hushem_1x_deit_tiny_sgd_001_fold3

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: 1.2767
• Accuracy: 0.3488

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	1.6231	0.2791
1.6502	2.0	12	1.5615	0.2791
1.6502	3.0	18	1.5208	0.2558
1.5138	4.0	24	1.4935	0.2093
1.441	5.0	30	1.4720	0.2093
1.441	6.0	36	1.4541	0.2326
1.3942	7.0	42	1.4402	0.3023
1.3942	8.0	48	1.4271	0.3023
1.3895	9.0	54	1.4159	0.2791
1.3382	10.0	60	1.4069	0.2791
1.3382	11.0	66	1.3983	0.2558
1.3326	12.0	72	1.3893	0.2558
1.3326	13.0	78	1.3800	0.2558
1.3102	14.0	84	1.3707	0.2558
1.3163	15.0	90	1.3619	0.2791
1.3163	16.0	96	1.3528	0.2791
1.295	17.0	102	1.3463	0.2791
1.295	18.0	108	1.3391	0.2791
1.2552	19.0	114	1.3325	0.3023
1.2682	20.0	120	1.3269	0.3023
1.2682	21.0	126	1.3221	0.3256
1.2578	22.0	132	1.3173	0.3488
1.2578	23.0	138	1.3126	0.3488
1.2124	24.0	144	1.3087	0.3488
1.2284	25.0	150	1.3049	0.3488
1.2284	26.0	156	1.3017	0.3488
1.2178	27.0	162	1.2982	0.3488
1.2178	28.0	168	1.2955	0.3488
1.2019	29.0	174	1.2931	0.3488
1.2029	30.0	180	1.2906	0.3488
1.2029	31.0	186	1.2886	0.3488
1.1935	32.0	192	1.2863	0.3488
1.1935	33.0	198	1.2843	0.3488
1.164	34.0	204	1.2826	0.3488
1.1999	35.0	210	1.2814	0.3488
1.1999	36.0	216	1.2801	0.3488
1.1813	37.0	222	1.2790	0.3488
1.1813	38.0	228	1.2781	0.3488
1.1753	39.0	234	1.2775	0.3488
1.1877	40.0	240	1.2770	0.3488
1.1877	41.0	246	1.2768	0.3488
1.1774	42.0	252	1.2767	0.3488
1.1774	43.0	258	1.2767	0.3488
1.1704	44.0	264	1.2767	0.3488
1.1843	45.0	270	1.2767	0.3488
1.1843	46.0	276	1.2767	0.3488
1.1726	47.0	282	1.2767	0.3488
1.1726	48.0	288	1.2767	0.3488
1.1541	49.0	294	1.2767	0.3488
1.1928	50.0	300	1.2767	0.3488

Framework versions

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