11-classifier-finetuned-padchest

This model is a fine-tuned version of microsoft/resnet-50 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 1.1743
• F1: 0.5098

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
2.0747	1.0	18	2.0666	0.1549
2.0661	2.0	36	2.0560	0.1777
2.0497	3.0	54	2.0385	0.2169
2.018	4.0	72	2.0047	0.2515
1.9792	5.0	90	1.9773	0.2329
1.9619	6.0	108	1.9421	0.2321
1.9186	7.0	126	1.9145	0.2055
1.8838	8.0	144	1.8976	0.2596
1.8402	9.0	162	1.8444	0.2337
1.7906	10.0	180	1.7951	0.2397
1.7716	11.0	198	1.7695	0.3373
1.7474	12.0	216	1.7940	0.3209
1.6957	13.0	234	1.7425	0.3314
1.6791	14.0	252	1.6727	0.3558
1.6483	15.0	270	1.6638	0.3895
1.614	16.0	288	1.6513	0.4186
1.6166	17.0	306	1.6002	0.4406
1.5654	18.0	324	1.5528	0.4627
1.5145	19.0	342	1.5571	0.4676
1.5049	20.0	360	1.4334	0.4364
1.457	21.0	378	1.4711	0.4535
1.4516	22.0	396	1.5013	0.4516
1.4172	23.0	414	1.3614	0.4682
1.3817	24.0	432	1.3519	0.4545
1.3987	25.0	450	1.3806	0.4759
1.4063	26.0	468	1.2961	0.4866
1.3684	27.0	486	1.3328	0.4768
1.3789	28.0	504	1.2810	0.4859
1.341	29.0	522	1.3227	0.4737
1.3574	30.0	540	1.2406	0.5025
1.357	31.0	558	1.2427	0.5033
1.3204	32.0	576	1.2478	0.5053
1.3122	33.0	594	1.2205	0.5133
1.334	34.0	612	1.2138	0.5204
1.2998	35.0	630	1.2122	0.5111
1.3097	36.0	648	1.2118	0.5102
1.2956	37.0	666	1.2077	0.5163
1.3058	38.0	684	1.2023	0.5157
1.2851	39.0	702	1.1968	0.5067
1.2728	40.0	720	1.1940	0.5169
1.2653	41.0	738	1.1700	0.5165
1.2837	42.0	756	1.1767	0.5262
1.2789	43.0	774	1.1885	0.5146
1.2343	44.0	792	1.1925	0.5101
1.2454	45.0	810	1.1874	0.5119
1.2922	46.0	828	1.1845	0.5216
1.2547	47.0	846	1.1920	0.5299
1.272	48.0	864	1.1732	0.5225
1.2506	49.0	882	1.1722	0.5117
1.2494	50.0	900	1.1743	0.5098

Framework versions

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