hushem_1x_deit_base_adamax_001_fold2

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: 2.6061
• Accuracy: 0.6

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
No log	1.0	6	1.4006	0.4667
2.0073	2.0	12	1.4185	0.2444
2.0073	3.0	18	1.8021	0.2444
1.3793	4.0	24	1.3019	0.2889
1.3619	5.0	30	1.2559	0.4222
1.3619	6.0	36	1.3762	0.3556
1.2354	7.0	42	1.1026	0.5333
1.2354	8.0	48	1.3770	0.3556
1.116	9.0	54	1.3199	0.3333
1.2825	10.0	60	1.2535	0.4444
1.2825	11.0	66	0.9621	0.5333
1.0651	12.0	72	1.0556	0.5778
1.0651	13.0	78	1.1244	0.4889
0.8879	14.0	84	1.1678	0.4889
0.7249	15.0	90	1.1215	0.5778
0.7249	16.0	96	1.2306	0.5333
0.5807	17.0	102	1.9201	0.5333
0.5807	18.0	108	2.0291	0.4667
0.5755	19.0	114	2.7334	0.5333
0.7966	20.0	120	1.6804	0.5111
0.7966	21.0	126	2.2911	0.4444
0.7407	22.0	132	1.3830	0.5333
0.7407	23.0	138	1.5155	0.5556
0.3047	24.0	144	1.6845	0.4889
0.2535	25.0	150	1.8110	0.4889
0.2535	26.0	156	1.9764	0.5111
0.4369	27.0	162	1.6350	0.4889
0.4369	28.0	168	2.4101	0.5111
0.2888	29.0	174	2.3032	0.4889
0.3277	30.0	180	1.7523	0.5556
0.3277	31.0	186	1.6541	0.6
0.1303	32.0	192	2.1471	0.5333
0.1303	33.0	198	2.1714	0.5556
0.0771	34.0	204	2.1399	0.5778
0.0588	35.0	210	2.1914	0.5778
0.0588	36.0	216	2.2720	0.5778
0.0221	37.0	222	2.4076	0.5778
0.0221	38.0	228	2.4716	0.5556
0.0111	39.0	234	2.5364	0.5556
0.0075	40.0	240	2.5792	0.6
0.0075	41.0	246	2.6027	0.6
0.0045	42.0	252	2.6061	0.6
0.0045	43.0	258	2.6061	0.6
0.003	44.0	264	2.6061	0.6
0.0047	45.0	270	2.6061	0.6
0.0047	46.0	276	2.6061	0.6
0.0042	47.0	282	2.6061	0.6
0.0042	48.0	288	2.6061	0.6
0.0037	49.0	294	2.6061	0.6
0.0043	50.0	300	2.6061	0.6

Framework versions

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