hushem_40x_deit_base_rms_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: 4.0356
• Accuracy: 0.5581

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
1.1943	1.0	217	1.3862	0.3488
1.2108	2.0	434	1.3456	0.3721
0.8764	3.0	651	1.3683	0.4884
0.7995	4.0	868	0.8441	0.5814
0.8665	5.0	1085	1.2083	0.5116
0.7433	6.0	1302	0.7858	0.7209
0.7205	7.0	1519	0.8439	0.6744
0.6415	8.0	1736	0.6198	0.6512
0.6773	9.0	1953	0.8169	0.6744
0.5449	10.0	2170	0.8224	0.6512
0.5225	11.0	2387	0.7556	0.7209
0.5268	12.0	2604	0.8703	0.6744
0.41	13.0	2821	0.7919	0.6512
0.4695	14.0	3038	0.9473	0.6744
0.3173	15.0	3255	1.2235	0.6512
0.3283	16.0	3472	1.3091	0.6512
0.3212	17.0	3689	1.0773	0.6047
0.3662	18.0	3906	0.9193	0.6279
0.3712	19.0	4123	0.9811	0.6744
0.3483	20.0	4340	1.5620	0.5814
0.2594	21.0	4557	1.8035	0.5814
0.3019	22.0	4774	1.3880	0.6744
0.2498	23.0	4991	1.6113	0.5814
0.2349	24.0	5208	1.2780	0.6047
0.1589	25.0	5425	1.6674	0.6512
0.2341	26.0	5642	1.6966	0.6512
0.1986	27.0	5859	1.4673	0.6047
0.1141	28.0	6076	1.6993	0.6512
0.1291	29.0	6293	2.0265	0.5581
0.1273	30.0	6510	1.8689	0.6279
0.0887	31.0	6727	1.4863	0.6977
0.101	32.0	6944	2.2258	0.6279
0.09	33.0	7161	1.6918	0.5814
0.063	34.0	7378	2.4040	0.5349
0.0263	35.0	7595	2.2869	0.5814
0.0357	36.0	7812	2.0118	0.6047
0.033	37.0	8029	2.5046	0.6279
0.0417	38.0	8246	2.0462	0.6512
0.0049	39.0	8463	3.1349	0.5814
0.0034	40.0	8680	2.4922	0.6279
0.0115	41.0	8897	2.7021	0.5581
0.0248	42.0	9114	3.1496	0.5116
0.0078	43.0	9331	2.6336	0.6279
0.0022	44.0	9548	3.2458	0.5349
0.0015	45.0	9765	3.3966	0.5349
0.0031	46.0	9982	4.1353	0.5116
0.0	47.0	10199	3.5481	0.5814
0.0002	48.0	10416	3.8712	0.5349
0.0	49.0	10633	4.0305	0.5581
0.0	50.0	10850	4.0356	0.5581

Framework versions

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