--- language: - eng license: wtfpl tags: - multilabel-image-classification - multilabel - generated_from_trainer base_model: facebook/dinov2-base model-index: - name: DinoVdeau-base-2024_09_03-batch-size32_epochs150_freeze results: [] --- DinoVd'eau is a fine-tuned version of [facebook/dinov2-base](https://huggingface.co/facebook/dinov2-base). It achieves the following results on the test set: - Loss: 0.1260 - F1 Micro: 0.8131 - F1 Macro: 0.6976 - Roc Auc: 0.8760 - Accuracy: 0.3014 --- # Model description 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. The source code for training the model can be found in this [Git repository](https://github.com/SeatizenDOI/DinoVdeau). - **Developed by:** [lombardata](https://huggingface.co/lombardata), credits to [César Leblanc](https://huggingface.co/CesarLeblanc) and [Victor Illien](https://huggingface.co/groderg) --- # Intended uses & limitations 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. --- # Training and evaluation data Details on the number of images for each class are given in the following table: | Class | train | val | test | Total | |:-------------------------|--------:|------:|-------:|--------:| | Acropore_branched | 1469 | 464 | 475 | 2408 | | Acropore_digitised | 568 | 160 | 160 | 888 | | Acropore_sub_massive | 150 | 50 | 43 | 243 | | Acropore_tabular | 999 | 297 | 293 | 1589 | | Algae_assembly | 2546 | 847 | 845 | 4238 | | Algae_drawn_up | 367 | 126 | 127 | 620 | | Algae_limestone | 1652 | 557 | 563 | 2772 | | Algae_sodding | 3148 | 984 | 985 | 5117 | | Atra/Leucospilota | 1084 | 348 | 360 | 1792 | | Bleached_coral | 219 | 71 | 70 | 360 | | Blurred | 191 | 67 | 62 | 320 | | Dead_coral | 1979 | 642 | 643 | 3264 | | Fish | 2018 | 656 | 647 | 3321 | | Homo_sapiens | 161 | 62 | 59 | 282 | | Human_object | 157 | 58 | 55 | 270 | | Living_coral | 406 | 154 | 141 | 701 | | Millepore | 385 | 127 | 125 | 637 | | No_acropore_encrusting | 441 | 130 | 154 | 725 | | No_acropore_foliaceous | 204 | 36 | 46 | 286 | | No_acropore_massive | 1031 | 336 | 338 | 1705 | | No_acropore_solitary | 202 | 53 | 48 | 303 | | No_acropore_sub_massive | 1401 | 433 | 422 | 2256 | | Rock | 4489 | 1495 | 1473 | 7457 | | Rubble | 3092 | 1030 | 1001 | 5123 | | Sand | 5842 | 1939 | 1938 | 9719 | | Sea_cucumber | 1408 | 439 | 447 | 2294 | | Sea_urchins | 327 | 107 | 111 | 545 | | Sponge | 269 | 96 | 105 | 470 | | Syringodium_isoetifolium | 1212 | 392 | 391 | 1995 | | Thalassodendron_ciliatum | 782 | 261 | 260 | 1303 | | Useless | 579 | 193 | 193 | 965 | --- # Training procedure ## Training hyperparameters The following hyperparameters were used during training: - **Number of Epochs**: 150 - **Learning Rate**: 0.001 - **Train Batch Size**: 32 - **Eval Batch Size**: 32 - **Optimizer**: Adam - **LR Scheduler Type**: ReduceLROnPlateau with a patience of 5 epochs and a factor of 0.1 - **Freeze Encoder**: Yes - **Data Augmentation**: Yes ## Data Augmentation Data were