7-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: 0.7844
• F1: 0.7449

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
2.0948	1.0	18	1.9801	0.1856
1.916	2.0	36	1.7571	0.3242
1.6873	3.0	54	1.5333	0.4149
1.4576	4.0	72	1.3515	0.4656
1.2824	5.0	90	1.2288	0.4936
1.2004	6.0	108	1.1050	0.5462
1.1264	7.0	126	1.0643	0.5713
1.0149	8.0	144	1.0612	0.5718
0.9839	9.0	162	0.9897	0.6266
0.9001	10.0	180	0.9542	0.6710
0.9093	11.0	198	0.8993	0.6811
0.8824	12.0	216	0.8877	0.7018
0.8237	13.0	234	0.8970	0.7071
0.8446	14.0	252	0.8619	0.7084
0.7766	15.0	270	0.8271	0.7331
0.7405	16.0	288	0.8516	0.7237
0.7672	17.0	306	0.8036	0.7223
0.7149	18.0	324	0.8188	0.7186
0.7	19.0	342	0.8391	0.7274
0.7011	20.0	360	0.7922	0.7424
0.695	21.0	378	0.8065	0.7394
0.6655	22.0	396	0.7783	0.7473
0.6377	23.0	414	0.7977	0.7296
0.6884	24.0	432	0.7724	0.7387
0.614	25.0	450	0.8372	0.7351
0.6008	26.0	468	0.8229	0.7277
0.6402	27.0	486	0.7958	0.7300
0.592	28.0	504	0.8222	0.7264
0.5774	29.0	522	0.7613	0.7511
0.584	30.0	540	0.7866	0.7377
0.558	31.0	558	0.8298	0.7351
0.5871	32.0	576	0.7727	0.7494
0.5608	33.0	594	0.7753	0.7695
0.5385	34.0	612	0.7585	0.7575
0.5461	35.0	630	0.7664	0.7521
0.506	36.0	648	0.7624	0.7581
0.5132	37.0	666	0.7914	0.7347
0.5083	38.0	684	0.7913	0.7425
0.5042	39.0	702	0.7704	0.7556
0.4539	40.0	720	0.7590	0.7578
0.4714	41.0	738	0.7912	0.7503
0.4681	42.0	756	0.7838	0.7420
0.4482	43.0	774	0.7781	0.7345
0.4535	44.0	792	0.7823	0.7415
0.4284	45.0	810	0.8104	0.7449
0.436	46.0	828	0.7829	0.7421
0.4526	47.0	846	0.7932	0.7567
0.4672	48.0	864	0.7827	0.7411
0.4171	49.0	882	0.7835	0.7447
0.4126	50.0	900	0.7844	0.7449

Framework versions

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