hushem_5x_deit_base_adamax_001_fold3

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

• Loss: 4.0748
• Accuracy: 0.4651

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
1.4157	1.0	28	1.4434	0.2558
1.3795	2.0	56	1.2758	0.4651
1.2039	3.0	84	1.4398	0.3256
1.2374	4.0	112	1.2828	0.3953
0.9763	5.0	140	1.0462	0.5814
0.9947	6.0	168	1.0570	0.5349
0.984	7.0	196	0.9874	0.6047
1.0476	8.0	224	1.1609	0.5581
0.8502	9.0	252	1.6864	0.5116
0.9652	10.0	280	0.9691	0.6512
1.0167	11.0	308	0.9040	0.6279
0.8838	12.0	336	1.0599	0.4884
0.8002	13.0	364	1.1020	0.3721
0.7509	14.0	392	0.8581	0.6744
0.7759	15.0	420	0.8348	0.6047
0.6809	16.0	448	0.8650	0.6279
0.6145	17.0	476	1.6492	0.3953
0.5427	18.0	504	1.4270	0.4419
0.5373	19.0	532	1.3497	0.4651
0.5816	20.0	560	1.2788	0.5116
0.5066	21.0	588	0.8857	0.6744
0.4263	22.0	616	1.0042	0.6744
0.4421	23.0	644	0.8870	0.7442
0.2947	24.0	672	1.5951	0.6279
0.3564	25.0	700	1.4096	0.5349
0.4309	26.0	728	1.5181	0.5116
0.2828	27.0	756	2.3004	0.3953
0.3386	28.0	784	2.1526	0.6047
0.1818	29.0	812	2.5940	0.5349
0.3444	30.0	840	2.3898	0.4884
0.2604	31.0	868	2.5996	0.5581
0.1135	32.0	896	2.5370	0.6047
0.1275	33.0	924	3.5621	0.5349
0.0534	34.0	952	3.3776	0.5116
0.0878	35.0	980	3.5763	0.5349
0.0297	36.0	1008	3.1205	0.5814
0.0316	37.0	1036	3.7999	0.5116
0.0083	38.0	1064	3.7418	0.5814
0.0256	39.0	1092	4.0409	0.5116
0.0017	40.0	1120	3.7639	0.4651
0.0016	41.0	1148	4.0848	0.4884
0.0008	42.0	1176	3.8952	0.4884
0.0002	43.0	1204	3.9086	0.4651
0.0002	44.0	1232	3.9976	0.4651
0.0001	45.0	1260	4.0424	0.4651
0.0001	46.0	1288	4.0591	0.4651
0.0001	47.0	1316	4.0703	0.4651
0.0001	48.0	1344	4.0747	0.4651
0.0001	49.0	1372	4.0748	0.4651
0.0001	50.0	1400	4.0748	0.4651

Framework versions

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