smids_5x_deit_tiny_adamax_001_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: 1.5310
• Accuracy: 0.865

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: 0.001
• 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
0.4168	1.0	375	0.3631	0.85
0.2785	2.0	750	0.4582	0.82
0.1977	3.0	1125	0.4757	0.845
0.2154	4.0	1500	0.4151	0.8567
0.2216	5.0	1875	0.4921	0.84
0.1277	6.0	2250	0.5208	0.84
0.1577	7.0	2625	0.6509	0.84
0.1043	8.0	3000	0.6131	0.8483
0.0606	9.0	3375	0.7321	0.85
0.0399	10.0	3750	0.7332	0.8483
0.0878	11.0	4125	0.7794	0.86
0.0753	12.0	4500	0.9361	0.855
0.0315	13.0	4875	0.7541	0.87
0.0322	14.0	5250	0.8827	0.855
0.0291	15.0	5625	0.8552	0.8667
0.0323	16.0	6000	1.0097	0.8533
0.0358	17.0	6375	1.0442	0.8367
0.0726	18.0	6750	1.0675	0.8533
0.0105	19.0	7125	1.0350	0.8567
0.0155	20.0	7500	1.0612	0.8467
0.0001	21.0	7875	1.1933	0.8467
0.001	22.0	8250	0.9964	0.86
0.0061	23.0	8625	1.0207	0.86
0.0139	24.0	9000	1.1598	0.8467
0.0232	25.0	9375	1.1652	0.8583
0.0001	26.0	9750	1.1454	0.8583
0.0011	27.0	10125	1.1331	0.865
0.0	28.0	10500	1.2646	0.8667
0.0	29.0	10875	1.1994	0.8683
0.0001	30.0	11250	1.2306	0.8533
0.004	31.0	11625	1.2452	0.8617
0.0	32.0	12000	1.2904	0.8633
0.0	33.0	12375	1.3971	0.86
0.0001	34.0	12750	1.2738	0.8633
0.0	35.0	13125	1.4099	0.865
0.0	36.0	13500	1.3138	0.8633
0.0	37.0	13875	1.3962	0.8617
0.0037	38.0	14250	1.4247	0.8633
0.0	39.0	14625	1.4177	0.865
0.0	40.0	15000	1.4033	0.8633
0.0	41.0	15375	1.4591	0.8633
0.0	42.0	15750	1.4725	0.8617
0.0	43.0	16125	1.4752	0.8633
0.0	44.0	16500	1.4834	0.8633
0.0	45.0	16875	1.4967	0.8633
0.0	46.0	17250	1.5039	0.8633
0.0	47.0	17625	1.5125	0.8633
0.0	48.0	18000	1.5211	0.8633
0.0	49.0	18375	1.5277	0.865
0.0	50.0	18750	1.5310	0.865

Framework versions

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