
Train Transforms
|
| 80 |
+
- **PreProcess**: No additional parameters
|
| 81 |
+
- **Resize**: probability=1.00
|
| 82 |
+
- **RandomHorizontalFlip**: probability=0.25
|
| 83 |
+
- **RandomVerticalFlip**: probability=0.25
|
| 84 |
+
- **ColorJiggle**: probability=0.25
|
| 85 |
+
- **RandomPerspective**: probability=0.25
|
| 86 |
+
- **Normalize**: probability=1.00
|
| 87 |
+
|
| 88 |
+
Val Transforms
|
| 89 |
+
- **PreProcess**: No additional parameters
|
| 90 |
+
- **Resize**: probability=1.00
|
| 91 |
+
- **Normalize**: probability=1.00
|
| 92 |
+
|
| 93 |
+
|
| 94 |
+
|
| 95 |
+
## Training results
|
| 96 |
+
Epoch | Validation Loss | Accuracy | F1 Macro | F1 Micro | Learning Rate
|
| 97 |
+
--- | --- | --- | --- | --- | ---
|
| 98 |
+
1 | 0.25564512610435486 | 0.5342987804878049 | 0.7703523693803159 | 0.42604577042963465 | 0.001
|
| 99 |
+
2 | 0.1855568140745163 | 0.5449695121951219 | 0.7607875994972768 | 0.40573815087342935 | 0.001
|
| 100 |
+
3 | 0.15800759196281433 | 0.5335365853658537 | 0.7772908366533864 | 0.47696929852236075 | 0.001
|
| 101 |
+
4 | 0.15843382477760315 | 0.5274390243902439 | 0.7700247729149464 | 0.39442465091480916 | 0.001
|
| 102 |
+
5 | 0.15615205466747284 | 0.5442073170731707 | 0.761375774407178 | 0.3747774830573479 | 0.001
|
| 103 |
+
6 | 0.1484147012233734 | 0.5548780487804879 | 0.7822349570200573 | 0.4447304719957427 | 0.001
|
| 104 |
+
7 | 0.1444726288318634 | 0.5647865853658537 | 0.7904456041750301 | 0.5284590192832462 | 0.001
|
| 105 |
+
8 | 0.14579781889915466 | 0.5487804878048781 | 0.7778469197261979 | 0.48048557941722936 | 0.001
|
| 106 |
+
9 | 0.1419043242931366 | 0.5655487804878049 | 0.7989700930877401 | 0.4775190765596756 | 0.001
|
| 107 |
+
10 | 0.14263293147087097 | 0.5663109756097561 | 0.7916750858759345 | 0.5507171237721333 | 0.001
|
| 108 |
+
11 | 0.1459268480539322 | 0.555640243902439 | 0.7765845441145505 | 0.43142510371678144 | 0.001
|
| 109 |
+
12 | 0.14237171411514282 | 0.5625 | 0.7899022801302932 | 0.47044459599035204 | 0.001
|
| 110 |
+
13 | 0.14195148646831512 | 0.5586890243902439 | 0.7957159857199523 | 0.5304167061533165 | 0.001
|
| 111 |
+
14 | 0.14145776629447937 | 0.5685975609756098 | 0.795869737887212 | 0.5302959882181708 | 0.001
|
| 112 |
+
15 | 0.14466659724712372 | 0.5746951219512195 | 0.7903159622486664 | 0.5020556893881151 | 0.001
|
| 113 |
+
16 | 0.15050330758094788 | 0.5548780487804879 | 0.7779618889809444 | 0.44075758480309546 | 0.001
|
| 114 |
+
17 | 0.15037894248962402 | 0.5625 | 0.7849117174959872 | 0.5071297581342885 | 0.001
|
| 115 |
+
18 | 0.15835434198379517 | 0.5632621951219512 | 0.7868521879411171 | 0.4937774793717716 | 0.001
|
| 116 |
+
19 | 0.13946771621704102 | 0.5678353658536586 | 0.795441147573197 | 0.52511737039556 | 0.001
|
| 117 |
+
20 | 0.1404852569103241 | 0.5678353658536586 | 0.8031007751937984 | 0.5902607630639873 | 0.001
|
| 118 |
+
21 | 0.14341644942760468 | 0.5640243902439024 | 0.7965933848286789 | 0.4816903721736541 | 0.001
|
| 119 |
+
22 | 0.14662735164165497 | 0.5510670731707317 | 0.7923046721633294 | 0.5288404087153705 | 0.001
|
