hushem_5x_deit_tiny_adamax_001_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: 1.8507
• Accuracy: 0.7209

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
1.4597	1.0	28	1.3552	0.3256
1.2609	2.0	56	0.9015	0.6744
1.1559	3.0	84	2.3181	0.3256
1.0822	4.0	112	1.6592	0.4186
1.0553	5.0	140	0.9222	0.4651
0.9651	6.0	168	0.7306	0.7442
0.9771	7.0	196	0.7827	0.6279
0.9016	8.0	224	0.9874	0.5581
0.8187	9.0	252	1.0233	0.5349
0.8405	10.0	280	0.6414	0.7674
0.7808	11.0	308	0.6637	0.6977
0.8152	12.0	336	0.5374	0.8605
0.8044	13.0	364	0.5684	0.7442
0.7275	14.0	392	0.6622	0.8140
0.6931	15.0	420	0.7767	0.7209
0.691	16.0	448	0.5520	0.7674
0.7585	17.0	476	0.6770	0.7674
0.6303	18.0	504	0.6834	0.7442
0.6578	19.0	532	0.7776	0.6977
0.4934	20.0	560	1.1067	0.7442
0.5397	21.0	588	0.7250	0.7674
0.5586	22.0	616	0.9824	0.6047
0.4808	23.0	644	0.9582	0.7442
0.4823	24.0	672	0.9114	0.6279
0.4124	25.0	700	1.1614	0.7209
0.3991	26.0	728	1.3579	0.6279
0.5138	27.0	756	1.5915	0.6512
0.3857	28.0	784	0.6799	0.8140
0.3797	29.0	812	1.1771	0.7907
0.3781	30.0	840	0.9809	0.7674
0.3225	31.0	868	0.7120	0.8140
0.3017	32.0	896	0.9984	0.7674
0.2468	33.0	924	1.0271	0.7674
0.1651	34.0	952	1.2194	0.7674
0.254	35.0	980	1.2984	0.7907
0.1644	36.0	1008	1.3708	0.7674
0.1972	37.0	1036	1.6461	0.7209
0.1252	38.0	1064	1.8473	0.7442
0.1252	39.0	1092	1.5823	0.7209
0.4334	40.0	1120	1.7311	0.7209
0.1029	41.0	1148	1.7390	0.7674
0.0992	42.0	1176	1.6959	0.7907
0.0993	43.0	1204	1.9237	0.7442
0.0435	44.0	1232	1.8569	0.7442
0.054	45.0	1260	1.8007	0.7907
0.0678	46.0	1288	1.8496	0.7442
0.0576	47.0	1316	1.8145	0.7442
0.0232	48.0	1344	1.8497	0.7209
0.0393	49.0	1372	1.8507	0.7209
0.0213	50.0	1400	1.8507	0.7209

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0