3-classifier-finetuned-padchest

This model is a fine-tuned version of nickmuchi/vit-finetuned-chest-xray-pneumonia on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.8505
• Accuracy: 0.7251

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.0499	1.0	16	1.8761	0.3686
1.704	2.0	32	1.5961	0.4955
1.5393	3.0	48	1.3570	0.5770
1.3161	4.0	64	1.2687	0.5770
1.1991	5.0	80	1.1740	0.6073
1.1459	6.0	96	1.1388	0.6073
1.071	7.0	112	1.0763	0.6405
0.9948	8.0	128	1.0419	0.6526
0.9902	9.0	144	0.9869	0.6979
0.9515	10.0	160	0.9825	0.6767
0.9277	11.0	176	0.9645	0.6858
0.9182	12.0	192	0.9264	0.7009
0.895	13.0	208	0.9138	0.6979
0.8765	14.0	224	0.9089	0.7100
0.8536	15.0	240	0.8941	0.7009
0.8385	16.0	256	0.8764	0.7221
0.8187	17.0	272	0.8659	0.7160
0.8172	18.0	288	0.8673	0.7069
0.8101	19.0	304	0.8530	0.7341
0.8127	20.0	320	0.8505	0.7251

Framework versions

• Transformers 4.28.0.dev0
• Pytorch 2.0.0+cu117
• Datasets 2.18.0
• Tokenizers 0.13.3