This model is a fine-tuned version of Sohaibsoussi/ViT-NIH-Chest-X-ray-dataset-small on the Sohaibsoussi/NIH-Chest-X-ray-dataset-small dataset. It achieves the following results on the evaluation set:

Loss: 0.6731
Accuracy: 0.2189

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: 16
eval_batch_size: 8
seed: 42
optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
num_epochs: 9
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.0271	0.3690	100	0.0347	0.8584
0.0334	0.7380	200	0.0291	0.8624
0.0438	1.1070	300	0.0352	0.8607
0.0215	1.4760	400	0.0319	0.8746
0.0267	1.8450	500	0.0277	0.8798
0.0266	2.2140	600	0.0177	0.9116
0.014	2.5830	700	0.0127	0.9497
0.0207	2.9520	800	0.0144	0.9410
0.0115	3.3210	900	0.0097	0.9653
0.0113	3.6900	1000	0.0077	0.9711
0.0054	4.0590	1100	0.0068	0.9844
0.0047	4.4280	1200	0.0046	0.9850
0.0056	4.7970	1300	0.0040	0.9902
0.0026	5.1661	1400	0.0032	0.9925
0.0037	5.5351	1500	0.0027	0.9936
0.0039	5.9041	1600	0.0023	0.9977
0.0019	6.2731	1700	0.0019	0.9971
0.0019	6.6421	1800	0.0017	0.9988
0.0016	7.0111	1900	0.0015	1.0
0.002	7.3801	2000	0.0014	1.0
0.0013	7.7491	2100	0.0014	1.0
0.0015	8.1181	2200	0.0013	1.0
0.0011	8.4871	2300	0.0013	1.0
0.0013	8.8561	2400	0.0013	1.0

Framework versions

Transformers 4.46.3
Pytorch 2.4.0
Datasets 3.1.0
Tokenizers 0.20.3