Commit
•
439b1fe
1
Parent(s):
5481b13
Upload README.md
Browse files
README.md
CHANGED
|
@@ -1,141 +1,186 @@
|
|
|
|
|
| 1 |
---
|
| 2 |
-
|
| 3 |
-
|
|
|
|
| 4 |
tags:
|
|
|
|
|
|
|
| 5 |
- generated_from_trainer
|
|
|
|
| 6 |
model-index:
|
| 7 |
- name: drone-DinoVdeau-from-probs-large-2024_11_15-batch-size64_freeze_probs
|
| 8 |
results: []
|
| 9 |
---
|
| 10 |
|
| 11 |
-
|
| 12 |
-
should probably proofread and complete it, then remove this comment. -->
|
| 13 |
|
| 14 |
-
# drone-DinoVdeau-from-probs-large-2024_11_15-batch-size64_freeze_probs
|
| 15 |
|
| 16 |
-
This model is a fine-tuned version of [facebook/dinov2-large](https://huggingface.co/facebook/dinov2-large) on the None dataset.
|
| 17 |
-
It achieves the following results on the evaluation set:
|
| 18 |
- Loss: 0.4672
|
| 19 |
-
-
|
| 20 |
-
-
|
| 21 |
-
-
|
| 22 |
-
|
| 23 |
-
|
|
|
|
|
|
|
|
|
|
| 24 |
|
| 25 |
-
|
| 26 |
|
| 27 |
-
|
| 28 |
|
| 29 |
-
|
|
|
|
|
|
|
|
|
|
| 30 |
|
| 31 |
-
|
| 32 |
|
| 33 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 34 |
|
| 35 |
-
|
| 36 |
|
| 37 |
-
|
| 38 |
|
| 39 |
-
|
| 40 |
|
| 41 |
The following hyperparameters were used during training:
|
| 42 |
-
|
| 43 |
-
-
|
| 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 |
-
| 0.
|
| 76 |
-
| 0.
|
| 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 |
-
|
| 125 |
-
| 0.
|
| 126 |
-
| 0.
|
| 127 |
-
| 0.
|
| 128 |
-
| 0.
|
| 129 |
-
| 0.
|
| 130 |
-
| 0.
|
| 131 |
-
| 0.
|
| 132 |
-
| 0.
|
| 133 |
-
| 0.
|
| 134 |
-
|
| 135 |
-
|
| 136 |
-
|
| 137 |
-
|
| 138 |
-
|
| 139 |
-
|
| 140 |
-
|
| 141 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
|
| 2 |
---
|
| 3 |
+
language:
|
| 4 |
+
- eng
|
| 5 |
+
license: cc0-1.0
|
| 6 |
tags:
|
| 7 |
+
- multilabel-image-classification
|
| 8 |
+
- multilabel
|
| 9 |
- generated_from_trainer
|
| 10 |
+
base_model: drone-DinoVdeau-from-probs-large-2024_11_15-batch-size64_freeze_probs
|
| 11 |
model-index:
|
| 12 |
- name: drone-DinoVdeau-from-probs-large-2024_11_15-batch-size64_freeze_probs
|
| 13 |
results: []
|
| 14 |
---
|
| 15 |
|
| 16 |
+
drone-DinoVdeau-from-probs is a fine-tuned version of [drone-DinoVdeau-from-probs-large-2024_11_15-batch-size64_freeze_probs](https://huggingface.co/drone-DinoVdeau-from-probs-large-2024_11_15-batch-size64_freeze_probs). It achieves the following results on the test set:
|
|
|
|
| 17 |
|
|
|
|
| 18 |
|
|
|
|
|
|
|
| 19 |
- Loss: 0.4672
|
| 20 |
+
- RMSE: 0.1553
|
| 21 |
+
- MAE: 0.1147
|
| 22 |
+
- KL Divergence: 0.3577
|
| 23 |
+
|
| 24 |
+
---
|
| 25 |
+
|
| 26 |
+
# Model description
|
| 27 |
+
drone-DinoVdeau-from-probs is a model built on top of drone-DinoVdeau-from-probs-large-2024_11_15-batch-size64_freeze_probs 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 estimated number of images for each class are given in the following table:
