hushem_1x_deit_small_rms_0001_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: 2.4160
• Accuracy: 0.5952

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
No log	1.0	6	2.4234	0.2619
2.1235	2.0	12	1.6901	0.2381
2.1235	3.0	18	1.6241	0.2619
1.5609	4.0	24	1.4045	0.2381
1.4894	5.0	30	1.3850	0.2619
1.4894	6.0	36	1.4265	0.2619
1.443	7.0	42	1.3761	0.4286
1.443	8.0	48	1.3894	0.2857
1.4117	9.0	54	1.3579	0.2619
1.4179	10.0	60	1.3731	0.2381
1.4179	11.0	66	1.2327	0.3571
1.3285	12.0	72	1.3528	0.3095
1.3285	13.0	78	1.0099	0.6429
1.188	14.0	84	1.0249	0.5476
1.013	15.0	90	1.0156	0.4762
1.013	16.0	96	1.1277	0.5476
0.7705	17.0	102	1.0504	0.5238
0.7705	18.0	108	1.2933	0.5238
0.8013	19.0	114	1.2536	0.5238
0.4428	20.0	120	1.7547	0.5476
0.4428	21.0	126	1.4286	0.5
0.3007	22.0	132	1.5184	0.5714
0.3007	23.0	138	1.6958	0.5714
0.209	24.0	144	2.1697	0.5238
0.0896	25.0	150	2.0909	0.6190
0.0896	26.0	156	2.2304	0.5952
0.0492	27.0	162	2.2787	0.5952
0.0492	28.0	168	2.0038	0.6667
0.0152	29.0	174	2.0666	0.6667
0.0011	30.0	180	2.1829	0.6190
0.0011	31.0	186	2.2393	0.6429
0.0006	32.0	192	2.2827	0.6190
0.0006	33.0	198	2.3132	0.6190
0.0005	34.0	204	2.3363	0.6190
0.0004	35.0	210	2.3564	0.6190
0.0004	36.0	216	2.3732	0.6190
0.0004	37.0	222	2.3884	0.6190
0.0004	38.0	228	2.3988	0.6190
0.0004	39.0	234	2.4073	0.5952
0.0003	40.0	240	2.4126	0.5952
0.0003	41.0	246	2.4153	0.5952
0.0003	42.0	252	2.4160	0.5952
0.0003	43.0	258	2.4160	0.5952
0.0003	44.0	264	2.4160	0.5952
0.0003	45.0	270	2.4160	0.5952
0.0003	46.0	276	2.4160	0.5952
0.0003	47.0	282	2.4160	0.5952
0.0003	48.0	288	2.4160	0.5952
0.0003	49.0	294	2.4160	0.5952
0.0003	50.0	300	2.4160	0.5952

Framework versions

• Transformers 4.35.1
• Pytorch 2.1.0+cu118
• Datasets 2.14.7
• Tokenizers 0.14.1