hushem_40x_deit_small_adamax_001_fold5

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: 0.6357
• Accuracy: 0.9268

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
0.2206	1.0	220	1.1411	0.6829
0.1873	2.0	440	0.5940	0.8780
0.0203	3.0	660	0.9936	0.7805
0.0624	4.0	880	0.3597	0.9024
0.0108	5.0	1100	1.3539	0.7805
0.0858	6.0	1320	0.8241	0.8049
0.0246	7.0	1540	1.0359	0.8049
0.0131	8.0	1760	0.7509	0.8049
0.0013	9.0	1980	1.4351	0.7805
0.0095	10.0	2200	1.1916	0.7561
0.0002	11.0	2420	0.7203	0.8293
0.0011	12.0	2640	1.0391	0.8293
0.007	13.0	2860	1.8906	0.7317
0.0002	14.0	3080	0.4058	0.9512
0.0	15.0	3300	0.3547	0.9268
0.0	16.0	3520	0.3764	0.9268
0.0	17.0	3740	0.3894	0.9268
0.0	18.0	3960	0.4031	0.9268
0.0	19.0	4180	0.4138	0.9268
0.0	20.0	4400	0.4231	0.9268
0.0	21.0	4620	0.4326	0.9268
0.0	22.0	4840	0.4413	0.9268
0.0	23.0	5060	0.4490	0.9268
0.0	24.0	5280	0.4564	0.9268
0.0	25.0	5500	0.4638	0.9268
0.0	26.0	5720	0.4710	0.9268
0.0	27.0	5940	0.4779	0.9268
0.0	28.0	6160	0.4851	0.9268
0.0	29.0	6380	0.4923	0.9268
0.0	30.0	6600	0.4998	0.9268
0.0	31.0	6820	0.5069	0.9268
0.0	32.0	7040	0.5143	0.9268
0.0	33.0	7260	0.5224	0.9268
0.0	34.0	7480	0.5303	0.9268
0.0	35.0	7700	0.5381	0.9268
0.0	36.0	7920	0.5458	0.9268
0.0	37.0	8140	0.5543	0.9268
0.0	38.0	8360	0.5622	0.9268
0.0	39.0	8580	0.5706	0.9268
0.0	40.0	8800	0.5791	0.9268
0.0	41.0	9020	0.5871	0.9268
0.0	42.0	9240	0.5951	0.9268
0.0	43.0	9460	0.6028	0.9268
0.0	44.0	9680	0.6101	0.9268
0.0	45.0	9900	0.6166	0.9268
0.0	46.0	10120	0.6227	0.9268
0.0	47.0	10340	0.6281	0.9268
0.0	48.0	10560	0.6322	0.9268
0.0	49.0	10780	0.6350	0.9268
0.0	50.0	11000	0.6357	0.9268

Framework versions

• Transformers 4.32.1
• Pytorch 2.1.0+cu121
• Datasets 2.12.0
• Tokenizers 0.13.2