This model is a fine-tuned version of microsoft/swin-tiny-patch4-window7-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

Loss: 0.5516
Crack: {'precision': 0.575, 'recall': 0.71875, 'f1-score': 0.6388888888888888, 'support': 32}
Environment - ground: {'precision': 0.9714285714285714, 'recall': 0.9714285714285714, 'f1-score': 0.9714285714285714, 'support': 35}
Environment - other: {'precision': 0.8571428571428571, 'recall': 0.8888888888888888, 'f1-score': 0.8727272727272727, 'support': 27}
Environment - sky: {'precision': 0.9761904761904762, 'recall': 0.9318181818181818, 'f1-score': 0.9534883720930233, 'support': 44}
Environment - vegetation: {'precision': 0.9791666666666666, 'recall': 0.9791666666666666, 'f1-score': 0.9791666666666666, 'support': 48}
Joint defect: {'precision': 0.9166666666666666, 'recall': 0.7096774193548387, 'f1-score': 0.7999999999999999, 'support': 31}
Loss of section: {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 2}
Spalling: {'precision': 0.6041666666666666, 'recall': 0.6041666666666666, 'f1-score': 0.6041666666666666, 'support': 48}
Vegetation: {'precision': 0.8309859154929577, 'recall': 0.8939393939393939, 'f1-score': 0.8613138686131386, 'support': 66}
Wall - grafitti: {'precision': 0.7, 'recall': 0.9545454545454546, 'f1-score': 0.8076923076923077, 'support': 22}
Wall - normal: {'precision': 0.6976744186046512, 'recall': 0.7317073170731707, 'f1-score': 0.7142857142857143, 'support': 41}
Wall - other: {'precision': 0.7910447761194029, 'recall': 0.7794117647058824, 'f1-score': 0.7851851851851852, 'support': 68}
Wall - stain: {'precision': 0.8222222222222222, 'recall': 0.6491228070175439, 'f1-score': 0.7254901960784313, 'support': 57}
Accuracy: 0.8061
Macro avg: {'precision': 0.7478222490154723, 'recall': 0.754817164008097, 'f1-score': 0.7472179777173742, 'support': 521}
Weighted avg: {'precision': 0.8107856771401473, 'recall': 0.8061420345489443, 'f1-score': 0.8050072232872345, 'support': 521}

