hushem_5x_deit_small_adamax_001_fold1

This model is a fine-tuned version of facebook/deit-small-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 4.9670
• Accuracy: 0.4667

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
1.4243	1.0	27	1.4114	0.2667
1.0749	2.0	54	1.3320	0.4222
0.8047	3.0	81	1.4606	0.4
1.1443	4.0	108	1.1606	0.3778
0.6749	5.0	135	1.3783	0.3778
0.6355	6.0	162	1.1666	0.5111
0.6766	7.0	189	2.0044	0.3778
0.5072	8.0	216	2.1218	0.3778
0.5923	9.0	243	2.3668	0.3778
0.5554	10.0	270	1.8883	0.3556
0.5809	11.0	297	1.7992	0.3778
0.3925	12.0	324	2.0376	0.3778
0.468	13.0	351	2.2940	0.3556
0.388	14.0	378	2.2184	0.4
0.3576	15.0	405	2.7767	0.3556
0.2702	16.0	432	1.8836	0.4444
0.2448	17.0	459	2.0997	0.4
0.3121	18.0	486	1.6999	0.4667
0.1734	19.0	513	2.3268	0.4444
0.318	20.0	540	3.2732	0.3556
0.2011	21.0	567	2.9255	0.4
0.1594	22.0	594	3.3998	0.3778
0.1002	23.0	621	2.6724	0.5556
0.1302	24.0	648	2.4689	0.5111
0.1301	25.0	675	2.3542	0.4889
0.0517	26.0	702	3.6395	0.4222
0.0477	27.0	729	3.0379	0.4667
0.1158	28.0	756	3.8148	0.4
0.0196	29.0	783	4.2909	0.4
0.0022	30.0	810	4.5812	0.4222
0.0023	31.0	837	4.8297	0.4667
0.0086	32.0	864	4.7485	0.4667
0.0049	33.0	891	4.6840	0.4889
0.0002	34.0	918	4.4345	0.4667
0.0001	35.0	945	5.3275	0.4889
0.0001	36.0	972	4.4575	0.4667
0.0001	37.0	999	4.7352	0.4889
0.0	38.0	1026	4.7979	0.4667
0.0	39.0	1053	4.8394	0.4667
0.0	40.0	1080	4.8675	0.4667
0.0	41.0	1107	4.8943	0.4667
0.0	42.0	1134	4.9142	0.4667
0.0	43.0	1161	4.9318	0.4667
0.0	44.0	1188	4.9439	0.4667
0.0	45.0	1215	4.9539	0.4667
0.0	46.0	1242	4.9606	0.4667
0.0	47.0	1269	4.9652	0.4667
0.0	48.0	1296	4.9670	0.4667
0.0	49.0	1323	4.9670	0.4667
0.0	50.0	1350	4.9670	0.4667

Framework versions

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