smids_5x_deit_tiny_sgd_00001_fold4

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

• Loss: 1.0682
• Accuracy: 0.4133

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: 32
• eval_batch_size: 32
• seed: 42
• 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	Accuracy
1.3789	1.0	375	1.3304	0.3467
1.3522	2.0	750	1.2980	0.3417
1.2851	3.0	1125	1.2700	0.3467
1.3268	4.0	1500	1.2462	0.35
1.2083	5.0	1875	1.2264	0.3533
1.2564	6.0	2250	1.2096	0.3667
1.2076	7.0	2625	1.1953	0.375
1.1738	8.0	3000	1.1833	0.375
1.1964	9.0	3375	1.1730	0.3767
1.1824	10.0	3750	1.1642	0.375
1.1746	11.0	4125	1.1567	0.375
1.0941	12.0	4500	1.1499	0.3783
1.1561	13.0	4875	1.1439	0.3817
1.1702	14.0	5250	1.1384	0.3817
1.1181	15.0	5625	1.1334	0.3867
1.149	16.0	6000	1.1288	0.3833
1.1131	17.0	6375	1.1244	0.3867
1.1335	18.0	6750	1.1203	0.39
1.105	19.0	7125	1.1164	0.3933
1.0655	20.0	7500	1.1128	0.3933
1.1098	21.0	7875	1.1094	0.395
1.0972	22.0	8250	1.1061	0.3933
1.112	23.0	8625	1.1030	0.3917
1.0932	24.0	9000	1.1001	0.395
1.0801	25.0	9375	1.0974	0.3933
1.1085	26.0	9750	1.0947	0.4
1.1153	27.0	10125	1.0922	0.4
1.0883	28.0	10500	1.0899	0.4
1.0621	29.0	10875	1.0877	0.4017
1.0559	30.0	11250	1.0856	0.4017
1.0795	31.0	11625	1.0837	0.4
1.1076	32.0	12000	1.0819	0.4017
1.1027	33.0	12375	1.0802	0.405
1.0471	34.0	12750	1.0787	0.41
1.032	35.0	13125	1.0772	0.4117
1.0529	36.0	13500	1.0759	0.4083
1.0365	37.0	13875	1.0747	0.4067
1.0659	38.0	14250	1.0736	0.4067
1.073	39.0	14625	1.0726	0.4117
1.1034	40.0	15000	1.0717	0.4117
1.0918	41.0	15375	1.0710	0.4117
1.0873	42.0	15750	1.0703	0.4133
1.0582	43.0	16125	1.0697	0.4133
1.0527	44.0	16500	1.0693	0.4133
1.0394	45.0	16875	1.0689	0.4133
1.0718	46.0	17250	1.0686	0.4133
1.0719	47.0	17625	1.0684	0.4133
1.0655	48.0	18000	1.0683	0.4133
1.0516	49.0	18375	1.0682	0.4133
1.0396	50.0	18750	1.0682	0.4133

Framework versions

• Transformers 4.32.1
• Pytorch 2.1.1+cu121
• Datasets 2.12.0
• Tokenizers 0.13.2