|
| 42 |
+
| Class | train | test | val | Total |
|
| 43 |
+
|:------------------------|--------:|-------:|------:|--------:|
|
| 44 |
+
| Acropore_branched | 1220 | 363 | 362 | 1945 |
|
| 45 |
+
| Acropore_digitised | 586 | 195 | 189 | 970 |
|
| 46 |
+
| Acropore_tabular | 308 | 133 | 119 | 560 |
|
| 47 |
+
| Algae | 4777 | 1372 | 1384 | 7533 |
|
| 48 |
+
| Dead_coral | 2513 | 671 | 693 | 3877 |
|
| 49 |
+
| Millepore | 136 | 55 | 59 | 250 |
|
| 50 |
+
| No_acropore_encrusting | 252 | 88 | 93 | 433 |
|
| 51 |
+
| No_acropore_massive | 2158 | 725 | 726 | 3609 |
|
| 52 |
+
| No_acropore_sub_massive | 2036 | 582 | 612 | 3230 |
|
| 53 |
+
| Rock | 5976 | 1941 | 1928 | 9845 |
|
| 54 |
+
| Rubble | 4851 | 1486 | 1474 | 7811 |
|
| 55 |
+
| Sand | 6155 | 2019 | 1990 | 10164 |
|
| 56 |
|
| 57 |
+
---
|
| 58 |
|
| 59 |
+
# Training procedure
|
| 60 |
|
| 61 |
+
## Training hyperparameters
|
| 62 |
|
| 63 |
The following hyperparameters were used during training:
|
| 64 |
+
|
| 65 |
+
- **Number of Epochs**: 79.0
|
| 66 |
+
- **Learning Rate**: 0.001
|
| 67 |
+
- **Train Batch Size**: 64
|
| 68 |
+
- **Eval Batch Size**: 64
|
| 69 |
+
- **Optimizer**: Adam
|
| 70 |
+
- **LR Scheduler Type**: ReduceLROnPlateau with a patience of 5 epochs and a factor of 0.1
|
| 71 |
+
- **Freeze Encoder**: Yes
|
| 72 |
+
- **Data Augmentation**: Yes
|
| 73 |
+
|
| 74 |
+
|
| 75 |
+
## Data Augmentation
|
| 76 |
+
Data were augmented using the following transformations :
|
| 77 |
+
|
| 78 |
+
Train Transforms
|
| 79 |
+
- **PreProcess**: No additional parameters
|
| 80 |
+
- **Resize**: probability=1.00
|
| 81 |
+
- **RandomHorizontalFlip**: probability=0.25
|
| 82 |
+
- **RandomVerticalFlip**: probability=0.25
|
| 83 |
+
- **ColorJiggle**: probability=0.25
|
| 84 |
+
- **RandomPerspective**: probability=0.25
|
| 85 |
+
- **Normalize**: probability=1.00
|
| 86 |
+
|
| 87 |
+
Val Transforms
|
| 88 |
+
- **PreProcess**: No additional parameters
|
| 89 |
+
- **Resize**: probability=1.00
|
| 90 |
+
- **Normalize**: probability=1.00
|
| 91 |
+
|
| 92 |
+
|
| 93 |
+
|
| 94 |
+
## Training results
|
| 95 |
+
Epoch | Validation Loss | MAE | RMSE | KL div | Learning Rate
|
| 96 |
+
--- | --- | --- | --- | --- | ---
|
| 97 |
+
1 | 0.5005590319633484 | 0.1552 | 0.1904 | 0.1025 | 0.001
|
| 98 |
+
2 | 0.47547808289527893 | 0.1245 | 0.1681 | 0.5180 | 0.001
|
| 99 |
+
3 | 0.47452571988105774 | 0.1227 | 0.1675 | 0.6862 | 0.001
|
| 100 |
+
4 | 0.47420722246170044 | 0.1255 | 0.1672 | 0.3212 | 0.001
|
| 101 |
+
5 | 0.47245556116104126 | 0.1224 | 0.1653 | 0.5072 | 0.001
|
| 102 |
+
6 | 0.4725925624370575 | 0.1216 | 0.1657 | 0.6710 | 0.001
|
| 103 |
+
