9-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: 0.1585
• F1: 0.9563

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	F1
2.5332	1.0	18	0.5695	0.7920
0.4488	2.0	36	0.3419	0.7934
0.3259	3.0	54	0.2451	0.7934
0.2795	4.0	72	0.1954	0.9443
0.2348	5.0	90	0.1698	0.9343
0.1937	6.0	108	0.1829	0.9297
0.1851	7.0	126	0.1484	0.9454
0.1925	8.0	144	0.1330	0.9545
0.1614	9.0	162	0.1403	0.9387
0.1734	10.0	180	0.1221	0.9531
0.1697	11.0	198	0.1142	0.9524
0.1824	12.0	216	0.1129	0.9586
0.1336	13.0	234	0.1369	0.9441
0.1596	14.0	252	0.1181	0.9540
0.1474	15.0	270	0.1116	0.9646
0.1256	16.0	288	0.1035	0.9598
0.1398	17.0	306	0.1195	0.9519
0.1219	18.0	324	0.1123	0.9588
0.1114	19.0	342	0.1126	0.9586
0.1089	20.0	360	0.1083	0.9584
0.1123	21.0	378	0.1038	0.9554
0.1241	22.0	396	0.0927	0.9657
0.099	23.0	414	0.1397	0.9559
0.1025	24.0	432	0.1201	0.9584
0.1088	25.0	450	0.0894	0.9627
0.0953	26.0	468	0.1083	0.9632
0.0953	27.0	486	0.1061	0.9592
0.0831	28.0	504	0.1129	0.9570
0.0836	29.0	522	0.1123	0.9598
0.0705	30.0	540	0.1611	0.9499
0.1047	31.0	558	0.1191	0.9570
0.0803	32.0	576	0.1440	0.9563
0.0852	33.0	594	0.1149	0.9541
0.0588	34.0	612	0.1830	0.9489
0.0701	35.0	630	0.1475	0.9592
0.0607	36.0	648	0.1350	0.9627
0.0749	37.0	666	0.1389	0.9563
0.073	38.0	684	0.1463	0.9559
0.0579	39.0	702	0.1289	0.9595
0.0757	40.0	720	0.1585	0.9584
0.0538	41.0	738	0.1565	0.9588
0.0461	42.0	756	0.1630	0.9559
0.072	43.0	774	0.1704	0.9554
0.0517	44.0	792	0.1657	0.9559
0.0524	45.0	810	0.1358	0.9570
0.0569	46.0	828	0.1538	0.9533
0.0506	47.0	846	0.1579	0.9588
0.0506	48.0	864	0.1505	0.9566
0.0538	49.0	882	0.1593	0.9588
0.0532	50.0	900	0.1585	0.9563

Framework versions

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