lpcv_seg

This model is a fine-tuned version of google/deeplabv3_mobilenet_v2_1.0_513 on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.7747
  • Mean Iou: 0.3647
  • Mean Accuracy: 0.4742
  • Overall Accuracy: 0.7441
  • Per Category Iou: [0.7508396000072782, 0.461963906773346, 0.41562431632163865, 0.2643890336606752, 0.20882280410355394, 0.21001420486640948, 0.45776923048905305, 0.6263265430221951, 0.5990132199881534, nan, 0.0, 0.4389678627683, 0.30687238462719907, 0.0]
  • Per Category Accuracy: [0.8794950417196551, 0.5122587212045141, 0.47963636761360323, 0.28307093261894156, 0.22726847453969443, 0.817351469679542, 0.5476940642254209, 0.672940072704085, 0.8955934160757268, nan, 0.0, 0.44608425759467085, 0.40330970104326763, 0.0]

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 6e-05
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 30

Training results

Training Loss Epoch Step Validation Loss Mean Iou Mean Accuracy Overall Accuracy Per Category Iou Per Category Accuracy
1.0299 3.12 400 1.1445 0.1503 0.2293 0.6846 [0.7016834423401825, 0.0, 0.28392253175773663, 0.09630649211770352, 0.16489618177561766, 0.05923195228949304, 0.0, 0.47146079872020313, 0.0, nan, 0.0, 0.002537363927640351, 0.17407988046100897, 0.0] [0.891422563505799, 0.0, 0.7160794493379222, 0.101650104427665, 0.20155241518296468, 0.08566222638071554, 0.0, 0.7801995050268263, 0.0, nan, 0.0, 0.002538359085235112, 0.20133152963007558, 0.0]
1.2946 6.25 800 0.9463 0.1514 0.2217 0.6990 [0.7200410800543089, 0.025577984693804623, 0.272186783532416, 0.03900872477465889, 0.1420562982408946, 0.13672818472813886, 0.0, 0.3747836709762802, 1.4597782434650593e-05, nan, 0.0, 0.0, 0.2576138826730331, 0.0] [0.9197800777719178, 0.025581056237151574, 0.3936555737035525, 0.03917595636589459, 0.1525480793198657, 0.16772954472705784, 0.0, 0.8514708072975876, 1.4597782434650593e-05, nan, 0.0, 0.0, 0.33275562889985905, 0.0]
0.8869 9.38 1200 0.9005 0.2553 0.3463 0.7230 [0.7311911080696011, 0.1716353741804461, 0.32989770226558307, 0.2484733739307282, 0.20678309581325105, 0.16925674004774066, 0.026560396219635213, 0.43309262889720945, 0.4791539954039438, nan, 0.0, 0.1996331281686418, 0.3234492318785704, 0.0] [0.8817514590505282, 0.1734977169471996, 0.5536498133348888, 0.26666168042981364, 0.2199222972903758, 0.30495278898499545, 0.026735171458080034, 0.8794234634009936, 0.5647817144933271, nan, 0.0, 0.19998287851911137, 0.43016977002783463, 0.0]
0.8056 12.5 1600 0.7305 0.3106 0.3897 0.7595 [0.7538928589848243, 0.25918801874572633, 0.4203485338620474, 0.2917137053557758, 0.22289503658523782, 0.1500749815626878, 0.08861529960760275, 0.624086456268012, 0.48451939021185103, nan, 0.0, 0.3968018568653773, 0.34541558718052773, 0.0] [0.9269385933729979, 0.2666068145623601, 0.5690834086799277, 0.323866282726429, 0.23199974868520706, 0.27844356662715314, 0.08945714746397179, 0.7853752936779852, 0.7445501612243971, nan, 0.0, 0.4213536720599331, 0.42861443009823574, 0.0]
0.7534 15.62 2000 0.8091 0.2762 0.3774 0.7408 [0.7450525081446471, 0.08805507548598117, 0.39296675038083556, 0.17888143033371248, 0.23466648108570068, 0.16079006682931019, 0.03714210117676656, 0.6219865494576018, 0.5133743295405404, nan, 0.0, 0.25318448016278294, 0.3644475822818699, 0.0] [0.8925789949539187, 0.08813412970942185, 0.6256571559820335, 0.18979883934004887, 0.24639981315291482, 0.4060874245051725, 0.03731593695118816, 0.8003825980603533, 0.8838065177476595, nan, 0.0, 0.25812424222515545, 0.47821461707899227, 0.0]
0.909 18.75 2400 0.7351 0.3258 0.4109 0.7698 [0.7635658546626002, 0.40487164496967293, 0.3833268397798514, 0.2752089531995694, 0.2867291055853458, 0.019354003421208514, 0.16890635042927862, 0.5748136637050552, 0.6549805047839075, nan, 0.0, 0.41018196642738697, 0.29410209406700305, 0.0] [0.936836241285118, 0.4435412622122596, 0.6115797993350055, 0.29433416172859234, 0.3084226650431367, 0.01980599690692518, 0.1729223045848531, 0.8637196946532967, 0.9040471540812156, nan, 0.0, 0.44469064868513, 0.3419148529297193, 0.0]
0.6622 21.88 2800 0.6963 0.3398 0.3986 0.7827 [0.7663761109958838, 0.1332027535954636, 0.4428707076945945, 0.35775923195424675, 0.3733786220955952, 0.24949901778334058, 0.07535404639899677, 0.6174965473057359, 0.5443824880761584, nan, 0.0, 0.4962856002887922, 0.3608184550398603, 0.0] [0.9493652245312661, 0.13338632341294235, 0.6300075103540804, 0.4173944872618647, 0.41213577281689057, 0.31108969330720987, 0.0759392172499597, 0.6593100168039995, 0.5955570838172228, nan, 0.0, 0.5200351189445205, 0.4777004199984418, 0.0]
1.1202 25.0 3200 1.6352 0.1684 0.2322 0.7026 [0.6963172586859363, 0.011559652438449069, 0.22669839158942193, 0.05922256175168542, 0.06532726727054458, 0.0023358067869460536, 0.0, 0.509361415684211, 0.38727970470408707, nan, 0.0, 0.09125458640751972, 0.13993367687629527, 0.0] [0.9487412199772098, 0.011559652438449069, 0.24983867321938985, 0.05932148648786475, 0.06540672362680026, 0.003050864548414594, 0.0, 0.7020197466594538, 0.7380590139684559, nan, 0.0, 0.09130527401337467, 0.14901587211649467, 0.0]
0.6529 28.12 3600 0.7747 0.3647 0.4742 0.7441 [0.7508396000072782, 0.461963906773346, 0.41562431632163865, 0.2643890336606752, 0.20882280410355394, 0.21001420486640948, 0.45776923048905305, 0.6263265430221951, 0.5990132199881534, nan, 0.0, 0.4389678627683, 0.30687238462719907, 0.0] [0.8794950417196551, 0.5122587212045141, 0.47963636761360323, 0.28307093261894156, 0.22726847453969443, 0.817351469679542, 0.5476940642254209, 0.672940072704085, 0.8955934160757268, nan, 0.0, 0.44608425759467085, 0.40330970104326763, 0.0]

Framework versions

  • Transformers 4.32.1
  • Pytorch 2.0.1+cu118
  • Datasets 2.14.4
  • Tokenizers 0.13.3
Downloads last month
31
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API: The model has no pipeline_tag.

Model tree for gabryland/lpcv_seg

Finetuned
(3)
this model