deit_fold_5_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: 1.1352
Accuracy: 0.9614
F1 Score: 0.9625
Recall: 0.9610

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
2.7385	1.0	20	2.6882	0.3312	0.3035	0.3403
2.6061	2.0	40	2.5293	0.4984	0.4626	0.4734
2.3085	3.0	60	2.2662	0.6592	0.6420	0.6298
1.9367	4.0	80	1.9558	0.7781	0.7760	0.7657
1.5973	5.0	100	1.6854	0.8167	0.8183	0.8156
1.3828	6.0	120	1.5037	0.8714	0.8702	0.8708
1.2393	7.0	140	1.3885	0.8875	0.8869	0.8860
1.1855	8.0	160	1.3267	0.9003	0.8997	0.8986
1.1754	9.0	180	1.2754	0.9132	0.9134	0.9157
1.1145	10.0	200	1.2655	0.9164	0.9174	0.9213
1.0929	11.0	220	1.2461	0.9196	0.9201	0.9167
1.0782	12.0	240	1.2546	0.9196	0.9192	0.9190
1.0536	13.0	260	1.2059	0.9421	0.9435	0.9425
1.0342	14.0	280	1.2038	0.9421	0.9429	0.9413
1.0225	15.0	300	1.1835	0.9453	0.9461	0.9450
1.0106	16.0	320	1.1949	0.9421	0.9422	0.9413
1.0228	17.0	340	1.1882	0.9293	0.9301	0.9326
0.9900	18.0	360	1.1842	0.9325	0.9337	0.9382
1.0063	19.0	380	1.1777	0.9453	0.9467	0.9501
1.0024	20.0	400	1.1728	0.9453	0.9458	0.9469
0.9878	21.0	420	1.1696	0.9421	0.9439	0.9476
0.9881	22.0	440	1.1523	0.9550	0.9558	0.9575
0.9722	23.0	460	1.1996	0.9389	0.9405	0.9457
0.9808	24.0	480	1.1487	0.9582	0.9589	0.9587
0.9756	25.0	500	1.1437	0.9582	0.9590	0.9591
0.9692	26.0	520	1.1600	0.9550	0.9555	0.9568
0.9734	27.0	540	1.1790	0.9421	0.9435	0.9482
0.9663	28.0	560	1.1624	0.9550	0.9559	0.9566
0.9848	29.0	580	1.1690	0.9453	0.9471	0.9431
0.9686	30.0	600	1.1608	0.9421	0.9435	0.9430
0.9671	31.0	620	1.1463	0.9518	0.9524	0.9523
0.9803	32.0	640	1.1536	0.9550	0.9556	0.9537
0.9627	33.0	660	1.1391	0.9582	0.9593	0.9585
0.9641	34.0	680	1.1441	0.9518	0.9524	0.9535
0.9636	35.0	700	1.1466	0.9550	0.9556	0.9541
0.9636	36.0	720	1.1427	0.9582	0.9585	0.9572
0.9682	37.0	740	1.1423	0.9550	0.9559	0.9566
0.9612	38.0	760	1.1486	0.9550	0.9558	0.9554
0.9603	39.0	780	1.1517	0.9486	0.9501	0.9517
0.9759	40.0	800	1.1551	0.9453	0.9464	0.9423
0.9698	41.0	820	1.1441	0.9518	0.9529	0.9542
0.9636	42.0	840	1.1352	0.9614	0.9625	0.9610
0.9544	43.0	860	1.1449	0.9582	0.9589	0.9579
0.9598	44.0	880	1.1404	0.9614	0.9625	0.9610
0.9576	45.0	900	1.1384	0.9582	0.9589	0.9579
0.9576	46.0	920	1.1506	0.9518	0.9529	0.9542
0.9564	47.0	940	1.1472	0.9614	0.9625	0.9610
0.9534	48.0	960	1.1514	0.9518	0.9525	0.9516
0.9808	49.0	980	1.1484	0.9582	0.9592	0.9560
0.9647	50.0	1000	1.1430	0.9614	0.9623	0.9616
0.9650	51.0	1020	1.1385	0.9614	0.9625	0.9610
0.9587	52.0	1040	1.1374	0.9614	0.9625	0.9610

Framework versions

Transformers 5.0.0
Pytorch 2.10.0+cu128
Datasets 5.0.0
Tokenizers 0.22.2

nqvii/deit_fold_5_v3

deit_fold_5_v3

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

为此模型做种