deit_fold_3_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: 1.0615
• Accuracy: 0.9776
• F1 Score: 0.9779
• Recall: 0.9783

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
2.8067	1.0	20	2.8042	0.2308	0.1857	0.2025
2.6818	2.0	40	2.6787	0.3301	0.2595	0.2840
2.4775	3.0	60	2.4942	0.4744	0.4403	0.4350
2.1304	4.0	80	2.1863	0.6571	0.6641	0.6486
1.7918	5.0	100	1.8476	0.8045	0.8103	0.8106
1.5272	6.0	120	1.5614	0.8590	0.8606	0.8641
1.3652	7.0	140	1.3730	0.9199	0.9191	0.9261
1.2514	8.0	160	1.2684	0.9263	0.9257	0.9337
1.1792	9.0	180	1.2029	0.9551	0.9547	0.9597
1.1409	10.0	200	1.1749	0.9551	0.9552	0.9597
1.0983	11.0	220	1.1696	0.9583	0.9587	0.9634
1.1016	12.0	240	1.1319	0.9647	0.9644	0.9677
1.0771	13.0	260	1.1243	0.9647	0.9644	0.9677
1.0477	14.0	280	1.1087	0.9615	0.9608	0.9627
1.0469	15.0	300	1.1094	0.9551	0.9544	0.9552
1.0373	16.0	320	1.1093	0.9487	0.9486	0.9516
1.0312	17.0	340	1.1033	0.9551	0.9544	0.9552
1.0166	18.0	360	1.0903	0.9615	0.9607	0.9614
1.0146	19.0	380	1.0961	0.9679	0.9672	0.9676
1.0268	20.0	400	1.0846	0.9712	0.9709	0.9720
1.0042	21.0	420	1.0741	0.9712	0.9717	0.9727
1.0118	22.0	440	1.0857	0.9679	0.9688	0.9715
0.9751	23.0	460	1.0877	0.9712	0.9717	0.9727
0.9916	24.0	480	1.0910	0.9647	0.9650	0.9658
0.9838	25.0	500	1.0876	0.9647	0.9647	0.9652
0.9865	26.0	520	1.0825	0.9647	0.9647	0.9652
0.9818	27.0	540	1.0722	0.9712	0.9710	0.9714
0.9899	28.0	560	1.0736	0.9744	0.9746	0.9764
0.9816	29.0	580	1.0659	0.9744	0.9745	0.9751
0.9795	30.0	600	1.0697	0.9712	0.9714	0.9714
0.9688	31.0	620	1.0696	0.9712	0.9714	0.9714
0.9783	32.0	640	1.0826	0.9712	0.9714	0.9714
0.9596	33.0	660	1.0759	0.9679	0.9687	0.9702
0.9724	34.0	680	1.0751	0.9647	0.9651	0.9652
0.9837	35.0	700	1.0623	0.9744	0.9745	0.9751
0.9736	36.0	720	1.0704	0.9679	0.9674	0.9683
0.9763	37.0	740	1.0599	0.9712	0.9717	0.9727
0.9683	38.0	760	1.0733	0.9679	0.9687	0.9702
0.9641	39.0	780	1.0615	0.9776	0.9779	0.9783
0.9677	40.0	800	1.0642	0.9744	0.9745	0.9751
0.9718	41.0	820	1.0638	0.9744	0.9744	0.9745
0.9649	42.0	840	1.0689	0.9712	0.9716	0.9721
0.9724	43.0	860	1.0657	0.9776	0.9779	0.9783
0.9703	44.0	880	1.0595	0.9744	0.9745	0.9751
0.9675	45.0	900	1.0616	0.9776	0.9779	0.9783
0.9737	46.0	920	1.0590	0.9776	0.9779	0.9783
0.9640	47.0	940	1.0640	0.9776	0.9779	0.9783
0.9640	48.0	960	1.0613	0.9744	0.9745	0.9751
0.9581	49.0	980	1.0638	0.9744	0.9744	0.9745
0.9784	50.0	1000	1.0627	0.9744	0.9745	0.9751
0.9603	51.0	1020	1.0665	0.9744	0.9745	0.9751
0.9601	52.0	1040	1.0627	0.9744	0.9748	0.9764
0.9693	53.0	1060	1.0637	0.9744	0.9745	0.9751
0.9636	54.0	1080	1.0576	0.9744	0.9751	0.9758
0.9686	55.0	1100	1.0593	0.9744	0.9745	0.9751
0.9585	56.0	1120	1.0593	0.9744	0.9745	0.9751
0.9615	57.0	1140	1.0586	0.9776	0.9779	0.9783
0.9671	58.0	1160	1.0549	0.9744	0.9751	0.9758
0.9652	59.0	1180	1.0562	0.9776	0.9779	0.9783
0.9651	60.0	1200	1.0559	0.9744	0.9748	0.9764
0.9645	61.0	1220	1.0538	0.9776	0.9779	0.9783
0.9570	62.0	1240	1.0575	0.9744	0.9745	0.9751
0.9532	63.0	1260	1.0546	0.9776	0.9779	0.9783
0.9587	64.0	1280	1.0525	0.9776	0.9779	0.9783
0.9541	65.0	1300	1.0583	0.9744	0.9745	0.9751
0.9571	66.0	1320	1.0579	0.9744	0.9745	0.9751
0.9548	67.0	1340	1.0539	0.9712	0.9717	0.9727
0.9559	68.0	1360	1.0544	0.9712	0.9717	0.9727
0.9594	69.0	1380	1.0542	0.9744	0.9751	0.9758
0.9594	70.0	1400	1.0559	0.9744	0.9751	0.9758
0.9599	71.0	1420	1.0551	0.9712	0.9717	0.9727
0.9614	72.0	1440	1.0526	0.9712	0.9717	0.9727
0.9568	73.0	1460	1.0535	0.9712	0.9717	0.9727
0.9561	74.0	1480	1.0546	0.9712	0.9717	0.9727

Framework versions

• Transformers 5.0.0
• Pytorch 2.10.0+cu128
• Datasets 5.0.0
• Tokenizers 0.22.2