augmented using the following transformations : Train Transforms - **PreProcess**: No additional parameters - **Resize**: probability=1.00 - **RandomHorizontalFlip**: probability=0.25 - **RandomVerticalFlip**: probability=0.25 - **ColorJiggle**: probability=0.25 - **RandomPerspective**: probability=0.25 - **Normalize**: probability=1.00 Val Transforms - **PreProcess**: No additional parameters - **Resize**: probability=1.00 - **Normalize**: probability=1.00 ## Training results Epoch | Validation Loss | Accuracy | F1 Macro | F1 Micro | Learning Rate --- | --- | --- | --- | --- | --- 1 | 0.17516958713531494 | 0.2079002079002079 | 0.73108765167112 | 0.5105390450682302 | 0.001 2 | 0.1577771008014679 | 0.23492723492723494 | 0.7582569600553347 | 0.5498069096094584 | 0.001 3 | 0.15162432193756104 | 0.23146223146223147 | 0.7721545657578696 | 0.6037272443934714 | 0.001 4 | 0.15218119323253632 | 0.24220374220374222 | 0.7649537378914902 | 0.613953187695023 | 0.001 5 | 0.14836864173412323 | 0.24220374220374222 | 0.7719928186714542 | 0.6161642626912543 | 0.001 6 | 0.14818257093429565 | 0.2560637560637561 | 0.775030471878809 | 0.6051867487843677 | 0.001 7 | 0.1486394852399826 | 0.24185724185724186 | 0.7729166666666668 | 0.617739220969942 | 0.001 8 | 0.14861202239990234 | 0.2512127512127512 | 0.7767065175472426 | 0.6171646674895677 | 0.001 9 | 0.14834754168987274 | 0.2512127512127512 | 0.7805490458654168 | 0.6366264906922544 | 0.001 10 | 0.15029709041118622 | 0.24532224532224534 | 0.7682759232167399 | 0.6081428044829309 | 0.001 11 | 0.14407172799110413 | 0.2609147609147609 | 0.7756647297059341 | 0.6199915554248129 | 0.001 12 | 0.14866559207439423 | 0.2494802494802495 | 0.781485559413907 | 0.6299207659511814 | 0.001 13 | 0.14902691543102264 | 0.25190575190575193 | 0.7779037321241716 | 0.6241659824597257 | 0.001 14 | 0.14337006211280823 | 0.26056826056826055 | 0.7826389795829524 | 0.6378982802249643 | 0.001 15 | 0.14354591071605682 | 0.2553707553707554 | 0.7873585308562887 | 0.639716503598517 | 0.001 16 | 0.1439499706029892 | 0.25675675675675674 | 0.7792974686292388 | 0.6343613127126344 | 0.001 17 | 0.14478015899658203 | 0.2543312543312543 | 0.787784461363732 | 0.6422270697798029 | 0.001 18 | 0.14397625625133514 | 0.25675675675675674 | 0.786493860845839 | 0.6417123667888478 | 0.001 19 | 0.14199253916740417 | 0.253984753984754 | 0.7863510343356792 | 0.6317583185615991 | 0.001 20 | 0.14092272520065308 | 0.2588357588357588 | 0.7868513006341401 | 0.6408966299078661 | 0.001 21 | 0.1425119787454605 | 0.26195426195426197 | 0.7864882090503504 | 0.6412583916380257 | 0.001 22 | 0.15379400551319122 | 0.23700623700623702 | 0.7854284761587195 | 0.6371452798177432 | 0.001 23 | 0.1418805718421936 | 0.25571725571725573 | 0.7841676771176165 | 0.6390434486158698 | 0.001 24 | 0.14135514199733734 | 0.2598752598752599 | 0.7869535635312129 | 0.6458978920546691 | 0.001 25 | 0.13985148072242737 | 0.26853776853776856 | 0.786773581652009 | 0.6262981090846956 | 0.001 26 | 0.14649754762649536 | 0.2591822591822592 | 0.7846557710221018 | 0.6237830069375186 | 0.001 27 | 0.15506784617900848 | 0.23804573804573806 | 0.7719951506754418 | 0.6344307952131357 | 0.001 28 | 0.14431345462799072 | 0.2616077616077616 | 0.7891238152420981 | 0.6429949936408241 | 0.001 29 | 0.14275498688220978 | 0.25675675675675674 | 0.7873995663818392 | 0.6415824285032449 | 0.001 30 | 0.14164045453071594 | 0.2525987525987526 | 0.7798808735936467 | 0.6308133523221491 | 0.001 31 | 0.13976627588272095 | 0.26888426888426886 | 0.7895365707945718 | 0.6431010910213645 | 0.001 32 | 0.1448184847831726 | 0.25675675675675674 | 0.7891036166898235 | 0.6520927708015384 | 0.001 33 | 0.14042973518371582 | 0.26403326403326405 | 0.7895652173913044 | 0.6496848321151188 | 0.001 34 | 0.1426127403974533 | 0.25571725571725573 | 0.7870906828033133 | 0.6448790211155284 | 0.001 35 | 0.14135821163654327 | 0.262993762993763 | 0.7846327880264532 | 0.6428423378015612 | 0.001 36 | 0.14652539789676666 | 0.26784476784476785 | 0.7834209497328063 | 0.6434020884943297 | 0.001 37 | 0.13795886933803558 | 0.2668052668052668 | 0.792425408224331 | 0.6438477431550106 | 0.001 38 | 0.13921019434928894 | 0.2636867636867637 | 0.7892280686732029 | 0.6475331965590188 | 0.001 39 | 0.14584119617938995 | 0.24601524601524602 | 0.7871620243872598 | 0.659217552215385 | 0.001 40 | 0.1389026641845703 | 0.26992376992376993 | 0.79463243873979 | 0.6469476365862663 | 0.001 41 | 0.14020991325378418 | 0.2616077616077616 | 0.784842032071618 | 0.6509894683187031 | 0.001 42 | 0.14042720198631287 | 0.27165627165627165 | 0.7927685516081564 | 0.6608924914997423 | 0.001 43 | 0.13943640887737274 | 0.2695772695772696 | 0.7930726352070125 | 0.6427022769326964 | 0.001 44 | 0.1367315948009491 | 0.27546777546777546 | 0.7989137353078458 | 0.6567716426576066 | 0.0001 45 | 0.13616175949573517 | 0.28274428274428276 | 0.8018308187828446 | 0.6686203083248894 | 0.0001 46 | 0.13375289738178253 | 0.2851697851697852 | 0.8021852369457503 | 0.6640104860714046 | 0.0001 47 | 0.14095526933670044 | 0.2785862785862786 | 0.7998804746862461 | 0.65726703563479 | 0.0001 48 | 0.13375185430049896 | 0.28482328482328484 | 0.8044442566853957 | 0.6728387979723557 | 0.0001 49 | 0.13221527636051178 | 0.2855162855162855 | 0.8058309037900874 | 0.674164075762875 | 0.0001 50 | 0.13315953314304352 | 0.28967428967428965 | 0.8062985513331933 | 0.6738599949249782 | 0.0001 51 | 0.13057135045528412 | 0.28967428967428965 | 0.8062836021505377 | 0.6770873238469556 | 0.0001 52 | 0.13108478486537933 | 0.2872487872487873 | 0.8043922369765066 | 0.6726562275384118 | 0.0001 53 | 0.13161474466323853 | 0.2872487872487873 | 0.8070734160241367 | 0.6702824874792834 | 0.0001 54 | 0.1315840184688568 | 0.2882882882882883 | 0.8064162093710426 | 0.6787531928667037 | 0.0001 55 | 0.13084293901920319 | 0.2875952875952876 | 0.8061478697800111 | 0.6698514928377199 | 0.0001 56 | 0.12969879806041718 | 0.29417879417879417 | 0.8094286190238215 | 0.6799502024965028 | 0.0001 57 | 0.1296372264623642 | 0.2934857934857935 | 0.8086806577785254 | 0.6716759101412201 | 0.0001 58 | 0.12973745167255402 | 0.29244629244629244 | 0.8068982880161129 | 0.6784509633805341 | 0.0001 59 | 0.12995606660842896 | 0.2910602910602911 | 0.8087436297013858 | 0.6811347101829983 | 0.0001 60 | 0.13024823367595673 | 0.28794178794178793 | 0.8056052474657126 | 0.6725887638706813 | 0.0001 61 | 0.12872998416423798 | 0.2948717948717949 | 0.8095537925534148 | 0.6842961167409227 | 0.0001 62 | 0.12909561395645142 | 0.29002079002079 | 0.8079526226734349 | 0.6821531683206365 | 0.0001 63 | 0.12872986495494843 | 0.29799029799029797 | 0.8075538806791719 | 0.6812919501021206 | 0.0001 64 | 0.12864243984222412 | 0.2959112959112959 | 0.8090726144558109 | 0.6805602232602442 | 0.0001 65 | 0.12800218164920807 | 0.29313929313929316 | 0.809268560334276 | 0.6837997472607307 | 0.0001 66 | 0.12777170538902283 | 0.2959112959112959 | 0.8107521495951249 | 0.685457875933014 | 0.0001 67 | 0.12816764414310455 | 0.2948717948717949 | 0.8098450774612694 | 0.6849396578990685 | 0.0001 68 | 0.12804801762104034 | 0.29799029799029797 | 0.8123470107455503 | 0.6903099963278952 | 0.0001 69 | 0.12803924083709717 | 0.29521829521829523 | 0.8104663431103608 | 0.6800351861453543 | 0.0001 70 | 0.12764029204845428 | 0.29313929313929316 | 0.8096462751380749 | 0.684802818649885 | 0.0001 71 | 0.12794704735279083 | 0.29036729036729036 | 0.8072724183339705 | 0.6796736257485385 | 0.0001 72 | 0.12780210375785828 | 0.2938322938322938 | 0.8102650399663442 | 0.6802343842914587 | 0.0001 73 | 0.12723641097545624 | 0.29764379764379767 | 0.8091473263623224 | 0.6805723882610378 | 0.0001 74 | 0.12804573774337769 | 0.2934857934857935 | 0.8064391831142698 | 0.6777188921642516 | 0.0001 75 | 0.1273234635591507 | 0.29244629244629244 | 0.8109922383050138 | 0.6885203936930924 | 0.0001 76 | 0.1272992193698883 | 0.2972972972972973 | 0.8088975345709815 | 0.6810578369044884 | 0.0001 77 | 0.12745273113250732 | 0.29244629244629244 | 0.8102101349375445 | 0.6863183190306963 | 0.0001 78 | 0.12705788016319275 | 0.2945252945252945 | 0.8121675531914894 | 0.6897104532016692 | 0.0001 79 | 0.12710121273994446 | 0.2934857934857935 | 0.809842452990005 | 0.6881838490414868 | 0.0001 80 | 0.12715762853622437 | 0.2983367983367983 | 0.8123911420751431 | 0.6914032136958002 | 0.0001 81 | 0.12650521099567413 | 0.2966042966042966 | 0.810378232667846 | 0.6899389708752343 | 0.0001 82 | 0.12635371088981628 | 0.29140679140679143 | 0.8105446364138047 | 0.6844864031747653 | 0.0001 83 | 0.1272997260093689 | 0.2934857934857935 | 0.8099670022844573 | 0.6832392344549459 | 0.0001 84 | 0.12640425562858582 | 0.2959112959112959 | 0.8124478558318038 | 0.6944491344986764 | 0.0001 85 | 0.12647400796413422 | 0.2972972972972973 | 0.812659392115055 | 0.6879519222426981 | 0.0001 86 | 0.12585221230983734 | 0.29521829521829523 | 