smids_10x_deit_tiny_sgd_0001_fold3

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: 0.3822
• Accuracy: 0.85

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: 0.0001
• 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
1.0766	1.0	750	1.0931	0.4283
0.9667	2.0	1500	1.0054	0.5083
0.9193	3.0	2250	0.9314	0.5533
0.7955	4.0	3000	0.8645	0.595
0.78	5.0	3750	0.7998	0.6283
0.6828	6.0	4500	0.7419	0.6833
0.6256	7.0	5250	0.6910	0.715
0.6533	8.0	6000	0.6501	0.7417
0.5725	9.0	6750	0.6153	0.76
0.5585	10.0	7500	0.5875	0.7733
0.5119	11.0	8250	0.5647	0.7917
0.4751	12.0	9000	0.5443	0.8033
0.4748	13.0	9750	0.5264	0.815
0.543	14.0	10500	0.5129	0.8167
0.4792	15.0	11250	0.4995	0.8233
0.4638	16.0	12000	0.4900	0.8267
0.4373	17.0	12750	0.4791	0.8267
0.4624	18.0	13500	0.4695	0.8333
0.4581	19.0	14250	0.4625	0.8317
0.4031	20.0	15000	0.4549	0.8283
0.3798	21.0	15750	0.4466	0.8367
0.4127	22.0	16500	0.4410	0.84
0.4351	23.0	17250	0.4350	0.8433
0.4464	24.0	18000	0.4308	0.835
0.4334	25.0	18750	0.4260	0.8333
0.4382	26.0	19500	0.4212	0.8367
0.3088	27.0	20250	0.4169	0.8367
0.3982	28.0	21000	0.4136	0.8367
0.4061	29.0	21750	0.4101	0.8367
0.4007	30.0	22500	0.4079	0.84
0.3333	31.0	23250	0.4046	0.8417
0.3804	32.0	24000	0.4012	0.84
0.4007	33.0	24750	0.3989	0.8417
0.4048	34.0	25500	0.3970	0.8417
0.3319	35.0	26250	0.3948	0.8433
0.3736	36.0	27000	0.3932	0.8467
0.3994	37.0	27750	0.3918	0.8483
0.3998	38.0	28500	0.3900	0.8483
0.3526	39.0	29250	0.3888	0.85
0.4438	40.0	30000	0.3872	0.8483
0.3369	41.0	30750	0.3860	0.8517
0.3716	42.0	31500	0.3855	0.8517
0.3661	43.0	32250	0.3844	0.8517
0.3454	44.0	33000	0.3840	0.8517
0.3872	45.0	33750	0.3835	0.8517
0.3238	46.0	34500	0.3830	0.8517
0.404	47.0	35250	0.3826	0.8517
0.3607	48.0	36000	0.3823	0.8517
0.3506	49.0	36750	0.3823	0.8517
0.3442	50.0	37500	0.3822	0.85

Framework versions

• Transformers 4.32.1
• Pytorch 2.1.0+cu121
• Datasets 2.12.0
• Tokenizers 0.13.2