smids_5x_deit_tiny_adamax_00001_fold4

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: 1.3354
• Accuracy: 0.8717

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
0.2847	1.0	375	0.3588	0.86
0.2399	2.0	750	0.3497	0.8783
0.1069	3.0	1125	0.3839	0.8683
0.1588	4.0	1500	0.3944	0.8767
0.0651	5.0	1875	0.4658	0.8717
0.0408	6.0	2250	0.5268	0.8767
0.026	7.0	2625	0.6145	0.865
0.0213	8.0	3000	0.6639	0.8717
0.0032	9.0	3375	0.8259	0.8583
0.0148	10.0	3750	0.8114	0.8717
0.015	11.0	4125	0.8747	0.87
0.0014	12.0	4500	0.9000	0.8783
0.0001	13.0	4875	0.9761	0.87
0.015	14.0	5250	0.9964	0.87
0.0001	15.0	5625	1.0366	0.87
0.0	16.0	6000	1.0464	0.8733
0.0	17.0	6375	1.0647	0.8733
0.0197	18.0	6750	1.1060	0.87
0.0	19.0	7125	1.1223	0.87
0.0	20.0	7500	1.1307	0.8717
0.0	21.0	7875	1.1563	0.8667
0.0	22.0	8250	1.1659	0.8717
0.0	23.0	8625	1.1873	0.8717
0.0	24.0	9000	1.1966	0.875
0.0	25.0	9375	1.2159	0.8667
0.0	26.0	9750	1.2192	0.8667
0.0	27.0	10125	1.2446	0.865
0.0	28.0	10500	1.2249	0.8683
0.0	29.0	10875	1.2382	0.8717
0.0	30.0	11250	1.2391	0.87
0.0	31.0	11625	1.2547	0.8733
0.0	32.0	12000	1.2771	0.87
0.0	33.0	12375	1.2797	0.8717
0.0	34.0	12750	1.2938	0.87
0.0	35.0	13125	1.2859	0.87
0.0	36.0	13500	1.2998	0.8717
0.0	37.0	13875	1.2941	0.8683
0.0015	38.0	14250	1.3001	0.8683
0.0	39.0	14625	1.3082	0.8717
0.0	40.0	15000	1.3151	0.8717
0.0	41.0	15375	1.3188	0.8717
0.0	42.0	15750	1.3223	0.8717
0.0	43.0	16125	1.3285	0.87
0.0	44.0	16500	1.3298	0.87
0.0	45.0	16875	1.3317	0.87
0.0	46.0	17250	1.3330	0.87
0.0	47.0	17625	1.3343	0.87
0.0	48.0	18000	1.3352	0.87
0.0	49.0	18375	1.3354	0.8717
0.0	50.0	18750	1.3354	0.8717

Framework versions

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