0.8135877542461731 | 0.6933253774763921 | 0.0001 87 | 0.12641744315624237 | 0.2966042966042966 | 0.8111366966715512 | 0.6882459007361815 | 0.0001 88 | 0.1263686865568161 | 0.29902979902979904 | 0.8126931106471816 | 0.6859575429209334 | 0.0001 89 | 0.12690132856369019 | 0.2983367983367983 | 0.8140188460902628 | 0.6990366097632199 | 0.0001 90 | 0.12612390518188477 | 0.29799029799029797 | 0.8155163144617673 | 0.6994167448254883 | 0.0001 91 | 0.1268243044614792 | 0.28932778932778935 | 0.8108811552831535 | 0.6827913109763548 | 0.0001 92 | 0.12613284587860107 | 0.29521829521829523 | 0.8123787840458724 | 0.6858483939371968 | 0.0001 93 | 0.1258293092250824 | 0.2966042966042966 | 0.8138213420238991 | 0.6897216822080747 | 1e-05 94 | 0.12682591378688812 | 0.29764379764379767 | 0.8137706015226304 | 0.6940665827082791 | 1e-05 95 | 0.1256789118051529 | 0.2948717948717949 | 0.8133975298304374 | 0.6913394393323408 | 1e-05 96 | 0.12587758898735046 | 0.29764379764379767 | 0.8147281313996739 | 0.6957055849225957 | 1e-05 97 | 0.1256256103515625 | 0.2945252945252945 | 0.8126029480086159 | 0.6940781337907567 | 1e-05 98 | 0.1253080666065216 | 0.2993762993762994 | 0.8158955813276801 | 0.6951390304078455 | 1e-05 99 | 0.12485036998987198 | 0.2993762993762994 | 0.8141971169963125 | 0.6968244403216463 | 1e-05 100 | 0.12519583106040955 | 0.2972972972972973 | 0.8134507606084869 | 0.693647218520028 | 1e-05 101 | 0.12475299090147018 | 0.29902979902979904 | 0.8148550421923302 | 0.6961023046950545 | 1e-05 102 | 0.12659381330013275 | 0.29625779625779625 | 0.81366198367965 | 0.692743439816851 | 1e-05 103 | 0.124935083091259 | 0.29902979902979904 | 0.8146347596496376 | 0.6954353634647259 | 1e-05 104 | 0.12519653141498566 | 0.29764379764379767 | 0.8148796863922599 | 0.692659001716947 | 1e-05 105 | 0.12513257563114166 | 0.29902979902979904 | 0.8152223750573132 | 0.6961790886935857 | 1e-05 106 | 0.12511174380779266 | 0.29972279972279975 | 0.8147252563995664 | 0.6963861386142265 | 1e-05 107 | 0.12498941272497177 | 0.29799029799029797 | 0.8144894800685992 | 0.694620567930595 | 1e-05 108 | 0.1248873621225357 | 0.29972279972279975 | 0.8144792584203683 | 0.6934713387989168 | 1.0000000000000002e-06 109 | 0.12527066469192505 | 0.29521829521829523 | 0.8116150302210575 | 0.6900779361953018 | 1.0000000000000002e-06 110 | 0.125152125954628 | 0.29799029799029797 | 0.8143917285082964 | 0.69491512245201 | 1.0000000000000002e-06 111 | 0.12495684623718262 | 0.2983367983367983 | 0.8137025263510123 | 0.6932228755688746 | 1.0000000000000002e-06 --- # CO2 Emissions The estimated CO2 emissions for training this model are documented below: - **Emissions**: 1.3368314147555413 grams of CO2 - **Source**: Code Carbon - **Training Type**: fine-tuning - **Geographical Location**: Brest, France - **Hardware Used**: NVIDIA Tesla V100 PCIe 32 Go --- # Framework Versions - **Transformers**: 4.41.1 - **Pytorch**: 2.3.0+cu121 - **Datasets**: 2.19.1 - **Tokenizers**: 0.19.1