nj1867/Shingle-Classifier

nj1867/roof_classification_35

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

• Train Loss: 0.2855
• Validation Loss: 0.5466
• Train Accuracy: 0.8413
• Epoch: 29

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:

• optimizer: {'name': 'AdamWeightDecay', 'learning_rate': {'module': 'keras.optimizers.schedules', 'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 3e-05, 'decay_steps': 42240, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, 'registered_name': None}, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight_decay_rate': 0.0001}
• training_precision: float32

Training results

Train Loss	Validation Loss	Train Accuracy	Epoch
3.4052	3.2118	0.3472	0
2.9771	2.8091	0.5122	1
2.5834	2.4653	0.6379	2
2.2565	2.2264	0.6794	3
1.9808	2.0136	0.6869	4
1.7365	1.7525	0.7934	5
1.5394	1.6366	0.7668	6
1.3621	1.5575	0.7519	7
1.1855	1.4560	0.7412	8
1.0507	1.0795	0.8477	9
0.9409	1.0760	0.8413	10
0.8775	1.1174	0.7827	11
0.7837	1.2030	0.7487	12
0.6933	0.8608	0.8413	13
0.6005	0.8497	0.8264	14
0.5648	0.7881	0.8381	15
0.5482	0.8070	0.8083	16
0.4783	0.6537	0.8530	17
0.4632	0.7237	0.8232	18
0.4047	0.5131	0.8818	19
0.3885	0.4548	0.9042	20
0.3793	0.5766	0.8584	21
0.3670	0.6578	0.8147	22
0.3202	0.8783	0.7551	23
0.3094	0.4762	0.8733	24
0.3193	0.5481	0.8498	25
0.3190	0.4589	0.8711	26
0.2742	0.4616	0.8637	27
0.2583	0.4354	0.8807	28
0.2855	0.5466	0.8413	29

Framework versions

• Transformers 4.35.2
• TensorFlow 2.14.0
• Datasets 2.15.0
• Tokenizers 0.15.0