deit-base-distilled-patch16-224-hasta-75-fold1

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

• Loss: 0.1907
• Accuracy: 1.0

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: 5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	1	1.0839	0.4167
No log	2.0	2	0.8911	0.6667
No log	3.0	3	0.5837	0.75
No log	4.0	4	0.3481	0.9167
No log	5.0	5	0.2815	0.9167
No log	6.0	6	0.2839	0.9167
No log	7.0	7	0.2838	0.9167
No log	8.0	8	0.2692	0.9167
No log	9.0	9	0.2701	0.9167
0.3107	10.0	10	0.3363	0.9167
0.3107	11.0	11	0.3816	0.9167
0.3107	12.0	12	0.3427	0.9167
0.3107	13.0	13	0.2728	0.9167
0.3107	14.0	14	0.2273	0.9167
0.3107	15.0	15	0.2052	0.9167
0.3107	16.0	16	0.1840	0.9167
0.3107	17.0	17	0.1907	1.0
0.3107	18.0	18	0.1761	1.0
0.3107	19.0	19	0.1302	1.0
0.1503	20.0	20	0.0937	1.0
0.1503	21.0	21	0.0767	1.0
0.1503	22.0	22	0.0642	1.0
0.1503	23.0	23	0.0730	1.0
0.1503	24.0	24	0.0937	1.0
0.1503	25.0	25	0.0713	1.0
0.1503	26.0	26	0.0476	1.0
0.1503	27.0	27	0.0462	1.0
0.1503	28.0	28	0.0553	1.0
0.1503	29.0	29	0.0689	1.0
0.068	30.0	30	0.0676	1.0
0.068	31.0	31	0.0534	1.0
0.068	32.0	32	0.0472	1.0
0.068	33.0	33	0.0575	1.0
0.068	34.0	34	0.0614	1.0
0.068	35.0	35	0.0596	1.0
0.068	36.0	36	0.0503	1.0
0.068	37.0	37	0.0571	1.0
0.068	38.0	38	0.0694	1.0
0.068	39.0	39	0.0869	0.9167
0.0416	40.0	40	0.1039	0.9167
0.0416	41.0	41	0.1130	0.9167
0.0416	42.0	42	0.1127	0.9167
0.0416	43.0	43	0.1028	0.9167
0.0416	44.0	44	0.0840	0.9167
0.0416	45.0	45	0.0703	0.9167
0.0416	46.0	46	0.0564	1.0
0.0416	47.0	47	0.0572	1.0
0.0416	48.0	48	0.0689	0.9167
0.0416	49.0	49	0.0967	0.9167
0.048	50.0	50	0.1440	0.9167
0.048	51.0	51	0.1667	0.9167
0.048	52.0	52	0.1823	0.9167
0.048	53.0	53	0.1734	0.9167
0.048	54.0	54	0.1543	0.9167
0.048	55.0	55	0.1383	0.9167
0.048	56.0	56	0.1266	0.9167
0.048	57.0	57	0.1013	0.9167
0.048	58.0	58	0.0830	0.9167
0.048	59.0	59	0.0780	0.9167
0.0118	60.0	60	0.0756	0.9167
0.0118	61.0	61	0.0723	0.9167
0.0118	62.0	62	0.0563	1.0
0.0118	63.0	63	0.0470	1.0
0.0118	64.0	64	0.0469	1.0
0.0118	65.0	65	0.0522	1.0
0.0118	66.0	66	0.0576	1.0
0.0118	67.0	67	0.0639	1.0
0.0118	68.0	68	0.0827	0.9167
0.0118	69.0	69	0.1089	0.9167
0.0271	70.0	70	0.1343	0.9167
0.0271	71.0	71	0.1514	0.9167
0.0271	72.0	72	0.1552	0.9167
0.0271	73.0	73	0.1500	0.9167
0.0271	74.0	74	0.1392	0.9167
0.0271	75.0	75	0.1229	0.9167
0.0271	76.0	76	0.1009	0.9167
0.0271	77.0	77	0.0858	0.9167
0.0271	78.0	78	0.0844	0.9167
0.0271	79.0	79	0.0855	0.9167
0.0462	80.0	80	0.0972	0.9167
0.0462	81.0	81	0.1140	0.9167
0.0462	82.0	82	0.1398	0.9167
0.0462	83.0	83	0.1639	0.9167
0.0462	84.0	84	0.1842	0.9167
0.0462	85.0	85	0.1938	0.9167
0.0462	86.0	86	0.2000	0.9167
0.0462	87.0	87	0.2008	0.9167
0.0462	88.0	88	0.1949	0.9167
0.0462	89.0	89	0.1896	0.9167
0.022	90.0	90	0.1798	0.9167
0.022	91.0	91	0.1700	0.9167
0.022	92.0	92	0.1617	0.9167
0.022	93.0	93	0.1549	0.9167
0.022	94.0	94	0.1492	0.9167
0.022	95.0	95	0.1447	0.9167
0.022	96.0	96	0.1435	0.9167
0.022	97.0	97	0.1431	0.9167
0.022	98.0	98	0.1418	0.9167
0.022	99.0	99	0.1411	0.9167
0.0236	100.0	100	0.1408	0.9167

Framework versions

• Transformers 4.41.0
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1