7 | 0.4731809198856354 | 0.1255 | 0.1655 | 0.3162 | 0.001
|
| 104 |
+
8 | 0.47284314036369324 | 0.1260 | 0.1651 | 0.2719 | 0.001
|
| 105 |
+
9 | 0.4707973003387451 | 0.1206 | 0.1639 | 0.6393 | 0.001
|
| 106 |
+
10 | 0.4732784628868103 | 0.1230 | 0.1654 | 0.5359 | 0.001
|
| 107 |
+
11 | 0.47162503004074097 | 0.1253 | 0.1647 | 0.2479 | 0.001
|
| 108 |
+
12 | 0.47083696722984314 | 0.1244 | 0.1631 | 0.3119 | 0.001
|
| 109 |
+
13 | 0.47152063250541687 | 0.1230 | 0.1635 | 0.3694 | 0.001
|
| 110 |
+
14 | 0.47212228178977966 | 0.1216 | 0.1653 | 0.5592 | 0.001
|
| 111 |
+
15 | 0.47012239694595337 | 0.1213 | 0.1628 | 0.4936 | 0.001
|
| 112 |
+
16 | 0.4718552827835083 | 0.1229 | 0.1646 | 0.2820 | 0.001
|
| 113 |
+
17 | 0.46933484077453613 | 0.1200 | 0.1621 | 0.5294 | 0.001
|
| 114 |
+
18 | 0.4710436165332794 | 0.1216 | 0.1635 | 0.4093 | 0.001
|
| 115 |
+
19 | 0.4698491394519806 | 0.1219 | 0.1622 | 0.2918 | 0.001
|
| 116 |
+
20 | 0.4691685736179352 | 0.1190 | 0.1617 | 0.4772 | 0.001
|
| 117 |
+
21 | 0.46830564737319946 | 0.1204 | 0.1606 | 0.4336 | 0.001
|
| 118 |
+
22 | 0.47239789366722107 | 0.1183 | 0.1650 | 0.7962 | 0.001
|
| 119 |
+
23 | 0.47136834263801575 | 0.1223 | 0.1641 | 0.2854 | 0.001
|
| 120 |
+
24 | 0.4706868529319763 | 0.1207 | 0.1633 | 0.4206 | 0.001
|
| 121 |
+
25 | 0.46786901354789734 | 0.1185 | 0.1606 | 0.5436 | 0.001
|
| 122 |
+
26 | 0.47084224224090576 | 0.1192 | 0.1634 | 0.4964 | 0.001
|
| 123 |
+
27 | 0.4695045053958893 | 0.1185 | 0.1625 | 0.6399 | 0.001
|
| 124 |
+
28 | 0.4700873792171478 | 0.1184 | 0.1624 | 0.5737 | 0.001
|
| 125 |
+
29 | 0.4698559045791626 | 0.1200 | 0.1624 | 0.4459 | 0.001
|
| 126 |
+
30 | 0.4722815454006195 | 0.1254 | 0.1643 | 0.2726 | 0.001
|
| 127 |
+
31 | 0.46958214044570923 | 0.1184 | 0.1622 | 0.5308 | 0.001
|
| 128 |
+
32 | 0.46677276492118835 | 0.1175 | 0.1593 | 0.4200 | 0.0001
|
| 129 |
+
33 | 0.46626824140548706 | 0.1177 | 0.1587 | 0.3529 | 0.0001
|
| 130 |
+
34 | 0.46665358543395996 | 0.1181 | 0.1592 | 0.3588 | 0.0001
|
| 131 |
+
35 | 0.46587392687797546 | 0.1160 | 0.1584 | 0.4813 | 0.0001
|
| 132 |
+
36 | 0.46578526496887207 | 0.1173 | 0.1581 | 0.3504 | 0.0001
|
| 133 |
+
37 | 0.4654408395290375 | 0.1158 | 0.1578 | 0.3919 | 0.0001
|
| 134 |
+
38 | 0.46546319127082825 | 0.1166 | 0.1580 | 0.4058 | 0.0001
|
| 135 |
+
39 | 0.465843141078949 | 0.1174 | 0.1585 | 0.4118 | 0.0001
|
| 136 |
+
40 | 0.46561121940612793 | 0.1170 | 0.1579 | 0.3564 | 0.0001
|
| 137 |
+
41 | 0.4657152593135834 | 0.1171 | 0.1582 | 0.3573 | 0.0001
|
| 138 |
+
42 | 0.4651602804660797 | 0.1155 | 0.1579 | 0.5042 | 0.0001
|
| 139 |
+
43 | 0.4651065468788147 | 0.1157 | 0.1575 | 0.4462 | 0.0001
|
| 140 |
+
44 | 0.46537330746650696 | 0.1166 | 0.1579 | 0.4236 | 0.0001
|
| 141 |
+
45 | 0.46489208936691284 | 0.1151 | 0.1574 | 0.4510 | 0.0001
