metadata

license: apache-2.0
tags:
  - image-classification
  - generated_from_trainer
datasets:
  - imagefolder
metrics:
  - accuracy
model-index:
  - name: raildefectfft2
    results:
      - task:
          name: Image Classification
          type: image-classification
        dataset:
          name: defect
          type: imagefolder
          config: Dhika--defectfft
          split: validation
          args: Dhika--defectfft
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.7542857142857143

raildefectfft2

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

Loss: 0.7207
Accuracy: 0.7543

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: 0.0002
train_batch_size: 30
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	Accuracy
1.3922	0.67	10	1.1690	0.6114
0.8518	1.33	20	0.8874	0.6829
0.5386	2.0	30	0.7207	0.7543
0.3125	2.67	40	0.8383	0.7286
0.2264	3.33	50	0.8440	0.7429
0.1613	4.0	60	0.8516	0.7457
0.119	4.67	70	1.3625	0.6
0.0972	5.33	80	0.9110	0.7429
0.0844	6.0	90	0.8272	0.78
0.0725	6.67	100	0.8958	0.74
0.0708	7.33	110	1.0972	0.7371
0.041	8.0	120	1.0089	0.7629
0.0312	8.67	130	1.0348	0.7629
0.0401	9.33	140	1.2427	0.7257
0.0271	10.0	150	1.0154	0.7543
0.0328	10.67	160	1.0373	0.7714
0.023	11.33	170	1.0051	0.7686
0.0199	12.0	180	0.9775	0.7657
0.0189	12.67	190	1.0088	0.7657
0.0188	13.33	200	1.1904	0.7343
0.0167	14.0	210	1.2999	0.7286
0.0159	14.67	220	1.1326	0.7514
0.0145	15.33	230	1.1386	0.7543
0.015	16.0	240	1.1441	0.7543
0.0133	16.67	250	1.1544	0.7514
0.0132	17.33	260	1.1629	0.7514
0.0121	18.0	270	1.1708	0.7514
0.0121	18.67	280	1.1773	0.7514
0.0114	19.33	290	1.1831	0.7514
0.0111	20.0	300	1.1883	0.7514
0.011	20.67	310	1.1937	0.7514
0.0103	21.33	320	1.1993	0.7514
0.0103	22.0	330	1.2046	0.7514
0.0103	22.67	340	1.2089	0.7514
0.0096	23.33	350	1.2133	0.7514
0.0095	24.0	360	1.2171	0.7514
0.0096	24.67	370	1.2204	0.7514
0.0093	25.33	380	1.2235	0.7486
0.0091	26.0	390	1.2262	0.7486
0.0092	26.67	400	1.2280	0.7514
0.0089	27.33	410	1.2296	0.7514
0.0092	28.0	420	1.2310	0.7514
0.0089	28.67	430	1.2319	0.7486
0.0089	29.33	440	1.2325	0.7486
0.0088	30.0	450	1.2327	0.7486

Framework versions

Transformers 4.30.2
Pytorch 2.0.1+cu118
Datasets 2.12.0
Tokenizers 0.13.3