| 120 |
+
23 | 0.14562036097049713 | 0.5746951219512195 | 0.7918968692449356 | 0.4974177137762122 | 0.001
|
| 121 |
+
24 | 0.13980671763420105 | 0.5586890243902439 | 0.7888934258881176 | 0.5008214078450646 | 0.001
|
| 122 |
+
25 | 0.13920389115810394 | 0.5807926829268293 | 0.8018232263178755 | 0.5881062115996991 | 0.001
|
| 123 |
+
26 | 0.14581459760665894 | 0.5846036585365854 | 0.8024120603015076 | 0.5378532463138629 | 0.001
|
| 124 |
+
27 | 0.1388114094734192 | 0.5716463414634146 | 0.7989535117729925 | 0.530975959272302 | 0.001
|
| 125 |
+
28 | 0.14750176668167114 | 0.5647865853658537 | 0.7952771662997797 | 0.4926175534546719 | 0.001
|
| 126 |
+
29 | 0.14280082285404205 | 0.5754573170731707 | 0.7916238965304866 | 0.4730643714031951 | 0.001
|
| 127 |
+
30 | 0.14457112550735474 | 0.5647865853658537 | 0.7960474308300396 | 0.5269721557360777 | 0.001
|
| 128 |
+
31 | 0.2518298327922821 | 0.555640243902439 | 0.7859719438877755 | 0.5163333467519122 | 0.001
|
| 129 |
+
32 | 0.13625293970108032 | 0.573170731707317 | 0.7984836392657622 | 0.5223572225800281 | 0.001
|
| 130 |
+
33 | 0.14134813845157623 | 0.583079268292683 | 0.797995991983968 | 0.5094906036573911 | 0.001
|
| 131 |
+
34 | 0.13918223977088928 | 0.5617378048780488 | 0.7939271255060729 | 0.539855430008089 | 0.001
|
| 132 |
+
35 | 0.14003774523735046 | 0.5876524390243902 | 0.8020833333333333 | 0.48732799827507844 | 0.001
|
| 133 |
+
36 | 0.15099692344665527 | 0.5853658536585366 | 0.8016096579476861 | 0.514951340134722 | 0.001
|
| 134 |
+
37 | 0.14428909122943878 | 0.5571646341463414 | 0.7900541407659917 | 0.48501044595638915 | 0.001
|
| 135 |
+
38 | 0.14405055344104767 | 0.5708841463414634 | 0.7945869521308826 | 0.5062759605333714 | 0.001
|
| 136 |
+
39 | 0.13543353974819183 | 0.5891768292682927 | 0.8024193548387097 | 0.535367946456459 | 0.0001
|
| 137 |
+
40 | 0.1358059197664261 | 0.59375 | 0.8035498184751916 | 0.5341458086147945 | 0.0001
|
| 138 |
+
41 | 0.13504844903945923 | 0.5945121951219512 | 0.8055001992825829 | 0.5376190537551451 | 0.0001
|
| 139 |
+
42 | 0.13691695034503937 | 0.5891768292682927 | 0.8035073734555599 | 0.5479834256786992 | 0.0001
|
| 140 |
+
43 | 0.13574542105197906 | 0.586890243902439 | 0.8031840796019901 | 0.5561858306297028 | 0.0001
|
| 141 |
+
44 | 0.13488240540027618 | 0.5891768292682927 | 0.8039920159680638 | 0.5433240630212841 | 0.0001
|
| 142 |
+
45 | 0.1361350268125534 | 0.5815548780487805 | 0.8024593415311385 | 0.5564064426334538 | 0.0001
|
| 143 |
+
46 | 0.1349516659975052 | 0.59375 | 0.8072669826224328 | 0.5738450359977211 | 0.0001
|
| 144 |
+
47 | 0.13875022530555725 | 0.586890243902439 | 0.8035892323030908 | 0.5398873319319681 | 0.0001
|
| 145 |
+
48 | 0.1370573788881302 | 0.586890243902439 | 0.8029458598726115 | 0.5475297928828294 | 0.0001
|
| 146 |
+
49 | 0.13690504431724548 | 0.5769817073170732 | 0.800396432111001 | 0.5558247567089973 | 0.0001
|
| 147 |
+
50 | 0.13561294972896576 | 0.5853658536585366 | 0.8055390702274975 | 0.5774903390899776 | 0.0001
|
| 148 |
+
51 | 0.13570135831832886 | 0.5899390243902439 | 0.8056984566679858 | 0.5639229458996943 | 1e-05