|
| 142 |
+
46 | 0.46484702825546265 | 0.1157 | 0.1575 | 0.4490 | 0.0001
|
| 143 |
+
47 | 0.4648602306842804 | 0.1152 | 0.1574 | 0.4751 | 0.0001
|
| 144 |
+
48 | 0.4647873342037201 | 0.1151 | 0.1575 | 0.5305 | 0.0001
|
| 145 |
+
49 | 0.4647849500179291 | 0.1154 | 0.1574 | 0.4799 | 0.0001
|
| 146 |
+
50 | N/A | 0.0000 | 0.0000 | 0.0000 | 0.0001
|
| 147 |
+
51 | 0.465638667345047 | 0.1151 | 0.1582 | 0.4879 | 0.0001
|
| 148 |
+
52 | 0.46429532766342163 | 0.1155 | 0.1566 | 0.4199 | 0.0001
|
| 149 |
+
53 | 0.46441230177879333 | 0.1156 | 0.1569 | 0.3880 | 0.0001
|
| 150 |
+
54 | 0.4646008610725403 | 0.1148 | 0.1569 | 0.4229 | 0.0001
|
| 151 |
+
55 | 0.4644174873828888 | 0.1159 | 0.1569 | 0.4009 | 0.0001
|
| 152 |
+
56 | 0.464743047952652 | 0.1164 | 0.1572 | 0.3405 | 0.0001
|
| 153 |
+
57 | 0.4645179808139801 | 0.1152 | 0.1569 | 0.4188 | 0.0001
|
| 154 |
+
58 | 0.465102881193161 | 0.1164 | 0.1576 | 0.3079 | 0.0001
|
| 155 |
+
59 | 0.4644688367843628 | 0.1150 | 0.1570 | 0.4339 | 1e-05
|
| 156 |
+
60 | 0.46417686343193054 | 0.1150 | 0.1566 | 0.3894 | 1e-05
|
| 157 |
+
61 | 0.4639436900615692 | 0.1146 | 0.1563 | 0.4145 | 1e-05
|
| 158 |
+
62 | 0.4641311764717102 | 0.1148 | 0.1565 | 0.4064 | 1e-05
|
| 159 |
+
63 | 0.4643491506576538 | 0.1149 | 0.1565 | 0.3542 | 1e-05
|
| 160 |
+
64 | 0.46402981877326965 | 0.1150 | 0.1564 | 0.3718 | 1e-05
|
| 161 |
+
65 | 0.4640822410583496 | 0.1152 | 0.1565 | 0.4128 | 1e-05
|
| 162 |
+
66 | 0.46441909670829773 | 0.1145 | 0.1570 | 0.4988 | 1e-05
|
| 163 |
+
67 | 0.46383005380630493 | 0.1151 | 0.1562 | 0.4122 | 1e-05
|
| 164 |
+
68 | 0.4639807641506195 | 0.1144 | 0.1565 | 0.4579 | 1e-05
|
| 165 |
+
69 | 0.4637599587440491 | 0.1143 | 0.1561 | 0.4197 | 1e-05
|
| 166 |
+
70 | 0.46392253041267395 | 0.1145 | 0.1563 | 0.4286 | 1e-05
|
| 167 |
+
71 | 0.46406444907188416 | 0.1153 | 0.1563 | 0.3542 | 1e-05
|
| 168 |
+
72 | 0.46417826414108276 | 0.1147 | 0.1566 | 0.4250 | 1e-05
|
| 169 |
+
73 | 0.4637835919857025 | 0.1140 | 0.1561 | 0.4397 | 1e-05
|
| 170 |
+
74 | 0.463798850774765 | 0.1145 | 0.1563 | 0.4437 | 1e-05
|
| 171 |
+
75 | 0.46379053592681885 | 0.1145 | 0.1561 | 0.4049 | 1e-05
|
| 172 |
+
76 | 0.4639701247215271 | 0.1141 | 0.1565 | 0.4926 | 1.0000000000000002e-06
|
| 173 |
+
77 | 0.463869571685791 | 0.1142 | 0.1562 | 0.4427 | 1.0000000000000002e-06
|
| 174 |
+
78 | 0.46388140320777893 | 0.1145 | 0.1563 | 0.4293 | 1.0000000000000002e-06
|
| 175 |
+
79 | 0.46412238478660583 | 0.1147 | 0.1564 | 0.3765 | 1.0000000000000002e-06
|
| 176 |
+
|
| 177 |
+
|
| 178 |
+
---
|
| 179 |
+
|
| 180 |
+
# Framework Versions
|
| 181 |
+
|
| 182 |
+
- **Transformers**: 4.41.0
|
| 183 |
+
- **Pytorch**: 2.5.0+cu124
|
| 184 |
+
- **Datasets**: 3.0.2
|
| 185 |
+
- **Tokenizers**: 0.19.1
|
| 186 |
+
|