hushem_5x_deit_tiny_adamax_001_fold4

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: 1.7375
• Accuracy: 0.7857

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
1.4656	1.0	28	1.3873	0.3095
1.2586	2.0	56	1.2764	0.5
1.1469	3.0	84	0.9328	0.5
1.0441	4.0	112	0.8472	0.5952
0.811	5.0	140	0.7157	0.6667
0.9304	6.0	168	1.0112	0.5476
0.6688	7.0	196	1.1025	0.6429
0.7046	8.0	224	0.9337	0.7143
0.6117	9.0	252	0.8164	0.7143
0.4458	10.0	280	0.9534	0.7381
0.3231	11.0	308	0.8226	0.7381
0.2594	12.0	336	1.4042	0.6905
0.3579	13.0	364	1.0151	0.6905
0.3051	14.0	392	1.1256	0.7619
0.212	15.0	420	1.4604	0.6905
0.1011	16.0	448	1.4776	0.7381
0.0658	17.0	476	1.5151	0.7143
0.0916	18.0	504	2.0212	0.6667
0.12	19.0	532	1.3757	0.7143
0.1417	20.0	560	1.9592	0.6905
0.1037	21.0	588	1.5184	0.7143
0.0812	22.0	616	1.5083	0.7381
0.0404	23.0	644	1.7932	0.7381
0.0758	24.0	672	1.5450	0.7143
0.0384	25.0	700	2.0953	0.6667
0.0277	26.0	728	1.9894	0.6667
0.0016	27.0	756	1.8938	0.7143
0.0008	28.0	784	1.7999	0.7619
0.0118	29.0	812	1.7512	0.7619
0.0001	30.0	840	1.8297	0.7619
0.0002	31.0	868	1.7978	0.7381
0.0105	32.0	896	1.6941	0.7857
0.0001	33.0	924	1.6973	0.7619
0.0	34.0	952	1.6981	0.7381
0.0	35.0	980	1.7026	0.7381
0.0	36.0	1008	1.7088	0.7619
0.0	37.0	1036	1.7123	0.7619
0.0	38.0	1064	1.7165	0.7619
0.0	39.0	1092	1.7201	0.7619
0.0	40.0	1120	1.7234	0.7619
0.0	41.0	1148	1.7263	0.7619
0.0	42.0	1176	1.7294	0.7619
0.0	43.0	1204	1.7316	0.7619
0.0	44.0	1232	1.7334	0.7857
0.0	45.0	1260	1.7350	0.7857
0.0	46.0	1288	1.7363	0.7857
0.0	47.0	1316	1.7371	0.7857
0.0	48.0	1344	1.7375	0.7857
0.0	49.0	1372	1.7375	0.7857
0.0	50.0	1400	1.7375	0.7857

Framework versions

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