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: 0.0255
Accuracy: 1.0

Model description

Image classification model using a pretrained Vision Transformer to categorize different types of pipe failures.

Intended uses & limitations

Diagnostic for Failure on Pipe through image recognition

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: 16
eval_batch_size: 16
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	15	0.0516	0.9867
No log	2.0	30	0.0441	0.9867
No log	3.0	45	0.0497	0.9733
No log	4.0	60	0.0464	0.9867
No log	5.0	75	0.0677	0.9867
No log	6.0	90	0.0208	1.0
No log	7.0	105	0.0183	1.0
No log	8.0	120	0.0943	0.9733
No log	9.0	135	0.0216	1.0
No log	10.0	150	0.0148	1.0
No log	11.0	165	0.0144	1.0
No log	12.0	180	0.0188	1.0
No log	13.0	195	0.0602	0.9867
No log	14.0	210	0.0882	0.9733
No log	15.0	225	0.0314	0.9867
No log	16.0	240	0.0127	1.0
No log	17.0	255	0.0119	1.0
No log	18.0	270	0.0117	1.0
No log	19.0	285	0.0114	1.0
No log	20.0	300	0.0131	1.0

Framework versions

Transformers 4.38.2
Pytorch 2.2.1+cu121
Datasets 2.18.0
Tokenizers 0.15.2

Ghiffari4869/pipe-failure_classification

pipe-failure_classification