hushem_5x_deit_tiny_rms_0001_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: 2.7509
• Accuracy: 0.6977

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: 0.0001
• 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
1.5103	1.0	28	2.0778	0.2558
1.4169	2.0	56	1.4920	0.2558
1.1717	3.0	84	1.4368	0.3488
0.9912	4.0	112	0.9988	0.4651
0.8022	5.0	140	1.6709	0.3953
0.7789	6.0	168	0.6692	0.7907
0.7753	7.0	196	0.7299	0.7209
0.7094	8.0	224	0.9947	0.7209
0.5393	9.0	252	1.1069	0.6279
0.4827	10.0	280	1.3153	0.5581
0.4471	11.0	308	0.7571	0.7209
0.2911	12.0	336	1.0945	0.6977
0.2341	13.0	364	1.4428	0.7209
0.1731	14.0	392	1.1663	0.7442
0.1668	15.0	420	2.1058	0.5581
0.0808	16.0	448	1.5095	0.6977
0.0267	17.0	476	2.3464	0.5349
0.0601	18.0	504	1.3157	0.7442
0.0193	19.0	532	1.9786	0.6279
0.0007	20.0	560	1.8771	0.7209
0.0002	21.0	588	1.8199	0.6744
0.0002	22.0	616	2.2093	0.6279
0.0001	23.0	644	2.3026	0.6512
0.0	24.0	672	2.3149	0.6744
0.0	25.0	700	2.3349	0.6744
0.0	26.0	728	2.3579	0.6744
0.0	27.0	756	2.3790	0.6744
0.0	28.0	784	2.4090	0.6744
0.0	29.0	812	2.4324	0.6744
0.0	30.0	840	2.4483	0.6977
0.0	31.0	868	2.4871	0.6977
0.0	32.0	896	2.5064	0.6977
0.0	33.0	924	2.5268	0.6977
0.0	34.0	952	2.5458	0.6977
0.0	35.0	980	2.5702	0.6977
0.0	36.0	1008	2.5945	0.6977
0.0	37.0	1036	2.6129	0.6977
0.0	38.0	1064	2.6351	0.6977
0.0	39.0	1092	2.6496	0.6977
0.0	40.0	1120	2.6665	0.6977
0.0	41.0	1148	2.6790	0.6977
0.0	42.0	1176	2.6948	0.6977
0.0	43.0	1204	2.7095	0.6977
0.0	44.0	1232	2.7229	0.6977
0.0	45.0	1260	2.7315	0.6977
0.0	46.0	1288	2.7407	0.6977
0.0	47.0	1316	2.7476	0.6977
0.0	48.0	1344	2.7508	0.6977
0.0	49.0	1372	2.7509	0.6977
0.0	50.0	1400	2.7509	0.6977

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0