deit_fold_4_v3

This model is a fine-tuned version of facebook/deit-small-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

Loss: 1.0823
Accuracy: 0.9744
F1 Score: 0.9761
Recall: 0.9759

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: 1e-05
train_batch_size: 64
eval_batch_size: 64
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 150
num_epochs: 100
label_smoothing_factor: 0.15

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Score	Recall
2.7733	1.0	20	2.7802	0.3141	0.2358	0.2698
2.6592	2.0	40	2.6547	0.4071	0.3282	0.3507
2.4295	3.0	60	2.4600	0.5032	0.4527	0.4513
2.0844	4.0	80	2.1512	0.6827	0.6782	0.6559
1.7587	5.0	100	1.7780	0.8013	0.8102	0.7977
1.4649	6.0	120	1.4535	0.8942	0.8959	0.8968
1.3221	7.0	140	1.3009	0.9071	0.9081	0.9058
1.2698	8.0	160	1.2227	0.9327	0.9343	0.9362
1.1416	9.0	180	1.1886	0.9263	0.9272	0.9306
1.1385	10.0	200	1.1925	0.9327	0.9345	0.9401
1.1101	11.0	220	1.1562	0.9423	0.9433	0.9468
1.0563	12.0	240	1.1363	0.9583	0.9593	0.9616
1.0536	13.0	260	1.1351	0.9487	0.9499	0.9524
1.0680	14.0	280	1.1316	0.9583	0.9598	0.9604
1.0535	15.0	300	1.1226	0.9551	0.9562	0.9554
1.0392	16.0	320	1.1156	0.9583	0.9593	0.9591
1.0192	17.0	340	1.1333	0.9487	0.9510	0.9543
1.0271	18.0	360	1.1237	0.9551	0.9572	0.9605
0.9999	19.0	380	1.1219	0.9615	0.9637	0.9648
1.0084	20.0	400	1.1194	0.9519	0.9542	0.9562
1.0009	21.0	420	1.1206	0.9487	0.9514	0.9556
1.0102	22.0	440	1.1420	0.9551	0.9580	0.9625
0.9913	23.0	460	1.1073	0.9647	0.9664	0.9679
0.9960	24.0	480	1.1068	0.9615	0.9625	0.9626
0.9866	25.0	500	1.0989	0.9647	0.9657	0.9663
0.9960	26.0	520	1.1028	0.9647	0.9668	0.9672
0.9841	27.0	540	1.1211	0.9519	0.9543	0.9568
0.9760	28.0	560	1.0944	0.9679	0.9695	0.9703
0.9858	29.0	580	1.0954	0.9679	0.9695	0.9703
0.9820	30.0	600	1.1262	0.9551	0.9574	0.9612
0.9766	31.0	620	1.1109	0.9583	0.9599	0.9614
0.9781	32.0	640	1.0823	0.9744	0.9761	0.9759
0.9671	33.0	660	1.1072	0.9583	0.9606	0.9621
0.9746	34.0	680	1.0900	0.9679	0.9700	0.9710
0.9660	35.0	700	1.0965	0.9647	0.9662	0.9679
0.9698	36.0	720	1.0958	0.9647	0.9670	0.9685
0.9746	37.0	740	1.0999	0.9615	0.9634	0.9654
0.9725	38.0	760	1.1208	0.9583	0.9602	0.9617
0.9580	39.0	780	1.1081	0.9583	0.9604	0.9629
0.9630	40.0	800	1.0995	0.9647	0.9660	0.9666
0.9691	41.0	820	1.0995	0.9679	0.9695	0.9703
0.9609	42.0	840	1.0744	0.9712	0.9723	0.9728
0.9732	43.0	860	1.0931	0.9712	0.9723	0.9728
0.9836	44.0	880	1.1029	0.9647	0.9664	0.9679
0.9653	45.0	900	1.1136	0.9583	0.9604	0.9629
0.9682	46.0	920	1.0677	0.9712	0.9723	0.9728
0.9703	47.0	940	1.0935	0.9615	0.9634	0.9654
0.9766	48.0	960	1.0852	0.9712	0.9723	0.9728
0.9651	49.0	980	1.0944	0.9615	0.9629	0.9639
0.9641	50.0	1000	1.0817	0.9679	0.9695	0.9703
0.9671	51.0	1020	1.0981	0.9679	0.9695	0.9695
0.9714	52.0	1040	1.0929	0.9679	0.9695	0.9703
0.9682	53.0	1060	1.0976	0.9615	0.9629	0.9639
0.9576	54.0	1080	1.1010	0.9679	0.9700	0.9710
0.9591	55.0	1100	1.0903	0.9744	0.9757	0.9759
0.9672	56.0	1120	1.0976	0.9679	0.9695	0.9703

Framework versions

Transformers 5.0.0
Pytorch 2.10.0+cu128
Datasets 5.0.0
Tokenizers 0.22.2

nqvii/deit_fold_4_v3

deit_fold_4_v3

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

为此模型做种