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: 8
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 32
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Crack	Environment - ground	Environment - other	Environment - sky	Environment - vegetation	Joint defect	Loss of section	Spalling	Vegetation	Wall - grafitti	Wall - normal	Wall - other	Wall - stain	Accuracy	Macro avg	Weighted avg
0.9193	1.0	146	0.7596	{'precision': 0.5681818181818182, 'recall': 0.78125, 'f1-score': 0.6578947368421052, 'support': 32}	{'precision': 0.9444444444444444, 'recall': 0.9714285714285714, 'f1-score': 0.9577464788732395, 'support': 35}	{'precision': 0.8846153846153846, 'recall': 0.8518518518518519, 'f1-score': 0.8679245283018868, 'support': 27}	{'precision': 0.9736842105263158, 'recall': 0.8409090909090909, 'f1-score': 0.9024390243902439, 'support': 44}	{'precision': 1.0, 'recall': 1.0, 'f1-score': 1.0, 'support': 48}	{'precision': 0.7419354838709677, 'recall': 0.7419354838709677, 'f1-score': 0.7419354838709677, 'support': 31}	{'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 2}	{'precision': 0.5769230769230769, 'recall': 0.3125, 'f1-score': 0.4054054054054054, 'support': 48}	{'precision': 0.75, 'recall': 0.9090909090909091, 'f1-score': 0.821917808219178, 'support': 66}	{'precision': 0.5142857142857142, 'recall': 0.8181818181818182, 'f1-score': 0.6315789473684209, 'support': 22}	{'precision': 0.7692307692307693, 'recall': 0.4878048780487805, 'f1-score': 0.5970149253731344, 'support': 41}	{'precision': 0.7540983606557377, 'recall': 0.6764705882352942, 'f1-score': 0.7131782945736433, 'support': 68}	{'precision': 0.6428571428571429, 'recall': 0.7894736842105263, 'f1-score': 0.7086614173228346, 'support': 57}	0.7562	{'precision': 0.7015581850454902, 'recall': 0.7062228366021391, 'f1-score': 0.692745926964697, 'support': 521}	{'precision': 0.7618631381912654, 'recall': 0.7562380038387716, 'f1-score': 0.7479524876767193, 'support': 521}
0.7347	2.0	293	0.6495	{'precision': 0.5526315789473685, 'recall': 0.65625, 'f1-score': 0.6, 'support': 32}	{'precision': 1.0, 'recall': 0.9714285714285714, 'f1-score': 0.9855072463768115, 'support': 35}	{'precision': 0.8461538461538461, 'recall': 0.8148148148148148, 'f1-score': 0.830188679245283, 'support': 27}	{'precision': 0.9761904761904762, 'recall': 0.9318181818181818, 'f1-score': 0.9534883720930233, 'support': 44}	{'precision': 0.9591836734693877, 'recall': 0.9791666666666666, 'f1-score': 0.9690721649484536, 'support': 48}	{'precision': 0.9130434782608695, 'recall': 0.6774193548387096, 'f1-score': 0.7777777777777777, 'support': 31}	{'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 2}	{'precision': 0.5306122448979592, 'recall': 0.5416666666666666, 'f1-score': 0.5360824742268041, 'support': 48}	{'precision': 0.7058823529411765, 'recall': 0.9090909090909091, 'f1-score': 0.794701986754967, 'support': 66}	{'precision': 0.6333333333333333, 'recall': 0.8636363636363636, 'f1-score': 0.7307692307692307, 'support': 22}	{'precision': 0.5510204081632653, 'recall': 0.6585365853658537, 'f1-score': 0.6, 'support': 41}	{'precision': 0.8095238095238095, 'recall': 0.75, 'f1-score': 0.7786259541984734, 'support': 68}	{'precision': 0.9393939393939394, 'recall': 0.543859649122807, 'f1-score': 0.688888888888889, 'support': 57}	0.7678	{'precision': 0.7243822416365717, 'recall': 0.7152067510345803, 'f1-score': 0.7111617519445933, 'support': 521}	{'precision': 0.7869554245446998, 'recall': 0.7677543186180422, 'f1-score': 0.7672943491004631, 'support': 521}
0.7515	2.99	438	0.5516	{'precision': 0.575, 'recall': 0.71875, 'f1-score': 0.6388888888888888, 'support': 32}	{'precision': 0.9714285714285714, 'recall': 0.9714285714285714, 'f1-score': 0.9714285714285714, 'support': 35}	{'precision': 0.8571428571428571, 'recall': 0.8888888888888888, 'f1-score': 0.8727272727272727, 'support': 27}	{'precision': 0.9761904761904762, 'recall': 0.9318181818181818, 'f1-score': 0.9534883720930233, 'support': 44}	{'precision': 0.9791666666666666, 'recall': 0.9791666666666666, 'f1-score': 0.9791666666666666, 'support': 48}	{'precision': 0.9166666666666666, 'recall': 0.7096774193548387, 'f1-score': 0.7999999999999999, 'support': 31}	{'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 2}	{'precision': 0.6041666666666666, 'recall': 0.6041666666666666, 'f1-score': 0.6041666666666666, 'support': 48}	{'precision': 0.8309859154929577, 'recall': 0.8939393939393939, 'f1-score': 0.8613138686131386, 'support': 66}	{'precision': 0.7, 'recall': 0.9545454545454546, 'f1-score': 0.8076923076923077, 'support': 22}	{'precision': 0.6976744186046512, 'recall': 0.7317073170731707, 'f1-score': 0.7142857142857143, 'support': 41}	{'precision': 0.7910447761194029, 'recall': 0.7794117647058824, 'f1-score': 0.7851851851851852, 'support': 68}	{'precision': 0.8222222222222222, 'recall': 0.6491228070175439, 'f1-score': 0.7254901960784313, 'support': 57}	0.8061	{'precision': 0.7478222490154723, 'recall': 0.754817164008097, 'f1-score': 0.7472179777173742, 'support': 521}	{'precision': 0.8107856771401473, 'recall': 0.8061420345489443, 'f1-score': 0.8050072232872345, 'support': 521}

Framework versions

Transformers 4.33.2
Pytorch 2.0.1+cu118
Datasets 2.14.5
Tokenizers 0.13.3

ziauldin/swin-tiny-patch4-window7-224-finetuned-vit

swin-tiny-patch4-window7-224-finetuned-vit

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

ziauldin/
swin-tiny-patch4-window7-224-finetuned-vit