hushem_1x_deit_tiny_adamax_00001_fold2

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: 1.3630
• Accuracy: 0.5333

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: 1e-05
• 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	1.3933	0.2889
1.4502	2.0	12	1.3758	0.2889
1.4502	3.0	18	1.3846	0.1556
1.1864	4.0	24	1.3867	0.2
1.0417	5.0	30	1.4200	0.2222
1.0417	6.0	36	1.4398	0.2667
0.8998	7.0	42	1.4309	0.2667
0.8998	8.0	48	1.4422	0.2889
0.802	9.0	54	1.4525	0.3111
0.7173	10.0	60	1.4451	0.3333
0.7173	11.0	66	1.4170	0.3556
0.6327	12.0	72	1.4262	0.3778
0.6327	13.0	78	1.4500	0.3778
0.5705	14.0	84	1.4362	0.3778
0.4928	15.0	90	1.4119	0.3778
0.4928	16.0	96	1.4031	0.4
0.4272	17.0	102	1.4009	0.4
0.4272	18.0	108	1.4134	0.4
0.3882	19.0	114	1.4007	0.4
0.3396	20.0	120	1.3936	0.4
0.3396	21.0	126	1.3916	0.4222
0.2975	22.0	132	1.3801	0.4222
0.2975	23.0	138	1.3854	0.4222
0.2664	24.0	144	1.3827	0.4444
0.2292	25.0	150	1.3826	0.4444
0.2292	26.0	156	1.3717	0.4667
0.2136	27.0	162	1.3670	0.4667
0.2136	28.0	168	1.3720	0.4667
0.1873	29.0	174	1.3622	0.4667
0.1666	30.0	180	1.3494	0.5111
0.1666	31.0	186	1.3586	0.4889
0.1595	32.0	192	1.3677	0.5111
0.1595	33.0	198	1.3760	0.5111
0.1486	34.0	204	1.3711	0.5111
0.1401	35.0	210	1.3652	0.5111
0.1401	36.0	216	1.3610	0.5333
0.1317	37.0	222	1.3597	0.5333
0.1317	38.0	228	1.3618	0.5333
0.1202	39.0	234	1.3633	0.5333
0.122	40.0	240	1.3628	0.5333
0.122	41.0	246	1.3631	0.5333
0.1214	42.0	252	1.3630	0.5333
0.1214	43.0	258	1.3630	0.5333
0.1203	44.0	264	1.3630	0.5333
0.1185	45.0	270	1.3630	0.5333
0.1185	46.0	276	1.3630	0.5333
0.1174	47.0	282	1.3630	0.5333
0.1174	48.0	288	1.3630	0.5333
0.1152	49.0	294	1.3630	0.5333
0.1204	50.0	300	1.3630	0.5333

Framework versions

• Transformers 4.35.0
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1