vit-base-patch16-224-in21k-FINALAsphaltLaneClassifier-detectorVIT30epochsTrainVal

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.0886
• Accuracy: {'accuracy': 0.9563318777292577}
• F1: {'f1': 0.9318274318274318}
• Precision: {'precision': 0.9367965367965368}
• Recall: {'recall': 0.927993839045052}

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: 5e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
1.7198	0.9956	57	1.5576	{'accuracy': 0.7467248908296943}	{'f1': 0.49276057276057283}	{'precision': 0.5726499690785405}	{'recall': 0.5304274269791511}
0.9081	1.9913	114	0.7542	{'accuracy': 0.9126637554585153}	{'f1': 0.6815669614826331}	{'precision': 0.6667189132706375}	{'recall': 0.700220596772321}
0.4761	2.9869	171	0.3690	{'accuracy': 0.9606986899563319}	{'f1': 0.8055100468538725}	{'precision': 0.8327510665283063}	{'recall': 0.7959183673469388}
0.4783	4.0	229	0.3196	{'accuracy': 0.9344978165938864}	{'f1': 0.7876659055230484}	{'precision': 0.8222068488412314}	{'recall': 0.7732035342872782}
0.3541	4.9956	286	0.2280	{'accuracy': 0.9563318777292577}	{'f1': 0.8029204504768414}	{'precision': 0.8327627953577691}	{'recall': 0.7909922589725545}
0.3096	5.9913	343	0.1897	{'accuracy': 0.9737991266375546}	{'f1': 0.9533858998144712}	{'precision': 0.9854721549636805}	{'recall': 0.9387755102040816}
0.4737	6.9869	400	0.1658	{'accuracy': 0.9781659388646288}	{'f1': 0.9625090929438755}	{'precision': 0.9876847290640394}	{'recall': 0.9489795918367347}
0.2629	8.0	458	0.1555	{'accuracy': 0.9606986899563319}	{'f1': 0.9470988692620754}	{'precision': 0.9443563789152024}	{'recall': 0.9509845577788695}
0.2431	8.9956	515	0.1322	{'accuracy': 0.9694323144104804}	{'f1': 0.9613275613275613}	{'precision': 0.9523809523809524}	{'recall': 0.9811320754716981}
0.2829	9.9913	572	0.1333	{'accuracy': 0.9694323144104804}	{'f1': 0.9613275613275613}	{'precision': 0.9523809523809524}	{'recall': 0.9811320754716981}
0.2283	10.9869	629	0.1119	{'accuracy': 0.9563318777292577}	{'f1': 0.9413515406162466}	{'precision': 0.9356980887593132}	{'recall': 0.9505198305737389}
0.1808	12.0	687	0.1128	{'accuracy': 0.9519650655021834}	{'f1': 0.9340521676882466}	{'precision': 0.9297478991596638}	{'recall': 0.9403157489410859}
0.2545	12.9956	744	0.1193	{'accuracy': 0.9650655021834061}	{'f1': 0.947008806419719}	{'precision': 0.9586734693877551}	{'recall': 0.9386626478828356}
0.1341	13.9913	801	0.1022	{'accuracy': 0.9650655021834061}	{'f1': 0.941991341991342}	{'precision': 0.9563909774436091}	{'recall': 0.9333846746245668}
0.1833	14.9869	858	0.1094	{'accuracy': 0.9563318777292577}	{'f1': 0.9386446886446886}	{'precision': 0.9352240896358543}	{'recall': 0.9430111667308434}
0.2267	16.0	916	0.1074	{'accuracy': 0.9650655021834061}	{'f1': 0.9454619454619454}	{'precision': 0.9512987012987013}	{'recall': 0.9408933384674626}
0.2109	16.9956	973	0.1028	{'accuracy': 0.9650655021834061}	{'f1': 0.9454619454619454}	{'precision': 0.9512987012987013}	{'recall': 0.9408933384674626}
0.2643	17.9913	1030	0.1110	{'accuracy': 0.9475982532751092}	{'f1': 0.9370300751879699}	{'precision': 0.9312015503875969}	{'recall': 0.9601463226800154}
0.1624	18.9869	1087	0.0891	{'accuracy': 0.9781659388646288}	{'f1': 0.9625090929438755}	{'precision': 0.9876847290640394}	{'recall': 0.9489795918367347}
0.144	20.0	1145	0.0990	{'accuracy': 0.9475982532751092}	{'f1': 0.9296218487394958}	{'precision': 0.9248461289277615}	{'recall': 0.9376203311513285}
0.1473	20.9956	1202	0.0907	{'accuracy': 0.9650655021834061}	{'f1': 0.941991341991342}	{'precision': 0.9563909774436091}	{'recall': 0.9333846746245668}
0.1364	21.9913	1259	0.0935	{'accuracy': 0.9519650655021834}	{'f1': 0.9271126934678336}	{'precision': 0.9291819291819292}	{'recall': 0.9252984212552945}
0.184	22.9869	1316	0.0906	{'accuracy': 0.9563318777292577}	{'f1': 0.9386446886446886}	{'precision': 0.9352240896358543}	{'recall': 0.9430111667308434}
0.149	24.0	1374	0.0950	{'accuracy': 0.9475982532751092}	{'f1': 0.932436974789916}	{'precision': 0.926595744680851}	{'recall': 0.9451289949942241}
0.213	24.9956	1431	0.0884	{'accuracy': 0.9563318777292577}	{'f1': 0.9318274318274318}	{'precision': 0.9367965367965368}	{'recall': 0.927993839045052}
0.1058	25.9913	1488	0.0876	{'accuracy': 0.9694323144104804}	{'f1': 0.9475127301214259}	{'precision': 0.9693486590038314}	{'recall': 0.9360800924143241}
0.1216	26.9869	1545	0.0874	{'accuracy': 0.9563318777292577}	{'f1': 0.9318274318274318}	{'precision': 0.9367965367965368}	{'recall': 0.927993839045052}
0.1126	28.0	1603	0.0895	{'accuracy': 0.9650655021834061}	{'f1': 0.941991341991342}	{'precision': 0.9563909774436091}	{'recall': 0.9333846746245668}
0.1325	28.9956	1660	0.0894	{'accuracy': 0.9563318777292577}	{'f1': 0.9318274318274318}	{'precision': 0.9367965367965368}	{'recall': 0.927993839045052}
0.1197	29.8690	1710	0.0886	{'accuracy': 0.9563318777292577}	{'f1': 0.9318274318274318}	{'precision': 0.9367965367965368}	{'recall': 0.927993839045052}

Framework versions

• Transformers 4.43.3
• Pytorch 2.3.1
• Datasets 2.20.0
• Tokenizers 0.19.1