5-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: 0.7307
• Accuracy: 0.7644

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: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.0492	1.0	16	1.9604	0.3142
1.8545	2.0	32	1.7361	0.4079
1.724	3.0	48	1.5064	0.5166
1.4761	4.0	64	1.3116	0.5710
1.3215	5.0	80	1.2030	0.6344
1.2325	6.0	96	1.0904	0.6254
1.124	7.0	112	1.0145	0.6677
1.0516	8.0	128	0.9864	0.6707
0.9858	9.0	144	0.9372	0.6767
0.9518	10.0	160	0.9161	0.6949
0.9612	11.0	176	0.8916	0.6949
0.8994	12.0	192	0.8579	0.7069
0.8194	13.0	208	0.8281	0.7100
0.8141	14.0	224	0.8064	0.7341
0.8056	15.0	240	0.8272	0.7221
0.7953	16.0	256	0.7751	0.7251
0.7679	17.0	272	0.7638	0.7523
0.7262	18.0	288	0.7867	0.7432
0.7302	19.0	304	0.7835	0.7311
0.7237	20.0	320	0.7698	0.7492
0.6496	21.0	336	0.7618	0.7523
0.6708	22.0	352	0.7595	0.7492
0.6719	23.0	368	0.7455	0.7553
0.6361	24.0	384	0.7993	0.7221
0.6125	25.0	400	0.7372	0.7432
0.6392	26.0	416	0.7321	0.7613
0.6175	27.0	432	0.7310	0.7704
0.5613	28.0	448	0.7244	0.7462
0.5831	29.0	464	0.7535	0.7523
0.5892	30.0	480	0.7299	0.7583
0.5259	31.0	496	0.7211	0.7674
0.5553	32.0	512	0.7564	0.7341
0.5497	33.0	528	0.7233	0.7704
0.5699	34.0	544	0.7314	0.7523
0.5263	35.0	560	0.7334	0.7583
0.4953	36.0	576	0.6991	0.7674
0.5029	37.0	592	0.7191	0.7674
0.5253	38.0	608	0.7233	0.7704
0.4657	39.0	624	0.7204	0.7644
0.498	40.0	640	0.7236	0.7674
0.4768	41.0	656	0.7242	0.7734
0.5016	42.0	672	0.7405	0.7553
0.4774	43.0	688	0.7363	0.7674
0.4859	44.0	704	0.7208	0.7734
0.4628	45.0	720	0.7393	0.7674
0.4515	46.0	736	0.7078	0.7734
0.4297	47.0	752	0.7287	0.7674
0.4023	48.0	768	0.7138	0.7734
0.4404	49.0	784	0.7272	0.7674
0.4236	50.0	800	0.7307	0.7644

Framework versions

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