|
| 149 |
+
52 | 0.13534915447235107 | 0.5876524390243902 | 0.803262383131092 | 0.5523965140127979 | 1e-05
|
| 150 |
+
53 | 0.13460643589496613 | 0.5891768292682927 | 0.8047619047619048 | 0.5553165219596603 | 1e-05
|
| 151 |
+
54 | 0.13505160808563232 | 0.586890243902439 | 0.80398406374502 | 0.5513512052592927 | 1e-05
|
| 152 |
+
55 | 0.13664333522319794 | 0.5899390243902439 | 0.8044295036582955 | 0.5586243767362515 | 1e-05
|
| 153 |
+
56 | 0.13582760095596313 | 0.5876524390243902 | 0.8056215360253365 | 0.559616045636613 | 1e-05
|
| 154 |
+
57 | 0.1356770098209381 | 0.5891768292682927 | 0.8046205935072694 | 0.5592413950447728 | 1e-05
|
| 155 |
+
58 | 0.13557715713977814 | 0.586890243902439 | 0.804201347602061 | 0.5652648109787017 | 1e-05
|
| 156 |
+
59 | 0.13481777906417847 | 0.5884146341463414 | 0.806522171405846 | 0.564174198827338 | 1e-05
|
| 157 |
+
60 | 0.13674791157245636 | 0.5899390243902439 | 0.8047030689517736 | 0.5543815900756193 | 1.0000000000000002e-06
|
| 158 |
+
61 | 0.13405902683734894 | 0.5876524390243902 | 0.8045563549160671 | 0.5561036196553861 | 1.0000000000000002e-06
|
| 159 |
+
62 | 0.134480819106102 | 0.5861280487804879 | 0.8052360174533915 | 0.5646122700508593 | 1.0000000000000002e-06
|
| 160 |
+
63 | 0.13709656894207 | 0.5899390243902439 | 0.805958291956306 | 0.5660740924208384 | 1.0000000000000002e-06
|
| 161 |
+
64 | 0.13475064933300018 | 0.5929878048780488 | 0.8056551174830745 | 0.5521602551489271 | 1.0000000000000002e-06
|
| 162 |
+
65 | 0.13584908843040466 | 0.5899390243902439 | 0.8048345551812959 | 0.5594424812390786 | 1.0000000000000002e-06
|
| 163 |
+
66 | 0.13522611558437347 | 0.5853658536585366 | 0.8040500297796307 | 0.5680592850157375 | 1.0000000000000002e-06
|
| 164 |
+
67 | 0.1348220705986023 | 0.5876524390243902 | 0.8059288537549407 | 0.5689511664868826 | 1.0000000000000002e-06
|
| 165 |
+
68 | 0.13475003838539124 | 0.5853658536585366 | 0.8046843985708614 | 0.5680973012829017 | 1.0000000000000002e-07
|
| 166 |
+
69 | 0.13508079946041107 | 0.5884146341463414 | 0.8036036036036036 | 0.5523522321966495 | 1.0000000000000002e-07
|
| 167 |
+
70 | 0.13654960691928864 | 0.5853658536585366 | 0.8053744319304486 | 0.567977027563922 | 1.0000000000000002e-07
|
| 168 |
+
71 | 0.13474920392036438 | 0.5907012195121951 | 0.8072885719944544 | 0.5700988754371263 | 1.0000000000000002e-07
|
| 169 |
+
|
| 170 |
+
|
| 171 |
+
---
|
| 172 |
+
|
| 173 |
+
# CO2 Emissions
|
| 174 |
+
|
| 175 |
+
The estimated CO2 emissions for training this model are documented below:
|
| 176 |
+
|
| 177 |
+
- **Emissions**: 0.2101236289384968 grams of CO2
|
| 178 |
+
- **Source**: Code Carbon
|
| 179 |
+
- **Training Type**: fine-tuning
|
| 180 |
+
- **Geographical Location**: Brest, France
|
| 181 |
+
- **Hardware Used**: NVIDIA Tesla V100 PCIe 32 Go
|
| 182 |
+
|
| 183 |
+
|
| 184 |
+
---
|
| 185 |
+
|
| 186 |
+
# Framework Versions
|
| 187 |
+
|
| 188 |
+
- **Transformers**: 4.41.1
|
| 189 |
+
- **Pytorch**: 2.3.0+cu121
|
| 190 |
+
- **Datasets**: 2.19.1
|
| 191 |
+
- **Tokenizers**: 0.19.1
|
| 192 |
+
|